Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland

1. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland Keith N. Eshleman & Andrew Elmore (PIs) Appalachian Laboratory University of Maryland Center for Environmental Science Frostburg, MD 21532 Final Report submitted to: Maryland Department of the Environment Baltimore, MD February 18, 2013

2. AcknowledgmentsWe solicited and received the help of quite a few people during the completion of this project.While the project investigators (KNE and AJE) are responsible for the content of the final report,we especially acknowledge three individuals who helped us with the organization, content, andwriting of several chapters: Russell R. Dickerson (Department of Atmospheric and OceanicScience, University of Maryland; Chapter 2) Jeanne M. VanBriesen (Department of Civil andEnvironmental Engineering, Carnegie Mellon University; Chapter 4); and Todd R. Lookingbill(Department Geography and the Environment, University of Richmond; Chapter 5). Previoustime commitments precluded them from assisting with other sections of the report, so theyare listed as chapter co-authors. We thank the following individuals for providing us withdata: Scott Bearer, Deborah Carpenter, Jennifer Chadwick-Moore, Kevin Coyne, LynnDavidson, Gregory Day, Dan Feller, Dave Foreman, Mike Garner, John Grace, Jeff Halka, KaraHawkins, Jaron Hawkins, Greg Hildreth, Zoe Johnson, Catherine McCall, Phung Pham, GenePiotrowski, Matt Rowe, Matt Sell, Frank Siano, Matthew Stover, and Kevin Wagner. RobertSabo compiled much of the extant literature on best practices that underpinned our analyses of bestpractices. Steven Guinn produced all of the maps used in this report and performed many of thegeographic analyses on which our recommendations are based. Robert Hilderbrand contributed textand comments to Chapter 6. Finally, Christine Conn (Maryland DNR), Brigid Kenney (MDE), RonKlauda (Maryland DNR) and several other anonymous reviewers provided constructive criticism ofan earlier draft of the report. ii

3. Table of ContentsCover page............................................................................................................................................ iAcknowledgments ............................................................................................................................... iiTable of Contents ............................................................................................................................... iiiList of Acronyms and Abbreviations ................................................................................................. viSummary of key findings and major recommendations.......................................................................1Chapter 1. General, planning, and permitting BMPs ................................................................. 1-1 A. Pre-development environmental assessment ......................................................................... 1-3 B. On-site and off-site monitoring ............................................................................................. 1-5 C. Comprehensive drilling plans (CDP) ................................................................................... 1-7 D. Well pad spacing ................................................................................................................ 1-10 E. Setback requirements .......................................................................................................... 1-11 F. Identification of freshwater aquifers and groundwater flowpaths ...................................... 1-12 G. Stakeholder engagement..................................................................................................... 1-15 H. Special siting criteria .......................................................................................................... 1-16 I. Reclamation planning .......................................................................................................... 1-20 J. Well permitting, county and state coordination, and financial assurance ............................ 1-21 K. Key recommendations ........................................................................................................ 1-25 L. Literature cited ..................................................................................................................... 1-27Chapter 2. Protecting air quality .................................................................................................. 2-1 A. Reducing pollutant and pollutant precursor emissions from MSGD operations................... 2-3 B. On-site and off-site air quality monitoring ............................................................................ 2-6 C. Key recommendations ........................................................................................................... 2-6 D. Literature cited ...................................................................................................................... 2-7Chapter 3. Well engineering and construction practices to ensure integrity and isolation .... 3-1 A. Well planning ....................................................................................................................... 3-2 B. Drilling.................................................................................................................................. 3-4 C. Casing and cementing ........................................................................................................... 3-7 D. Well logging ....................................................................................................................... 3-10 E. Pressure testing .................................................................................................................... 3-12 F. Blow-out prevention ............................................................................................................ 3-12 G. Completing and hydraulic fracturing ................................................................................. 3-13 H. Use of well development techniques other than hydraulic fracturing ................................ 3-14 I. Determining the extent of induced fractures ........................................................................ 3-15 J. Plugging ............................................................................................................................... 3-16 K. Key recommendations ......................................................................................................... 3-18 L. Literature cited ..................................................................................................................... 3-19Chapter 4. Protecting water resources ......................................................................................... 4-1 A. Siting requirements: setbacks and restrictions ..................................................................... 4-3 B. Monitoring of water resources prior to, during, and following development ....................... 4-6 C. Water pollution, stormwater management, and erosion and sediment control .................... 4-8 D. Water withdrawals ............................................................................................................... 4-15 E. Comprehensive basin-scale water management planning. .................................................. 4-19 F. Water storage and delivery .................................................................................................. 4-20 G. Alternative water sourcing .................................................................................................. 4-21 H. Chemical delivery, storage on-site, and transfers................................................................ 4-22 iii

4. I. Identification of chemicals .................................................................................................... 4-23 J. Drilling and drilling wastes .................................................................................................. 4-23 K. On site management of produced waters and wastewaters ................................................. 4-24 L. Management of produced water (including recordkeeping, manifesting) ........................... 4-25 M. Treatment of produced water .............................................................................................. 4-27 N. Disposal of produced water or residual treatment wastes ................................................... 4-29 O. Reclamation and closure (decommissioning)...................................................................... 4-30 P. Naturally occurring radioactive materials (NORM) ............................................................ 4-31 Q. Key recommendations ......................................................................................................... 4-32 R. Literature cited ..................................................................................................................... 4-33Chapter 5. Protecting terrestrial habitat and wildlife ................................................................ 5-1 A. Well pad spacing and siting................................................................................................... 5-1 B. Impoundments ....................................................................................................................... 5-4 C. Riparian setbacks ................................................................................................................... 5-6 D. Special protection areas ......................................................................................................... 5-7 E. Noise and light ..................................................................................................................... 5-10 F. Construction of roads and pipelines ..................................................................................... 5-12 G. Invasive plants and wildlife ................................................................................................. 5-13 H. Reclamation ......................................................................................................................... 5-14 I. Key recommendations .......................................................................................................... 5-15 I. Literature cited ...................................................................................................................... 5-16 Appendix 5A ............................................................................................................................ 5-19Chapter 6. Protecting aquatic habitat, wildlife, and biodiversity ............................................. 6-1 A. Buffers and setbacks .............................................................................................................. 6-3 B. Special protection of high-value assets ................................................................................ 6-5 C. Design and construction of well pads, access roads, pits, utility corridors, and pipelines ... 6-9 D. Erosion and sediment controls ............................................................................................ 6-12 E. Invasive species controls .................................................................................................... 6-13 F. Key recommendations.......................................................................................................... 6-13 G. Literature cited .................................................................................................................... 6-14Chapter 7. Protecting public safety .............................................................................................. 7-1 A. Spill prevention and emergency response ............................................................................. 7-1 B. Site security ........................................................................................................................... 7-3 C. Transportation planning ......................................................................................................... 7-4 D. Key recommendations ........................................................................................................... 7-5 E. Literature cited ....................................................................................................................... 7-6Chapter 8. Protecting cultural, historical, and recreational resources ..................................... 8-1 A. Identification of sites ............................................................................................................. 8-1 B. Setback requirements and mitigation .................................................................................... 8-2 C. Key recommendations ........................................................................................................... 8-5 D. Literature cited ...................................................................................................................... 8-6Chapter 9. Protecting quality of life and aesthetic values ......................................................... 9-1 A. Hours of operation ................................................................................................................ 9-1 B. Noise control......................................................................................................................... 9-2 C. Road impacts and transportation planning ........................................................................... 9-2 D. Visual pollution/viewsheds .................................................................................................. 9-4 E. Key recommendations ........................................................................................................... 9-5 F. Literature cited ....................................................................................................................... 9-6 iv

5. Chapter 10. Protecting agriculture and grazing ....................................................................... 10-1 A. Protection of prime farmland .............................................................................................. 10-2 B. Key recommendations ......................................................................................................... 10-4 C. Literature cited ..................................................................................................................... 10-4 v

6. List of Acronyms and AbbreviationsAMD acid mine drainageAPI American Petroleum InstituteBCF billion cubic feetbhp-hr brake horsepower-hourBioNET Biodiversity Conservation NetworkBMP best management practiceBOPE blow-out prevention equipmentBTEX benzene, toluene, ethylbenzene, and xyleneCBL cement bond loggingCBM coal bed methaneCDP comprehensive drilling planCNHI County Natural Heritage InventoryCO carbon monoxideCO2 carbon dioxideCOD chemical oxygen demandCOGCC Colorado Oil and Gas Conservation CommissionCOMAR Code of Maryland RegulationsCWA Clean Water ActDMRM Division of Mineral Resources Management (OH)DNR Department of Natural Resources (MD)EAF environmental assessment form vi

7. ELAP Environmental Laboratory Accreditation ProgramERP emergency response planERT Environmental Review ToolESC erosion and sediment controlFCA Forest Conservation Act (MD)FEMA Federal Emergency Management Agencyft feetGHG greenhouse gasHAB harmful algal bloomsHAP hazardous air pollutantHC hydrocarbonHDPE high density polyethyleneHDT Hoffman Drainage TunnelHF hydraulic fracturingHP horsepowerHVHF high volume hydraulic fracturingHQW high quality watersHUC hydrologic unit codeICPRB Interstate Commission on the Potomac River BasinIR infrared radiationkg kilogramkm kilometerskm2 square kilometers vii

8. LLRW low-level radioactive wasteLPG liquefied petroleum gasL/s liters per secondLWD logging while drillingMALPF Maryland Agricultural Land Preservation FoundationMBSS Maryland Biological Stream SurveyMDA Maryland Department of AgricultureMDE Maryland Department of the EnvironmentMG million gallonsMGD million gallons per daymg/L milligrams per literMGS Maryland Geological SurveyMHT Maryland Historical TrustMMCF million cubic feetMSDS material safety data sheetMSGD Marcellus shale gas developmentMSSCS Maryland Synoptic Stream Chemistry SurveyMWD measurement while drillingNAAQS National Ambient Air Quality StandardNESHAP National Emission Standards for Hazardous Air PollutantsNORM naturally occurring radioactive materialsNOx nitrous oxideNPDES National Pollutant Discharge Elimination System viii

9. NRC National Research CouncilNSPS New Source Performance StandardsNWI National Wetlands InventoryNYSDEC New York State Department of Environmental ConservationO3 ozone (photochemical smog)ONRW outstanding national resource watersOVM organic vapor meterORSANCO Ohio River Valley Water Sanitation CommissionPA DCNR Pennsylvania Department of Conservation and Natural ResourcesPA DEP Pennsylvania Department of Environmental ProtectionpCi/g picocuries per gramPE professional engineerPM2.5 fine particulate matter (< 2.5 μm)PNDI Pennsylvania natural diversity inventoryPNHP Pennsylvania Natural Heritage ProgramPOTW publically owned treatment worksPPC prevention, preparedness, and contingencypsi pounds per square inchQA/QC quality assurance/quality controlRAIN River Alert Information NetworkRCRA Resource Conservation and Recovery ActRGGI Regional Greenhouse Gas InitiativeRP recommended practice ix

10. RW radioactive wasteSAV submerged aquatic vegetationSHA State Highway Administration (MD)SRBC Susquehanna River Basin CommissionSRCBL segmented radial cement bond loggingSTRONGER State Review of Oil and Natural Gas Environmental Regulations, Inc.SWPPP stormwater pollution prevention programTCEQ Texas Commission on Environmental QualityTCF trillion cubic feetTDS total dissolved solidsTENORM technologically-enhanced, naturally-occurring radioactive materialsTNC The Nature ConservancyTSS total suspended solidsTVD true vertical depthUIC Underground Injection ControlULSD ultra low sulfur dieselUSACE United States Army Corps of EngineersUSDA United States Department of AgricultureUSDW underground source of drinking waterUSEPA United States Environmental Protection AgencyUSFWS United States Fish and Wildlife ServiceUSGS United States Geological SurveyVDL variable density log x

11. VOC volatile organic compoundWOC wait on cementWQS water quality standardsWSSC Wetlands of Special State ConcernWVDEP West Virginia Department of Environmental Protection xi

12. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandSummary of key findings and major recommendations1The Marcellus shale formation underlying numerous Appalachian states is considered the largestgas-bearing shale formation in the United States. The thousands of new gas wells that have beendrilled in this region since 2004 are testimony to a revolution in domestic natural gas productionin the U.S. through so-called “unconventional” development that includes both modernhorizontal drilling and high volume hydraulic fracturing technologies. Unlike neighboringPennsylvania that participated fully in the initial boom in exploration and production between2005 and 2009 (drilling has occurred extensively both on private and public lands inPennsylvania), Maryland (with a significantly smaller resource) has chosen to stay on thesidelines with an unofficial moratorium on unconventional Marcellus shale gas development(MSGD) while it studies the lessons from other states, determines whether development can goforward safely, and evaluates it options. The present study of best management practices (BMPs)for Marcellus shale gas development represents an effort to determine what actual practiceswould provide the maximum protection of Maryland’s environment, natural resources, andpublic safety should the state decide to move forward with development of this resource in thenear future.We carefully reviewed the current regulations governing unconventional shale gas developmentin five other states (Colorado, New York, Ohio, Pennsylvania, and West Virginia), as well as therecommendations of many other expert panels and organizations that have reviewed bothregulations and BMPs in these and other states. We visited several well pads as part of threeorganized field trips that allowed us to gain an important visual perspective of the operations,practices, and challenges involved in conducting MSGD. Wherever possible, we also reviewedthe scientific literature to evaluate the proven effectiveness of different practices, but the lack ofcomprehensive, data-driven studies of the impacts of MSGD both on-site and off-site present asignificant impediment to recommending best practices on the basis of this criterion alone. Forthis reason, we have explicitly chosen to identify and recommend specific BMPs that—mostlyon the basis of our professional judgment—would provide as much protection of Maryland’snatural, cultural, historical and recreational resources; the environment; and public safety as canreasonably be provided while allowing MSGD to move forward.We believe that it is inevitable that there will be negative impacts from MSGD in westernMaryland (and perhaps beyond the state’s borders) and that a significant portion of these “costs”will be borne by local communities. Heavy truck traffic on local roads, noise and odorsemanating from drilling sites, conflicts with outdoor recreation, diminished tourism, reducedbiodiversity, and deterioration of air and water quality are some examples of the types of impactsthat are likely even under the best of circumstances. While difficult to quantify in economicterms, these “costs” will ideally be greatly outweighed by the benefits of increased economicactivity—otherwise it is very difficult to make a case that MSGD should occur at all. Our goalwas to identify and recommend specific BMPs that would provide maximum protection of1 Chapter co-authors: Keith N. Eshleman, Ph.D. and Andrew J. Elmore, Ph.D. (both at: Appalachian Laboratory, University of Maryland Center for Environmental Science, Frostburg, MD 21532) i

13. K.N. Eshleman & A.J. Elmore (2013) Summary of key findings and major recommendationsMaryland’s environment, natural resources, and public safety. There are a variety of types ofresources and hazards—in some cases overlapping—distributed across the western Marylandlandscape that present important constraints on MSGD. For this reason, implementation of someBMPs will effectively result in the exclusion of MSGD from select areas of the region to reducethe risk of impacts, thus limiting to some degree the total volume of gas eventually extracted.Due to the nature of this activity in which well bores can be drilled horizontally 8,000 ft from thewell pad, it will often be possible to drill under the most valuable and at-risk resources ofwestern Maryland. This potential is enhanced through the use of multi-well pads that are capableof draining between one and two square miles of the target formation. Further, locating multi-well pads in dense clusters—with clusters spaced as far apart as is technically feasible—makesmaximum use of horizontal drilling technology and could be an important BMP in terms ofminimizing surface development impacts. With careful and thoughtful planning (e.g., co-locationof associated infrastructure wherever possible), it may be possible to develop much of the gasresource in a way that converts less than 1-2% of the land surface, even when accounting for theneed for ancillary infrastructure such as access roads, pipelines, and compressor facilities. Whilethis build-out scenario would occupy much less surface area than other forms of development,even with the most protective BMPs in place it would certainly not be expected to occur withoutsome significant negative impacts on the western Maryland region.Maryland already has a reasonably well-developed set of regulations that pertains mostly toconventional oil and gas development, but the state lacks a regulatory/enforcement structure toaddress unconventional gas development. Clearly, a regulatory program would have to quickly“ramp up” to effectively address the myriad issues that would be presented by MSGD in the stateand to avoid some of the problems that have occurred elsewhere. An important best managementpractice is therefore to “go slow” and allow a new regulatory structure and experience ininspection and enforcement to evolve over time and effectively “catch up” to the new technologyas MSGD proceeds. If and when MSGD moves forward in western Maryland, we believe thateffective planning by local and state governments that moderates the rate at which the gasresource is developed across the region would help mitigate some of the negative effects of“boom-bust” cycles that have occurred elsewhere. There are a number of specificrecommendations throughout this report that provide guidance in this area.In particular, perhaps the single most important among these recommendations is that the stateshould develop regulations to support the design and implementation of comprehensive drillingplans (CDPs) for MSGD. We envision a voluntary program similar to Colorado’s approach (andthe program that has been used to develop the Marcellus gas resource in Pennsylvania stateforests), but one that provides strong incentives for operators willing to consider this option.After identifying foreseeable oil and gas activities in a defined geographic area upfront, energycompanies would work cooperatively with other stakeholders (including state natural resourceagencies, counties, citizen groups, etc.) to develop an integrated plan for efficiently exploitingthe resource while minimizing impacts on communities, ecosystems, and natural resources. TheCDP approach offers many advantages, but the most important one is that it provides a way ofeffectively channeling this industrial activity into those areas where fewer of the most sensitiveresources are “in harm’s way” and where new infrastructure needs (e.g., roads, pipelines) arelower. Logically, the first approved CDP would most likely result in permitting an area fordrilling where major drilling hazards, risks to public safety, and impacts on sensitive ecological, ii

14. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandrecreational, historical and cultural resources can be largely avoided. Since we expect that theplanning process for a CDP would be longer than for individual well drilling permits, anothermajor advantage of this approach is that it could enable MSGD to move forward at a somewhatslower, more manageable rate. One way the state might incentivize comprehensive gasdevelopment planning could be by reducing permit fees and bonding requirements for wellscovered under a CDP. Over time, monitoring data collected both on-site and off-site to documentimpacts (or non-impacts) would be used by the industry to improve BMPs (this is the way theBMP process is supposed to work). Additional CDPs would presumably be dependent upon theindustry demonstrating that any impacts from earlier drilling were within acceptable limits orthat newer practices were significantly better at reducing any unacceptable impacts observed inprior phases.A critically important consideration influencing the success of CDPs in Maryland would becareful site selection based on pre-development environmental assessment for well pads andrelated infrastructure. A careful pre-drilling environmental assessment would include, at aminimum, an assessment of all existing data combined with two years of pre-drilling monitoring,including surface and groundwater testing, inventories of rare, threatened and endangeredspecies, and an assessment of the potential to introduce invasive species during site developmentor water procurement. Should any changes in observed water quality occur during drilling orproduction, pre-drilling assessment should make possible a defendable determination of liability.It is important to remember that western Maryland is a geographically small, rural, andmountainous landscape, offering residents a high quality of life, in part due to abundantbiological, recreational, and cultural resources with exceptional value. Because of itsmountainous landscape and history of coal mining, there are also many hazards in westernMaryland that must be avoided in the interest of long-term well integrity and public safety. Thegoal of best management practices for siting MSGD-related infrastructure would be to provide asafe environment for all residents, avoid conflicts with existing land uses, and observe all on-going efforts to conserve biological diversity. Throughout this report we have recommendedspecific setbacks from irreplaceable natural areas, aquatic habitat, and hotspots for biodiversity(e.g., caves). Maryland has recently placed an emphasis on mapping valuable resources; thisactivity should continue and the resulting data should be made available to prospective drillingoperators to optimize the placement of well pads and related infrastructure.One BMP we have highlighted throughout this report is the avoidance of underground voids,which can often be justified based on caves’ conservation value for many rare threatened andendangered species. Additionally, complications from encountering a cave (or deep coal mine)during drilling can jeopardize the integrity of the well, leading to an increased chance of leaks,methane contamination of underground sources of drinking water, and even blowouts. Althoughit is standard practice in many states to drill down through subterranean voids, our researchsuggests that this practice comes with important risks and at least one state has begun looking atthe technique with greater scrutiny. A best practice for Maryland would be to avoid allunderground voids by employing the best mapping and detection technologies and then applyingadditional setbacks when siting the borehole. Similarly, there are several clusters of historicconventional gas wells throughout western Maryland. Because these boreholes provide apotential conduit for gas and possibly brines to migrate upwards into underground sources ofdrinking water, we recommend that all portions of new unconventional gas wells (vertical and iii

15. K.N. Eshleman & A.J. Elmore (2013) Summary of key findings and major recommendationshorizontal) be positioned at least ¼ mile from such boreholes. Finally, at least until it can beshown that hydraulic fracturing can be done safely within relatively close proximity tounderground sources of drinking water, we recommend that Maryland prudently follow guidancefrom New York’s experience in regulating unconventional shale gas development and not permitMSGD (or any other unconventional gas development) where the Marcellus formation is locatedwithin 2,000 vertical ft of the ground surface.Despite best management practices designed to keep MSGD infrastructure away from our mosttreasured assets, there will always be impacts, which left unmitigated would adversely affecttourism, public safety, and the quality of life for residents and visitors alike. We, therefore,recommend implementing a suite of state-of-the-art mitigative techniques that would aim to limitthe impact of MSGD should Maryland decide to invest fully in this new industry. The first ofthese techniques would seek to limit total impervious surface (e.g., pavement, buildings, gravelroads, well pads) to 2% for any watershed currently below this threshold. There is abundantscientific evidence that watershed impervious surface area is a robust indicator of cumulativeimpacts to watershed structure and functioning. Secondly, we recommend imposing a “no-net-loss-of-forest” requirement on MSGD. This would tend to push well pad development into non-forest areas, but also require mitigation plantings of trees whenever forests are cut to make wayfor MSGD infrastructure. There are many other important mitigative techniques that could beemployed to reduce the overall impact of MSGD on biological, recreational, and culturalresources, and that ultimately help to maintain a high quality of life in western Maryland. Theseinclude the use of line power instead of diesel generators to protect air quality, sound barriersand visual screens to reduce the impacts of drill rigs and compressor stations, limits on hours ofdrilling operations to avoid peak tourism periods (e.g., hunting season for white tail deer), andthoughtful truck traffic regulation to reduce the impact of water hauling convoys on quiet ruralroads. Finally, although many landowners might earn substantial profits from MSGD on theirland, their neighbors who opt out should be protected from the worst impacts. Sensible zoningordinances and reasonable property line setbacks are certainly one way to reduce conflicts, butwe also recommend enhanced transparency and increased public advertising of planned drilling;no one should be surprised and concerns of all parties should be addressed fully before drillingbegins.Our review of well engineering practices revealed that the gas development industry hasresponded to pressure to reduce its environmental footprint by developing a suite of bestmanagement practices to maintain the integrity of each well system, isolate the well from thesurrounding subsurface environment, and effectively contain the produced gas and other fluidswithin the well’s innermost production conduit; in so doing, the gas can ultimately be transportedthrough ancillary pipelines for processing and delivery to market, while the wastewater (i.e.,“flowback”, brines) that is returned to the surface can be efficiently captured, contained, treated,and ultimately recycled (while things are rapidly changing, the industry still relies very heavilyon underground injection as the ultimate disposal process). The American Petroleum Institute(API)—as the technical arm of the oil and gas industry—has taken the lead in reviewing andevaluating the industry’s practices for drilling, completing, and operating oil and natural gaswells; on the basis of its on-going technical reviews of various practices, API has published anextensive number of documents describing so-called “recommended practices” (RPs) which itcommunicates and shares with the industry. Many of these RP’s explicitly address problems in iv

16. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandmaintaining well integrity and provide standards that have been expressly adopted by stateregulatory authorities. If Maryland decides to begin permitting MSGD, we recommend that anyoperator who proposes drilling in the state should be required—at a minimum—to adopt API’sRPs and standards for well planning, well design, well construction, well completion, and welldecommissioning. These practices can certainly be improved upon (for example, through morewidespread field testing), and we believe it is very likely that API will gradually refine its RPspertaining specifically to unconventional shale gas development. Maryland should require alloperators to employ drilling, completion, and environmental control technologies and practicesthat fully meet these evolving standards and that are considered up-to-date.The current BMP for handling drilling fluids, hydraulic fracturing chemicals, wastewaters, andsolid wastes on-site is through the use of a “closed-loop drilling system” in which all fluids arekept stored in watertight tanks that sit within secondary containment on lined and bermed “zero-discharge” well pads that can provide tertiary containment of contaminants and 100% retentionof stormwater. All transfers of materials must be performed carefully on the pad so that anyspills that occur can be quickly and fully contained. This type of drilling system—if properlydesigned and operated—would be expected to provide the lowest risk of contaminant leakageoff-site such as might occur during extreme weather events. Under no circumstances should openpits for storage of wastes or wastewaters be allowed in Maryland. Maryland will need tocarefully review its stormwater regulations as they pertain to oil and gas extraction and find away to treat these industrial sites in the same way that other “hotspots” are treated. Operatorswill need to employ both “active” and “passive” stormwater management to effectively collectall water on a pad site over the entire life of drilling, completing, and producing operations tominimize soil erosion and downstream sedimentation (and avoid any inadvertent contaminantreleases to the environment), although we explicitly recommend against employing any BMPson-site that rely on soil infiltration due to the risks of groundwater contamination.Marcellus shale gas development produces large volumes of wastewater (flowback and producedwater, commonly considered brines) that must be effectively contained, treated, and either safelydisposed of or reused. First of all, under no circumstances should Maryland allow discharge ofany untreated or partially-treated brine, or residuals from brine treatment facilities, into thewaters of the state. To protect its water supplies, Maryland should establish a goal of 100%recycling of wastewater in permitting any MSGD within the state and have a very strongpreference for on-site recycling of wastewater. Development of brine treatment plants thatrecycle water to drillers should be discouraged in favor of on-site treatment by mobile units andimmediate reuse for hydraulic fracturing at the same site (or at a nearby site). On-site watertreatment and reuse would be expected to minimize overall freshwater use for MSGD and reducethe volume of waste, while dramatically decreasing truck transport and associated impacts acrossthe region. Along these lines, the state should also explore the use of non-potable water sources(e.g., acid mine drainage that represents a legacy of past coal mining practices in the region) as away of supplementing needed water withdrawals from the region’s rivers and reservoirs. Finally,before permitting any development in the state, Maryland should carefully review the relevantregulations surrounding development and use of underground injection wells for produced waterfrom MSGD, and at the same time evaluate the capacity of nearby states to accept producedwater or residual concentrated brine from treatment of produced water. v

17. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland1. General, planning, and permitting BMPs1The Marcellus shale formation underlying numerous Appalachian states is considered the largestgas-bearing shale formation in the United States. The thousands of new gas wells that have beendrilled in this region since 2004 are testimony to a revolution in domestic natural gas productionin the US through so-called “unconventional” development that includes both modern horizontaldrilling and high volume hydraulic fracturing technologies (Soeder and Kappel 2009). Unlikeneighboring Pennsylvania that participated fully in the initial boom in exploration and productionbetween 2005 and 2009 (drilling has occurred extensively both on private and public lands inPennsylvania), Maryland (with a significantly smaller resource) chose to stay on the sidelineswith an unofficial moratorium on unconventional Marcellus shale gas development (MSGD)while it studies the lessons from other states, determines whether development can go forwardsafely, and evaluates it options2. The present study of best management practices (BMPs) forMarcellus shale gas development represents an effort to determine what actual practices wouldprovide the maximum protection of Maryland’s environment, natural resources, and publicsafety should the state decide to move forward with development of this resource in the nearfuture.Only about 1.1% of the Marcellus shale gas play is in Maryland—by far the smallest portion ofthe 95,000 square miles of land underlain by this Devonian sedimentary formation that wasdeposited about 380 million years before present (USEIA 2012). We found many estimates ofthe gas resource contained in the Marcellus formation: (1) in 2002, the U.S. Geological Survey(USGS) estimated that the formation contained 1.9 trillion cubic feet (TCF); in 2008, Englanderprovided an estimate of 500 TCF; and in 2012, the U.S. Energy Information Administration(USEIA) estimated that 141 TCF remained that were technically recoverable as of January 1,20093. Obviously, no one knows exactly how much gas exists within the Marcellus Shaleunderlying western Maryland, nor the value of the gas given uncertainties about future suppliesand demands that would in part determine pricing. It has been estimated, however, that there is a50% chance that there is at least 1,286 billion cubic feet (BCF) present in Maryland (a “mid-casescenario”) and development of this resource could support aggregate production of 710 BCFfrom 365 wells on private land over a 30-year period from 2016 to 2045—valued in total at morethan $4B (in constant 2011 US dollars; Sage Policy Group, Inc. 2012). Regardless of whetherthese estimates are at all realistic, it is obvious from Pennsylvania’s experience that very realeconomic benefits have been realized from MSGD (including generation of $413M in lease saleson 139,000 acres of state forest from 2008-2010, plus $88M in royalties from gas production ofabout 250 wells)4.As part of our research, we have carefully reviewed the current regulations governing MSGD infive other states (Colorado, New York, Ohio, Pennsylvania, and West Virginia), as well as the1 Chapter co-authors: Keith N. Eshleman, Ph.D. and Andrew J. Elmore, Ph.D. (both at: Appalachian Laboratory, University of Maryland Center for Environmental Science, Frostburg, MD 21532)2 Governor Martin O’Malley issued an Executive Order on June 6, 2011 establishing the Marcellus Shale Safe Drilling Initiative and Advisory Commission.3 Statistics are: MD (1.09%), NY (20.06%), OH (18.19%), PA (35.35%), VA (3.85%), WV (21.33%); USEIA 20124 Ellen Shutzbarger (PADCNR), personal communication (August 17, 2012) 1-1

18. K.N. Eshleman & A.J. Elmore (2013) Chapter 1recommendations of many other expert panels and organizations that have reviewed bothregulations and BMPs in these and other states. Where possible, we also reviewed the scientificliterature to determine the proven effectiveness of different BMPs, particularly those that areused in road construction and the protection of terrestrial and aquatic habitat and biodiversity.Finally, we visited several well pads as part of some organized field trips that allowed us to gainan important visual perspective of the operations, practices, and challenges involved inconducting MSGD.It is obvious that MSGD—if and when it comes to western Maryland—will be associated withboth benefits and costs. Christopherson and Rightor (2011) describe recent MSGD inPennsylvania and elsewhere as a classic “boom-bust cycle” that is characteristic of otherextractive industries. The most evident impacts of the “boom” phase of the cycle are a verysudden and rapid increase in local economic activity due to drilling companies, crews, and gas-related businesses moving into an area to extract the gas resource. During the “boom” period,there may be some local population growth, as well as increased hiring by the construction, retailand services sectors. Local business income, tax revenues, and royalty payments to owners ofmineral rights also increase dramatically during the “boom” phase of the cycle; costs tocommunities can rise significantly at this time, for everything from road maintenance to publicsafety to schools. When drilling declines or ceases entirely as the commercially recoverableresource is depleted, the cycle enters the “bust” phase in which population and jobs may quicklydepart the area—leaving fewer people to support the boomtown infrastructure. Communitieswhere drilling-related benefits have effectively ended continue to be affected by a legacy ofregional industrialization (e.g., truck traffic, gas storage facilities, compressor plants, andpipelines) and the impacts that are attendant thereto. Effective planning by local and stategovernment that moderates the rate of MSGD in a region may mitigate the negative effects of theboom-bust cycle to a considerable degree (Christopherson and Rightor 2011).It is inevitable that there will be environmental impacts from MSGD in western Marylandthroughout the “boom-bust” cycle (and perhaps beyond) and that a significant portion of these“costs” will be borne by local communities. Heavy truck traffic on local roads, noise and odorsemanating from drilling sites, conflicts with outdoor recreation, diminished tourism, reducedbiodiversity, and deterioration of air and water quality are some examples of the types of impactsthat are likely even under the best of circumstances. While difficult to quantify in economicterms, these “costs” will ideally be greatly outweighed by the benefits of increased economicactivity from the “boom-bust” cycle—otherwise it is very difficult to make a case that MSGDshould occur at all. These impacts (“externalities”, in economic terms) must be expected even ifbest practices are implemented, local ordinances and state gas development regulations arecarefully revised, and high standards of enforcement and inspection are put in place. Since theseimpacts are difficult to quantify in economic terms, we have explicitly chosen to identify andrecommend specific BMPs that—largely on the basis of our professional judgment—wouldprovide as much protection of Maryland’s natural, cultural, historical and recreational resources;the environment; and public safety as can reasonably be provided while allowing MSGD tomove forward. The hope is that through implementation of these BMPs many of the mostegregious environmental impacts can be prevented s (i.e., allowing the external costs toeffectively be “internalized”). 1-2

19. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandWe have also concluded from our review and from a simple geographic observation thatMaryland is definitely not in control of its own environmental destiny when it comes toMarcellus shale gas development. The fact of the matter is that air and water pollutants (andeven highway vehicles and U.S. dollars) are not observant of state boundaries. Since westernMaryland (just two counties: Garrett and Allegany) is a relatively small “panhandle” sandwichedbetween Pennsylvania and West Virginia, in essence it cannot be truly isolated from activities inthese and other states (e.g., some surface waters that originate in other states flow throughMaryland; emissions of air pollutants from other states impact Maryland air quality; traffic, thehuman environment, and the economics of small towns in western Maryland are not immunefrom what is occurring in neighboring states). This also means that even if Maryland were todecide not to permit MSGD, there will no doubt be impacts felt in Maryland (both positive andnegative) attributable to development of the resource in neighboring states that would mostly bebeyond Maryland’s ability to control.Finally, we should note that the federal government has not played a major role in regulatingunconventional gas development in Appalachia or elsewhere. There are several examples wherefederal statutes explicitly exempt unconventional gas development from federal environmentalregulation. In particular, we note that oil and gas wastes are exempt from hazardous wasteprovisions of the Resource Conservation and Recovery Act (RCRA)—based on a determinationby the U.S. Environmental Protection Agency (USEPA) that existing federal and stateregulations were adequate to manage these wastes and apply RCRA Subtitle C regulation tothese wastes would impose excessive costs on the energy industry (Hammer et al. 2012).Therefore, natural gas operators along with companies hauling or receiving these wastes aredoing so without any requirement to meet the “cradle to grave” safeguards established underRCRA. An amendment to the Safe Drinking Water Act of 2005 excluded hydraulic fracturingactivities under the definition of “underground injection” (with an exception made for fracturingfluid containing diesel fuel). Oil and gas operations are also exempt from NPDES stormwaterpermitting requirements under the Clean Water Act (Hammer et al. 2012). USEPA recentlydeveloped a federal rule mandating a BMP known as “green completion” as a way of capturingmethane gas and reducing emissions of volatile organic compounds (VOCs) during thecompletion process—a practice that has been effectively used in Colorado and Fort Worth, Texasfor several years.Implementation of BMPs for Marcellus shale gas development in Maryland must begin well inadvance of actual exploration, site development, and drilling to properly address a variety ofissues related to environmental assessment, planning, permitting, and bonding. For purposes ofthis report, we have explicitly defined the term ‘BMP’ in the most general way here to includevirtually all aspects of shale gas development (USDOE 2011). Also, while we have focused ourreport on Marcellus shale gas development, our recommendations are likely applicable tounconventional development of other shale formations such as the Utica as well. In this chapter,we make specific recommendations of some critical actions that must be taken if MSGD is goingto go forward in Maryland in as safe a way as possible.A. Pre-development environmental assessmentPre-development environmental assessment for MSGD should be used to identify (1) specificenvironmental conditions or features that would be expected to affect development of a 1-3

20. K.N. Eshleman & A.J. Elmore (2013) Chapter 1particular site or region and (2) the environmental resources that are likely at risk from any futuredevelopment activities. The ultimate goal of the assessment is to prevent conductingdevelopment activities that would cause temporary or lasting environmental damage fromMSGD. It has been proposed both in New York (NYSDEC 2011) and Pennsylvania (MarcellusShale Advisory Commission 2011) that state regulators of MSGD develop an environmentalassessment form or “checklist” as part of the permit application process that would be used to:(1) identify the environmental resources (e.g., areas with high ecological value, exceptional valuewaters, etc.) or features that would be relevant to developing a particular site; (2) identify theappropriate setbacks or restrictions that would control development of a particular site; and (3)determine the environmental assessment activities or baseline monitoring that would benecessary for development to go forward. In Ohio, the Department of Mineral ResourcesManagement (DMRM) conducts a site review prior to issuing a permit to evaluate any site-specific conditions that might be attached to a permit to drill in an urban area (Ohio LegislativeService Commission 2010); Colorado Oil and Gas Conservation Commission (COGCC)maintains a website with maps of “Sensitive Wildlife Habitat” and “Restricted SurfaceOccupancy” areas that operators can use to determine whether a proposed oil or gas drilling sitefalls within such an area5. Maryland regulations governing oil and gas development require areasonably extensive environmental assessment6, although it doesn’t appear to require anybaseline monitoring activities as part of the process.Pre-development activities are essential to ensuring that MSGD in Maryland is conducted assafely as possible; some of these activities can, at least in part, be based on digital maps of themost sensitive ecological resources and those habitats in greatest need of protection (seeChapters 5 and 6). These maps are a product of the state’s long-term investments inenvironmental monitoring and resource assessment [e.g., Maryland Synoptic Stream ChemistrySurvey (MSSCS); Maryland Biological Stream Survey (MBSS); etc.] and should be used as suchby making them available to the public and to the industry at a dedicated website. Once thesedata layers are made available, a prospective shale gas developer—prior to submitting a drillingapplication or comprehensive drilling plan for review and approval—should be required toconsult maps of (1) irreplaceable natural areas, (2) Maryland stronghold watersheds, (3)Maryland brook trout streams, (4) Tier II streams and drainages, (5) the entire stream network,and (6) other priority conservation areas to determine whether a proposed shale gas developmentwould fall within an area that contains any “high-value assets”. Such an exercise would furtherallow a prospective operator to quickly determine the applicable setbacks and other BMPsgoverning MSGD at a proposed site—thus saving considerable time and money during theplanning stages of a particular project.Given the relatively high density of sensitive ecological, recreational, historical, and culturalresources in western Maryland and a legacy of underground coal mining in the region, pre-development environmental assessment should be conducted on a site-specific basis and include:(1) identification of all on-site drilling hazards such as underground mine workings (both active5 COGCC Rule 1201, Identification of wildlife species and habitats6 COMAR 26.19.01.06.C(3); see also www.mde.state.md.us/programs/Land/SolidWaste/ApplicationsFormsandInstructions/Documents/www.mde.state. md.us/assets/document/permit/MDE-LMA-PER066.pdf for more details (webpage accessed February 6, 2013) 1-4

21. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandand abandoned), orphaned gas or oil wells, caves, caverns, Karst features, etc.; (2) identificationof all ecological, recreational, historical, and cultural resources in the vicinity of a proposed site(includes well pad and all ancillary development such as cleared areas around a well pad, roads,bridges, culverts, compressor stations, pipelines, etc.); (3) identification of all appropriatesetbacks and buffers for the proposed site; and (4) collection of two years of pre-developmentbaseline data on underground sources of drinking water, downstream surface water, and bothaquatic and terrestrial ecological resources. Several of these aspects of environmental assessmentare already required under Maryland’s existing oil and gas regulations, but other elements willneed to be added.7 Additional details on on-site and off-site monitoring to address MSGDimpacts are provided in the following section.B. On-site and off-site monitoringOn-site and off-site monitoring is an important aspect of MSGD that has not yet received theattention that the subject deserves. Environmental monitoring in the context of MSGD could playone or several important and legitimate roles, although generic monitoring would be unlikely toserve any particular purpose (except the purpose of making the citizenry of the state feel thatresources are being adequately protected because they are being “monitored”). Too often,monitoring systems are put in place at great expense without carefully considering howmonitoring data would actually be used. Depending on the specific types and ways that data arecollected, monitoring can clearly address: • environmental regulation (ensuring compliance with or documenting violations of standards and regulations); • environmental remediation (establishing a benchmark for assessing damages and performing reclamation or restoration activities); • environmental science (increasing process-level understanding, especially when combined with research); and • environmental control (detecting problems and providing feedback to the process of defining best management practices)There are virtually no comprehensive, carefully-designed, experimental studies of the impacts ofMSGD on environmental resources that have been published in the literature, so scientificobservations of actual impacts (or no impacts) associated with MSGD through case studies couldplay an important role in gaining process-level understanding (USEPA 2011)8. To date, mostmonitoring efforts have been associated with obtaining baseline water quality data from nearbygroundwater wells that could be used to assess future damages from development activities,particularly hydraulic fracturing. Given that the risks to surface water quality from chemical orwastewater spills or releases are considered at least as great as those to groundwater, greaterattention should be paid to benchmarking surface water quality (and continued monitoring toenable detection of water quality deterioration). However, almost no attention has been paid to7 COMAR 26.19.018 The U.S. Environmental Protection Agency’s ongoing study of the possible impacts of hydraulic fracturing comes closest to a systematic study, but it is addressing a limited number of possible impacts (drinking water resources) and has not been completed or published. Sadly, the recently released progress report in December 2012 (USEPA 2012) described a series of case studies in which many of the empirical data that would be used to test and parameterize impact models were collected after MSGD had already occurred (i.e., little or no pre-development data are available). 1-5

22. K.N. Eshleman & A.J. Elmore (2013) Chapter 1the use of monitoring data in improving best practices for shale gas development (USDOE2011)9. In fact, many of the BMPs that we have identified in this report are based primarily onprofessional judgment rather than on systematic experimental testing with replication under avariety of field conditions. Our review revealed that relatively little monitoring has been done toestablish baseline resource conditions prior to MSGD and subsequent monitoring of impacts maybe only marginally useful. The best example of monitoring that we found is the program beingdeveloped and implemented by Pennsylvania Department of Conservation and NaturalResources (PADCNR) to address impacts of MSGD in the Pennsylvania state forests (PADCNR2011). While this program certainly has some significant merits relative to what is being doneelsewhere, it is obvious that MSGD was well underway before this program was ever fullyimplemented (in fact, it has still not been fully implemented even today).Most of the baseline data that are presently being collected are for groundwater wells within adefined radius of a proposed gas well primarily to provide a benchmark for assessing damages(or as defense from presumed liability in the event that contamination is detected in the future).In Pennsylvania, for example, private water wells located within 1,000 ft of a proposed gas wellare tested before drilling as part of the permitting process. Well monitoring in Pennsylvaniashowed post-drilling increases in bromide (Br) concentrations, suggesting that 3,000 ft is a morereasonable distance than the 1,000 ft that is currently required for both presumed responsibilityand certified mail notification related to Marcellus gas well drilling (Boyer et al. 2011). Few, ifany, hydrogeologists would disagree with the conclusion that sampling water wells within a3,000 ft radius of a gas well is a pretty marginal groundwater monitoring program if the intent isto be able to detect a subsurface contaminant plume associated with a particular well integrityissue (especially in rural areas where the number of water wells may be very low or zero).Other resources that could be impacted by development of a particular site (e.g., surface waterquantity and quality, air quality, forest interior bird populations, etc.) have received even lessattention, however. It is, therefore, recommended that Maryland require as part of its permitapplication at least two years of site-specific data collection prior to any site development thatwould be used to characterize the resources at risk and provide a solid baseline dataset thatwould ultimately be used to understand process and feedback useful information for refinementof BMPs. These data should be collected at operator’s expense and reported to MarylandDepartment of the Environment (MDE) as part of the permit application process. Althoughproviding a detailed site-specific monitoring plan for MSGD is well beyond the scope of thisproject, we can provide some rough guidelines for what might constitute a realistic plan: (1) themonitoring system should be designed in a way that characterizes the extent of any site-specificimpacts on- and off-site (e.g., downstream of a particular well pad; groundwater well sampling atleast to the periphery of the area defined by the lateral boreholes); and (2) frequency of datacollection should be adequate to quantify natural variability of conditions (e.g., monthlysampling of surface water may be appropriate, but annual sampling of groundwater quality maysuffice). A draft plan that we obtained from Maryland DNR contains many of the elements that asolid, site-specific water and macro-invertebrate monitoring plan would likely include (Klauda et9 The report explains that developing reliable metrics for best practices for shale gas development is a major ongoing task, and further advised that the industry set a goal of “continuous improvement” in best practices that would be “validated by measurement and disclosure of key operating metrics”. Such validation would likely be heavily based on the collection and analysis of on-site and off-site monitoring data of specific parameters. 1-6

23. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandal. 2012). We envision that on-site and off-site monitoring would be continued through the life ofthe project as a means of assessing impacts, improving BMPs, and providing some process-levelunderstanding of how resources are being affected.Regional monitoring of environmental resources by the state is also recommended. In particular,both air quality and water quality may be impacted by cumulative MSGD over the entire region(or within a portion of the region), so a monitoring network will need to be established to addresscumulative impacts both before and after development begins. As examples, the proposals tosample methane (and other constituents) in a sample of drinking water wells in westernMaryland is an excellent idea that should be funded; comparable surveys of surface water qualityin specific western Maryland watersheds that are likely to experience MSGD would be equallyuseful in establishing a regional baseline. Finally, air quality impacts are likely to occur at theregional scale, so MDE should ensure the one existing air quality monitoring station in the regionis equipped with instrumentation to address primary MSGD impacts (e.g., NOx, VOC, and fineparticulate concentrations). While the design and implementation of this monitoring network iscrucial, it may not be necessary to build such a system from scratch. Many of the monitoringcomponents can probably be piggy-backed onto existing monitoring and resource assessmentactivities (e.g., MBSS) that the state is presently conducting for other purposes.C. Comprehensive drilling plans (CDP)One way of attempting to minimize some of the most significant negative impacts associatedwith developing gas resources within an area (and possibly moderating the rate at which theresource is developed) is through a process known as comprehensive planning. It is thought thatby carefully mapping the “constraints” on gas development presented by a variety ofenvironmental and socioeconomic factors and also identifying the foreseeable oil and gasactivities in a defined geographic area upfront, energy companies working cooperatively withother stakeholders (including state natural resource agencies) can come up with an integratedplan for efficiently exploiting the resource while minimizing impacts on local communities,ecosystems, and other natural resources. Under a COGCC rule10, gas operators in Colorado havethe option of proposing a Comprehensive Drilling Plan (CDP)11 that covers multiple drillinglocations within an area as a way of addressing some of these constraints; while voluntary, CDPsare definitely encouraged in Colorado and it has been concluded that the process would workbetter if operators would work together to develop a joint CDP to cover proposed activities ofmultiple operators where appropriate. Presently, one major operator is in the process ofdeveloping a CDP for 11 well pads and 200 gas wells in the Battlement Mesa area in GarfieldCounty, CO—a community that is home for about 5,000 residents12. Given the fact that westernMaryland is a largely intact landscape with areas of high terrestrial and aquatic biodiversity andknown surface resources that are in need of special protection, a comprehensive gas developmentplan makes a lot of sense. Comprehensive planning could potentially be used to effectively10 Rule 216, 2 CCR 404-1 Practice and Procedure11 The term comprehensive drilling plan (CDP) is actually somewhat of a misnomer. A better term would be “comprehensive gas development plan” because it would logically include all aspects of the activity (i.e., constraints mapping for resource protection, exploration, environmental monitoring, drilling/fracking, gas transmission, transportation, planned mitigation of impacts, etc.).12 See http://www.oilandgasbmps.org/resources/casestudies/battlementmesa.php (webpage accessed February 6, 2013) 1-7

24. K.N. Eshle eman & A.J. Elmore (2013) E Ch hapter 1channel MSGD into areas that would be less disruptive f western M M for Maryland ressidents andvisitors and less sens a sitive to impa while al acts llowing for c considerable and efficien exploitatio of e nt onthe gas re esource. On way that th might be done effect ne his e tively is by p evelopment in permitting dedensely “clustered” well pads in areas where sensitive re “ w e esources (and communiti can be m d ies) moreeasily avoided (e.g., see Figure 1-1).Figure 1-1 Idealized sch 1. hematic (plan view) showing a “clustered” Marcellus sha gas development area v g alecomprised of nine multi-w pads (soli green boxes) each pad wit six 8,000 ft l well id ), th laterals per pad (solid black l d lines)draining ab bout two squar miles of the target formatio (solid tan re re t on ectangles). It is estimated tha the total area of s at athe well pa in this exam is 36 acre (4 acres per pad) plus 44 ac for ancilla facilities (ac ads mple es p cres ary ccess road and co- dlocated pip pelines and util lities, solid cha artreuse line). Total land area disturbed is le than 1% of the total area T a ess fdrained. With respect to setback requir W rements, some setbacks shoul be measured from the indiv ld d vidual well pad (or dsdisturbed areas for each pad), while oth would be measured horiz a p hers m zontally from th farthest exte of hydrauli he ent icfracturing. This idealized example obv d viously represen a “best-case scenario” ina nts e asmuch as loca topography, alstreams, riv vers, and structures would na aturally require somewhat mo surface dist e ore turbance per un resource nitdeveloped. .Since Ma aryland has little recent experience permitting an regulating natural gas developme in l e p nd g s entthe state (and no experience with modern hig volume hy h gh ydraulic fraccturing), the state mightconsider putting in pllace regulati ions to suppo a volunta compreh ort ary, hensive gas d developmen ntplanning approach in western Ma n aryland that could effecttively allow MSGD to m move forward at a dsensible, manageable rate. We en e nvision a vol luntary prog gram similar to Colorado approach but o’s h,one that could also pr c rovide stron incentives for operator willing to consider thi option. O ng s rs o is Oneway the state might incentivize CDPs could be by reduci permit fe and bond s C b ing ees dingrequiremments for wel covered under a CDP. Since the ti to devel and have approved a lls u ime lop eCDP wou likely be longer than for a single well drillin permit, ov uld e n e ng verall MSGD in western DMaryland might be significantly slowed (thu avoiding s d us some of the “ “boom-bust” issues ”discussed earlier). Lo d ogically, the first approv CDP wou result in permitting a area for ved uld andrilling where the mo sensitive resources would be less of an issue. Over time, monitoring data w ost w scollected on-site, off- d f-site, and thrroughout we estern Marylland (see Sec ction B) wou documen uld ntimpacts (or non-impa ( acts) and wo ould be used by the indus to impro BMPs. Additional C stry ove CDPswould pr resumably be dependent upon the ind e dustry demoonstrating tha any impac from earl at cts lierdrilling were within acceptable li w a imits or that newer practtices were si ignificantly bbetter at reduucingany unac cceptable imppacts observ in prior phases. The p ved p phasing of M MSGD in thi way would is dalso allow the regulatory enforce w ement arm of MDE to “r f ramp up” as developmen proceeds— nt — 1-8

25. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandgradually developing the appropriate experience over time that plays and essential role inensuring that development is conducted as safely as possible. In Pennsylvania, for example, ittook several years to staff a regulatory program within the Department of EnvironmentalProtection (PADEP) and PADCNR to effectively address MSGD on both private and publiclands. As of summer 2012, PADCNR alone had a 50 person “gas management team” that isresponsible for managing the program in the state forests13, in addition to the large number ofinspectors in PA DEP that enforces permit conditions throughout the state. It is generallyaccepted that many of the problems that have occurred in Pennsylvania and elsewhere can beexplained in part by the excessively rapid rate of MSGD before the necessary regulatorystructure had been put in place.Comprehensive drilling plans are also being used in Pennsylvania state forests and have beenproposed for private lands in the state (Lien and Manner 2010); these plans involve significantinteractions (“give and take”) between the energy companies, state regulators, local authorities,and the public at large to get all of the various stakeholders on the same page. Through this giveand take process, gas development infrastructure should be planned for in advance, even if fullimplementation ultimately takes many years. While we favor this approach in general, we havesome reservations as to whether Maryland’s regulatory structure and culture are sufficientlyflexible to enable such an approach to be effectively implemented.Another major impediment to comprehensive gas development planning in Maryland is that thestate lacks the power to do “forced pooling” (or “compulsory integration” or “unitization”)14.With forced pooling, a gas company could force one or more entities with ownership of themineral rights of some portion(s) of a gas “unit” into a lease in order to enable more efficientexploitation of the resource (perhaps while providing greater protection of some specific surfaceresources overlying a portion of the unit). The practice of forced pooling is controversial and hasbeen considered an infringement of property rights (the current Governor of Pennsylvania hascalled forced pooling “private eminent domain”15). Thirty-nine states have some type of forcedpooling law, but Maryland does not. This power is particularly important given the practice ofhorizontal drilling, since the technology itself makes it possible to capture gas thousands of feet(horizontal direction) from a wellhead (e.g., gas resources underlying sensitive surface resourcesthat would otherwise be impossible to extract without causing undesirable disturbances).Drilling companies have argued that forced pooling effectively enables more gas to be extractedfrom fewer well pads—thus reducing costs and environmental impacts. Without the power toenforce a forced pooling arrangement proposed by a drilling company, however, Marylandeffectively lacks a planning tool that could be used to provide greater resource protection whileallowing for efficient resource exploitation. It is not clear to us whether forced pooling would beacceptable in Maryland, given the state’s legal approach to mineral rights; nonetheless, it is atopic that should be further examined.13 Ibid., 514 Brigid Kenney (Maryland Department of the Environment), personal communication (December 3, 2012).15 Reported by ProPublica: http://www.propublica.org/article/forced-pooling-when-landowners-cant-say-no-to- drilling (webpage accessed December 3, 2012) 1-9

26. K.N. Eshleman & A.J. Elmore (2013) Chapter 1D. Well pad spacingOur research suggests that modern horizontal drilling and hydraulic fracturing from multi-wellpads are presently capable of draining at least one or perhaps as much as two square miles of thetarget formation (see hypothetical example in Figure 1-1)—thus enabling the siting of well padsat locations that can avoid sensitive resources and greatly minimize disturbances and associatedimpacts on both terrestrial and downstream aquatic ecosystems from development. Spacingmulti-well pads in dense clusters—with well pads located as far apart as is technically feasible—makes maximum use of horizontal drilling technology and could be an important BMP in termsof minimizing development impacts. Figure 1-2 shows an air photo of such a multi-well, multi-pad development in Pennsylvania, illustrating how the extent of surface disturbance can beminimized using this BMP. Further, our analysis suggests that—with careful and thoughtfulplanning (e.g., co-location of infrastructure wherever possible)—it may be possible to developmuch of the gas resource in a way that disturbs less than 1-2% of the land surface, even whenaccounting for the need for ancillary infrastructure such as access roads, pipelines, andcompressor facilities. While this may be a “best case” scenario and there is probably no definablethreshold of land disturbance below which zero impacts would be expected, it should beemphasized that disturbances of 1-2% of the land surface are quite low compared to other typesof development (e.g., suburban residential, surface mining, etc.).Figure 1-2. Air photo showing a densely-clustered well pad development in Pennsylvania. Drilled Marcellus gaswells are identified as red dots. Screen shot from web-based map viewer at http://maps.tnc.org/paenergy/ developedby The Nature Conservancy (website accessed February 8, 2013). 1-10

27. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandWe also note that clustered well pad development can only be expected to reduce surfaceimpacts if operators are held to reasonable spacing dimensions over time. In the Pennsylvaniastate forests, operators have agreed to drill wells as reasonably prudent as possible—although notall leases had well spacing limitations. Newer leases hold operators to a maximum number ofwell pad locations, or total disturbance of a predefined acreage, whichever occurs first. In theseleases, if an operator wishes to deviate from the well pad numbers or acreage, a waiver and stateforest approval is required (PADCNR 2011). Minimizing the number and density of well padsthrough coordinated planning and consultation (i.e., a CDP), as well as utilization of existingrights of way, can greatly mitigate the cumulative impacts on the landscape (Marcellus ShaleAdvisory Commission 2011). Given that the well pad and ancillary infrastructure will likely bein place for at least a 30-year period before final reclamation can be completed, we recommendthat Maryland guard against any tendency for infilling (i.e., drilling from new pads that expandsthe density of the surface infrastructure within an area) by incentivizing drilling of any new wellsfrom existing pads once these are permitted. Our concern here is for minimizing cumulativeimpacts that may likely prove to be a function of the total amount of surface development withinan area.E. Setback requirementsSetbacks are a primary tool by which regulatory agencies can restrict shale gas development inan effort to provide some additional protection of the most sensitive ecological resources, waterresources, personal property, public property, and the health and safety of the public at largeparticularly in the event of an accident (e.g., pollutant spill, blow-out, etc.) during the conduct ofshale gas development operations. How much protection (if any) these setbacks can provide canclearly be debated; many setbacks do not seem to be based on solid scientific reasoning orempirical data. Nevertheless, both industry and the state benefit when setbacks are clearly statedin statutes or regulations. Setbacks that are vague or that depend on subjective site analysisintroduce uncertainty into the decision-making process, leading to hidden costs (redundantanalyses at best and legal fees at worst). Setbacks can sometimes be voided if landownerpermission is obtained (e.g., setbacks from property lines), however they are sometimes used toprotect the rights of other leaseholders. Variances from setback requirements can also be grantedby regulatory authorities (typically if operators propose more stringent protective drilling and/oroperational practices). It should be noted that the efficacy of setbacks in providing protection forstreams may be especially questionable, given the fact that the network of “blue-line” streamsthat appears on 7½ minute topographic maps may significantly underestimate the surface waterresources at risk, especially small streams (Elmore et al. in review).Table 1-1 provides a summary of the recommended setbacks to provide protection of specificresources in Maryland, with justification and explanation following in the appropriate chapters ofthe report: special siting criteria (Chapter 1); water resources (Chapter 4); terrestrial habitat andwildlife (Chapter 5); aquatic habitat and wildlife (Chapter 6); public safety (Chapter 7); culturaland historic values (Chapter 8); quality of life and aesthetics (Chapter 9); and agriculture andgrazing (Chapter 10). In each case, wherever two or more setbacks apply, the most restrictivesetback would take precedence. 1-11

28. K.N. Eshleman & A.J. Elmore (2013) Chapter 1 Table 1-1. Summary of recommended setbacks for resource protection and public safety. From To Distance Chapter Aquatic habitat (defined as all streams, Edge of drill pad 300 ft Chapter 5 and 6 rivers, seeps, springs, wetlands, lakes, disturbance ponds, reservoirs, and floodplains) Special conservation areas (e.g., Edge of drill pad 600 ft Chapter 5 irreplaceable natural areas, wildlands) disturbance All cultural and historical sites, state and Edge of drill pad 300 ft Chapter 8 federal parks, trails, wildlife disturbance management areas, scenic and wild rivers, and scenic byways Mapped limestone outcrops or known Borehole 1,000 ft Chapter 1 and 5 caves Mapped underground coal mines Borehole 1,000 ft Chapter 1 and 3 Historic gas wells Any portion of the 1,320 ft Chapter 1 and 3 borehole, including laterals Any occupied building Compressor stations 1,000 ft Chapter 9 Any occupied building Borehole 1,000 ft Chapter 9 Private groundwater wells Borehole 500 ft Chapter 4 Public groundwater wells or surface Borehole 2,000 ft Chapter 4 water intakesF. Identification of freshwater aquifers and groundwater flowpathsDrilling for gas in the Marcellus shale formation (located 0 to 9,000 ft below the surface inwestern Maryland) will obviously require that operators drill vertical boreholes through thefreshwater zone. Many western Maryland residents are dependent on groundwater for theirdrinking water16—underlining a critical need to identify and understand the hydrogeologicalsetting and dynamics of the principal aquifers underlying this region prior to MSGD so that safedrilling practices that are protective of these systems can be implemented. The USGS reportedthat there are currently ten permitted water wells (mostly public supply wells) in the Deep Creekwatershed in Garrett County with a reported combined average annual withdrawal of 0.28 MGDin 2007, plus an additional 2,900 permit-exempted individual wells with an estimated combinedaverage annual withdrawal between 0.43 and 0.87 MGD.17 Withdrawals of groundwater forpublic supply increased by more than 2,000% between 1988 and 2007, reflecting both populationgrowth and expanded public service in the watershed.18 An important issue is the depth that asurface well casing string must be placed and cemented to ensure that the fresh groundwater16 Estimated average daily withdrawals for self-supplied domestic use and public supply distribution in these two counties in 2000 (the most recent year for which data are available) were 2.41 million gallons per day (MGD) of surface water and 4.01 MGD of groundwater. See http://md.water.usgs.gov/freshwater/withdrawals/#ga (webpage accessed December 4, 2012).17 McPherson, W.S. Water use in the Deep Creek Lake watershed, Garrett County, Maryland, 1980-2007; http://md.water.usgs.gov/deepcreek/wateruse/index.html (webpage accessed December 4, 2012).18 Ibid. 1-12

29. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandresources can be adequately protected in the Deep Creek watershed and elsewhere.Subsequently, gas wells will have to be properly cased and cemented to protect freshgroundwater supplies (see Chapter 3). While many neighboring states have mapped the interfacebetween saline and freshwater aquifers, Maryland has not. Water quality data from eightdifferent projects conducted in the Northern Appalachian Coal Basin indicated total dissolvedsolids (TDS) levels between 2,000 and 5,000 milligrams per liter (mg/L) at depths ranging from500 to 1,025 feet below the ground surface—values that are within USEPAs water qualitycriterion (< 10,000 mg/L TDS) for underground sources of drinking water (Zebrowski et al.1991). It was reported that one deep well drilled in southern Garrett County encountered afreshwater/saltwater interface at a depth of 940 feet (Duigon and Smigaj 1985).We were unable to find detailed digital maps of the principal freshwater aquifers of westernMaryland, nor hydrogeological cross-sections or quantitative data that could be used to developflow nets or models to infer groundwater flowpaths and other important features such as rechargeareas, discharge areas, hydrologic residence times, and depth of the freshwater zone across thearea. The best resource appears to be a USGS report that includes a fairly detailed descriptionand map of the principal aquifers of the area, plus some qualitative analysis of groundwaterflowpaths and quality (Trapp, Jr. and Horn 1997). There is a definite need for a comparablehydrogeological analysis focused strictly on western Maryland that could be based onmeasurements of static water levels from domestic and commercial wells, well water qualitydata, and observations reported in well logs; a map of the freshwater/saltwater interface would bea useful product from such an analysis. Provided with adequate resources, the MarylandGeological Survey (MGS) might be the logical group to undertake such an investigation.The principal aquifers of unglaciated western Maryland include: (1) Appalachian Plateauaquifers in Paleozoic sedimentary rocks that are usually flat-lying or gently folded (especiallysandstones of Pennsylvanian and Mississippian age and carbonate rocks of Mississippian age);and (2) Valley and Ridge aquifers that are often heavily folded (Paleozoic fractured sandstonesand limestones are typically the most productive aquifers in these rocks). In the AppalachianPlateau in western Maryland, the principal aquifers have been identified as belonging to theMonongahela Formation, the Conemaugh Formation, the Allegheny-Pottsville Group, the MauchChunk Formation, the Pocono Formation, and the Greenbrier Formation (the latter is a limestoneformation that is only locally water-yielding). The Monongahela, Conemaugh, and AlleghenyFormations each contain multiple seams of mineable coal—most of which are economicallyimportant (Staubitz and Sobashinski 1983). In the Valley and Ridge of western Maryland, theprincipal aquifers are the Hampshire Formation, the Chemung Formation, and the RomneyGroup overlying the Oriskany Sandstone which is commonly saline (Trapp, Jr. and Horn 1997);coal seams are not considered mineable.In the Appalachian Plateau Province it has been suggested that groundwater flow is “step-like”—following vertical pathways through fractures and horizontal pathways through fracturedsandstone aquifers and coal beds. Groundwater recharge in the province is thought to be fairlylow, owing to the relatively steep topography and shallow regolith. The low permeability shalesunderlying the Appalachian Plateau Province function as confining beds that can often give riseto flowing Artesian conditions in wells penetrating aquifers below the shales; this often occurs inwells located in the synclinal valleys of the province. Even in areas where groundwater has not 1-13

30. K.N. Eshleman & A.J. Elmore (2013) Chapter 1been impacted by earlier oil and gas development that began as early as the 19th century, salinewater can be encountered in aquifers within just a few hundred feet (or less in some cases) of theground surface. Hydrogeologists have attributed this situation to (1) the presence of nearly flat-lying low permeability strata of shales, siltstones, clays, and dense limestone; and (2) the lack ofintensely fractured formations that effectively prevent deeper circulation of freshwater to greatdepths below the surface. In other areas, however, shallow groundwater has been contaminatedby brines that flowed upward under pressure through improperly cased or plugged oil and gaswells that penetrated deeper saline aquifers; there are other examples of fresh groundwatercontamination from infiltration and percolation of brines stored in open pits (Trapp, Jr. and Horn1997).In the Valley and Ridge Province, it is thought that groundwater moves mostly along fracturesand bedding planes, and in solution channels within carbonate rocks; the alternately foldedsedimentary rocks, combined with a fluvially-dissected topography, have created a series of localgroundwater flow systems that exist within the upper few hundred feet of the land surface andare effectively isolated from the intermediate and regional flow systems below. Springs(including both gravity springs from unconfined aquifers and Artesian springs from confinedaquifers) are also very common in the Valley and Ridge. Thermal springs are also well known inthe province (e.g., including famous Berkeley Springs, WV and Warm Springs, VA) wheredeeper heated groundwater is effectively channeled back to the surface by folding, faulting, andfracturing of the confined aquifers (Trapp, Jr. and Horn 1997). It appears that these “warm”springs form on the crests or limbs of anticlines, particularly where vertical permeability isenhanced at openings along bedding planes, tension fractures, open faults, or other fracturescommon to folded structures (Hobba, Jr. et al. 1979). Thus, there definitely are hydrogeologicalpathways by which groundwater heated at great depths (i.e., several thousand feet below thesurface) flows upward through fault or fracture systems under pressure and discharges at surfacesprings in the Valley and Ridge Province. White Sulfur Spring and Black Sulfur Spring in GreenRidge State Forest may be less well-known examples of this same phenomenon.A major concern with hydraulic fracturing is that fractures may propagate vertically upwardsinto the freshwater zone (or intersect with natural fractures), thus increasing the communicationbetween the target gas formation and underground sources of drinking water (USDW), thusproviding a mechanism for contamination of drinking water sources. We couldn’t find anyconclusive studies documenting contamination of USDW by this mechanism; one fairlyextensive study investigated hundreds of alleged occurrences of this phenomenon fromhydraulically-fractured coal bed methane (CBM) deposits without drawing much of a conclusion(USEPA 2004). With respect to CBM wells, the consensus seems to be that there are two distinctmechanisms by which groundwater contamination could occur: (1) direct injection of hydraulicfracturing chemicals into a CBM formation that is in direct communication with USDW; or (2)creation of a hydraulic communication between a coal seam and an overlying or underlyingaquifer by breaching a confining layer that provides natural isolation of the CBM from USDW(USEPA 2004).Therefore, at least until it can be shown that hydraulic fracturing can be done safely withinrelatively close proximity to USDW, we recommend that Maryland follow guidance from NewYork’s experience in regulating unconventional shale gas development and effectively not 1-14

31. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandpermit MSGD (or any other unconventional gas development) where the target formation occurswithin 1,000 vertical feet of USDW or within 2,000 vertical feet of the ground surface(NYSDEC 2011). While the stratigraphy of New York’s Marcellus region is certainly notidentical to Maryland’s (the most obvious example is that New York’s landscape was mostlyglaciated during the Pleistocene, effectively removing many of the Pennsylvanian sedimentaryformations from the profile), the basic stratigraphy of the Devonian formations is quite similar(Kostelnik and Carter 2009). Since the freshwater/saltwater interface has not been mapped inMaryland, the prudent approach would be to rely on the 2,000 ft criterion to provide a reasonablemargin of safety. As discussed in Chapter 3, a recent report shows data from several thousandhydraulic fracturing treatments that had been mapped in the Barnett shale in Texas using amicro-seismic method that purportedly indicate that the closest a vertical fracture came to aUSDW was 2,800 ft and the typical distance was nearly 5,000 ft; data from hundreds offracturing treatments in the Marcellus shale displayed in the same report shows a similar result(Fisher 2010). An important limitation of this interpretation is that neither analysis included anymicro-seismic measurements where the target shale formation was within 4,000 ft of the surface.On the other hand, rock mechanics theory suggests that a hydraulic fracture will propagate in adirection that is perpendicular to the least principal stress. In shallow (i.e., < 1,000 ft) formations,the least principal stress is likely the overburden stress—so hydraulic fractures would bepredicted to propagate primarily in the horizontal direction. In deeper reservoirs, however, theleast principal stress would likely be horizontal and hydraulic fractures would thus be expectedto propagate vertically (USEPA 2004). In Chapter 4 we discuss methods by which verticalpropagation of fractures can be estimated.G. Stakeholder engagement (e.g., education, town hall meetings, localcommunity interactions, landowner and lessor protections)Despite recommendations by the American Petroleum Institute (API) that operators proactivelyengage both surface owners and surface users before MSGD operations are initiated to fosterunderstanding and alleviate concerns about hydraulic fracturing and other activities (API 2009),we found very few examples of novel approaches to ensuring that such engagement actuallytakes place. API recommends that operators communicate with land owners and/or surface usersconcerning activities planned for a particular site and provide information on the measures to betaken for safety, protection of the environment, and minimizing impacts to surface uses. Wedefinitely support these recommendations, but feel strongly that this activity should go farbeyond posting a notice in a local newspaper, which may not have the circulation of other media.State agency websites can be informative, but better approaches to stakeholder engagementwould be through public forums or perhaps even via social media (e.g., Facebook). The goals ofany interactions should be for transparency and increasing the flow of timely and relevantinformation to surface owners, users, and other stakeholders. As recommended for Pennsylvania,Maryland might consider developing a standardized stakeholder process that could beimplemented as part of a comprehensive planning strategy; the goal of such a process would beto engage stakeholders and the community in the most effective ways possible, while allowingthe permit review process to be expedited (Ubinger et al. 2010). 1-15

32. K.N. Eshleman & A.J. Elmore (2013) Chapter 1H. Special siting criteriaThe practice of identifying and using special criteria for siting well pads within a region is basedon the idea that some natural (e.g., geological, hydrological, etc.) or man-made (e.g.,underground mining) factors may significantly increase specific risks associated with shale gasdevelopment; efforts could be made to identify such criteria prior to developing the gas resourceso that operators and regulatory authorities can make use of such information during the permitapplication and review process. One such restriction is the topography; well pads should be sitedon land with a slope of <15%. Much of western Maryland exhibits steep and rugged terrain thatexceeds this slope recommendation (Figure 1-3). Steep topography increases risks associatedwith spills, sediment and erosion pollution, and natural hazards (landslides). While increasing thecost associated with MSGD, some operators might be tempted to drill on steep slopes. However,in conversations with industry representatives, we learned that most MSGD operators avoidslopes greater than 15%, likely making this recommendation moot19.Highly permeable subsurface zones—including both natural subsurface reservoirs (e.g., caves,caverns, and fractures) and man-made voids (e.g., underground mines and abandoned wells)—can provide preferential pathways by which aqueous or gas-phase contaminants could rapidlymigrate away from a site in the event of a casing or cementing failure. Moreover, such voidspresent technical challenges and safety issues both in drilling (i.e., lost circulation of drillingfluids that could cause borehole collapse), in maintaining well control (i.e., avoiding a blowout),and in ensuring well integrity during and following cementing operations (Abbas et al.2003/2004). Voids are very commonly encountered when drilling in southwesternPennsylvania20, necessitating the use of remedial cementing (i.e., cement “squeezes” from thesurface, rather than normal cementing in which cement is pumped under pressure down thecasing and back up to the surface through the annular space, displacing non-cement fluids andestablishing a bond with the casing and the borehole wall). Cement squeezes (i.e., grouting of theannular space by pouring cement from the surface) are very time-consuming, expensive, and—most importantly—have a very low success rate and can leave a portion of the surface casingstring unprotected from corrosive fluids (Abbas et al. 2003/2004).To address these issues, the recently enacted Horizontal Well Act (H.B. 401) required theSecretary of the West Virginia Department of Environmental Protection to:propose emergency and legislative rules pertaining to drilling in karst formations, establish designatedgeographic regions of the state where these provisions of the act are applicable, and establish standardsfor drilling horizontal wells in naturally occurring karst terrain. Drilling horizontal wells in naturallyoccurring karst terrain may require precautions not necessary in other parts of the state; such additionalsafeguards may include changing proposed well locations to avoid damage to water resources, specialcasing programs, and additional or special review of drilling procedures. At a minimum, the act requiresoperators to perform certain predrilling testing to identify the location of caves and other voids, faultsand relevant features in the strata and the location of surface features prevalent in naturally occurringkarst terrain such as sink holes; and provide any other requirements deemed necessary by the secretary19 Scott Rotruck, Chesapeake Energy, personal communication (April 27, 2012)20 Archie Miller (Chevron Appalachia), personal communication (July 20, 2012) 1-16

33. Recomm mended Best Manageme Practices for Marcel Shale Ga Developm t ent s llus as ment in Mary ylandto protect the unique ch t haracteristics of naturally occurring ka terrain in s arst ncluding base eline water tes sting 21within an established distance from a drilling site d e.Figure 1-3 Topographic slope is a spec siting crite 3. c cial erion because it influences the effectiveness of sediment a t e s anderosion con ntrol plans and BMPs designed to protect public safety. A BMP would b restricting w pad sites to d be well olocations with a slope < 15%. w 1While rel latively little work has been done to identify cav in wester Maryland a systematic e b ves rn d,analysis and descript a tion of know caves was completed by Maryland Geologica Survey in t wn s d al the1970s. As in many ot A ther environ nments, most of the know caves in w t wn western Mar ryland areconsidere solution caves and are associated with either the Tonolow and Held ed c e way derbergLimeston in Allega County (east of Cum nes any ( mberland) or the Greenbr Limesto in Allega r rier one anyCounty (west of Cum mberland) an in Garrett County. In A nd Allegany Co ounty, in parrticular, thedistributi patterns form lines (o ion oriented sou uthwest to no ortheast) that are parallel to the folds in t l sthe regional structure of the Ridg and Valley province. A band of G e ge y Greenbrier Li imestone cro opsout along the eastern flank of Da Mountain west of Cu g ans n umberland; n caves hav been repor no ve rtedin this area, probably due to the rugged and remote terrai Similarly only a few caves (e.g., y r r in. y, w ,Crabtree and John Fr riends) are known to exist in Garrett County, but many more large caves k t e slikely exi the relati sparsene of popula ist; ive ess ation and roa is primar responsib for our ads rily bleignorance of caves in this area (F n Franz and Sli 1976). A ifer Appalachian caves—incl luding those in ewestern Maryland—o M often contain unique cav n ve-dwelling species (including some that rare an e ndendanger red) that wou best be avoided.22 uld a21 West Vir rginia Horizon Well Act (H 401); ntal H.B.http://www w.legis.state.wv v.us/Bill_Statu us/bills_text.cfm m?billdoc=hb4 401%20enr.htm m&yr=2011&se esstype=4X&i i=40122 Dan Fell (Maryland DNR), persona communication. ler D al 1-17

34. K.N. Eshleman & A.J. Elmore (2013) Chapter 1At the time of this report, spatial locations for 33 caves in Garrett and Allegany counties wereavailable. These locations are aligned along relatively narrow bands in the vicinity of wherelimestone units crop out at the surface (Figure 1-4) with two locations falling outside of a 1000 ftbuffer surrounding outcropping limestone units. Because cave systems are inherently difficult tofind and map, estimates are that only 10% of caves in western Maryland are known23. Drilling inthe vicinity (i.e., within 1,000 ft) of outcropping limestone should be considered a serious riskthroughout the entire extent of limestone in western Maryland24. However, due to horizontaldrilling techniques it should be possible to avoid drilling through these highly permeableformations and voids (both mapped and unmapped)—thus avoiding some of the most seriousrisks associated with poor well cementing that compromises well isolation and integrity.Consistent with the suggested requirements under H.B. 401, an obvious best practice forMaryland would be to site well pads so as to avoid vertical drilling (i.e., surface boreholes) inareas where shallow caves and caverns have been mapped or where there is a high probabilitythat such systems might be present. In cases where caves or underground voids are unexpectedlyencountered during drilling, the technical approaches outlined by Abbas et al. (2003-04) toensure well isolation should be carefully applied. The technical capability to drill horizontallateral wells many thousands of feet long under such cave systems may allow the shale gasresource to be extracted in a way that significantly minimizes the risks described above.We have comparable concerns about MSGD in areas with extensive underground coal mineworkings (both abandoned and active), gas storage fields (e.g., the Accident Gas Storage Field),and/or existing and/or orphaned oil and gas wells (see Figure 1-5). Western Maryland has a longhistory of underground coal mining in each of five different fields (the Lower Youghioghenyfield, the Upper Youghiogheny field, the Potomac field, the Georges Creek field and theCasselman field) that has left a legacy of underground voids that present real challenges in termsof well isolation. Drilling in the vicinity of active underground coal mines represents an extremehazard for a variety of reasons, most importantly the safety of workers who could be exposed toflammable and poisonous gases released into mine workings during the drilling process. Further,casing through these voids would require cement “squeezes” from the surface that are subject tofailure. Finally, steel casing could be subjected to corrosive acid mine drainage (AMD) thatmight be present in abandoned underground workings, possibly leading to a catastrophic casingfailure over time. For these reasons MSGD should be avoided in areas with known undergroundmine workings; we recommend the same 1,000 ft buffer around known workings to provide theadditional margin of safety that was recommended for drilling in Karst terrain25.23 Ibid.24 Our recommended 1,000 ft setback was based on two observations: (1) the setback would protect known caves in the mapped limestone formations; and (2) since caves in the Greenbrier formation in Garrett County are expected to be confined to a weathered zone within 200 ft of the surface and the dip of these formations is approximately 20° (from the horizontal), a setback on the down-dip side of 550 ft [L = 200 ft/tan(20°)] should be adequate. Since the dip of limestone beds in Allegany County is even steeper, a setback of 550 ft would suffice. The 1,000 ft buffer on both sides provides an additional margin of safety due to uncertainties about the exact location of these outcrops.25 This additional margin of safety is justified, in part, due to the fact that the digital layer of underground mines that was used to create Figure 1-5 is likely to be only 70% complete (Jaron Hawkins, Maryland Bureau of Mines, personal communication, September 21, 2012). 1-18

35. Recomm mended Best Manageme Practices for Marcel Shale Ga Developm t ent s llus as ment in Mary ylandFigure 1-4 Limestone ou 4. utcrops in west tern Maryland exist as narrow bands orient southwest t northeast acr w ted to rossthe region. Mapped caves (not shown) are generally, but not always , located within a 1,000 ft bu s a b n uffer surroundinngthese outcr rops. Rare and endemic speci are at risk of disturbance b visitors to c ies o by caves; therefore specific cave e elocations are considered sensitive inform a mation and are not included i the report. L e in Limestone outc crops are based on a dpreliminary digital geological map obta y ained from USG 26 GS.The Acci ident Gas Storage Field is situated between the C Casselman a Upper Y and Youghioghen nyCoal Fiel covering an area of 34,000 acres; the field w discovere by the firm of Snee an lds g 3 was ed m ndEberly in 1953 and is currently owned by Sp n s o pectra Energy 27 The prin y. ncipal gas foormations tha atprovide gas storage are the Hunte g a erville chert and Oriskan sandstone of the Devo ny e onian series at anaverage local depth of 7,350 ft. An older refe l o A erence indica ated that 18 original producing wells swere rewworked for storage servic 35 additio wells ha been drill by 1969, and that the ce, onal ad led , erewas a pla to increase the total nu an umber of we to 83 for the full dev ells r velopment of the field.28 fPresumab Spectra Energy knows the geog bly, graphic exten of the mod nt dern storage facility, the elocations of individual storage wells, and the locations of any abando s w e f oned or orph haned wellswithin th Accident Field that wo he F ould present additional M MSGD hazar These d could be rds. data eused by regulators to restrict MSGD develop r o pment in the vicinity of t Accident Field. Give the t enhow wide espread thes obstacles are to MSGD in western Maryland, we highly re se a D n ecommend t thatMaryland follow Col d lorado’s regu ulation requiiring identifi fication of all potential co l onduits for f fluidmigration prior to dri n illing, includ ding plugged and abando d oned wells wwithin ¼ mile of propose ed26 Preliminnary integrated geologic map databases for the United Stat t tes: Delaware, Maryland, Ne York, , ew Pennsylvvania, and Virg ginia; backgrou information and documen und n ntation (version 1.0); USGS. n27 http://ww ww.spectraener rgy.com/Opera ations/Overvieww-of-Operationns/We-Store-N Natural-Gas/28 Purgator Jr., B.R. 1969. Design and operation of the Accident gas-storage fi rio, a o t ield. American Petroleum Institute n conference paper. http:/ //www.onepetr ro.org/mslib/se ervlet/onepetro opreview?id=A API-69-136 (weebpage accesse ed Decembe 4, 2012). er 1-19

36. K.N. Eshleman & A.J. Elmore (2013) Chapter 1coal bed methane wells, and gas seeps and springs within two miles of such wells. COGCCmaintains a GIS map system that has a data layer showing bottomhole locations that the staffincludes in their review of historic plugged and abandoned wells within ¼ mile (STRONGER2011). An important best practice will be for Maryland to require setbacks from areas ofprevious deep coal and gas extraction. Maryland should develop a GIS of both active andabandoned oil and gas wells (including gas storage wells) and active and abandoned coal mineworkings prior to permitting any new Marcellus wells. All underground hazards within ¼ mile ofany section of a proposed Marcellus well should be identified as part of the permit reviewprocess. We recommend a 1,320 ft (¼ mile) setback from all historic gas wells.Figure 1-5. Mineral resource extraction in Maryland includes deep coal mines (both active and abandoned) andactive and historic conventional gas wells. Data on abandoned and active deep coal mines, and active and historicalgas wells was provided by Maryland Department of the Environment. The boundary of the Accident Storage Fieldwas digitized from paper maps provided by Spectra Energy of Texas Eastern Transmission, LP. The size of thesymbols representing the locations of gas wells have been adjusted to closely match the recommended setback of1,320 ft provided in Table 1-1.In addition to avoiding underground voids through implementation of these setbacks, Marylandmight also consider mandating the use of surface geophysical techniques (e.g., seismic surveys)or “pilot hole” boring as part of an exploration/drilling hazard assessment program that is aimedat identifying other subsurface MSGD hazards that are not well mapped.I. Reclamation planningAnother very important conclusion from our review of the literature and of activities in otherstates is that for planning purposes, MSGD infrastructure should be considered a quasi- 1-20

37. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandpermanent (i.e., at least 30 years) industrial addition onto a mostly rural Appalachian landscape.We have drawn this conclusion because: a) Marcellus wells are expected to produce for at least30 years; b) it may be possible to refracture these wells in the future to enhance diminishing gasproduction; c) wells on multi-well pads may not be drilled in rapid succession to allowcompanies the ability to determine if additional drilling is warranted and justified financially;and d) established pads and associated infrastructure could possibly support futureunconventional drilling into the Utica formation. In the Tiadaghton State Forest in Pennsylvania,for example, we observed that no permanent site restoration or reclamation has occurred or isplanned (Figure 1-6) despite the fact that the drilling/hydraulic fracturing equipment had allmoved on by the time wellhead gas prices had plummeted to less than $3 per thousand cubic feetin early 2012 (from a peak of nearly $11 in July 2008; see Figure 1-7). The thinking there is thatmany of the most serious impacts (i.e., erosion and stream sedimentation) are associated withearth moving and construction activities, so it makes sense both economically (for the gascompanies) and environmentally (for the state) to maintain the established infrastructure ratherthan imposing conditions that would require multiple reclamation efforts over time at the samesites.The quasi-permanent superposition of this industrial infrastructure onto the landscape andassociated time delays until any permanent site restoration, reclamation, and well plugging takesplace has important implications for how states regulate MSGD now to ensure that liabilities forreclamation and closure are properly addressed by the gas industry. The best practice forMaryland would be to develop regulations that force rapid partial reclamation (includingrevegetating disturbed areas surrounding wells pads, corridors, and ancillary infrastructure) of allland not needed for drilling and production as quickly as possible, while allowing the remainingportion to exist unreclaimed only until such time as drilling is completed, production ends, andfinal reclamation can be performed. We feel strongly that the costs of reclamation should beborne directly by the operators (using resources set aside or accumulated for this specificpurpose)—as opposed to ultimately passing these costs on to Maryland residents in the form offuture tax liabilities and diminished natural and environmental resources (see discussion offinancial assurance in Section J below).J. Well permitting, county and state coordination, and financial assuranceBased on our review of practices in other states, it is obvious that MSGD in Maryland shouldrequire approval of a drilling permit issued by the state that addresses all possible issuesassociated with developing a particular site, drilling and completing a well (or wells), preventingerosion and sedimentation impacts, controlling stormwater pollution, protecting public safety,and responding to emergencies) and a mechanism for providing adequate financial assurance fordecommissioning (plugging a well or wells), site reclamation, and any legacy responsibilitiesassociated with a particular operation. All five states that we reviewed require permits andbonding for drilling gas wells, but the permitting and bonding requirements vary drasticallyamong the different states. We found that Maryland’s current oil and gas regulations governingpermitting for conventional development require many of the elements that would be needed toproperly address MSGD or unconventional development in general including: (1) an 1-21

38. K.N. Eshleman & A.J. Elmore (2013) Chapter 1 Figure 1-6. Marcellus shale natural gas infrastructure in Tiadaghton State Forest near Waterville, Pennsylvania: well pad, multiple producing wells, and produced water tanks (upper left); shallow impoundment for freshwater (lower left); and access road, utility corridor, and compressor station (right). Photos by K.N. Eshleman.environmental assessment; (2) a certificate of liability insurance; (3) a performance bond; (4) acopy of the oil and gas lease; (5) written approval by the local zoning authority that all planningand zoning requirements have been met; (6) an approved erosion and sediment control plan; (7)an approved stormwater management plan; (8) a reclamation plan; (9) a spill prevention, control,and countermeasures plan; and (10) a drilling and operating permit plat.29 Maryland’s currentregulations even allow for directional drilling (although they were clearly not written to addressthe practice of hydraulic fracturing); the current statutes allow for the use of pits for temporarystorage of drilling fluids, but do not explicitly address impacts of water withdrawals orwastewater treatment and disposal issues. The state should consider revising its oil and gaspermitting regulations to explicitly address water withdrawal and storage issues, drilling wasteand wastewater treatment and disposal issues, as well as transportation planning issues.29 COMAR 26.19.01.06 (Drilling and Operating Permit Application Procedures for the Applicant) 1-22

39. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandFigure 1-7. Monthly U.S. natural gas wellhead prices (1975-2012);http://www.eia.gov/dnav/ng/hist/n9190us3m.htm (webpage accessed February 8, 2013).Local zoning could be used to avoid the most problematic conflicts among competing land usesin western Maryland, although zoning in Garrett County is not county-wide; it is restricted to afew municipal zoning districts (e.g., Deep Creek watershed). It does not appear to us that countyzoning ordinances for Allegany County30 have been modified to explicitly address MSGD.Zoning ordinances for Deep Creek watershed in Garrett County surprisingly allow for “drillingfor, or removal or underground storage of natural gas” in all nine zoning subdistricts (subject toprescribed minimum setbacks and regulations of MDE, Maryland Public Service Commission,and Federal Energy Regulatory Commission)31. It is not clear to us that Garrett County hascarefully weighed the impacts of MSGD within its zoning districts, although current restrictionsfor Deep Creek watershed that restrict gas development within 1,000 ft of a property boundaryand within 2,000 ft of the shoreline seem reasonably restrictive.32 Local zoning ordinances forboth counties should be amended to spell out in which zoning districts MSGD would bepermitted as a way of minimizing some of the major conflicts and public safety issues that wehave identified in this report.With respect to performance bonding, Maryland’s requirements under current regulations($100,000 per well or $500,000 blanket bond for all of an applicant’s wells33) are relatively highcompared to other states that we reviewed (e.g., Section 215 of Pennsylvania’s Oil and Gas Actset bonding limits at $2,500 per well or $25,000 for a blanket bond for drilling on private land(Ubinger, et al. 2010), although limits are higher for drilling in Pennsylvania state forests).Performance bonding has been deemed inadequate for providing financial assurance foraddressing decommissioning, site reclamation, and legacy responsibilities associated withMSGD (Mitchell and Casman 2011). Mitchell & Casman (2011) used Pennsylvania’s experience30 Code of Allegany County Maryland, Part 4 Zoning (published November 25, 2002)31 Garrett County, Maryland; Deep Creek Watershed Zoning Ordinance (amended May 25, 2010)32 Ibid.33 COMAR 26.19.01.06.C(5)(a) 1-23

40. K.N. Eshleman & A.J. Elmore (2013) Chapter 1with bonding of surface coal mining sites to speculate what might be expected to occur withMSGD infrastructure in the absence of new regulations: from 1985 to 1999, bonds for surfacemining permits covering about 10% of the total acreage of mineland in the state were forfeited.Since the costs to reclaim this mineland was in most cases higher than the bond amountsforfeited, the costs of bringing these abandoned minelands into compliance are inadequate andthe difference must be made up by the responsible entity (in this case, the state, in some caseswith help from the federal government’s abandoned mineland funds).While Maryland’s performance bonding limits are comparatively high, another concern is thatsteep declines in gas production of these wells in tandem with increasing liabilities fordecommissioning and reclamation may drive divestment of shale gas assets before expectedclosure occurs. The transfer of marginally-productive assets to smaller independent operators oreven to surface owners is a common practice in the oil and gas industry, with the primaryexploration and production companies using these divestments to fund new drilling operations.At least in Pennsylvania (not sure about Maryland), there is no formal regulatory mechanism toprevent transfers of shale gas assets to entities under conditions in which the accumulatedreclamation liabilities dwarf the financial wherewithal of the new asset owners—even thoughthese new owners would also require bonding. In some cases, these firms can obtain suretybonds for only a percentage of a bond’s face value—putting much of the financial obligation onthe backs of banks or surety companies who themselves would be liable for the reclamation costsdown the road. Another problem with surety bonding is that underwriting firms will only marketsuch bonds when the amount and terms of the liability are strictly defined; bonds are thus notwell suited to covering unforeseen liabilities (e.g., legacy issues such as long-term replacementor treatment of a community’s water supply in the event that an existing supply is lost orirrevocably contaminated). Even in the event that the costs are covered prior to release of a bond,environmental problems that arise later would be difficult for individuals, communities, or a stateto address without pursing a civil action (Mitchell and Casman 2011).We believe that Maryland’s relatively high bonding limits on oil and gas well drilling maylargely prevent such divestments from occurring, and may also provide adequate fundingthrough performance bonding to address all but the most catastrophic environmental impacts(e.g., loss of a community’s water supply, etc.). Typical costs of plugging Devonian shale wellsand reclaiming sites in Pennsylvania are estimated to be somewhere in the range of $60,000 to$700,000 (mean of around $100,000) per well (with some economies of scale for pluggingmultiple wells on the same pad), so Maryland’s current bond limits appear reasonable (Mitchelland Casman 2011). Nonetheless, the state might wish to reexamine whether current bondingamounts (especially the blanket amount of $500,000) are adequate to address the full range oflikely decommissioning costs. If the state is going to enforce the best practice of drillingmultiple (e.g., six) wells from a single pad, it isn’t unlikely that a single operator could developfive pads (total of 30 wells) with a single blanket bond of $500,000 (less than $17,000 per well).Maryland might also consider alternate mechanisms of covering decommissioning andreclamation costs through a trust fund mechanism (i.e., investing revenue from pre-drilling feesand a five-year severance tax on production) as an alternative to bonding. The obvious downsideto the state for such a mechanism is the case of the underperforming well (or dry hole) thatwould produce inadequate funding of the trust account. This problem could be solved fairly 1-24

41. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandeasily through the use of a pooled trust funded through revenue from multiple operators and byregularly adjusting the severance tax rate to ensure that the pooled fund is always adequate tocover the expected cumulative liability (Mitchell and Casman 2011).K. Key recommendations1-A Pre-development environmental assessment should be conducted on a site-specific basis and include: (1) identification of all on-site drilling hazards such as underground mine workings, orphaned gas or oil wells, caves, caverns, Karst features, etc.; (2) identification of all ecological, recreational, historical, and cultural resources in the vicinity of a proposed site (includes well pad and all ancillary development such as cleared areas around a well pad, roads, bridges, culverts, compressor stations, pipelines, etc.); (3) identification of the appropriate setbacks and buffers for the proposed site; and (4) collection of two years of pre-development baseline data on underground drinking water, surface water, and both aquatic and terrestrial ecological resources.1-B Maryland should require as part of its permit application at least two years of site specific data collection prior to any site development that would be used to characterize the resources at risk and provide a solid baseline dataset that would ultimately be used to understand process and feedback to the refinement of BMPs.1-C Comprehensive planning (a.k.a., comprehensive drilling plans) could potentially be used to effectively channel MSGD into areas that would be less sensitive to impacts while allowing for considerable and efficient exploitation of the gas resource. Spacing multi- well pads in clusters—as far apart as is technically feasible—makes maximum use of horizontal drilling technology and could be an important BMP in terms of minimizing development impacts. With careful and thoughtful planning (e.g., co-location of infrastructure wherever possible), it may be possible to develop much of the gas resource in a way that disturbs less than 1-2% of the land surface, even when accounting for the need for ancillary infrastructure such as access roads, pipelines, and compressor facilities. Comprehensive gas development plans could also moderate the rate at which the resource is developed in Maryland, thus allowing the regulatory enforcement arm of MDE (with little recent experience in gas well permitting and no experience in unconventional gas) to ramp up over time.1-D Maryland should consider legislation that would enable the state to implement “forced pooling” as a way of providing greater resource protection while allowing for efficient resource exploitation.1-E Maryland should impose by regulation sensible setbacks (see Table 1.1) that are adequate to protect public safety, as well as ecological, recreational, historical, cultural, and aesthetic resources.1-F There is a definite need for an analysis of extant hydrogeological data from western Maryland that could be used to develop flow nets or models and infer groundwater flowpaths and other important features such as recharge areas, discharge areas, hydrologic residence times, and depth of the freshwater zone across the area.1-G Maryland might consider developing a standardized stakeholder process that could be implemented as part of comprehensive planning strategy; the goal of such a process 1-25

42. K.N. Eshleman & A.J. Elmore (2013) Chapter 1 would be to engage stakeholders and the community in the most effective ways possible, while allowing the permit review process to be expedited.1-H We recommend that Maryland follow guidance from New York’s experience with unconventional shale gas development and effectively not permit MSGD (or any other unconventional gas development) where the target formation occurs within 1,000 vertical feet of USDW or within 2,000 vertical feet of the ground surface. Since the freshwater/saltwater interface has not been mapped in Maryland, the prudent approach would be to rely on the 2,000 ft criterion to provide an adequate margin of safety.1-I An obvious best practice would be to site well pads so as to avoid vertical drilling (i.e., surface boreholes) in areas where shallow caves and caverns have been mapped or where there is a high probability that such systems might be present. Maryland should develop a GIS map system of both active and abandoned oil and gas wells (including gas storage wells) and active and abandoned coal mine workings prior to permitting any new Marcellus wells; all underground hazards with ¼ mile of any section of a proposed Marcellus well should be identified as part of the permit review process and avoided wherever possible.1-J Maryland should require a 1,000 ft setback from all deep mine workings and ¼ mile setback from all historic gas wells. The gas well setback should be measured from any portion of the borehole (vertical or horizontal) to the historic well.1-K Maryland should develop regulations that force rapid partial reclamation (including revegetating disturbed areas surrounding wells pads, corridors, and ancillary infrastructure) of all land not needed for drilling and production as quickly as possible, while allowing the remaining portion to exist unreclaimed only until such time as drilling is completed, production ends, and final reclamation can be performed.1-L We found that Maryland’s current oil and gas regulations governing permitting for conventional development require many of the elements that would be needed to properly address MSGD or unconventional development in general; however, the state should consider revising its oil and gas permitting regulations to explicitly address water withdrawal and storage issues, drilling waste and wastewater treatment and disposal issues, as well as transportation planning issues.1-M Local zoning ordinances for both counties should be amended to spell out in which zoning districts MSGD would be permitted as a way of minimizing some of the major conflicts and public safety issues that we addressed in this report.1-N Maryland’s requirements for performance bonding under current regulations ($100,000 per well or $500,000 blanket bond for all of an applicant’s wells) are relatively high compared to other states; thus, the state might be to avoid some of the problems associated with divestment of MSGD assets from primary to secondary firms that are predicted as gas production declines. Nonetheless, Maryland might want to consider alternate mechanisms of covering decommissioning and reclamation costs through a trust fund mechanism (i.e., investing revenue from pre-drilling fees and a five-year severance tax on production) as an alternative to performance bonding. 1-26

43. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandL. Literature citedAbbas, R., et al. 2003/2004. A safety net for controlling lost circulation. Oilfield Review 15:20-27.API. 2009. Environmental Protection for Onshore Oil and Gas Production Operations and Leases. API RP-51R. Washington, DC: American Petroleum Institute Publishing Services.Boyer, E. W., B. R. Swistock, J. Clark, M. Madden, and D. E. Rizzo. 2011. The Impact of Marcellus Gas Drilling on Rural Drinking Water Supplies. Harrisburg, PA: The Center for Rural Pennsylvania.Christopherson, S., and N. Rightor. 2011. The Boom-Bust Cycle of Shale Gas Extraction Economies. Ithaca, NY: CaRDI, Cornell University.Duigon, M. T., and M. J. Smigaj. 1985. First report on the hydrologic effects of underground coal mining in southern Garrett County, Maryland. Report of Investigations No. 41, U.S. Geological Survey.Elmore, A. J., J. P. Julian, S. M. Guinn, and M. C. Fitzpatrick. In review. Potential stream density in mid- Atlantic U.S. watersheds. Plos One.Fisher, K. 2010. Data confirm safety of well fracturing. The American Oil and Gas Reporter, July issue.Franz, R., and D. Slifer. 1976. Caves of Maryland. Educational Series No. 3, Maryland Geological Survey.Hammer, R. J. Van Briesen, and L. Levine. 2012. In frackings wake: new rules are needed to protect our health and environment from contaminated wastewater. Natural Resources Defense Council.Hobba, Jr., W. A., D. W. Fisher, F. J. Pearson, Jr., and J. C. Chemerys. 1979. Hydrology and geochemistry of thermal springs in the Appalachians. Geological Survey Professional Paper 1044-E, Washington, DC: U.S. Geological Survey.Klauda, R., A. Prochaska, and M. Kashiwagi. 2012. Recommended monitoring plan for surface waters associated with Marcellus shale gas development in western Maryland. Annapolis, MD (May 8, 2012 Draft): Maryland Department of Natural Resources.Kostelnik, J., and K. M. Carter. 2009. Unraveling the stratigraphy of the Oriskany Sandstone: A necessity in assessing its site-specific carbon sequestration potential. Environmental Geosciences 16:187- 200.Lien, A. M., and W. J. Manner. 2010. The Marcellus shale: resources for stakeholders in the Upper Delaware watershed region. Pinchot Institute for Conservation.Marcellus Shale Advisory Commission. 2011. [Pennsylvania] Governors Marcellus Shale Advisory Commission Report.Mitchell, A. L., and E. A. Casman. 2011. Economic incentives and regulatory framework for shale gas well site reclamation in Pennsylvania. Environmental Science and Technology 45: 9506-9514.NYSDEC. 2011. Revised Draft Supplemental Generic Environmental Impact Statement on the Oil, Gas, and Solution Mining Regulatory Program.Ohio Legislative Service Commission. 2010. Final analysis: Sub S.B. 165 (128th General Assembly).PADCNR. 2011. Guidelines for Administering Oil and Gas Activity on State Forest Lands. Commonwealth of Pennsylvania Department of Conservation and Natural Resources Bureau of Forestry.Sage Policy Group, Inc. 2012. The Potential Economic & Fiscal Impacts of Natural Gas Production on Western Maryland. Submitted on behalf of American Petroleum Council.Soeder, D. J., and W. M. Kappel. 2009. Water resources and natural gas production from the Marcellus shale. Fact Sheet 2009-3032, Baltimore, MD: U.S. Geological Survey.Staubitz, W. W., and J. R. Sobashinski. 1983. Hydrology of area 6, eastern coal province, Maryland, West Virginia, and Pennsylvania. Water-Resources Investigations Open-File Report 83-33, Towson, MD: U.S. Geological Survey.STRONGER. 2011. Colorado Hydraulic Fracturing State Review. Oklahoma City, OK: STRONGER, Inc.Trapp, Jr., H., and M. A. Horn. 1997. Delaware, Maryland, New Jersey, North Carolina, Pennsylvania, Virginia, and West Virginia. In Ground Water Atlas of the United States, by J. A. Miller, L2 - L24. Reston, VA: U.S. Geological Survey. 1-27

44. K.N. Eshleman & A.J. Elmore (2013) Chapter 1Ubinger, J. W., J. J. Walliser, C. Hall, and R. Oltmanns. 2010. Developing the Marcellus Shale: Environmental Policy and Planning Recommendations for the Development of the Marcellus Shale Play in Pennsylvania. Harrisburg, PA: Pennsylvania Environmental Council.USDOE. 2011. Secretary of Energy Advisory Board Shale Gas Production Subcommittee Second Ninety Day Report. U.S. Department of Energy.USEIA. 2012. Review of Emerging Resources: U.S. Shale Gas and Shale Oil Plays. U.S. Energy Information Administration.USEPA. 2004. Evaluation of Impacts to Underground Sources of Drinking Water by Hydraulic Fracturing of Coalbed Methane Reservoirs Study. EPA 816-R-04-003. Washington, DC: U.S. Environmental Protection Agency, Office of Water.USEPA. 2011. Plan to Study the Potential Impacts of Hydraulic Fracturing on Drinking Water Resources; EPA/600/R-11/122. Washington, DC: U.S. Environmental Protection Agency, Office of Research and Development.USEPA. 2012. Study of the Potential Impacts of Hydraulic Fracturing on Drinking Water Resources (Progress Report). EPA/601/R-12/011. Washington, DC: U.S. Environmental Protection Agency, Office of Research & Development.Zebrowski, M. J., J. R. Kelafant, and C. M. Boyer. 1991. Reservoir characterization and production potential of the coal seams in Northern and Central Appalachian basins. Proceedings of the 1991 Coalbed Methane Symposium. University of Alabama Tuscaloosa. 1-28

45. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland2. Protecting air quality1Natural gas from MSGD has the potential to provide substantial energy economically and at amuch lower cost to the atmospheric environment than the same amount of energy generated fromcoal combustion. In particular, natural gas (predominantly methane) has advantages over coalwith respect to trapping of infrared radiation (IR) by greenhouse gases (GHGs) and contributingto planetary warming. The GHG advantage of natural gas arises from the relative heat (availablefor energy production) per unit of CO2 released in combustion. For each molecule of CO2produced, roughly twice as much energy is available from natural gas than from coal. Thisadvantage is only realized if the gas is combusted completely, however. In trapping IR radiationand warming the planet, methane is 30 times more potent than carbon dioxide (IPCC 2007).Therefore, if 1/30th (~3%) or more of natural gas is lost in production, processing, and transportto market, there is no climate advantage over coal2. Actual emission rates are a hotly debatedsubject (Armendariz 2009, Howarth et al. 2011a, Howarth et al. 2011b, Cathles et al. 2012) andcan only be estimated for local operations. Nonetheless, there is some evidence that emissionsrates may be significantly higher than initial estimates (Petron et al. 2012).The specifics of the GHG calculation are as follows. Coal is roughly 80% carbon by mass;natural gas is about 90% methane. The combustion of a molecule of carbon or methane producesone molecule of CO2, but the methane produces roughly twice as much heat represented by theenthalpy of combustion, DH°. C + O2 → CO2 DH° = -94 kcal/mole CH4 + 2O2 → CO2 + 2H2O DH° = -193 kcal/moleThe ratio depends on details such as the exact composition of coal and gas but can beapproximated as 193/94 = 2.05 or ~ 2. Substantial amounts of methane are released in coalmining and processing as well. If natural gas is used as a substitute for coal in electricitygeneration, it offers the additional advantage of higher efficiency by approximately a factor oftwo. But the general rule holds: natural gas is better for climate than coal as long as losses canbe kept below 3% of total production.Maryland’s primary air quality issues from among all of USEPA’s criteria pollutants are ozone(O3, also called photochemical smog or Los Angeles type smog) and fine particulate matter(PM2.5, the mass of particles less than 2.5 μm in diameter in a cubic meter of air). Maryland is inviolation of the National Ambient Air Quality Standard (NAAQS: 75 parts per billion for an 8-hraverage) for ozone and in or near compliance for PM2.5, although both standards are likely to betightened in coming years. Maryland must also comply with the Regional Haze Rule to improve1 Chapter co-authors: Keith N. Eshleman, Ph.D. (Appalachian Laboratory, University of Maryland Center for Environmental Science, Frostburg, MD 21532); Russell R. Dickerson, Ph.D. (Department of Atmospheric and Oceanic Science, University of Maryland, College Park, MD 20742)2 It should be noted, however, that natural gas has several other air quality benefits relative to coal, including lower emissions factors (lb/MMBTU) for nitrous oxides, sulfur dioxide, particulate matter, carbon monoxide, and mercury. 2-1

46. K.N. Eshleman & A.J. Elmore (2013) Chapter 2visibility or visual range. The limit to visual range is generally fine particles, so comments onPM2.5 also apply to haze. Maryland is a member of the Regional Greenhouse Gas Initiative(RGGI) that seeks to limit emissions of pollutants that disrupt the Earth’s radiative balanceincluding carbon dioxide, nitrous oxide, and methane. On a local scale, hazardous air pollutants(HAPs), malodorous gases such as hydrogen sulfide (H2S), and mercaptans (organic sulfurcompounds) can be of concern, and radon (an α particle-emitting, respirable radioactive materialproduced as a decay product of radium present in the Marcellus shale formation)3.In terms of emissions, Maryland’s top priorities are the precursors to O3 and PM2.5 (i.e., thechemical species that form these pollutants in the atmosphere). In the eastern U.S., both ozoneand haze are considered to be secondary pollutants (made in the atmosphere by photochemicalreactions of precursor gases) rather than primary pollutants (released directly into theatmosphere). Ozone forms by atmospheric reactions involving two main classes of precursorpollutants: volatile organic compounds (VOCs) and nitrous oxides (NOX); carbon monoxide(CO) is also important for O3 formation in polluted areas and in the remote troposphere. Theformation of O3 from these precursors is a complex, nonlinear function of many factorsincluding: (1) the intensity and spectral distribution of sunlight; (2) atmospheric mixing; (3)concentrations of precursors in the ambient air and the rates of chemical reactions of theseprecursors; and (4) processing on cloud and aerosol particles (USEPA 2012). Fine particularmatter (PM2.5) are those particles (such as those found in smoke and haze) that can be deeplyrespired into the lungs; while the sources of these particles can be from forest fires, wood stoves,and other direct combustion sources (e.g., soot or “black carbon” emitted from the tailpipes ofcars, trucks, and other on-road vehicles), they are commonly formed when gases emitted frompower plants, industrial plants, and automobiles react in the atmosphere. The most commongases cited as precursors of PM2.5 formation include: sulfur dioxide (SO2), NOx, VOCs, andammonia (NH3). Secondary formation of O3 and PM2.5 is thus linked by virtue of involving someof the same precursor pollutants.High quality (dry) natural gas is composed primarily of methane (CH4), but contains appreciableamounts (percentages) of other light alkanes such as ethane (C2H6), propane (C3H8), butane(C4H10) and pentane (C5H12). The amounts decrease with increasing carbon number, but recentevidence indicates wide variability among MSGD wells in neighboring states. Methane and thelight alkanes themselves do not contribute significantly to ground level ozone or fine particulatematter. Heavier and unsaturated volatile organic compounds (VOCs), particularly biogenicisoprene, do contribute substantially to ozone formation, however. Hazardous air pollutants(HAPs) may also arise from natural gas production. Recently, evidence indicates that somenatural gas operations could be non-negligible sources of benzene (C6H6), a variety of otherHAPs, and heavier hydrocarbons (McKenzie et al. 2012, Petron et al. 2012, personalcommunication: R. Schnell, Global Monitoring Division, NOAA). These could pose a health riskto individuals living within ~1000 m of a gas operation.3 Among different types of rocks, granites and rhyolites (igneous rocks) are most commonly enriched in uranium, but some sedimentary rocks—such as the Marcellus shale—that are rich in organic matter can be significantly enriched in uranium (and thus radium). 2-2

47. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandThe purpose of this chapter is to examine the sources of pollutants and air pollution precursorsassociated with MSGD operations and make recommendations of best management practicesthat should be use to control such emissions and protect air quality in Maryland locally,regionally, and globally. Ozone and PM2.5 are predominantly regional problems, as pollutantformation continues to occur well downwind of precursor emissions. Pollution events tend tohave spatial scales of ~1000 km and temporal scales of 1-5 d. Releases of primary pollutants(particularly NOx) in western Maryland where MSGD could occur could certainly have adverseeffects on eastern Maryland where NAAQS ozone violations occur. The regional scale of theseproblems would suggest that even if MSGD does not go forward in Maryland, the state’s airquality would be expected to be affected to some degree by activities in surrounding states; inparticular, we are concerned that greater regional emissions of NOx into a regional atmosphereupwind (i.e., west) of Maryland would be expected to make it more difficult for the state to meetthe NAAQS for ozone in the future. While no studies have been published on regional air qualityimpacts from MSGD, one numerical atmospheric modeling study of Texas and Louisianaindicated increases in the 8-hr ozone values of up to 5 ppb as a result of natural gas developmentof the Haynesville Shale (Kemball-Cook et al. 2010).A. Reducing pollutant and pollutant precursor emissions from MSGD operationsImplementation of BMPs to control air pollution emissions in Maryland—as in neighboringMarcellus shale states—would be driven largely in an effort to comply with USEPA regulationsunder the Clean Air Act that mandate both New Source Performance Standards (NSPS) andNational Emissions Standard for Hazardous Air Pollutants (NESHAP) for oil and natural gasproduction. In the following subsections we describe the BMPs that could be deployed to reduceemissions of the key air pollutants (and air pollution precursors) described above.Methane and VOCs. Determining BMPs for reducing methane and VOC emissions from MSGDis in part dependent on identifying and inventorying the specific sources of these gases within thegas sector. Again, no such studies have been conducted for the MSGD region, but a recent studyof the Barnett shale region of Texas (Armendariz 2009) provides useful information that may berelevant to western Maryland. Emission sources in the gas industry can be classified as follows:(1) fugitive emissions; (2) vented emissions; and (3) combustion emissions. Fugitive emissionsare unintentional leaks around seals and gaskets, leaks from underground pipelines due to fromcorrosion or faulty connections, or emissions that occur during the well completion process.Vented emissions are releases to the atmosphere by design or operational practice. Examples ofvented emissions include: emissions from continuous process vents, such as dehydrator reboilervents; maintenance practices, such as blowdowns; and small individual sources, such as gasoperated pneumatic device vents. Combustion emissions are exhaust emissions from combustionsources such as compressor engines (Kirchgessner et al. 1997). Although there is quite a bit ofuncertainty in methane emission rates both in absolute terms and expressed as a percentage oftotal production (Cathles et al. 2012), among the most significant emission sources of VOCs andmethane are: (1) fugitive emissions during completion (Howarth et al. 2010); (2) fugitiveemissions from compressor station and transmission systems (Kirchgessner et al. 1997, Howarthet al. 2010); and (3) routine venting emissions (Kirchgessner et al. 1997, Howarth et al. 2010). 2-3

48. K.N. Eshleman & A.J. Elmore (2013) Chapter 2One best practice that can dramatically limit both VOC and methane emissions during the wellcompletion phase is a procedure known as a “green completion” or “green flowback process”(Armendariz 2009). In this process, performed using special equipment brought onto the wellpad, gases and liquids brought to the surface during the 3- to 10-day completion process arecollected, filtered, and transported into production pipelines and tanks, instead of being dumped,vented to the atmosphere, or flared. After the completion process has ended, the produced gasesand liquids can be directed to permanent on-site separators (that separate gas from water and anyhydrocarbon liquids), condensate tanks, and piping that had been installed at the well site;condensate tanks are sources of emissions through venting to the atmosphere. “Greencompletions” are considered highly cost-effective in reducing VOC and methane emissions inthe Barnett Shale in Texas (Armendariz 2009) and will be required by USEPA nationwide afterJanuary 1, 2015 under NSPS for VOCs and NESHAP for oil and natural gas production.Unfortunately, a major loophole in implementing green completions is that the process is notapplicable to “exploratory” or “wildcat” drilling, because the well must be near an operationalpipeline; therefore, such well completions have been exempted by USEPA from complying withthis requirement. In a phased comprehensive gas development plan, Maryland could work withindustry to site early pads at specific locations where “wildcatting” would be permitted; duringthis phase, green completions would not occur and gases would likely be flared during thecompletion process. Pending the outcome of this exploratory phase, construction of additionalwell pads and the associated pipeline and compressor infrastructure to transport gas wouldsubsequently be coordinated in a second phase (during which green completions would berequired).Two other final rules governing VOC and methane emissions were recently instituted byUSEPA: (1) use of modified (“low bleed”) pneumatic controllers for many functions between thewellhead and the point where natural gas enters a transmission pipeline; and (2) use of newstorage tanks for condensate which are capable of routing VOC emissions to a combustion orflaring device. Enclosed flaring devices are highly efficient (98%) devices that can dramaticallyreduce VOC and methane emissions from tanks (Armendariz 2009). USEPA is also trying toaddress VOC emissions for natural gas processing plants through the NSPS process—inparticular controlling fugitive emissions from separators, glycol dehydrators, storage tanks, andmetering stations. Many of these standards promote aggressive leak detection and repairs. Leakdetection at processing plants is covered by NSPS and can be performed using handheld organicvapor meters (OVMs); inspections are performed at a specified frequency under the NSPS.Natural gas that is low in high molecular weight hydrocarbons may not require such processing,and we are not sure whether such plants will be required in Maryland.Expansion of a comparable leak detection and repair program that governs operations fromwellhead to the transmission line would be considered a BMP for reducing emissions inMaryland and elsewhere, regardless of whether processing plants are necessary. Thermalimaging cameras (e.g., FLIR Commercial Systems B.V., Breda, The Netherlands) have beenused to great effect in identifying hydrocarbon leaks in Houston refineries. These cameras can bemounted on aircraft to survey broad areas to identify major leaks of hydrocarbons; this approachmight be applicable to identifying hydrocarbon leaks from well sites, compressor sites, andpipeline networks in the Marcellus region. The Texas Commission on Environmental Quality 2-4

49. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland(TCEQ) has identified leaks that, when repaired, saved the refinery operators substantial productloss that more than paid for the monitoring and repair actions.4 These and many other BMPshave been advocated by USEPA’s Natural Gas STAR program aimed at implementing cost-effective strategies for reducing methane emissions by the industry5; as proposed for New YorkState, best practice in Maryland would be that all operators voluntarily participate in thisprogram and implement as many of the recommended strategies as possible (NYSDEC 2011),including: (1) reducing methane emissions from pneumatic devices in the natural gas industry;(2) reducing methane emissions from compressor rod packing systems; (3) reducing emissionswhen taking compressors off-line; (4) replacing glycol dehydrators with desiccant dehydrators;(5) replacing gas-assisted glycol pumps with electric pumps; (6) optimizing glycol circulationand installing flash tank separators in glycol dehydrators; (7) using efficient compressor engines;(8) using efficient line heaters; (9) using efficient glycol dehydrators; (10) re-using productionbrines; (11) ensuring all flow connections are tight and sealed; (12) performing leak detectionsurveys and taking corrective actions; (13) using efficient exterior lighting; and 14) using solar-powered telemetry devices.NOx. Unlike VOCs and methane that are principally emitted through fugitive and ventingmechanisms, NOx is primarily a product of operating internal combustion engines. Large (1,000horsepower, HP) diesel internal combustion engines are often used to operate drilling rigs andpower hydraulic fracturing pumps, although electric drill rigs can be powered off of the electricalgrid where NOx is effectively capped at the electrical generating plant under the Clean Air ActAmendments. Smaller combustion engines are used to power compressors that produce andtransport the gas through pipelines; these engines can be powered by either diesel fuel, naturalgas, or electricity. An obvious best practice for controlling NOx emissions from MSGD inMaryland would be through the use of electrical drilling rigs, hydraulic fracturing pumps, andcompressor engines that are operated off of line power; Maryland should consider mandatingelectrically-powered equipment wherever line power is available (or could be made readilyavailable); this alternative might be reasonably cost-effective if MSGD can be conductedprimarily or exclusively in densely clustered multi-well pad developments as discussed inChapter 1. As an alternative to this practice that would be applicable to well pad locations noteasily served by line power, Maryland could require that all engines (i.e., diesel and/or sparkignited for drilling devices, pumps, compressors, trucking, etc.) used by MSGD operators meet“fleet average” standards for NOx emissions based on USEPA 1998 standards for heavy-dutydiesel highway vehicles of 4.0 g NOx/bhp-hr6 equivalent to 25 g NOx/kg fuel7. In Texas, TCEQhas taken a similar approach in regulating NOx emissions in the Dallas-Fort Worth metropolitanarea, although TCEQ used an even more stringent emissions standard of 0.5 g NOx/bhp-hr(Armendariz 2009). Operators would essentially have three options: (1) utilizing newer dieselengines that can meet these emission standards; (2) replacing internal combustion engines withelectrically-powered motors; or (3) some combination of the two options that would expectedlybe determined by cost. Either of these three options would have significant co-benefits in terms4 See http://www.tceq.texas.gov/airquality/barnettshale/bshale-faq and http://www.texassharon.com/2012/03/20/tceq-videos-show-voc-emissions-in-eagle-ford-shale/5 http://www.epa.gov/gasstar/tools/recommended.html6 http://www.epa.gov/otaq/standards/heavy-duty/hdci-exhaust.htm. A bhp-hr is a brake horsepower-hour (a unit of work).7 This is similar to observed emissions rates for in-use vehicles, but not as stringent as the 2007+ standard of 0.2 g NOx/bhp-hr. 2-5

50. K.N. Eshleman & A.J. Elmore (2013) Chapter 2of reducing VOCs as well, but engines (i.e., drilling devices, compressors, trucking etc.) couldalso be required to meet the “fleet average” of all engines set by the USEPA 1998 hydrocarbon(HC) standards for heavy-duty diesel highway vehicles of 1.3 g HC/bhp-hr8.PM2.5. We recommend that Maryland require the “fleet average” of all internal combustionengines (i.e., drilling devices, compressors, pumps, trucking etc.) used in MSGD meet USEPA1998 standards for heavy-duty diesel highway vehicles of 0.1 g PM/bhp-hr9. Restricting idlingtime and requiring use of ultra low sulfur diesel (ULSD) fuel would also be considered bestpractices.Hazardous air pollutants. Hazardous air pollutants (HAPs), particularly organic HAPs, havealso been reported to exist in concentrations that are a cause for concern in the vicinity of naturalgas production facilities and should be monitored near any Maryland sites. The compounds ofprimary concern as HAPs include benzene, toluene, ethylbenzene, and xylene (i.e., BTEX), aswell as formaldehyde, among others.Radon. As discussed in Chapter 4, production brine is likely to contain elevated levels ofnaturally occurring radioactive material (NORM), principally radium-226 (226Ra)—a radonprecursor. This material may pose a hazard to workers handling the drilling and recoveryequipment, so gamma and alpha radiation from production brine should be monitored at eachsite. The radon gas itself that is released is unlikely to pose either a health or safety hazard unlessit is contained in a confined space, however. There are no effective ways of controlling therelease of radon to the atmosphere other than reburying the radium source.B. On-site and off-site air quality monitoringIf and when drilling begins in Maryland, one way the state could attempt to address regional airquality issues (i.e., ozone) associated with developing the Marcellus shale would be to developand implement an air emissions monitoring program throughout the region as has been proposedfor Pennsylvania (Lien and Manner, 2010). The program would be focused on assessing bothpoint sources and fugitive sources of pollutants (and pollutant precursors) at well pads and atother sources resulting from natural gas production.C. Key recommendations2-A Require that operators in Maryland establish a methane leak detection and repair program that governs operations from wellhead to the transmission line, regardless of whether processing plants are necessary. All operators in Maryland should voluntarily participate in USEPA’s Natural Gas STAR program aimed at implementing cost-effective strategies for reducing methane emissions by the industry.8 http://www.epa.gov/otaq/standards/heavy-duty/hdci-exhaust.htm; this is not as stringent as the 2007+ standard of 0.14 g HC/bhp-hr and is practicable with current technology at reasonable cost.9 http://www.epa.gov/otaq/standards/heavy-duty/hdci-exhaust.htm; this not as stringent as the 2007+ standard of 0.01 g PM/bhp-hr and is practicable with current technology at reasonable cost. 2-6

51. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland2-B Encourage operators to either use newer internal combustion engines or convert from diesel internal combustion engines to electric motors for operating drilling rigs, pumps, and compressors wherever possible by implementing “fleet average” emission standards for NOx, VOCs, and PM2.5.2-C Require monitoring of hazardous air pollutants at well pad sites.2-D Monitor gamma and alpha radiation of production brines.2-E Implement an air emissions monitoring program throughout the region, focusing on sources and fugitive sources of pollutants (and pollutant precursors) at well pads and at other sources resulting from natural gas production.D. Literature citedArmendariz, A. 2009. Emissions from natural gas production in the Barnett Shale area and opportunities for cost-effective improvements, report to EDF.Burgard, D. A., G. A. Bishop, and D. H. Stedman. 2006. Remote sensing of ammonia and sulfur dioxide from on-road light duty vehicles. Environmental Science & Technology 40:7018-7022.Cathles, L. M., L. Brown, M. Taam, and A. Hunter. 2012. A commentary on "The greenhouse-gas footprint of natural gas in shale formations" by RW Howarth, R. Santoro, and Anthony Ingraffea. Climatic Change 113:525-535.Howarth, R. W., A. Ingraffea, and T. Engelder. 2011a. Natural gas: should fracking stop?, Nature 477:271-275.Howarth, R. W., R. Santoro, and A. Ingraffea. 2011b. Methane and the greenhouse-gas footprint of natural gas from shale formations. Climatic Change 106:679-690.IPCC. 2007. Intergovernmental Panel on Climate Change: Fourth Assessment Report, Geneva, Switzerland.Kemball-Cook, S., et al. 2010. Ozone impacts of natural gas development in the Haynesville shale. Environmental Science & Technology 44:9357-9363.Kirchgessner, D.A., R.A. Lott, R.M. Cowgill, M.R. Harrison, and T.M. Shires. 1997. Estimate of methane emissions from the natural gas industry. Chemosphere 35:1365-90.Lien, A. M., and W.J. Manner. 2010. The Marcellus shale: resources for stakeholders in the Upper Delaware watershed region. Pinchot Institute for Conservation.McKenzie, L. M., R. Z. Witter, L. S. Newman, and J. L. Adgate. 2012. Human health risk assessment of air emissions from development of unconventional natural gas resources. Science of the Total Environment, 424:79-87.NYSDEC. 2011. Revised Draft Supplemental Generic Environmental Impact Statement on the Oil, Gas, and Solution Mining Regulatory Program.Petron, G., et al. 2012. Hydrocarbon emissions characterization in the Colorado Front Range: a pilot study. Journal of Geophysical Research-Atmospheres 117.USEPA. 2012. Integrated Science Assessment for Ozone and Related Photochemical Oxidants (Third External Review Draft). EPA/600/R-10/076C. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research and Development 2-7

52. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland3. Well engineering and construction practices to ensure integrity andisolation1The primary goal of the oil and natural gas industry is to cost-effectively explore for and extractpetroleum and natural gas from subsurface environments where such substances have formedand accumulated over geologic time—typically hundreds of millions of years. The most commonapproach to extracting these substances from onshore reservoirs is through the drilling ofboreholes from the land surface to the target zone within which these substances are thought tobe concentrated and then completing a well by hydraulic fracturing that provides a pathway forthese substances to be brought to the surface in an efficient, safe, and controlled way. Obviously,well engineering and construction practices have evolved over time as operators have gainedgreater experience and as technological improvements have allowed. For a century or more in anera in which environmental resources were not greatly considered, the industry made very little,if any, significant effort to explore and produce oil and natural gas in ways that would beconsidered environmentally sound by modern standards. For example, large volumes of brine(saline water) that were brought to the surface with the oil and gas were typically stored inunlined pits that overflowed into streams and rivers or seeped into groundwater causingwidespread water pollution.In recent decades, the industry has responded to pressure to reduce its environmental footprintand many best management practices (BMPs) have been developed and employed to ensure theintegrity of each well system, isolate the well from the surrounding subsurface environment, andeffectively contain the produced gas and other fluids within the well’s innermost productionconduit so it can be successfully transported through ancillary pipelines for processing anddelivery to market. Heightened environmental awareness and elevated environmental standardshave also forced the industry to make substantial progress in collecting, storing, treating, andrecycling of liquid drilling wastes (i.e., “flowback”, brines), although the industry still relies veryheavily on underground injection as the ultimate disposal process. API—as the technical arm ofthe oil and gas industry—has taken the lead in reviewing and evaluating the industry’s practicesfor drilling, completing, and operating oil and natural gas wells; on the basis of its on-goingtechnical reviews of various practices, API has published an extensive number of documentsdescribing so-called “recommended practices” (RPs) which it communicates and shares with theindustry. Many of these RP’s explicitly address problems in maintaining well integrity andprovide standards that have been expressly adopted by some state regulatory authorities.Obviously, not all well construction activities go according to plan and—despite significantexperience with hydraulic fracturing—there have been relatively few published data-drivenstudies that explicitly address the problem of transport of subsurface contaminants fromhydraulically fractured horizontally-drilled gas wells into aquifers over the lifetime of aproducing well (Myers 2012); the author of a white paper on the subject described the science ofunderstanding this problem as “recent, ongoing, and incomplete” (Ingraffea 2012). The recent1 Chapter author: Keith N. Eshleman, Ph.D. (Appalachian Laboratory, University of Maryland Center for Environmental Science, Frostburg, MD 21532) 3-1

53. K.N. Eshleman & A.J. Elmore (2013) Chapter 3modeling study by Myers (2012) addressed movement of contaminants to surficial aquifersthrough natural pathways—both advective transport through porous media overlying ahydraulically-fractured shale formation and preferential transport through fractures, butsignificant questions have been raised regarding the assumptions and conclusions of the study(Saiers and Barth 2012). Two recent peer-reviewed studies provided circumstantial experimentalevidence that methane gas and formation brine can seep out of shale formations and contaminateoverlying aquifers (Osborn et al. 2011, Warner et al. 2012), but the mechanism for suchcontamination is unknown and any relationship to hydraulic fracturing remains unproven(Osborn et al. 2011, Davies 2011).The Deepwater Horizon oil spill in the Gulf of Mexico in 2010 following the Macondo wellblow-out is a chilling reminder of what can go wrong when well fluids cannot be isolated,contained, and controlled (in this example, due principally to a faulty cement job). The finalreport from the U.S. government’s official investigation into the causes of this accident alsohighlighted a series of decisions that complicated the cementing operation, increased the risks offailure, and were major contributing factors in the blow-out and explosion on April 20, 2010 thatkilled 11 men working on the drilling platform and caused the subsequent spill of an estimated4.9 million barrels of oil into the Gulf of Mexico (National Commission on the BP DeepwaterHorizon Oil Spill and Offshore Drilling 2011). The Deepwater Horizon incident also underlinesthe importance of ensuring that the activities of all subcontractors working on a particular wellare coordinated and adequately supervised by the lead operator (or prime contractor).The purpose of this chapter is to review and recommend best management practices for ensuringwell integrity and isolation of unconventional Marcellus shale gas wells based on our review ofAPI recommended practices and regulations in place in Colorado, Ohio, Pennsylvania, and WestVirginia (and proposed regulations in New York State). Other best practices that are consideredancillary to well drilling, completion, and production (e.g., BMPs for containing, treating, anddisposing of drilling wastes—especially “flowback” and brines) are discussed in Chapter 4.A. Well planningAPI provides a very detailed explanation of the critical need for operators to perform adequatewell planning as a first step to ensuring well integrity and isolation (API 2010)2. The rationalefor such planning is very well established from experience, and optimum well planning forconstructing wells for developing the Marcellus shale gas resource in Maryland would likelyinclude the following elements: • evaluation of potential flow zones; • site selection; • hazard assessment and contingency planning; • well control planning for fluid influxes; • lost circulation control plans; • regulatory issues and communications plans; and2 API Standard 65-Part 2 was prepared based principally on experience in the U.S. outer continental shelf and deepwater areas of the Gulf of Mexico, but the recommendations may be applicable to other offshore and onshore areas (see p. iii). 3-2

54. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland • construction designs and plans for a specific well that would include: (1) an analysis of pore pressures, fracture gradients, and required drilling fluid weights; (2) a casing plan; (3) a cementing plan; (4) a drilling plan; (5) a hydraulics plan that provides for adequate wellbore cleaning and control of static and dynamic wellbore pressures; (6) a barrier design that provides for control of all pressures that may be encountered during the life of the well; and (7) a contingency plan that addresses wellbore instability and unintended gains and losses of fluids.Site selection is a critical aspect of well planning, and we discuss some of the primary constraintson siting a well pad and wells in Chapter 1 (other environmental criteria are discussed inChapters 5 through 10 of the report). We are particularly concerned about drilling in areas wherethere is a high probability of encountering large underground voids (e.g., caverns, caves, mineworkings, abandoned wells, etc.) that have the potential to cause a loss of fluid circulation duringdrilling and impose additional risks during the cementing process. Such hazards are relativelycommon in western Maryland and we recommend that sites with a high probability ofencountering such hazards be avoided.Another very important element of proper well planning includes appropriate regulatory review.Typically, the regulatory agency with jurisdiction for a particular well will need to review thewell plan before operations can begin. All four states that we reviewed with activeunconventional oil and gas development (Ohio, Pennsylvania, Colorado, and West Virginia)require some of the elements of the well plans recommended by API for offshore operations; insome cases, additional components are required. Pennsylvania’s and West Virginia’srequirements with respect to well planning are nearly identical, requiring information on3 • the anticipated depth and thickness of any producing formation, expected pressures, anticipated fresh groundwater zones and the method or information by which the depth of the deepest fresh groundwater was determined; • the diameter of the borehole; • casing type, whether the casing is new or used, depth, diameter, wall thickness and burst pressure rating; • cement type, yield, additives and estimated amount; • the estimated location of centralizers; • the proposed borehole conditioning procedures; and • alternate methods or materials as required by the state regulatory agency as a condition for the well permit.Both states’ well planning requirements lack any explicit attention to potential flow pathwaysthat are addressed in Ohio’s oil and gas regulations, however. Agency review of a well permitapplication in Ohio includes a review of wells and other potential pathways for contamination ofgroundwater within the minimum spacing distance for a proposed well (with the reviewextending along the entire lateral of a horizontal well and includes plugging records for pluggedwells and casing records for other offset wells)4. Ohio also requires that applicants meet3 25 Pa Code § 78.83a. Casing and Cementing Plan (provisions were adopted February 4, 2011, effective February 5, 2011, 41 Pa.B. 805).4 Ohio Administrative Code 1501:9-1-08 Well construction. 3-3

55. K.N. Eshleman & A.J. Elmore (2013) Chapter 3additional requirements when planning to drill in urban areas such as (1) photo imagery andlocation information for tanks and flow lines, and (2) notification of all property owners within a500 ft radius around the proposed well. Ohio also requires a pre-permit on-site review incooperation with local officials or their designees in urban areas. In general, we found that thestate requirements for well planning lack many of the essential elements recommended by APIsuch as: hazards assessment and contingency planning, plans for addressing lost circulation, andhydraulics plans for controlling all static and dynamic borehole pressures. API (2009a) alsorecommends that operators investigate and review the history of nearby wells for cementingproblems encountered (e.g., lost returns, irregular hole erosion, poor hole cleaning, poor cementdisplacement, etc.) prior to drilling; computer simulation and other planning should be carriedout to optimize casing and cement placement procedures. A BMP for anyone proposing tooperate in Maryland should be adoption of API’s extensive guidelines for well planning—at leastthose elements that are clearly relevant to onshore development. API may choose to eventuallydevelop BMPs for well planning that are specific to onshore operations, but until such practicescan be determined, the adoption of the practices advocated in API Standard 65—Part 2 (API2010) would at least ensure that a prospective operator has addressed in writing all of the majorhazards likely to be encountered and effectively communicated these, and contingencies foraddressing them, to all subcontractors and to the appropriate regulatory authorities prior tospudding the well. Ohio’s requirement for pre-permit on-site review by state regulatory staffshould also be adopted by Maryland, but this requirement should be expanded to all proposedgas wells (not just those proposed for urban areas).B. DrillingConstructing a Marcellus shale gas well typically requires several cycles of drilling, installing ofcasing strings, and cementing casing strings in place to ensure integrity and isolation. Duringeach cycle, lengths of steel casing are installed in sequentially smaller diameters inside apreviously installed and cemented casing string. Drilling the well utilizes a drill string, consistingof a drill bit, drill collars (heavy weight pipe to put weight on the bit), and sections of drill pipe.The drill string is assembled and run into the hole, and suspended at the surface from a drillingderrick or mast. The drill string is then rotated by the use of a turntable (rotary table), top driveunit, or downhole motor drive. During drilling, a fluid is normally circulated down the drill stringand up the space between the drill string and the hole that: (1) provides lubrication of the drillbit; (2) removes the formation cuttings; (3) maintains control of pressures in the well; and (4)stabilizes the hole being drilled. Drilling fluid is generally a mixture of water, clays, fluid losscontrol additives, density control additives, and viscosifiers. Drilling fluid is a carefullymonitored and controlled mixture designed to achieve best drilling results (API 2009a).The first hole to be drilled is a conductor pipe. In some cases the conductor pipe can be driveninto place like a structural piling, but in western Maryland any conductor hole would need to bedrilled. A conductor hole would logically be drilled to a depth that would provide isolation fromany nearby water wells or freshwater springs. The conductor hole would be followed bysequentially deeper (smaller diameter) holes drilled to install the surface casing, the intermediatecasing (if necessary), and the production casing (API 2009a). 3-4

56. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandPrior to drilling, best practice would be to either slightly crown the location around the wellboreto divert fluids to a flow ditch, or construct a liquid-tight cellar at least three ft in diameter toprevent surface infiltration of fluids adjacent to the wellbore. A Marcellus shale gas well wouldtypically begin by drilling vertically through the subsurface zone containing freshwater aquifers(both unconfined and confined) that can provide groundwater (or USDW5); in many areas, coalseams will also be encountered while drilling for the surface casing. Caution must be taken whiledrilling through this zone to adequately protect USDW from contamination, and state regulationsare meant to require operators to prudently drill through fresh groundwater zones so as tominimize disturbances to such zones6. One way that this can be accomplished is by drilling allintervals prior to reaching a “USDW protective depth” either on compressed air, fresh water, afreshwater-based drilling fluid, or a combination of the above. Ohio, for example, requires thatonly additives that are suitable for drilling through potable water supplies may be used whiledrilling these intervals, although the Chief of the Department of Mineral Resources Management(DMRM) has the authority to require the use of a freshwater-based drilling fluid and specify itscharacteristics while an operator isdrilling any interval prior to reachingthe USDW protective depth.7Maryland explicitly prohibits the useof any additives to drilling liquidswithout approval of MDE (exceptunder emergency conditions), andthis regulation should be retained.An intermediate hole (if needed) isalso drilled vertically after thesurface casing has been set andproperly cemented—in some casesto a kick-off point that would allow adownhole motor to gradually makethe turn from vertical to apredominantly horizontal directionduring drilling of the production hole Figure 3-1: Horizontal drilling technology requires the ability to make a turn from vertical to horizontal drilling at depths ranging from 2,000(Figure 3-1). An intermediate casing to 9,000 ft below the surface. These flexible drilling pipes are used tois typically used to isolate the well accomplish this task.from any subsurface formationsbelow the protective USDW depth that could cause well instability and provide protection fromany abnormally pressurized subsurface zones (API 2009a). The intermediate hole would notlikely be drilled “on air”. Downhole motors (which operate using the hydraulic pressure exertedby the drilling fluid) are “steerable,” meaning that the direction (in all dimensions) of drilling canbe controlled from the surface to stay within the target formation (API 2009a). New York Statehas recommended that both the intermediate and production wellbores can be drilled after allfreshwater aquifers have been properly sealed behind steel casing and cement (see Section C)5 A USDW is defined by federal statute (40 CFR 144.3). The term “groundwater” is more general and includes subsurface waters that do not necessarily meet the legal definition of USDW.6 e.g., §22-6A-24 (West Virginia Horizontal Well Act, H.B. 401)7 Ibid., 4 3-5

57. K.N. Eshleman & A.J. Elmore (2013) Chapter 3with a mud that may be either: (a) water-based; (b) potassium chloride/polymer-based with amineral oil lubricant; or (c) synthetic oil-based. Synthetic oil-based muds are described as“food-grade” or “environmentally friendly.” When drilling horizontally, mud is needed to: (a)power and cool the downhole motor and bit used for directional drilling; (b) operate thenavigational tools which require mud to transmit sensor readings; (c) provide stability to thehorizontal borehole; and (d) efficiently remove cuttings from the horizontal hole. Some operatorscan apparently drill the horizontal wellbore “on air” (i.e., with compressed air) using specialequipment to control fluids and gases that enter the wellbore (NYSDEC 2011).Air drilling has now been used extensively in the Appalachian region for both gas drilling andfor drilling water wells and should probably be considered a best practice. Air drilling is aprocess that utilizes high pressure air rather than water as the fluid to remove the rock fragmentsand cool the drill bit when drilling through rock. Its principal environmental benefit is that lesswater is utilized during the drilling process and dry rock fragments are returned to the surfacerather than a slurry of water, drilling mud, and rock fragments. Air drilling also reduceswastewater generation and subsequent treatment (Lien and Manner 2010), but it cannot alwaysbe done safely—especially under conditions in which excessive subsurface pressures and flowsmay be encountered that cannot be effectively be controlled without the use of a drilling liquid.In addition to cooling and cleaning the drill bit and bringing cuttings to the surface, the use ofdrilling mud serves another important purpose: the density (i.e., weight of the fluid volume) ofthe mud effectively controls the formation pressures; well pressures can be held in check as longas the mud weight is sufficient to prevent flows from the formations being drilled. As higherpressures are encountered in deeper formations, it is therefore necessary to increase the muddensity to offset those pressures (King 2012).Maryland’s current oil and gas regulations state that “drilling liquid may be required when thereis insufficient geological data to safely drill with air as the circulating medium”8. Maryland alsorequires that when drilling on air is permitting, sufficient liquid shall be available on-site to killany unexpected flow from a particular well9. Maryland should consider the experiences gainedby other states and permit air drilling of any holes (i.e., the conductor hole and surface hole)above the USDW protective depth [which API (2010) considers to be 100 ft below the deepestUSDW encountered while drilling10), although the current regulations should be retained so thatair drilling can be permitted on a case-by-case basis.11 If and until the freshwater/saline waterinterface is mapped in Maryland, the state will have to rely on operators to determine when theUSDW protective depth has been reached while drilling the surface hole (likely on the basis ofdata obtained from geophysical logs from a particular borehole). Casing setting depths should be8 COMAR 26.19.01.10.F9 COMAR 26.19.01.10.I10 Maryland’s current oil and gas regulations require that a string of surface casing be installed in a hole which is at least 100 ft below the deepest known stratum bearing freshwater or the deepest known workable coal bed, whichever is deeper (COMAR 26.19.01.10.O(4)).11 On the other hand, it is important to case and cement any surface hole prior to drilling into hydrocarbon-bearing flow zones or zones which contain waters with TDS concentrations exceeding 10,000 mg/L to avoid contamination of USDW as recommended in Ohio’s regulations. For this reason, in areas where USDW cannot be adequately mapped, Ohio’s oil and gas regulations also allow for use of a conductor casing through the deepest useable water zone that is first cemented to the surface, followed by setting and cementing of a surface casing string through water zones that may include brackish or brine bearing zones. 3-6

58. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandspecified in the drilling plan, but the actual lengths of these strings can be adjusted based on fieldmeasurements and data collected during the drilling operation (API 2009a).We recommend Maryland develop regulations on wellbore diameter to ensure adequate spacingfor equipment and instrumentation that will need to be run into the wellbore and for an adequatethickness of cement (i.e., a “sheath”) inside the annular space. Ohio has some excellentregulations that require that the diameter of each section of the wellbore in which casing will beset and cemented to be at least one inch greater than the outside diameter of casing collar to beinstalled, unless otherwise approved12. Ohio also requires that any wellbore diameter shall beconsistent with manufacturers recommendations for all float equipment, centralizers, packers,cement baskets, and any other equipment that will need to be run into the wellbore13.C. Casing and cementingCasing and cementing are critical elements of any well construction that must be properlydesigned and engineered to ultimately serve their primary purpose of providing well integrityand isolation from surrounding subsurface formations while providing a pathway by which thegas can be safely extracted over the life of the well. For this reason, both API and all five statesthat we reviewed have very lengthy descriptions of practices and standards that should governthese important well construction tasks. The steel casing must be capable of withstanding all theforces that are exerted on it while running it into a hole, as well as during subsequent cementingand hydraulic fracturing operations. Similarly, cementing is used to provide isolation ofsubsurface flow zones, provide structural support of the well, and protect the casing fromcorrosion. The cement must also be able to contain all pressurized fluids during all phases ofdrilling and operation of the well. Operators (including company engineers who design the wellcasings, their supervisors, and any drilling subcontractors involved in casing installation andcementing) bear the primary responsibility for ensuring that these critical tasks are carried outproperly. Therefore, as noted in Section A, a critical element of a properly-executed wellconstruction plan is a “casing and cementing plan” that is required by all five states that weexamined. In many cases, state regulations require that standards (e.g., compressive strength ofthe cement) must be consistent with those recommended by API.Without detailed geological characterization (“cross-sections”) of the subsurface strata inwestern Maryland (including depths that various formations will likely be encountered, depths ofthe USDW/saltwater interface, etc.), it is very difficult to make anything but generalrecommendations for setting and cementing casing strings in place. Based on anticipated depthsto the target formation, we believe it is likely that Marcellus shale gas wells in Maryland willnormally require four casing strings (i.e., conductor casing, surface casing, intermediate casing,and production casing). All steel casing used in a Marcellus shale gas well should bemanufactured to API specifications and meet strict requirements for compression, tension,collapse, and burst resistance, quality, and consistency (including API Spec 5CT); casing shouldalso be designed to withstand all anticipated hydraulic fracturing pressures, production pressures,and corrosive conditions expected to be encountered. Used or reconditioned casing would onlybe used if it is shown to meet API standards for new casing (API 2009a). Casing and coupling12 Ibid.13 Ibid. 3-7

59. K.N. Eshleman & A.J. Elmore (2013) Chapter 3threads should also meet API standards, and casing strings should be assembled to the correcttorque specifications to ensure leak-proof connections. Casing centralizers should be used toproperly center the casing in the hole and provide for good mud removal and cement placementin the form of a continuous sheath around the casing string. API lists recommended types ofcentralizers and has various formulae for determining the type, number, and best placement ofcentralizers along a particular casing string (API 2009a). Maryland should require that operatorsuse casing that meets the high standards recommended by API, as well as a sufficient number ofcentralizers to properly center the casing in each borehole.API also provides good recommendations of materials and practices for ensuring that the variouscasing strings are properly cemented in place and can provide the desired zonal isolation ofdifferent formations, including complete isolation of USDW. Best practices are for an operator to(a) provide notice to the appropriate regulatory agency at least 24 hours prior to thecommencement of any cementing operations; and (b) maintain a copy of the cementing recordsat the well site during the drilling and completion of the well. Cementing is best achieved bypumping a cement mixture (or “slurry”) down inside the casing string being cemented andcirculating the cement mixture back up the outside of the casing (i.e., between concentric rings ofcasing or between the outermost casing and the borehole wall). Top and bottom wiper plugs areused to minimize mixing of the cement with drilling fluids inside the casing while the cement isbeing pumped. Zonal isolation and integrity of the well to minimize migration of fluids throughthe annulus are highly contingent on complete displacement of the drilling fluid by the cementmixture; complete and tight filling of the annulus with the cement mixture to the proper heightabove the bottom of the hole; absence of voids; and good bonding with the casing strings andborehole walls (API 2009a). Appropriate testing of cement should always be carried out by theservice company to ensure that the mixture meets the criteria specified for the specificapplication. It is recommended that all surface casings be cemented with a continuous columnfrom the bottom of the casing to the surface.Most of the states that we reviewed have established recommended standards and minimumcompressive strength values for cement used in oil and gas wells, and describe how tests ofcement should be conducted (i.e., API RP 10 B-2 “Recommended Practice for Testing WellCements”). As in most states, New York has proposed cementing the surface casing by thepump and plug method with circulation to the surface, with a minimum of 25% excess cementpumped, with appropriate lost circulation materials; testing of the mixing water for pH andtemperature prior to mixing; cement slurry preparation to the manufacturer’s or contractor’sspecifications to minimize free water in the cement; and no casing disturbance after cementinguntil the cement achieves a calculated compressive strength of 500 pounds per square inch (psi)(NYSDEC 2011). Similarly, in Ohio, cemented casing strings shall remain static until all cementhas reached a compressive strength of at least 500 psi before drilling the plug, or initiating anyintegrity testing14.While cementing of both the conductor and surface casing strings should normally be completedfrom top to bottom, there are likely to be situations in which large underground voids areencountered during drilling that preclude circulation of cement back to the surface. Under thesecircumstances, it may be possible to perform a cementing operation from top to bottom (i.e., a14 Ibid., 4 3-8

60. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland“top job”) on a conductor casing, but this approach is normally not recommended because ofdifficulties in isolating the various water-bearing formations and thus protecting USDW asdiscussed in Chapter 1. In one of its publications, API notes that “a top job should be done onlyas a last resort” (API 2010). If it is determined that a “top job” will be necessary, then theconductor casing should be installed as deeply as possible to protect all USDW, and it should beabsolutely required that the surface casing be fully cemented from bottom to top.API (2009) recommends that cementing of intermediate casing should also be done in the normalmanner (i.e., bottom to the surface), but notes that there may be situations where this technique isunnecessary (e.g., where the surface casing string is fully protecting the USDW) or is inadvisable(e.g., where attempts to do so result in lost circulation of the slurry). According to King (2012),in cases of very long intermediate casing strings, cementing the full casing string may be ill-advised due to the risks of fracturing formations by the pressure exerted from the weight of acolumn of cement (nearly twice the weight of an equivalent column of water). On the other hand,cementing an intermediate string to an insufficient height may leave the annulus exposed tohigher pressures from non-isolated gas-containing shales and coals that could provide a pathwayfor migration of gas into the outer annulus and into overlying freshwater zones (King 2012).While we were able to find relatively few published studies that explicitly address the problem offreshwater contamination resulting from overpressurization and flow through the annulus, thereis one older study that addressed this problem using a modeling approach (Harrison 1985). Amore recent paper hypothesizes that high methane concentrations in drinking water aquifers inPennsylvania are most likely attributed to annular overpressurization resulting from leaky wellcasings rather than from hydraulic fracturing (Osborn et al. 2011), although other mechanismsare at least as plausible (Davies 2011, Jackson et al. 2011). Data from Warner et al. (2012) mayalso be consistent with the same mechanism whereby Marcellus Formation brine has seeped intoshallow aquifers in Pennsylvania. We cannot make a clear recommendation with respect toinstalling and cementing intermediate casing strings; this is a situation where the best design andconstruction practices will be determined, in part, by the specific geological conditions that areencountered while drilling in western Maryland. At a minimum, however, an absoluterequirement should be that all flow zones (including USDW) must be fully protected through theuse of cemented intermediate well casings. Where this cannot likely be accomplished with asingle casing string, the use of multiple strings should be favored in the well design (even if thisresults in greater costs in casing and cementing).Problems encountered in cementing of gas well casings have significant implications for upwardcontaminant migration into USDW; a recent white paper provides a description of severalmechanisms by which oil and gas wells develop fluid leaks and lose their structural integrity(Ingraffea, 2012): (1) repeated pressurization of casings with open-annulus sections; or (2) highgas pressures encountering curing cement or entering open-annulus sections. Related to theseproblems, loss of integrity due to poor cementing can also be attributed to: (1) poor cementplacement (i.e., failure to displace the mud prior to cementing or failure to generate a sufficientheight of cement within the annulus to fully cover flow zones); (2) lack of centralization of thecasing string; and (3) from gas migration through the cement as it sets in place (King 2012). Inparticular, the latter problem has apparently been known for decades, but many operators areunaware of the hazards that gases create if they are allowed to migrate sufficiently up a settingcement column, establishing a network of linked subchannels. Mud channels formed in setting 3-9

61. K.N. Eshleman & A.J. Elmore (2013) Chapter 3cement can allow for gas or fluid migration through the annulus if these voids are continuous.Fortunately, large voids or channels are not typically continuous over long distances, but micro-annulus (hairline) cracks that allow for such migration must be detected through well logging(see Section D) and addressed accordingly (King 2012).D. Well loggingUseful geophysical data to support MSGD in Maryland would likely be obtained from a varietyof sources, particularly: (1) published geological maps of the region from USGS or privatesources; (2) available well log records from previous gas and water well drilling in Maryland andsurrounding states; (3) results from seismic refraction tests conducted as part of explorationactivities; and (4) results from test (“pilot hole”) drilling (the latter providing information forimproving stratigraphic interpretation specifically through calibration of seismic data). Data fromthese sources would likely be sufficient to: (1) identify subsurface drilling hazards; (2) accuratelyassess the location of the target zone; and (3) enable the design of production wells and detailedwell plans. Once drilling for production actually begins, there are many types of data that wouldbe collected through well logging techniques to provide detailed records of subsurface propertiesactually encountered in the well construction process.Open-hole logging is a method used in borehole geophysics that is conducted after drilling thehole, but before any casing is installed. Open-hole logging can provide important information onthe specific depths of various formations encountered during the drilling process—and is thusvery important in optimizing the well design and drilling operations. Drilling each hole to thecorrect depth theoretically allows casing strings to be installed at optimal locations to achievemaximum well integrity (API 2009a). Logging while drilling (LWD) technology was initiallydeveloped in the 1970s, but the technology now allows for most “open-hole” measurements to bemade without lowering a suite of instruments into the borehole as part of a “wireline”. Witheither LWD or traditional wireline technology, it is possible to accurately determine formationproperties from gamma ray logs (lithology), electrical resistivity logs (hydrocarbons), neutronporosity logs (liquid-filled porosity), and density logs (bulk density)—among others; boreholecaliper logs provide measurements of the size (i.e., diameter) and shape of the borehole along itslength that are crucial in estimating cement volumes. Mud logging is another boreholegeophysical technique which is most commonly used in the petroleum industry to determine theconcentration of natural gas being brought to the surface with the drilling mud. Modernmeasurement while drilling (MWD) technology allows information on natural gas levels to beobtained near the drill bit, thus providing an additional level of safety for rig workers in the eventthat levels are observed to reach dangerous levels. It is likely that all of these types of welllogging would be used in MSGD in western Maryland.Other types of well logging occur after cementing the casing, including gamma ray logging andcement bond logging (CBL). The objectives of a cased-hole logging program are to determinethe exact location of the casing, the casing collars, and the integrity of the cement job (especiallyas a function of location relative to various subsurface formations). CBL is an acoustic techniquethat works by transmitting a vibration and then recording the amplitude of the arrival signal at adetector. Casing that is not encased in cement produces a relatively high amplitude acousticsignal because the sonic energy is not very well absorbed. Conversely, casing with a good sheathof cement throughout the annular space produces a much smaller amplitude signal because the 3-10

62. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandsonic energy would be much better absorbed. A variable density log (VDL) provides a graphicalrepresentation of the receiver waveform (API 2009a). Finally, newer equipment used by bondlogging service companies apparently has the capacity to do segmented radial cement bondlogging (SRCBL)—a technique for determining the presence and location of “mud channels” inthe cement that would be indicative of poor zonal isolation. Such channels, if extensive orcontinuous, could provide a pathway for unintended gas or liquid flow within the annulus (King,2012). An SRCBL can be combined with equipment for gamma ray logging, casing collarlogging, and neutron logging during a single descent.15 Additional information on the varioustypes of cement evaluation tools that are available can be found in API TR 10TR1.We found relatively little agreement among the states as to which well logging techniquesconstitute best practice. Apparently neither Pennsylvania nor Ohio require any well logging(either open-hole or cased-hole), while West Virginia requires only a CBL. Colorado requiresthat operators run a minimum of a: (1) resistivity log with gamma-ray or other approvedpetrophysical logs that adequately describe the stratigraphy of the wellbore; and (2) a CBL on allproduction casing or, if a production liner is used, on the intermediate casing. Colorado alsorequires that open hole logs shall be run at depths that adequately verify the setting depth ofsurface casing and any aquifer coverage and that all logs run shall be submitted with a wellcompletion or recompletion report to the regulatory authority. New York State has proposed thata radial cement bond evaluation log or other approved method should be use to verify the cementbond on the intermediate casing and the production casing (NYSDEC 2011). The best practicewould utilize modern open-hole well logging methods to help fine tune casing placement andcharacterize flow and hydrocarbon zones, perhaps mud logging to determine levels ofhydrocarbons in real-time during drilling, and SRCBL, casing collar logging, and gammalogging as part of a cased-hole program. We found virtually no information on possible remedialactions that can be taken by an operator in the event that problems with cement bond integrityare identified through the logging process. If remedial actions cannot fully resolve cement bondintegrity issues, the operator should have no recourse but to correctly plug and abandon the wellin accordance with state regulations.Maryland’s current regulations16 apparently require only electrical induction and gamma ray“open-hole” logging to determine the depth of freshwater, but they also require operators tomaintain a detailed driller’s log book17 and provide MDE with a completion report within 30days after drilling, stimulating, and well testing have been completed18. Such completion reportsinclude, among other items, information on the lithology of the penetrated strata, generalizedcore descriptions, estimates of porosity and permeability of formations, and copies of all logs runof the well. Maryland should consider amending its regulations to require SRCBL (or equivalentcasing integrity testing) and other types of logging (e.g., neutron logging) to assist withdetermining the depth of freshwater as part of a cased-hole program.15 e.g., Tetra Technologies (2012); http://tetratec.com/index.asp?page_ID=30916 COMAR 26.19.01.10.O(3)17 COMAR 26.19.01.10.R18 COMAR 26.19.01.10.V 3-11

63. K.N. Eshleman & A.J. Elmore (2013) Chapter 3E. Pressure testingAPI (API 2009a) and the five states that we reviewed all call for testing of the various casingstrings after the cement has achieved the appropriate compressive strength during a pre-specifiedwait-on-cement (WOC) period, but prior to drilling out. These tests are known as casing pressuretests and are performed to ensure that the integrity of each casing string is adequate to meet thewell design and construction objectives. Recommended pressures and holding times for thesetests were not consistent among the states, and API (2009) does not provide specificrecommendations. In West Virginia, for example, the regulations only state that an operatorshould conduct the test at a pressure more than 20% greater than the pressure expected to beexerted on the casing. Similarly, in Pennsylvania, to pass a casing pressure test, the casing musthold the anticipated maximum pressure to which the casing will be exposed for 30 minutes withnot more than a 10% decrease; certification of the pressure test shall be confirmed by entry andsignature of the person performing the test on the driller’s log. API also recommends thatformation integrity tests (also known as “shoe tests” or “leak-off tests”) be performed afterdrilling out both the surface and intermediate casings (API 2009a). Best practice would clearlycall for use of pressure testing of Marcellus shale gas wells in Maryland, with specific criteriaand technical details governing the conduct of such tests likely established through consultationwith industry. Maryland’s current regulations19 with regard to pressure testing of cementedcasings are even less specific than those established by neighboring states and appear to be inneed of revision.F. Blow-out preventionBlow-out prevention equipment (BOPE) on a rotary drilling rig is a pressure control systeminstalled at the top of the surface casing that is designed specifically to contain and control a“kick” (i.e., an unexpected pressure resulting in the flow of formation fluids into the wellboreduring drilling operations). BOPE consists of four parts: 1) a blow-out preventer stack, 2) anaccumulator unit, 3) a choke manifold, and 4) a kill line. Blow-out preventers are manually orhydraulically operated devices. Within the blow-out preventer there may be a combination ofdifferent types of devices to seal off the well. A suitable BOPE should have at least tworedundant (and operational) mechanisms for preventing a blow-out. Pipe rams contain two metalblocks with semi-circular notches that fit together around the outside of the drill pipe when it isin the hole to block movement of fluids around the pipe. Blind rams contain two rubber facedmetal blocks that can completely seal off the hole when there is no drill pipe in it. Annular or"bag" type blowout preventers contain a resilient packing element which expands inward to sealoff the hole with or without drill pipe. To be effective, BOPE systems must be maintained and inproper working order during operations; a BOPE testing program must be employed on a regularbasis to ensure that the system is functioning properly if and when it is needed (NYSDEC 2011).All BOPE should have a working pressure rating that exceeds the maximum expected surfacepressure; training exercises or drills should be held as necessary to ensure crew familiarity andthat the BOPE is in good working order (API 2009a).BOPE is an example of a temporary mechanical barrier for preventing loss of well controlthrough annular flows (API 2010). It should be kept in mind that BOPE is not the only type ofbarrier used during drilling and completion. Columns of fluids (e.g., drilling fluids, cement19 COMAR 26.19.01.10.S(3) 3-12

64. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandslurries, fracturing fluids, etc.) are considered hydrostatic barriers, because they can providehydrostatic pressure that exceeds the pore pressures of potential flow zone, thus maintainingcontrol of flow in the annulus. Set cement is usually the ultimate barrier element, but itscompetency should be carefully assessed prior to removing a mechanical barrier such as BOPE.There are many other types of mechanical barriers that are used by the oil and gas industry, butwe lack the technical capacity to make recommendations as to which specific types of barrierswould be employed in developing Maryland’s Marcellus shale gas resource.Pennsylvania requires the use of BOPE for drilling production wells for natural gas inunconventional formations20, but Ohio and West Virginia do not. In Colorado, the use of BOPEis required only when drilling in high density areas; otherwise it is at the discretion of theregulatory agency (COGCC). In Colorado, pressure testing of the casing string and eachcomponent of the blowout prevention equipment (if blowout prevention equipment is required)should be conducted prior to drilling out any string of casing except the conductor pipe. Theminimum test pressure should be 500 psi, and that pressure should hold for 15 minutes withoutpressure loss in order for the casing string to be considered serviceable. Use of BOPE with twoor more redundant mechanisms should be considered a best practice for MSGD in Maryland.G. Completing and hydraulic fracturingThe production casing is normally run to the total depth of the well and—once cemented—isintended to provide: (a) total zonal isolation between the production formation and all othersubsurface formations; (b) a continuous conduit to the surface for pumping hydraulic fracturingfluids into the production formation without affecting other subsurface formations; (c) acontinuous conduit for containing and transporting hydrocarbons between the production zoneand the surface; and (d) a secondary barrier for the production tubing and packer that are used inthe final completion step (API 2009a). In the absence of using an intermediate string, New YorkState recommends cementing of the production casing all the way to the surface (NYSDEC,2011). Similarly, in Ohio, when cementing the production string of a well that will be stimulatedby hydraulic fracturing, and the uppermost perforation is less than 500 ft below the base of thedeepest USDW, sufficient cement shall be used to fill the annular space outside the casing fromthe seat to the surface. Since we explicitly recommended against drilling in situations wherethere is less than 1,000 vertical ft between USDW and the production formation, this optionshould not apply in western Maryland. Recommendations by API and used by the state of Ohioboth call for cementing of the production casing to a depth at least 500 ft above the highestformation in which hydraulic fracturing will be performed, however. Ohio calls for use of acement slurry that is designed to control annular gas migration consistent with recommendedmethods in API (2010). However, both API (2009a) and Ohio regulations allow for “open-hole”completions and the use of production liners in some circumstances. Maryland regulators willhave to work with industry to carefully evaluate the pros and cons of these different completionoptions.Hydraulic fracturing (sometimes referred to as “fracking”) is a well stimulation techniqueemployed by the oil and gas industry to increase the permeability of a hydrocarbon-bearing20 25 Pa Code § 78.72. Use of safety devices—blow-out prevention equipment (provisions adopted July 28, 1989, effective July 29, 1989, 19 Pa.B. 3229; amended February 4, 2011, effective February 5, 2011, 41 Pa.B. 805). 3-13

65. K.N. Eshleman & A.J. Elmore (2013) Chapter 3formation of low permeability and provide a pathway for hydrocarbons and other fluids to flowmore easily out of a formation and, ultimately, into a wellbore. In horizontal gas wells inunconventional formations, high volume hydraulic fracturing (HVHF, or just HF for short) isnormally considered an essential part of the completion process (since gas production would betoo low to justify the costs of drilling and completing a well in such formations). During the HFprocess, fluid (normally comprised of water and a variety of chemical additives to reduce theviscosity of the water; prevent microbial growth; disinfect the water; reduce interfacial tension;inhibit scale; etc.) is pumped into the production casing of a completed well, throughperforations made in the casing, and into the target formation at pressures great enough tofracture the rock (King 2012). As fluid injection continues, fractures grow throughout the targetformation; as the fractures grow, a proppant (sand) is added to the fluid. Once pumping stops andexcess pressure is removed, the fractures attempt to close under the weight of the overlyingstrata, but the proppant keeps the fractures open—effectively increasing the permeability and,ultimately, the rate of fluid migration out of the formation (API 2009a).Technical concerns about hydraulic fracturing have tended to focus on three major issues: (1)transport of HF fluids and other contaminants (e.g., methane gas) from a fractured Marcellusformation into natural fractures where they could be transported long distances (thousands of ft)to USDW (Myers 2012, Saiers and Barth 2012); (2) induced seismic activity associated with theHF process; and (3) the specific chemical additives used in making up the HF fluid (and theirtoxicity). While well beyond the scope of our study of best practices to fully explore, we believethat there has been insufficient scientific study of the first issue to allow any firm conclusions tobe drawn. Such studies would undoubtedly need to consider the full gamut of pathways (e.g.,improperly cemented well casings) by which contaminants—both gases and liquids—eitherintroduced or native to overpressurized formations such as the Marcellus could impact USDW.Certainly this complexity warrants continued study, both of new methods in well engineeringand well completion as they become available, but also environmental data demonstrating wellisolation has been successful in protecting the USDW. With respect to the second issue ofinduced seismic activity, we cite the recently published National Research Council (NRC) reportwhich concluded that: (1) the process of hydraulic fracturing a well as presently implemented forshale gas recovery does not pose a high risk for inducing felt seismic events; and (2) injection fordisposal of waste water derived from energy technologies into the subsurface does poses somerisk for induced seismicity, but very few events have been documented over the past severaldecades relative to the large number of disposal wells in operation (NRC 2012 ). Best practicesfor selecting, handling, and disposing of HF chemicals are discussed in Chapter 4.H. Use of well development techniques other than hydraulic fracturingNew York State has performed a technical review of possible future alternative well stimulationtechniques to water-based hydraulic fracturing (which could largely eliminate the need fortrucking water to well sites and presumably produce less waste), but at present these techniquesappear to be limited to demonstration or pilot projects in the United States and none can beconsidered a best practice. Unfortunately, we lack the technical capacity and necessaryexperience to evaluate the potential of any of these methods to replace water-based HF in thefuture, but we provide the following material excerpted from New York’s draft assessmentdocument (NYSDEC 2011) as an aid to state regulators that may wish to explore these options inthe future: 3-14

66. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland • Liquid CO2. The use of a liquid CO2 and proppant mixture obviously reduces the use of other additives. Once CO2 vaporizes, it leaves only the proppant in the fractures. The appropriate level of environmental review for this alternative, if proposed in New York, would need to be determined at the time of application. • Nitrogen-based foam. Nitrogen-based foam fracturing was used in vertical shale wells in the Appalachian Basin until recently. Nitrogen gas is unable to carry appreciable amounts of proppant and the nitrogen foam was found to introduce liquid components that can cause formation damage. • Liquified petroleum gas (LPG). More recently, New York looked into the use of LPG (primarily propane) which has the advantages of carbon dioxide and nitrogen noted above; additionally, LPG is known to be a good carrier of proppant due to the higher viscosity of the propane gel. Further, mixing LPG with natural gas apparently does not “contaminate” natural gas and the mixture may, therefore, be flowed directly into a gas pipeline and separated at the gas plant and recycled. LPG’s high volatility, low weight, and high recovery potential make it a particularly good fracturing agent. Use of LPG as a hydraulic fracturing fluid also inhibits formation damage which can occur during hydraulic fracturing with conventional fluids. Using propane not only minimizes formation damage, but also eliminates the need to source water for hydraulic fracturing, recover flowback fluids at the surface, and dispose of the flowback fluids. As a result of the elimination of hydraulic fracturing source water, truck traffic to and from the well site could be greatly reduced. Finally, since LPG is less reactive with the formation matrix, it is less likely that this technique would mobilize constituents that are ultimately discharged with the flowback fluid (NYSDEC 2011).I. Determining the extent of induced fracturesThere are two methods that can be used for determining the extent of induced vertical fracturegrowth by hydraulic fracturing. The first technique is through the use of either surface ordownhole tiltmeters that are capable of measuring extremely small changes in the inclination ofthe Earth’s surface from level. Historically, tiltmeters have been used extensively for monitoringvolcanoes, the responses of dams to filling, and small movements of potential landslides, butextremely sensitive (nanoradian) surface and downhole tiltmeters developed in the 1970s byPinnacle (a Halliburton subsidiary) now allow for fracture mapping from either offset wells orfrom the surface; a new generation method can apparently map induced fractures from an activefracture well in real-time (API 2009a). The second technique produces a map of vertical fractureheight growth based on data from passive micro-seismic monitoring that is capable oftriangulating the sounds made by rock breaking up during shear fracturing. Micro-seismicmeasurements are typically made with: (1) a 200 to 400 ft long set of geophones placed in anoffset well located within a few hundred ft of the well being fractured; or (2) an array ofmicrophones placed at the surface. Micro-seismic monitoring makes it possible to determinesuch critical hydraulic fracturing parameters as vertical extent, lateral extent, azimuth, andfracture complexity (API 2009a).Micro-seismic data (from more than 3,000 HF applications) from Pinnacle (Fisher 2010,reprinted by King 2012) shows vertical fracture growth in hydraulically-fractured Barnett Shalewells typically extends hundreds to thousands of ft above and below the frac depth, but in no 3-15

67. K.N. Eshleman & A.J. Elmore (2013) Chapter 3case was fracture growth observed closer than 2,800 vertical ft from USDW. Comparable data(from more than 300 fracs with micro-seismic data) from Pinnacle (Fisher 2010, reprinted byKing 2012) suggest that the closest measured approach of Marcellus shale fractures inPennsylvania to USDW was 3,800 vertical ft. As discussed in Chapter 1, these results generallysupport our recommendation that Maryland follow guidance from New York’s experience withunconventional shale gas development and not permit MSGD (or any other unconventional gasdevelopment) where the target formation occurs within 1,000 vertical ft of USDW or within2,000 vertical ft of the ground surface (NYSDEC 2011).Best practice is not to employ tiltmeter surveys or microseismic on every well, rather it is mostcommonly used to evaluate new techniques, refine the effectiveness of fracturing in new areas orformations, and in calibrating computer models of the fracturing process (API 2009a). There areno micro-seismic monitoring protocols or criteria established by regulatory agencies that arespecific to HVHF. Nonetheless, operators can employ micro-seismic methods to monitor thehydraulic fracturing process and thus optimize the results for successful gas recovery. It is in theoperator‘s best interest to closely control the hydraulic fracturing process to ensure that fracturesare propagated in the desired direction and distance and to minimize the materials and costsassociated with the process. Best practice would suggest that operators place multiple receiverson a wireline array in one or more offset borings (e.g., a new, unperforated well or an older wellwith production isolated) or in the treatment well during the HVHF process. At least one receivershould be in the treatment zone, with another located above and one below this zone. Maximumobservation distances for microseisms should be within approximately 2,500 ft of the treatmentwell, with the distance dependent on formation properties and background noise levels.Locations are triangulated using the arrival times of the various p- and s-waves to the receivers inthese wells, and using the formation velocities to determine the location of the microseisms. Amulti-level vertical array of receivers can be used if only one offset observation well is available.The induced fracture is interpreted to lie within the envelope of mapped microseisms (NYSDEC2011). We highly recommend that a sufficient number (at least tens) of tiltmeter or seismicsurveys be performed as part of MSGD in Maryland, so that the extent, geometry, and locationof Marcellus fracturing can be adequately characterized. The goal would be to feed usefulinformation back to the operators, so that subsequent hydraulic fracturing could be conductedmore safely and effectively. Data from such surveys in Maryland (and other states) would also bedeemed crucial in evaluating whether HVHF might eventually be safely conducted in locationswhere the target formation is located within 2,000 ft of the surface.J. PluggingThe purpose of plugging a well is to: (1) prevent interzonal migration of fluids; (2) preventcontamination of freshwater aquifers, surface soils, and surface waters; and (3) conservehydrocarbon resources either in the production zone or in potential production zones. Generally,contamination by an improperly plugged or abandoned well can occur in two ways: (1) theabandoned well can act as a conduit for fluid flow between penetrated strata, into USDW, or tothe surface; or (2) contaminated water can enter the abandoned wellbore at the surface andmigrate into USDW. Such contamination can be prevented by properly plugged a well. It shouldbe noted that while plugging operations can prevent an abandoned well from becoming a conduitfor contamination, well construction and completion methods also contribute to the prevention of 3-16

68. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandcontamination (API 2009b). Plugging should be considered a critical element of the welldecommissioning process that also includes land reclamation.Well plugging operations are focused primarily on protecting USDW, isolating downholeformations productive of hydrocarbons or used for injection, and protecting surface soils andsurface waters. A surface plug prevents surface water runoff from seeping into the wellbore andmigrating into USDW cement plugs isolating hydrocarbon and injection/disposal intervals and aplug at the base of the lowermost USDW accomplishes this primary purpose. Surface water entryinto an abandoned well is a concern because the water may contain contaminants fromagricultural, industrial, or municipal activities. It is, therefore, recommended that operators set acement plug at the base of the lowermost freshwater aquifer or USDW during plugging andabandonment operations applicable to the well. (NOTE: The cement plugs also work to protectsurface soils and water from wellbore fluids by confining those fluids in the well.) In addition tothe cement plugs described herein, many state and federal regulatory agencies require cementplugs across the base of the surface casing and in, or between, each producing and potentialproducing zone (API 2009b).All formations bearing usable quality water, oil, gas, or geothermal resources (e.g., coal seams)should be protected and/or isolated. The prevention of gas or fluid migration to other zones or tothe surface is of primary importance. Open-hole plugs, casing plugs, or cement squeezed throughcasing perforations will isolate the target formations in most cases. However, special procedures,such as perforating casing and circulating cement, may be necessary to isolate that potentialproduction or injection formations behind any uncemented or poorly cemented casing. It isimportant to prevent interzonal flow in an abandoned well so that such cross-flow does notinterfere in the commercial exploitation of the zones through nearby wellbores. The operatorshould also: (1) set the required surface plugs; (2) remove the wellhead; (3) weld a steel plate onthe surface casing stub; (4) fill in any well cellar; and (5) level the area. Casing strings left in thewell should be cut off at least 3–6 ft below ground level (API 2009b).Pennsylvania21 and Colorado22 have enacted regulations governing plugging of gas wells thatappear to be consistent with API’s recommended practices, but West Virginia and Ohio have not.Maryland also has what appear to be excellent regulations23 that are consistent with APIrecommendation for plugging of wells. Given the long expected time lags (of the order of 30years or more) between drilling and well decommissioning, the biggest problem that weanticipate with plugging of Marcellus wells in Maryland will be ensuring that the appropriateparty is held accountable and has sufficient assets to do so. The costs associated with pluggingwells that were poorly constructed in the first place can be extremely high (Mitchell amdCasman 2011), reinforcing the need to ensure that any Marcellus shale gas wells in Maryland areconstructed to the highest standards.21 25 Pa. Code § 78.92 (relating to wells in coal areas—surface or coal protective casing is cemented); 25 Pa. Code § 78.93 (relating to wells in coal areas—surface or coal protective casing anchored with a packer or cement); 25 Pa. Code § 78.94 (relating to wells in noncoal areas—surface casing is not cemented or not present); 25 Pa. Code § 78.95 (relating to wells in noncoal areas—surface casing is cemented); and 25 Pa. Code § 78.407 (relating to plugging gas storage wells).22 COGCC Rule 319, Abandonment.23 COMAR 26.19.01.12 3-17

69. K.N. Eshleman & A.J. Elmore (2013) Chapter 3K. Key recommendations3-A A best practice for anyone proposing to operate in Maryland should be adoption of API’s extensive guidelines for well planning—at least those elements that are clearly relevant to onshore development. Pre-permit site review should also be required.3-B Site selection is a critical aspect of well planning for multiple reasons discussed throughout the report. As discussed in Chapter 1, we are particularly concerned about drilling in areas where there is a high probability of encountering large underground voids (e.g., caverns, caves, mine workings, abandoned wells, etc.) that have the potential to cause a loss of fluid circulation during drilling and impose additional risks during the cementing process. Such hazards are locally common in western Maryland and we recommend that sites with a high probability of encountering such hazards be avoided.3-C Surface casing must be fully cemented from the bottom to the surface to provide total protection of all USDW. There may be situations (e.g., very deep wells) where fully cementing the intermediate casing to the surface may not be required, however. At a minimum, an absolute requirement should be that all flow zones (including USDW) must be fully protected through the use of cemented intermediate well casings. Where this cannot be accomplished feasibly with a single casing string, the use of multiple casing strings should be favored in the well design.3-D Maryland should consider amending its regulations to require SRCBL (or equivalent casing integrity testing) and other types of logging (i.e., neutron logging) as part of a cased-hole program.3-E Best practice would clearly call for use of pressure testing of Marcellus shale gas wells in Maryland, with specific criteria and technical details governing the conduct of such tests likely established through consultation with industry. Maryland’s current regulations with regard to pressure testing of cemented casings are even less specific than those established by neighboring states and appear to be in need of revision.3-F Use of BOPE with two or more redundant mechanisms should be considered a best practice for MSGD in Maryland.3-G We recommend that a sufficient number of tiltmeter or micro-seismic surveys be performed as part of any MSGD in Maryland, so that the extent, geometry, and location of Marcellus fracturing can be adequately characterized across the entire region. The principal goal of this effort would be to feed useful information back to the operators, so that subsequent hydraulic fracturing can be conducted more safely and effectively. Data from such surveys in Maryland (and other states) would also be deemed crucial in evaluating whether HVHF might eventually be safely conducted in locations where the target formation is located within 2,000 ft of the surface.3-H Maryland also has what appear to be excellent regulations that are consistent with API recommendation for plugging of wells. Given the long expected time lags (of the order of 30 years) between drilling and well decommissioning, the biggest problem that we anticipate with plugging of Marcellus wells in Maryland will be establishing liability and ensuring that liable parties can be held accountable for performing this critical task. The costs associated with plugging wells that were poorly constructed in the first place can be 3-18

70. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland extremely high, which reinforces the need to ensure that any Marcellus shale gas wells in Maryland are constructed to the highest standards.L. Literature citedAPI. 2009a. Hydraulic Fracturing Operations--Well Construction and Integrity Guidelines (API Guidance Document HF1, First Edition), Washington, DC: American Petroleum Institute.API. 2009b. Environmental Protection for Onshore Oil and Gas Production Operations and Leases (API RP 51R, First Edition), Washington, DC: American Petroleum Institute.API. 2010. Isolating Potential Flow Zones During Well Construction (API Standard 65--Part 2, Second Edition), Washington, DC: American Petroleum Institute.Davies, R. J. 2011. Methane contamination of drinking water caused by hydraulic fracturing remains unproven. Proceedings of the National Academy of Sciences USA 108: E871.Fisher, K. 2010. Data confirm safety of well fracturing. American Oil and Gas Reporter, July issue.Harrison, S. S. 1985. Contamination of aquifers by overpressurizing the annulus of oil and gas wells. Ground Water 23:317-324.Ingraffea, A. R. 2012. Fluid migration mechanisms due to faulty well design and/or construction: an overview of recent experiences in the Pennsylvania Marcellus Play (unpublished white paper), Physicians Scientists & Engineers for Health Energy.Jackson, R. B., S. G. Osborn, A. Vengosh, and N. R. Warner. 2011. Reply to Davies: Hydraulic fracturing remains a possible mechanism for observed methane contamination of drinking water. Proceedings of the National Academy of Sciences, USA 103:E872.King, G. E. 2012. Estimating frac risk and improving frac performance in unconventional gas and oil wells (paper SPE 152596).Lien, A. M., and W.J. Manner. 2010. The Marcellus shale: resources for stakeholders in the Upper Delaware watershed region. Pinchot Institute for Conservation.Mitchell, A. L., and E. A. Casman. 2011. Economic incentives and regulatory framework for shale gas well site reclamation in Pennsylvania. Environmental Science & Technology 45:9506-9514.Myers, T. 2012. Potential contaminant pathways from hydraulically fractured shale to aquifers. Ground Water 50:872-882.National Commission on the BP Deepwater Horizon Oil Spill and Offshore Drilling. 2011. Deep Water: the Gulf Oil Disaster and the Future of Offshore Drilling (Report to the President).NRC. 2012. Induced Seismicity Potential in Energy Technologies (advanced copy released June 15, 2012), Washington, DC: National Research Council.NYSDEC. 2011. Revised Draft Supplemental Generic Environmental Impact Statement on the Oil, Gas, and Solution Mining Regulatory Program.Osborn, S. G., A. Vengosh, N. R. Warner, and R. B. Jackson. 2011. Methane contamination of drinking water accompanying gas-well drilling and hydraulic fracturing. Proceedings of the National Academy of Sciences, U.S.A. 108:8172-8176.Saiers, J. E. and E. Barth. 2012. Comment by James E. Saiers and Erica Barth on "Potential contaminant pathways from hydraulically fractured shale to aquifers". Ground Water 50:826-828.Warner, N. R., R. B. Jackson, T. H. Darrah, S. G. Osborn, A. Down, K. Zhao, A. White, and A. Vengosh. 2012. Geochemical evidence for possible natural migration of Marcellus Formation brine to shallow aquifers in Pennsylvania. Proceedings of the National Academy of Sciences, U.S.A. 109:11161-11166. 3-19

71. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland4. Protecting water resources1Water is central to the advancements in shale gas recovery that have revolutionized domesticnatural gas resources in the past decade. High volume hydraulic fracturing (HVHF) withchemically-amended water enables extraction of large reserves previously consideredeconomically unviable. Significant amounts of water are required for the process, and significantamounts of wastewater are produced. Wastewaters (commonly called flowback and productionwaters) are contaminated with anthropogenic chemicals associated with the hydraulic fracturingprocess and with naturally-occurring chemicals associated with the shale formation. The possibleimpacts of shale gas development on regional water resources (i.e., quantity and quality) must beconsidered at all phases of the life-cycle of well and gas field development. Figure 4-1 provides alife cycle representation for shale gas development at a single pad. The general concerns forfreshwater resources and the generation of wastewater are presented, aligned with when theyoccur in the life cycle of development. Some aspects of potential water impacts (e.g., landclearing and stormwater runoff) can be generalized as associated with the development of anyindustrial site. Other aspects (e.g., concerns with managing chemicals and preventing spills) arespecific to activities that use large quantities of chemicals in the open, frequently undersuboptimal climatic conditions. Finally, other concerns (e.g., concerns regarding casing andmanagement of produced brine) are specific to oil and gas drilling operations.A critical point shown in Figure 4-1 is that many potential impacts occur throughoutdevelopment until closure (e.g., generation and management of surface runoff from the site),while other concerns (e.g., the effect of drilling on groundwater resources) may occur during alimited period of time. Further, there are additional concerns at the level of the entire resourcedevelopment (the play) that must also be considered. The distributed nature of the activity (i.e.,potentially hundreds of locations in western Maryland) raises specific issues with respect towatershed-wide effects. For example, it is important to consider the cumulative impact of waterwithdrawals for multiple wells and multiple pads, as well as the total volume of wastewater thatwill be generated by the formation once many hundreds of wells are in production.This chapter provides a summary of recommendations based on a review of the actual andproposed best management practices (BMPs) for shale gas development (MSGD2) related towater acquisition and wastewater management in five states (WV, OH, PA, NY, and CO). Mostof the practices either adopted or considered by Pennsylvania and West Virginia are applicable toMaryland, where geology, hydrology, and topography are very similar. Some practices that havebeen routinely employed in western states (e.g., evaporative concentration of wastes in openimpoundments) are inappropriate for the mid-Atlantic region and cannot be recommended. Wehave also addressed some of the key regulatory and policy aspects in addressingwater/wastewater issues associated with MSGD in Maryland.1 Chapter co-authors: Jeanne M. VanBriesen, Ph.D., P.E. (Department of Civil and Environmental Engineering, Carnegie Mellon University, Pittsburgh, PA 15213); Keith N. Eshleman, Ph.D. (Appalachian Laboratory, University of Maryland Center for Environmental Science, Frostburg, MD 21532); and Andrew J. Elmore, Ph.D. (Appalachian Laboratory, University of Maryland Center for Environmental Science, Frostburg, MD 21532)2 As elsewhere in this report, MSGD refers to Marcellus shale gas development. However, our review of actual and proposed BMPs in the five states covered all shale gas development, regardless of the target formation. 4-1

72. K.N. Eshleman & A.J. Elmore (2013) Chapter 4 Life Cycle Phase Pre Pad Completion via Site Closure Drilling Production Development Development hydraulic fracturing and Restoration Surface water runoff may occur from stormwater and can affect fresh water resources at the surface and in aquifers. Erosion and Fresh Water and Drinking Water Issues Water is used in sediment controls will be employed to collect stormwater for use or treatment on site. chemical manufacture, sand extraction, and Water withdrawals from surface and Drill cuttings will be Temporary Fresh water energy generation ground water sources will occur. removed. impoundments will impoundments and necessary to Alternative waters may be used Temporary chemical be closed. Storage temporary water produce the (e.g., AMD, mine water, recycled storage tanks will be tanks will be transport systems equipment and produced water). removed. removed. may be built on site materials involved in or in the the process of shale development area. Fresh and recycled water will be The well bore will be gas extraction. trucked to the pad site. Partial reclamation sealed. of pad, including revegetation. Chemical storage Complete Drilling and completion equipment will tanks and other reclamation of pad be brought to the site and staged. Produced water will temporary facilities and all developed may be built. be stored on site area. Chemicals will be brought to the site. with regularly scheduled collection Removal of access via truck. roads and pipelines. Ground water resources will be isolated from drilling and completion activities through appropriate well casing. Produced water may be partially treated and reused on site. Waste Water Issues Wastewater is Stormwater and surface runoff will generate wastewater. Erosion and sediment controls will be employed to collect stormwater for use produced in chemical or treatment on site. manufacture, sand extraction and energy Early produced Produced water will Impoundment liners generation necessary Drill cuttings and water (flowback) and residuals will be be collected and to produce the drilling muds will be will be collected and collected and trucked to disposal equipment and collected for reuse may be treated on trucked to disposal. or treatment. materials involved in or disposal. site for reuse or the process of shale trucked off site for gas extraction. reuse, treatment or disposal. 12/6/2012 Figure 4-1. Water and wastewater issues across the life cycle of pad and well development for unconventional shale gas development. 4-2

73. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandA. Siting requirements: setbacks and restrictionsAs discussed in previous and subsequent chapters, site selection for well pads and wells is anextremely important aspect of MSGD. The first step to preventing negative environmentalimpacts is to make careful site selections and require adequate setbacks to reduce impacts tocritical water resources. While water has a number of different values that require consideration(some of these are discussed in Chapter 6), we primarily address water used for humans in thissection.In Allegany County (2011 estimated population of 74,692), 83% of the population uses surfacewater (community- provided) while 17% of the population uses groundwater (individual wells).In Garrett County (2011 estimated population of 30,051) 90% of the population uses groundwater (predominantly individual wells: 70%), while 10% of the population uses community-provided surface water. In Allegany County, two large drinking water systems (City ofCumberland and the City of Frostburg), one medium system (Midland-Lonaconing), and onevery small system (Rawlings Heights) treat their surface water. Three medium systems (EasternRegion Allegany, Lavale Sanitary Commission, Western Region Allegany) and two smallsystems (Bel-Air Pinto and Southern Region Allegany) purchase surface water from anotherprovider. Six very small systems (Barrelville, Green Ridge, Midlothian, Reckley Spring, andRocky Gap) use groundwater, and one other system (Mount Savage) uses groundwater under theinfluence of surface water. An additional three non-transient non-community systems and 36transient non-community water systems are predominantly on groundwater (only Rocky GapState Park uses surface water). For the City of Cumberland, the source water is the Lake Koonand Gordon reservoirs in Pennsylvania, part of the Evitts Creek watershed. The City ofFrostburg receives its water from the Piney Dam Reservoir in Garrett County, MD, as well asfrom two deep wells in the Pocono aquifer and a series of springhouses. These sources are mixedprior to treatment. The Midland-Lonaconing system uses several reservoirs (Midland Gilmore,Charlestown, Koontz) that are part of the Georges Creek watershed (Potomac River watershed)and all fed by headwater streams. This surface water is supplemented by several groundwaterwells that either pump into the reservoir or the plant.All public drinking water systems in Garrett County are small: two systems (Friendsville andOakland) are on surface water and three systems (McHenry, Mountain Lake Park, Grantsville)are on groundwater. Two very small systems (Bloomington and Kitzmiller) are on surface water,while the balance (Backbone Mountain, Crellin, Gorman, Meadow Mountain, Meadow Park,Savage Mountain, Accident, Deer Park, and White Oak) are on groundwater. An additional 11non-transient non-community systems and 75 transient non-community water systems are all ongroundwater.Sufficient water is impounded for the surface water plants in the region, but historical watersupply problems suggest vulnerability on quantity should additional withdrawals take place fromthe reservoirs or the tributaries that feed them at certain times of the year. Further, source waterassessment documents for Evitts Creek watershed indicate concerns with turbidity increasesassociated with rainfall events that would likely be exacerbated if development did not includeadequate sedimentation controls (PADEP 2003). Similarly, Piney Dam Reservoir exhibitselevated nutrients and sodium levels, likely due to agricultural runoff and development,increasing the risk of harmful algal blooms that challenge drinking water treatment systems 4-3

74. K.N. Eshleman & A.J. Elmore (2013) Chapter 4(Castro et al. 2001). As indicated by the assessment above, significant numbers of westernMaryland residents rely on groundwater for their domestic water use.Particularly in Garrett County, private well supply dependence (see Figure 4-2) suggests a strongneed for setbacks and siting criteria that can effectively reduce the risks to these resources posedby surface spills (or incorrect drilling and cementing techniques in well development). Therefore,public and private water supply well identification should be part of the initial permit applicationprocess. Setbacks from existing water wells should be incorporated into siting requirements.Setbacks should be selected based on the source (groundwater wells vs. surface water intakes)and based on the area of influence for a well (the region of the aquifer affected by the pumping)and the mixing zone for a surface water system. Large public system wells have more impact onthe aquifer, and thus, surface disturbance or accidental spills over a larger surface area couldaffect public system wells, necessitating larger setbacks. For surface waters, an upstream spillwill have the largest impact if it occurs close to the intake where natural dilution capacity will bethe smallest; thus surface water intake setbacks provide a buffer, usually called a mixing zone,for dilution of a spilled material upstream of an intake.Figure 4-2. Map of density of public and private wells in western Maryland. Note: 23 acre unit is equivalent to 106sq. ft.Setbacks for public and private wells in current regulations are variable, although it has beenrecommended that both Pennsylvania and New York establish 500 ft. setbacks for private wells.Such setbacks could be waived with owner’s permission. West Virginia and Pennsylvaniapresently enforce 1,000 ft. setbacks to surface intakes and groundwater wells used for publicwater supply systems, but it has been recommended that New York impose a 2,000 ft. setback 4-4

75. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandfor public system intakes (NYSDEC 2011). Therefore, based on our review of what is beingdone in other states, we recommend that a best practice for Maryland would be to establish aregulation of 500 ft. and 2,000 ft. setbacks (measured from the well pad, not from the individualwellbores) for private wells and public system wells, respectively, and a setback of 2000 ft.upstream from public surface water supply intakes.Both Pennsylvania and West Virginia have presumption of contamination rules for drinkingwater wells that contain contaminants after drilling has taken place. Currently, the zone ofpresumptive liability is 1,000 feet. The Pennsylvania Governor’s Marcellus Shale AdvisoryCommission recommended increasing the liability zone to 2,500 feet from public water supplywells (Marcellus Shale Advisory Commission 2011). The Center for Rural Pennsylvaniacompleted a report based on analysis of water quality in private water wells in proximity toMarcellus gas wells, which recommended increasing the zone of liability to 3,000 feet fromprivate water wells (Boyer et al. 2011). In 2012, the Maryland legislature established a rebuttablepresumption that drilling or fracking activities were the cause of drinking water contamination ifthe contamination occurred within 2500 feet of the vertical borehole and within 365 days3. Pre-drilling sampling is not required, and refusal of such sampling vacates the landowner’s right tocompensatory damages. We support this regulatory structure and recommend that all waterquality data collected through pre-drilling testing be provided to the appropriate Marylandagency as well as to landowners to increase the information available related to groundwaterresources regionally. Pre-development notification should be made to public and private drinkingwater well owners. Further, we recommend requiring post-development assessment of impacts todrinking water wells. These issues are discussed further in section B below.Due to the heavy reliance on impounded surface water from headwater streams in small, mostlyforested watersheds as a drinking water source for the majority of Allegany County residents,water withdrawal plans and drill pad siting plans should be assessed within the context ofwatershed protection plans previously developed by the drinking water providers. Source waterassessment and protection plans typically include source water delineation maps, transportationcorridors, and existing potential sources of water quality impairment information that can assistpermitting and siting decisions. Any drinking water provider that does not have a watershedprotection plan should be required to develop one in advance of any approved developmentwithin its source watershed. To avoid contamination of all streams and rivers, no drilling shouldbe conducted on floodplains, nor should materials or equipment be staged on floodplains.Setbacks should be extended for on-site staging and storage of hazardous materials and foreventual collection tanks for produced water. Setbacks from streams and wetlands are alsorecommended to reduce the potential for surface spills affecting source waters; consistent withrecommendations in Chapter 6, a 300 ft. buffer from all streams, wetlands, and springs should beenforced to protect surface water quality. As noted above, Maryland should enforce a 2,000 ft.setback from drinking water intakes for surface water plants to reduce direct contamination in theevent of spills on site. Watershed protection plans, specific for each water provider, may in somecases recommend greater setbacks due to unique conditions within source watersheds. Inparticular, both large community systems in western Maryland (City of Cumberland and City ofFrostburg) receive most of their source water from Pennsylvania watersheds, so an assessment ofcurrent oil and gas water withdrawals and permitted development within the upstream basins in3 Md. Env. Code § 14-110.1 (H.B. 1123). 4-5

76. K.N. Eshleman & A.J. Elmore (2013) Chapter 4Pennsylvania is clearly warranted. We recommend flexibility in the setback statute to enablerequirement of larger setbacks when warranted by analysis of watershed protection plans fromdrinking water systems.B. Monitoring of water resources prior to, during, and following developmentAs discussed in section A above, routine pre-drilling assessment of groundwater quality shouldbe required in Maryland. Pre-drilling public notification should also be part of the permitprocess, thus allowing well owners outside the pre-drilling survey area to pursue their own waterquality sampling, if desired. In Pennsylvania, citizens have to ‘opt-in’ for notification of drillingin their area. Maryland should proactively publicize planned activities during the pre-drillingphase of MSGD. The identification of all potentially affected groundwater wells and pre-drillingtesting of these wells is a best management practice that should be required in Maryland. Pre-drilling testing should be required to be conducted by the operator and the results provided to theMaryland Department of the Environment and to the well owner. Post-drilling testing is often atthe discretion of the well owner, but a best management practice that would enable improvedunderstanding of the potential for effects on groundwater would be to require post-drilling andcompletion testing by the operator for all wells within a pre-determined potentially affectedregion for a specified time period after completion of well construction activities. As notedabove, in Pennsylvania this is 1,000 ft., but longer distances are likely relevant for the moreintensive activities associated with horizontal drilling and completion and have beenrecommended by several Pennsylvania studies (Boyer et al. 2011, Marcellus Shale AdvisoryCommission 2011).More extensive groundwater testing (e.g., up to ½ mile from the planned activity) would likelyproduce a better baseline of water quality in the region. Since Maryland does not have extensiveinformation on groundwater in the western part of the state, extensive pre-drilling testing couldprovide important information to MDE to be used in addressing potential impacts ofdevelopment on groundwater resources. Testing should include, at a minimum, the well yieldand the following water quality parameters: conductivity, total suspended solids or turbidity,total dissolved solids (TDS), chloride, bromide, sulfate, barium, strontium, naturally occurringradioactive materials (NORM), chemical oxygen demand (COD), and BTEX (benzene, toluene,ethylbenzene, and xylene). Currently, MGS tests for a wide variety of natural and anthropogeniccompounds in well samples (see Table 4-1). We recommend using this same suite of analyses forpre- and post-drilling sampling to provide the most comprehensive information on conditions inthe subsurface and add to the repository of knowledge in Maryland about groundwater resources.We support the proposal that water samples be collected by qualified professionals and analyzedutilizing an approved analytical laboratory (i.e., one approved by the Environmental LaboratoryAccreditation Program, ELAP), including the use of proper sampling and laboratory protocols inaddition to the use of proper sample containers, preservation methods, holding times, chain ofcustody, analytical methods, and laboratory quality assurance/quality control (QA/QC)procedures (NYSDEC 2011). As noted above, all data should be shared with MDE and MGS. Inaddition, Maryland should require full disclosure of hydraulic fracturing chemicals well inadvance of their use (see Section I), thus enabling pre-development and post-development 4-6

77. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandgroundwater monitoring efforts to include some of these substances. Post-completion welltesting should include the same wells tested in the pre-drilling phase of development. Table 4-1. Recommended water-quality constituents to be analyzed in groundwater in pre- and post-drilling assessment of Marcellus shale area. MAJOR IONS AND INDICATORS Calcium Sulfate Fluoride Dissolved oxygen Magnesium Chloride Bromide Total dissolved solids Sodium Alkalinity Silica (residue at 180o C.) Potassium Specific conductance Color Total organic carbon Nitrate plus nitrite Methane pH Nitrite Ammonia TRACE ELEMENTS Aluminum Cadmium Iron Strontium Antimony Chromium Manganese Thallium Arsenic Cobalt Molybdenum Uranium Barium Copper Nickel Vanadium Beryllium Lead Selenium Zinc Boron Lithium Silver HYDROCARBONS AND METHANE ISOTOPES RADIONUCLIDES Methane Gross alpha-particle activity Ethane Gross beta-particle activity Ethene (both analyzed within three days of sample Propane collection and again at 30 days after sample 2 H-CH4 (if sufficient methane available) collection) 13 C-CH4 (if sufficient methane available) Radon-222New developments in sensing technology have enabled improved monitoring at a variety oflocations with the potential to be affected by MSGD in Maryland. Drinking water providers inPennsylvania and West Virginia have installed a network of source water monitoring equipmentfor early detection of changes in conductivity (that can indicate salt levels are rising). The RiverAlert Information Network (RAIN)4 enables early detection of changes in source waterconditions that affect drinking water treatment and finished water quality for consumers. RAINis a collaborative effort among drinking water plants, PADEP, West Virginia Department ofEnvironmental Protection (WVDEP), and USEPA, with joint funding for the sensors,deployment, and maintenance. RAIN should be extended into the Marcellus development area inwestern Maryland, with funding provided to drinking water utilities to install monitors near theirintakes. Drinking water utilities have the technical expertise to operate and maintain thesesensors and can provide early notification of any significant changes in water quality. Drinkingwater treatment plants operating in western Maryland should also increase their source watermonitoring and specifically include bromide in their routine analyses. In other shale states,universities and watershed groups have also been involved in enhanced water sampling programsto provide baseline information on water quality and to alert the public when changed indicate4 Information on RAIN is available at www.3rain.org. 4-7

78. K.N. Eshleman & A.J. Elmore (2013) Chapter 4problems that might be associated with shale gas development activities. These organizations canbe involved (and are already involved to some degree) within Maryland as well. Routinesampling and the installation of real time sensors can provide useful data for understandingassessing any impacts of MSGD on water resources in the state). MDE should considerleveraging existing monitoring networks run by universities, watershed groups and otherorganizations as a way of capitalizing on existing datasets for baseline characterization.C. Water pollution, stormwater management, and erosion and sedimentcontrol across the life cycleDevelopment of shale gas begins as many other types of construction projects do with clearingand leveling of land for the creation of a well pad; additional cleared acreage would likely beneeded for roads, impoundments, pipelines, and utility corridors (see Chapter 1). One of thechallenges for water associated with land clearing is stormwater runoff from drilling pads,including erosion and sedimentation and wash-off of any chemicals that have been spilled ontothe pad during the various phases of an operation. Runoff of this type has the potential to affectdownstream human water use as well as aquatic habitat, biodiversity, and wildlife (see Chapter6). For this reason, implementation of effective BMPs for stormwater pollution and erosion andsediment controls will be critical in managing potential water quality impacts of MSGD.Surface water pollution in the U.S. is primarily addressed by the Federal Clean Water Act. TheClean Water Act (CWA) prohibits the discharge of pollutants by point sources into waters of theU.S., except in compliance with certain provisions of the Act, specifically section 402, 33 U.S.C.1311(a). Section 402 establishes the National Pollutant Discharge Elimination System (NPDES)under which USEPA, or an authorized state agency, may issue permits that allow the dischargeof pollutants into U.S. waters. In developing effluent limitations for an NPDES permit, limitsbased on available technology (i.e., technology-based effluent limits) and on the water qualitystandards of the receiving water (i.e., water quality-based effluent limits) must be considered.Technology-based effluent limits for direct discharges from oil and gas extraction facilities intosurface waters are found in 40 CFR Part 35, Subpart C. The effluent guidelines thus establishbest practicable control technology currently available for on-shore oil and gas extractionfacilities are as follows: “there shall be no discharge of waste water pollutants into navigablewaters from any source associated with production, field exploration, drilling, well completion orwell treatment (i.e., produced waters, drilling muds, drill cuttings, and produced sand).” Theimportance of this “no discharge” limit is that oil and gas facilities are not required to apply foran NPDES permit and that states can use their own authority to ensure that the “no discharge”requirement in the effluent guidelines is properly applied and that operator compliance isdemonstrated5. The “no discharge” limit has obvious important implications for how the wastes(e.g., flowback, produced water, drilling muds, etc.) generated by onshore oil and gas facilitiesmust be handled under federal law (see Sections G and H below). In addition to regulating suchdirect discharges, USEPA’s regulations also address (1) indirect releases of wastewaters intoU.S. waters such as by publicly owned treatment works (POTWs) that have received oil and gas5 Attachment to memorandum from James Hanlon, Director of EPA’s Office of Wastewater Management, to the USEPA Regions titled “Natural Gas Drilling in the Marcellus Shale under the NPDES Program” (March 16, 2011). 4-8

79. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandwastes; or (2) direct discharges from centralized treatment facilities that are subject to their owntechnology-based and water quality-based effluent guidelines.The CWA also gives USEPA (and authorized state programs) the authority to regulatestormwater pollution under a separate NPDES permitting program. Impervious surfaces, such asbuildings, homes, roads, sidewalks, and parking lots, can significantly alter the natural hydrologyof the land by increasing the volume, velocity, and temperature of runoff and by decreasing itsinfiltration capacity. Increasing the volume and velocity of stormwater runoff can cause streambank erosion, flooding, and degradation of stream aquatic habitat. As stormwater runoff isproduced, it can pick up trash, debris, and various pollutants such as sediment, oil and grease,pesticides and other toxics. Changes in ambient water temperature, sediment, and pollutants instormwater runoff can be detrimental to aquatic life, wildlife, habitat, and human health. Soilexposed by construction activities is especially vulnerable to erosion. Excess sediment canincrease the turbidity of receiving surface waters, reduce the amount of sunlight reaching aquaticplants, clog fish gills, and smother aquatic habitat and spawning areas. Therefore, the primarystormwater pollutant of concern from construction is usually sediment, and practices must beimplemented to effectively control runoff and associated stormwater pollution. USEPAregulations require operators disturbing one acre or more of land6 (including smaller individualareas that are part of larger developments) to apply for coverage under a NPDES constructiongeneral permit for stormwater discharge and develop and implement a Stormwater PollutionPrevention Program (SWPPP). An SWPPP is a site-specific, written document that: (1) identifiespotential sources of stormwater pollution at the construction site; (2) describes practices (BMPs)that will be employed to reduce pollutants in stormwater discharges from the site; and (3)identifies procedures that an operator will implement to comply with the terms and conditions ofa construction general permit. Pollution reduction is most often achieved by controlling thevolume of stormwater runoff (e.g., taking steps to allow stormwater to infiltrate into the soil).As in point source permitting, Maryland is also authorized to issue coverage under the NPDESconstruction general permit for stormwater discharges and has issued its own guidancedocuments and technical design manuals to aid in development of SWPPPs and implementationof BMPs for stormwater, erosion and sediment controls.7 While Maryland appears to have arobust program for controlling stormwater pollution, we believe there is a significant regulatoryimpediment to effective implementation of BMPs to address stormwater pollution impactsassociated with MSGD in the state and elsewhere. First of all, as amended by the Energy PolicyAct of 2005, the CWA (section 402(1)(2) and 502(24)) specifically exempts oil and gasoperations from most industrial stormwater permitting requirements by USEPA or by thosestates with approved NPDES programs (such as Maryland)8,9. Specifically, the section of the actreads as follows: “All field activities or operations associated with oil and gas exploration,production, processing, or treatment operations or transmission facilities, including activitiesnecessary to prepare a site for drilling and for the movement and placement of drilling6 Some states grant variances for activities that disturb less than five acres of land.7 Maryland Stormwater Design Manual, Volumes I and II (effective October 2000, revised May 2009);http://www.mde.state.md.us/programs/Water/StormwaterManagementProgram/MarylandStormwaterDesign Manual/Pages/programs/waterprograms/sedimentandstormwater/stormwater_design/index.aspx8 40 CFR § 122.26 Storm water discharges.9 Ibid., 5 4-9

80. K.N. Eshleman & A.J. Elmore (2013) Chapter 4equipment, whether or not such field activities or operations may be considered to beconstruction activities, except in accordance with paragraph (c)(1)(iii) of this section. Dischargesof sediment from construction activities associated with oil and gas exploration, production,processing, or treatment operations or transmission facilities are not subject to the provisions ofparagraph (c)(1)(iii)(C) of this section.”10 While exempted from NPDES industrial stormwaterpermitting, it is noted in the same statutes that the USEPA “encourages operators of oil and gasfield activities or operations to implement and maintain best management practices (BMPs) tominimize discharges of pollutants, including sediment, in storm water both during and afterconstruction activities to help ensure protection of surface water quality during storm events.Appropriate controls would be those suitable to the site conditions and consistent with generallyaccepted engineering design criteria and manufacturer specifications. Selection of BMPs couldalso be affected by seasonal or climate conditions.” 11The importance of this exemption is that, unlike an entire suite of different types of industrialactivities and operations12 that are not exempted from industrial stormwater permitting under theCWA, USEPA lacks the authority to regulate stormwater pollution from oil and gas activities inthe same way that it would do so for these other industrial activities. In Maryland, which isauthorized by USEPA to do NPDES industrial stormwater permitting, oil and gas extraction sitesare not statutorily exempted from the sediment and erosion control program. However, oil andgas extraction sites are not considered “hotspots” for stormwater pollution impacts, althoughthey may meet the definition13 of “hotspots”. Unlike most other industrial operations withequivalent (or perhaps even lower) risks of impacting surface water quality that are required toobtain an NPDES industrial permit and implement stormwater pollution prevention plans thataddress pollution both during and after construction, oil and gas operators are merely encouragedto implement BMPs to control stormwater pollution and can be covered under the generalconstruction permit14. USEPA and the approved state NPDES programs such as those inMaryland may be hampered in their efforts to control stormwater pollution from MSGD due tothis exemption in federal law.Other than consistency with the federal exemption, there is no compelling reason for Marylandto exempt oil and gas extraction activities from its industrial stormwater permitting requirements.Designation as a hotspot has practical implications for management of stormwater. Typically, at10 40 CFR § 122.26(a)(2)(ii)11 Ibid.12 Including vehicle salvage yards and recycling facilities, vehicle service and maintenance facilities, vehicle and equipment cleaning facilities, fleet storage areas, industrial sites, marinas, outdoor liquid container storage facilities, and outdoor loading/unloading facilities; see Maryland Stormwater Design Manual, Volumes I and II (effective October 2000, revised May 2009); Appendix D.6 Industrial Stormwater Permit Requirements; http://www.mde.state.md.us/programs/Water/StormwaterManagementProgram/MarylandStormwaterDesignManu al/Pages/programs/waterprograms/sedimentandstormwater/stormwater_design/index.aspx13 A stormwater hotspot is defined as a land use or activity that generates higher concentrations of hydrocarbons, trace metals or toxicants than are found in typical stormwater runoff, based on monitoring studies . See Maryland Stormwater Design Manual, Volumes I and II (effective October 2000, revised May 2009); Chapter 2.8 Designation of Stormwater Hotspots; http://www.mde.state.md.us/programs/Water/StormwaterManagementProgram/MarylandStormwaterDesignManu al/Pages/programs/waterprograms/sedimentandstormwater/stormwater_design/index.aspx14 Operations that result in disturbances of less than five acres of total land area are also exempted under 40 CFR § 122.26(b)(14)(x). 4-10

81. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandconstruction sites where the primary concern is erosion and downstream sedimentation, SWPPPstend to emphasize BMPs that promote infiltration into the ground as a primary means ofreducing stormwater discharges and thus associated erosion and sedimentation problems.Performance standards developed in Maryland through largely urban stormwater controlemphasize site designs that maximize pervious areas for stormwater treatment (standard no. 1)and promote infiltration through the use of structural and non-structural methods (standard no.2)15. However, for hotspots, where untreated stormwater runoff cannot be allowed to infiltrateinto the ground, Maryland applies differential requirements to prevent groundwatercontamination. Since oil and gas development sites are more similar to hotspots than to urbandevelopment sites, Maryland should review its stormwater regulations to ensure oil and gasextraction operations are managed in accordance with their characteristics, rather than through astatutory exemption. The use of a generic SWPPP, such as is often developed for residentialsubdivisions, is not the correct approach for managing stormwater pollution from shale gasoperations.The primary goal, intent, and spirit of the CWA is found in the first sentence of the act [Section101(a)] where it states that the legislation is meant to “restore and maintain the chemical,physical, and biological integrity of the Nation’s waters”. While the restoration component ofthe CWA is largely being dealt with by the states through regulation of point and non-pointdischarges of water quality pollutants, the goal to maintain water quality in situations whereimpairment is not presently an issue is addressed under the federal anti-degradation policy. Thisregulatory policy, described in Section 303(d) of the CWA, is designed to prevent deteriorationof existing levels of high or exceptional water quality in areas where such conditions exist. Thefederal policy requires states to develop rules and implementation procedures to protect existinguses of such waters and to prevent such waters from being degraded (unless the actionresponsible for the deterioration provides an important social or economic benefit). Each state’santi-degradation rules and implementation procedures must be included in the state’s waterquality standards (WQS). In addition, the federal rules16 require that: • Existing in-stream water uses and the level of water quality necessary to protect the existing uses shall be maintained and protected. • Where the quality of the waters exceeds levels necessary to support propagation of fish, shellfish, and wildlife and recreation in and on the water, that quality shall be maintained and protected unless the State finds, after full satisfaction of the intergovernmental coordination and public participation provisions of the States continuing planning process, that allowing lower water quality is necessary to accommodate important economic or social development in the area in which the waters are located. In allowing such degradation or lower water quality, the State shall assure water quality adequate to protect existing uses fully. • The state shall assure that there shall be achieved the highest statutory and regulatory requirements for all new and existing point sources and all cost-effective and reasonable best management practices for non-point source control.15 Maryland Stormwater Design Manual, Volumes I and II (effective October 2000, revised May 2009); Chapter 1.0 Introduction; http://www.mde.state.md.us/programs/Water/StormwaterManagementProgram/MarylandStormwaterDesignManu al/Pages/programs/waterprograms/sedimentandstormwater/stormwater_design/index.aspx16 40 CFR §131.12 4-11

82. K.N. Eshleman & A.J. Elmore (2013) Chapter 4 • Where high quality waters constitute an outstanding national resource, such as waters of national and state parks and wildlife refuges and waters of exceptional recreational or ecological significance, that water quality shall be maintained and protected.Essentially, under federal rules, each state’s anti-degradation policy must be implemented undera three-tiered program: • “Tier 1”, that protects "existing uses" and provides the absolute floor of water quality for all waters of the United States [Section 131.13(a)(1)]; • “Tier 2”, that includes “high quality waters” (HQW) in which water quality exceeds that necessary to protect the Section 101(a)(2) goals (fishable and swimmable). Water quality may be lowered under certain conditions, but never below the level necessary to fully protect the “fishable and swimmable” and other existing uses [Section 131.12(a)(2)]; and • “Tier 3”, that are “outstanding national resource waters” (ONRW) in which only temporary reductions in water quality are allowed [Section 131.12(a)(3)].Maryland adopted its anti-degradation policy as part of the WQS in 1985 and revised its policyin 2001. In 2004, Maryland adopted its current Tier II implementation policy and promulgated alist of 87 Tier II (i.e., HQW) stream segments based on established criteria of biologicalintegrity; the majority of these Tier II segments are located in western Maryland (Figure 4-3).Maryland’s current Tier II policy states that “where water quality is better than the minimumrequirements specified by the WQS, that water quality shall be maintained”. MDE will enforcethe state Tier II policy by requiring that “applicants for proposed amendments to county plans ordischarge permits for discharge to Tier II waters that will result in a new, or an increased,permitted annual discharge of pollutants and a potential impact to water quality, shall evaluatealternatives to eliminate or reduce discharges or impacts”. A Tier II anti-degradation review isrequired for permits involving individual discharges of at least 5,000 gallons per day; however,lesser proposed point discharges—and presumably non-point source discharges—of pollutantsthat could potentially contribute to significant degradation of water quality (especially of smallstreams) are exempted from anti-degradation review. Given this necessary trigger and the factthat point discharges from oil and gas development cannot be permitted as discussed earlier, wedo not believe that MSGD would trigger a Tier II anti-degradation review in Maryland undercurrent policy. For this reason, Maryland might wish to consider ways of strengthening its anti-degradation policy to take account of the impacts of non-point source pollution that are a majorthreat to its high quality waters. One way that this might be accomplished would be by revisingthe WQS rules to require that any land development practices (e.g., forest management, MSGD,etc.) conducted in Tier II watersheds meet an anti-degradation standard.17Based on review of stormwater management practices in other states, we recommend theconstruction of properly bermed “zero-discharge” pads that effectively collect all water on a padsite and enable the reuse of this water during drilling and completion operations. This practicerequires careful grading during the pad construction process, so that water (i.e., mostly excessprecipitation onto the pad, but also any other liquids) can flow by gravity to a single location onthe pad where these liquids can be collected on a regular basis—typically using vacuum trucks.A berm around the entire pad should be designed to prevent any stormwater from being17 The state of Washington has a similar approach for applying anti-degradation rules to forest practices, and Oregon is considering such an approach (State of Oregon Department of Environmental Quality, 2010). 4-12

83. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylanddischarged from the pad, except for the entrance road that should be elevated slightly above thepad to prevent runoff by the road. The entire pad must be underlain with a heavy impervioussynthetic liner (comparable to liners used by landfills). Any areas where mechanized equipmentwill operate should be overlain with a composite decking material to protect the liner fromabrasion and prevent infiltration (Lien and Manner 2010).Figure 4-3. Map showing locations of Tier II (i.e., HQW) streams (and catchments associated with these streams)and waterbodies greater than 1.5 acres in western Maryland.One of the weaknesses of this approach is that the system relies on coordination of active(vacuum trucks, water reuse) and passive (berms, liners) stormwater pollution preventionmeasures. In Maryland’s seasonally wet climate, it’s likely that well pads would overflow onceactive management ends. Since personnel are only expected to be on site continuously duringactive drilling and completion phases of development of a particular well, it is imperative toconsider how these pads will function after well completion, or between different rounds ofactivities (wells completed at different times from the same pad). Since activities at the pad sitemay cycle through periods of active development and periods of production, pad reclamation tomanage stormwater may have to occur multiple times. There are two options for managingrunoff from drill pads between episodes of drilling: (1) the pads could be revegetated andrestored to original condition any time operations cease for a defined time period (i.e., this wouldavoid excess runoff that was not being managed); or (2) the developed area could remaindisturbed and a stormwater collection and management system could remain in place. The bestsolution for addressing both quality (i.e., suspended solids) and quantity (i.e., peak discharge)issues would be through construction of a below-grade lined pond adjacent to the bermed zero-discharge pad that could be used as a sump during active stormwater management phases and 4-13

84. K.N. Eshleman & A.J. Elmore (2013) Chapter 4easily converted into a retention pond prior to any passive phases. Regular periodic (annual)maintenance of the pond would also be needed to ensure that the system is functioning correctlyat all times. Additional water quality treatment could be obtained through operation of aconstructed wetland sited downstream of the pond outlet.Related to stormwater management, operators would be required to develop and implementerosion and sediment control plans. These plans usually include BMPs for: (a) grading andstabilization to minimize erosion during development; (b) water conveyance plans for clear-water diversions around the development area to reduce stormwater that picks up sediment onthe site; (c) erosion control that reduces the velocity of surface flows; (d) filtering and sedimenttrapping systems to collect sediment and prevent its discharge from the site; and (e) dewateringpractices, if applicable to a site. Some plans also specify reclamation requirements, includingrestoration of grades and re-vegetation to prevent post-development changes in sediment loadsfrom the site. Plans are typically certified by a registered professional engineer (PE). Each ofthese elements should be addressed in Maryland’s regulations. Soil erosion and sediment controlplans should also be required for the development of new roads to sites. Stream crossings anddevelopment through wetland areas should be avoided (see Chapter 6 also). In addition, asrecommended for New York and Pennsylvania, the design of all stormwater control structures toaddress erosion and sedimentation should be based on a 10yr/24hr rainstorm (i.e., the rainstormwith a duration of 24 hours that occurs, on average, once every ten years), as opposed to the2yr/24hr storm that occurs more frequently. Given the complexities in addressing how active andpassive stormwater management will occur, we also recommend that the state ensure that SoilConservation Districts, which currently review and approve sediment control plans and who aremost knowledgeable for their geographic area are on-site during all majorconstruction/deconstruction activities. Post-construction inspections of stormwater structures byMDE and the relevant Soil Conservation District personnel should occur prior to well drillingand completion.On January 27, 2012, Maryland enacted new regulations for soil erosion and sediment control(MDE 2011). As part of this new regulation, each county is required to draft erosion andsediment control ordinances by January 2013. A model ordinance was published by Maryland inFebruary 2012 (MDE 2012). The model ordinance includes an exemption for clearing or gradingactivities that disturb less than 5,000 square ft. of land area (~ 0.1 acre), which is a fairly typicalexemption that is unlikely to affect shale gas pads (typically on the order of 4-6 acres in size). Inthe Maryland model ordinance, Erosion and Sediment Control (ESC) plan review is requiredprior to permit approval. This is ideal and should be retained (some state regulations haveexemptions that pertain specifically to oil and gas development18). Garrett and Allegany Countyshould follow the Model Ordinance proposed by MDE, but should also require consideration inESC plans of the potential effects of multiple clearings in relatively close proximity. Exemptionsfor small sites should not be enacted. Maryland should also evaluate potential issues associated18 For example, it is our understanding of Pennsylvania’s regulations that gas exploration and extraction facilities that result in disturbance of fewer than five acres are not required to obtain an “Erosion and Sediment Control Permit”. For such facilities (e.g., well pads), a “Permit Application for Drilling or Altering a Well” (5500- PMOG0001) is sufficient. An Erosion and Sediment Control Plan must still be developed, but the plan is not subject to regulatory review and approval before construction. This is in contrast to most other construction activities, which are subject to erosion and sediment control requirements at one acre or greater under the Pennsylvania Chapter 102 requirements and NPDES requirements. 4-14

85. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandwith stormwater, sedimentation and erosion within the context of multiple simultaneous MSGDsites and use its discretionary authority to require individual stormwater permits when warranted.ESC plans will also be critical in managing potential impacts on downstream water users.Maryland should ensure that public water supplies downstream of permitted MSGD activitiesshould be notified prior to such activities.With respect to specific BMPs, operators should be strongly encouraged to consult Maryland’sStormwater Design Manual and an industry document entitled “Guidance Document, Reasonableand Prudent Practices for Stabilization (RAPPS) at Oil and Natural Gas Exploration andProduction Sites” published by API. The latter report describes a host of specific operatingpractices and control measures that have been used and tested by oil and natural gas operators toeffectively control erosion and sedimentation from stormwater runoff resulting from landclearing, grading, and excavation operations at exploration and production sites under variousconditions of location, climate and slope.D. Water withdrawalsAt the scale of a single well or multiple wells on a single pad, 2-6 million gallons of water mustbe acquired to facilitate the drilling and completion of each well. The amount varies basedprimarily as a function of the length of the drilled lateral. To support MSGD, water can beextracted from surface or groundwater sources (including non-potable sources, see Section G), oreven purchased from existing treatment plants if excess capacity exists. While 2-6 milliongallons is a large volume of water, it is important to keep it in context relative to otherwithdrawals and supplies of water in the state. It was estimated by USGS that in the year 2000,Allegany and Garrett County withdrew on average 48.9 and 9.6 million gallons per day (MGD),respectively, from all surface water and groundwater sources. Over the course of an entire year,this works out to a combined volume of about 21.4 billion gallons of water. Thus, the combinedannual withdrawals by these two counties alone would be equivalent to the amount of waterrequired to develop about 3,500 Marcellus shale gas wells in the state.There are very long gage records available from USGS for most of the major rivers that could beused to support MSGD in western Maryland. We computed the mean annual discharge inwestern Maryland’s three largest headwater rivers (North Branch Potomac River near Steyer,Youghiogheny River at Friendsville, and Savage River below Savage River Dam) based on theserecords as part of a preliminary analysis of supply and obtained values of 112, 413, and 110MGD, respectively. The combined long-term average discharge in these three rivers is 635MGD—producing a volume of water on an average day that is more than 100 times larger thanthe water requirement to develop a single Marcellus shale gas well. However, it must be notedthat the average discharge of water in these rivers varies dramatically throughout the year: in theYoughiogheny, for example, the long-term mean daily discharge in March (while normallyswelled by spring snowmelt) is 769 MGD based on 71 years of data (1941-2011), although inSeptember the long-term mean discharge is only about 22% as great (172 MGD). During lowflow periods, as in drought years, flows in all of these rivers can become critical low. Forexample, in the Youghiogheny, the annual seven-day minimum flow was 18.7 MGD inSeptember of 1972—illustrating that it is unlikely that flow conditions in even these major riverscan support withdrawals for MSGD at all times under all conditions. On the other hand, under 4-15

86. K.N. Eshleman & A.J. Elmore (2013) Chapter 4average conditions and particularly at the higher discharges normally reached during the springof the year, the data suggest that there may be adequate supplies provided by western Maryland’smajor rivers as source water for MSGD.With respect to groundwater sources, in Chapter 1 we summarized the available information onwestern Maryland’s principal aquifers from the perspective of identification of aquifers andflowpaths as part of the drilling and hazard assessment processes. Much of that information isclearly relevant to identifying sources of groundwater that could be used for HVHF. On theAppalachian Plateau in Garrett County and western Allegany County, water yields of wellscompleted in Pennsylvanian age sandstone formations (the principal aquifers) reportedly rangefrom 20 to 430 gallons per minute. In the Valley & Ridge west of the Great Valley, Ordovicianto Devonian age sandstones are considered the principal aquifers, but wells completed in theseformations commonly yield less than 120 gallons per minute; wells in limestone formations oflate Silurian through early Devonian age may locally yield as much as 100 gallons per minutewhere these rocks are fractured (Trapp, Jr. and Horn 1997). These individual well yields (0.03 –0.6 MGD) are certainly high enough to suggest that western Maryland’s groundwater resourcecould potentially be exploited to support HVHF, but research would be needed to assess whethersuch development would likely cause the safe yield of these aquifers to be exceeded. We canenvision that groundwater could play a role in supplying hydraulic fracturing operations duringdry summer periods when water levels in major rivers and reservoirs are too low to permitsurface withdrawals, or cases where a particular well pad (1) is located an excessive distancefrom a permitted surface water supply location; or (2) is not efficiently served by public roadsthat enable trucking of water. A centralized water well field (with suitable impoundment) couldalso potentially be used to supply (via buried pipeline) a group of multi-well pads that were partof a clustered development (see Chapter 1), with the caveat that the water wells would need to besited so as to observe recommended setbacks. In the four eastern states (with similarhydrogeologic settings) that we reviewed, we found no evidence of extensive groundwaterresource development to support MSGD19—presumably due to limited supplies, low well yields,and high costs (compared to the surface water alternative). Use of groundwater (in some casesdrawn from saline aquifers) has supported shale gas development in some western states andwestern Canada (King 2012), but the far greater supply of surface water at lower cost suggeststhat it is highly unlikely that MSGD in Maryland would be primarily supported by availablegroundwater resources.In Maryland, both surface and groundwater withdrawals are regulated by MDE.20 Permitapproval requires that the applicant provide satisfactory proof that the proposed withdrawal ofwater is reasonable and the impacts on the water resource and other users are acceptable. Further,the proposed use must be consistent with local planning and zoning requirements and the countywater and sewer plan. Additional permitting documentation is required for requests forwithdrawals in excess of 10,000 gallons per day, and public notification may be required.19 Recent data provided by SRBC to MDE indicates that Marcellus operators in Pennsylvania have obtained permits to supply 4% of their total water needs from groundwater sources computed by averaging over the entire basin (John Grace, MDE, personal communication, February 15, 2013).20 COMAR 26.17.06 and COMAR 26-17.07 Details are available at: http://www.mde.maryland.gov/programs/Water/Water_Supply/Documents/www.mde.state.md.us/assets/document /permit/2008PermitGuide/WMA/3.15.pdf 4-16

87. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandDetermining that a proposed withdrawal of water is “reasonable” can be done in several ways.In Maryland, the “Criteria for Approval of Water Appropriation or Use Permits”21, providesnarrative regarding reasonableness, including a consideration of: (a) the purpose of the use; (b)the suitability of the use to the watercourse, lake or aquifer; (c) the extent and the amount ofharm it may cause; (d) the practicality of avoiding the harm by adjusting the proposed use ormethod of use of the applicant or another permittee; (e) the practicality of adjusting the quantityof water used by each permittee; (f) aggregate changes and cumulative impact that this andfuture appropriations in an area may have on the waters of the state; (g) the contribution that theproposed appropriation may make to future degradation of the waters of the state; and (h)whether the proposed appropriation or use is located within a water management strategy area.For surface water sources, approved withdrawals are “conditioned on the maintenance, by thepermittee, of a required minimum flow past the point of appropriation to protect other users ofthe water and to protect flora and fauna within the watercourse.22” The most common requiredminimum flow is typically set at the lowest 7 day average in the past 10 years (Q7-10). The Q7-10 is specific to each source water and can be difficult to determine for streams without gages.Within Maryland, the required minimum flow is not always the Q7-10. In other states,withdrawals are often allowed up to a specific restricted flow point, for example, when the flowis less than 20% of the average daily flow. For gaged streams, average daily flow is easilydetermined; however, for ungaged streams, a reference gage approach using USGS regressiontools is required. Significant uncertainty is often observed in such predictions (Murphy et al.2012, Razavi and Coulibaly 2012, Shu and Ouarda 2012). A comparison of the use of differentmethods for flow prediction using Maryland-specific historical gage data will provide a clearassessment of the most appropriate method for flow prediction in ungaged streams in the state.We are confident that there are adequate long-term and site-specific data for a sufficient numberof gaged watersheds to support a rigorous analysis of stream discharges in western Marylandwatersheds to inform an analysis of minimum required flow levels for streams that might supportMSGD water withdrawals in the region. While we have not done such a quantitative analysis aspart of our review of best practices, our experience in Maryland watersheds as well as review ofother areas that have completed such analysis, suggest that in western Maryland, waterwithdrawals for proposed MSGD will need to occur solely from the region’s large rivers (andperhaps from one or more reservoirs). Small streams (1) have significant existing withdrawalsfor drinking water; (2) have small catchment areas and discharges under most conditions; (3) arevery unlikely to have excess flow capacity for new permitted withdrawals; and (4) can be readilydewatered.Determining that the “impacts on the water resource and other users are acceptable” can be evenmore challenging than determining minimum required flows for streams. Multiple withdrawalswithin a basin can have a cumulative effect that must be considered in overall basin-levelanalysis. One method is to require the permitee to assess the cumulative net withdrawals up-gradient of the proposed new withdrawal, and to consider cumulative impacts in the permitreview process as is commonly done by basin commissions (SRBC 2012). Extensive review andanalysis of watersheds in Pennsylvania has been undertaken to review and update requirements21 COMAR 26.17.06.05. http://www.dsd.state.md.us/comar/getfile.aspx?file=26.17.06.05.htm22 COMAR 26.17.06.05 4-17

88. K.N. Eshleman & A.J. Elmore (2013) Chapter 4related to water withdrawals as shale gas drilling has expanded throughout the Susquehanna andOhio River basins. Several critical reviews have been completed that should be considered asMaryland decides how to update water withdrawal permitting processes to consider thetemporally and spatially distributed withdrawals typical of this industry. For example, the TNCEcosystem Flows Study (TNC 2010) was a collaboration between The Nature Conservancy(TNC), the Susquehanna River Basin Commission (SRBC) and the United States Army Corps ofEngineers (USACE). This study focused specifically on the Susquehanna River and itstributaries; however, the approach and general conclusions provide important information forMDE in assessing ecosystem considerations for streamflow and the role of this component inwater withdrawal permitting for shale gas development.Recently, as part of proposed changes to the Low Flow Protection Policy of SRBC, MDEexpressed concerns regarding the use of a fixed ‘de minimus’ threshold for headwaterwatersheds and highly sensitive watersheds since this ‘de minimus’ characterization is not usedin Maryland statute (Kasraei 2012). As MDE notes in their response to SRBC, “it may not befeasible to appropriately quantify the flow regimes of certain ungaged smaller systems due tolack of relevant data” and other site specific conditions. As discussed above, the difficulty inassessing the impacts of water withdrawals with insufficient gaging data is well known (Murphyet al. 2012, Razavi and Coulibaly 2012, Shu and Ouarda 2012) and we strongly recommend thatwater withdrawals for MSGD be encouraged from larger rivers (and perhaps existing reservoirs).MSGD should totally avoid small headwater streams and watersheds out of concern fordewatering of these sensitive systems. Specific rather than generic minimum flow values shouldbe established for these creek and stream systems, and seasonal water conditions should also beevaluated in the process.Timing of water withdrawals is also critical, with low flow conditions typically occurringseasonally. Storage and water transport from storage will be necessary to enable continuedoperations during the dry periods when withdrawals are likely to be limited. The size and numberof any centralized water impoundments and pipelines constructed to support industry operationsis often predicated on the stability of water supply. When water supplies are restricted for moreof the year, gas development operations will either be restricted or larger and moreimpoundments will be required to enable continuous operations. Details of multiple uses for asingle permitted withdrawals and of the plans to construct impoundments and pipelines are oftencontained within the water management plan required for shale gas development inPennsylvania. These plans require identification of the water source at the time of drill permitapplication and lead to more comprehensive water sourcing plans for multi-well and multi-paddevelopment. Multiple drilling companies within a single region may present overlapping planswith little coordination on water withdrawals. A regional multi-operator approach to waterprovision for shale gas operation would likely reduce the number of impoundments andwithdrawal locations and enable smaller facilities, while still providing adequate and stable watersupplies. Water management plans, in addition to water withdrawal permits, should be requiredfor all drilling activities to ensure that development activities incorporate water resourceplanning. Coordinated multi-operational water provisioning should be planned to reduce thedisturbances associated with impoundments and water withdrawals. 4-18

89. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandE. Comprehensive basin-scale water management planningComprehensive basin-scale water management involves consideration of all uses and allactivities associated with water within a watershed, often at the hydrologic unit code (HUC)level of 1 (i.e., HUC 2). This level is typically hundreds of thousands of square miles andusually involves multiple states. HUC 2 watersheds are often managed through interstatecompacts. Maryland participates in several basin commissions (e.g., SRBC) and the InterstateCommission on the Potomac River Basin (ICPRB). No basin commissions exist in the mostwestern part of the state where shale gas is found, however. Figure 4-4 shows the extent ofrelevant basins in Maryland.A portion of western Maryland within Garrett County is part of the Ohio River Basin (part of thelarger Mississippi River Basin). The Ohio River Basin Commission was founded in 1981;however, the organization no longer operates, and it does not assert authority over waterwithdrawals within the basin. The Ohio River Valley Water Sanitation Commission(ORSANCO) is an interstate commission (established in 1948) charged with management ofwater quality in the Ohio River and its tributaries. ORSANCO does not have regulatoryauthority over issues of water quantity. The USACE Pittsburgh District has authority for theYoughiogheny Reservoir, a flood control and recreational reservoir that begins in Maryland andcontinues into southwestern Pennsylvania. Several groups have recommended the creation of anOhio River Basin Commission that would manage water withdrawals to ensure water quality andprotection of aquatic resources, especially during low flow conditions in the region (generallysummer time) and ORSANCO is currently evaluating an expanded role that would incorporatewater quantity authority. In general, the Youghiogheny River has been cleaner than theMonongahela River at the points of entry to Pennsylvania, providing important dilution of themain stem of the Monongahela River that travels north from West Virginia into Pennsylvania,terminating at its confluence with the Allegheny River to form the Ohio (at Pittsburgh, PA).Changes in the water quality within the Youghiogheny River would affect downstream waterusers in Pennsylvania.A major portion of Garrett County is part of the North Branch Potomac (0207002) watershed.Allegany County is located within the South and North Branch of the Potomac (0207001 and0207002) and the Cacapon-Town (0207003) subwatersheds. The USACE Baltimore Districtoperates both Jennings Randolph Lake and Savage River Reservoir (both part of the NorthBranch of the Potomac River Basin). The ICPRB is the relevant watershed managementcommission for this part of the region; however, the ICPRB does not manage water withdrawalpermitting within the basin. Ideally, comprehensive basin-scale planning and analysis would beused for water withdrawal permitting in western Maryland and elsewhere in the Potomac andOhio River Basins. In the absence of interstate basin commissions with water permittingauthority in this region, we recommend that MDE continue to take a comprehensive, basin-scaleapproach to all water withdrawals and to the assessment of water management plans submittedby any shale gas developers. MDE should also discuss the operational conditions ofYoughiogheny Reservoir, Savage River Reservoir, and Jennings Randolph Lakes with USACEto evaluate these systems as potential sources of water for MSGD, particularly during high flowconditions when recreational and other uses would not be negatively impacted. 4-19

90. K.N. Eshleman & A.J. Elmore (2013) Chapter 4Figure 4-4. Major watersheds (HUC 8 and HUC 10) in western Maryland.F. Water storage and deliveryWater must be staged on site at the well pad to support the hydraulic fracturing operation.During active development of the site, through drilling and the hydraulic fracturing operations,stormwater and rainfall will likely be collected for use in operations (see Section C), but thisvolume of water will not meet the high water needs for hydraulic fracturing. Well pad watermanagement generally includes staging water tanks or constructing ponds to hold water, andusing trucks to convey the 2-6 million gallons of water needed for the hydraulic fracturing.Alternative methods have been proposed and are currently being utilized that reduce or eliminatetruck traffic and decrease the size of well pads. In the Pennsylvania state forests, freshwater isbeing moved from centralized storage facilities to active location(s) through the use of temporarypiping. This practice significantly reduces the frequency of heavy hauling across state forestroads, minimizes the possibility of vehicular conflicts, and decreases air and dust pollution(PADCNR 2011). The piping of freshwater may involve above-ground or buried water pipelinenetworks, or a combination. Above-ground piping should be laid out in a manner to reduceaesthetic impacts and the potential for vandalism to the extent possible. Further, such pipingshould avoid interfering with existing infrastructure, including stormwater structures (e.g.,culverts). Where applicable, buried piping should minimize additional earth disturbance and beco-located with natural gas pipelines, buried in the ditchline or vegetated berm, or trenched andburied beneath the running surface of an access road (PADCNR 2011). For example, in theTiadaghton State Forest in north-central Pennsylvania, truck transport has been used to fillseveral constructed impoundments that provided gravity flow to an underground pipelinenetwork that fed a cluster of well pads constructed in reasonably close proximity. 4-20

91. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandFreshwater delivered by either trucks or pipeline must also be staged near the operations. Openlined shallow impoundments (~15 MG capacity) are often constructed in Pennsylvania and usedfor this purpose (see Figure 1-5). The size and number of impoundments would need to bedetermined by the number of wells to be drilled and the number of pads within close proximityof a suitable impoundment location. Locations for these impoundments would be dictated bytopography (i.e., suitable, reasonably flat locations where they can be constructed). Marylandhas existing pond standards/specifications related to livestock watering, recreation, agriculturalstorage, and stormwater management (MDE 2000), but no specific standards exist for storage ofwater for oil and gas development. We recommend that freshwater impoundments be subject tothe same standards regardless of water use. No impoundments constructed in Maryland shouldever be used for storage of any wastewater (i.e., flowback or produced water), however. Norshould water released from temporary impoundments be discharged into any Maryland streamsand rivers due to concerns for introduction of exotic species (e.g., golden algae) and impacts onwater temperatures. As discussed in Section D above, coordinated planning for water needsacross multiple operators and multiple well pads and development regions will reduce thenumber of impoundments needed to ensure reliable and sufficient water supply for this industry.G. Alternative water sourcingWastewater is produced in a number of industrial activities that has potential as a source of waterfor hydraulic fracturing. The most frequently discussed alternate source is acid mine drainage(AMD) that is commonly discharged throughout the Marcellus shale region. Maryland Bureau ofMines has constructed and maintains 33 active and passive AMD facilities in Garrett andAllegany County with typical flows of 1 liter per second (L/s; 1 L/s = 0.023 MGD). Some ofthese facilities (or the mine pools from which the discharge is derived) could serve as alternate,non-potable water sources for MSGD. Among other known AMD sources is the outflow fromthe Hoffman Drainage Tunnel (HDT) near Clarysville, MD that typically discharges intoBraddock Run (a tributary of Wills Creek) at a rate of about 7.3 MGD (recently reported range is7–30 MGD)—making this source a candidate to be evaluated as an alternate water source23. Asdiscussed by one recent report, the use of AMD-impaired water or treated wastewater could haveoverall positive benefits on water quality through removal of these inputs from receiving streams(Lien and Manner 2010). We must note, however, that the cold water discharge of mine waterfrom mine workings underlying Frostburg and the Upper Georges Creek Valley into BraddockRun via HDT appears to provide sustained baseflow in the receiving stream that exceeds bymore than an order of magnitude the natural flow regime. HDT discharge also allows BraddockRun to support a brook trout population24, despite the fact that the discharge is obviously ladenwith ferric hydroxide—commonly known as “yellow-boy”.Nearby West Virginia and Pennsylvania have even more extensive acid mine drainage issues,with some outfalls in the tens of L/s range (Ziemkiewicz et al. 2003). Treatment costs aresignificant and treated water is discharged into surface waters (e.g., Hansen et al. 2010).Similarly, active coal mines and coal bed methane extraction activities generate high volumes ofwater during dewatering activities that are ongoing for the duration of the mining activity. These23 Unpublished analysis of field data by K.N. Eshleman and R.P. Morgan II for Lavale Sanitary Commission (September 4, 2002).24 Jason Cessna (Appalachian Laboratory), personal communication (February 11, 2013) 4-21

92. K.N. Eshleman & A.J. Elmore (2013) Chapter 4waters are often salty, but considerably less salty than produced water from oil and gas activities.In many states (including Pennsylvania and West Virginia) coal bed methane produced water ispermitted to discharge to surface water, and discharges can be considerable (32.3 milliongallons/year in the Appalachian basin; USEPA 2010). Treatment is generally necessary for useof AMD-impaired waters in hydraulic fracturing, with removal of sulfate as a critical issue toprevent formation of barium sulfate precipitates that clog the well. Both abandoned and currentcoal mine discharges vary significantly from site to site, so site specific characterization andtreatment would be needed for use of this water. Other industrial wastewaters may also havepotential to be repurposed for shale gas development; however, this would require carefulconsideration of the impacts of wastewater diversion on in-stream flow in receiving waters.Further, the requirements for pre-treatment of different wastewaters may make this optionimpractical.Beyond water conditioning, issues of ownership and liability are a concern with use of impairedwaters. As noted for Braddock Run, AMD-impaired water can even play an important ecologicalrole as well. Drilling companies have expressed significant interest in use of AMD-impairedwaters, but they do not want to be liable for cleanup of the continuing source once their short-term need for water ends. An evaluation of the potential for use of coal mine drainage forhydraulic fracturing was completed by the Rand Corporation in late 2011(Curtright and Giglio2012). Conclusions included a need for new studies on sources of coal mine water that would beavailable for hydraulic fracturing, the evaluation of quantity and quality available across aregion, and a collaborative approach among regulators, industry and other stakeholders todevelop and analyze technical concepts and implementation mechanisms. Clearly, a bestpractice would be for Maryland to conduct a feasibility study on the potential use of knownAMD-impaired waters in Garrett and Allegany County as source water for potential hydraulicfracturing operations. This is particularly important since as noted above, small headwaterstreams supply reservoirs extensively used for drinking water supply and are unlikely to haveexcess capacity for withdrawals. Water resources for extraction activities may be limited in areastargeted for development. As part of this study, Maryland should evaluate any regulatorylimitations that would interfere with beneficial repurposing of mine water for hydraulicfracturing.H. Chemical delivery, storage on-site, and transfersChemicals will need to be delivered and stored on site prior to drilling and completionoperations. Some of these chemicals are hazardous and attention must be paid to their propermanagement. Closed storage tanks are necessary for all chemicals used on site. All tanks shouldbe maintained in secondary containment to prevent contamination of the environment in theevent of a spill. Adequate secondary containment should also be used in all areas where blendingor transfer of chemicals takes place (NYSDEC 2011). Spill prevention, response and remediationplans (see Chapter 7 for details) should be developed and approved during well permitting andfully implemented when construction begins. Residual chemicals are not exempt wastes andmust be managed based on their hazard classification. No blending of residual chemicals withproduction wastes is permitted under federal law. Operator training should be specificallyrequired regarding the exempt vs. non-exempt wastes classification at the well pad, as this is anarea of common confusion. 4-22

93. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandI. Identification of chemicalsThe composition of chemicals used on site must be clearly identified for safety and to enableremediation in the event of accidental releases. While some industry participants have beenproactive in disclosure of chemical use, including chemical use in drill plans and spillprevention, response and remediation plans, this openness is not universal. Many, but not all,operators provide chemical disclosure through the web site, FracFocus (www.fracfocus.org ).Disclosures are permitted to include chemicals listed as proprietary if they represent a tradesecret, as defined by applicable U.S. law. A best practice would be a requirement by the state ofMaryland that operators provide full disclosure of chemicals used during completions. Detailedinventories including Material Data Safety Sheets (MSDS) should be required on site and on alltruck manifests. To support preparations and training by first responders and well pad staff forany chemical emergencies, lists of chemicals to be used on site (plus appropriate toxicologicaldata, chemical characterizations, MSDS, and spill clean-up procedures) should be included inpermit applications. It must be kept in mind that MSDS may not contain information on specificchemical compounds, so it isn’t clear to us whether this information is sufficient to fully protecthuman health in the event of a spill or other emergency, however.Drillers report using fewer chemicals for MSGD than for other shale gas plays and the economicincentive to reduce chemical use even further is a strong motivator. To encourage advancementsin “greener” (i.e., use of more benign chemicals) completions, Maryland should requirecompletion plan alternatives during the permitting process. These recommendations areconsistent with proposed practices for MSGD in New York (NYSDEC 2011).J. Drilling and drilling wastesOil and gas development produces drilling wastes that must be temporarily stored on site,processed, and disposed of. Until very recently, storage was accomplished using lined open pits,but these can no longer be considered best practices. Closed-loop drilling systems that sit withinsecondary (and perhaps tertiary) containment are preferable to open pit systems and should beconsidered a best practice for Maryland. As with all waste handling and processing, adequateplans for spill mitigation must be in place in the event that an accidental release occurs. Sincemost drilling muds contain polymer additives, cuttings generally represent a mixture of nativeand amended materials that should be managed in accordance with their chemical characteristics.While oil and gas production wastes have a federal statutory exemption under RCRA and are,therefore, not categorized as hazardous wastes, they should be managed as wastes and theirdisposal should be based on their characteristics. Drill cuttings should be separated, recycled, orproperly landfilled. Due to the potential for cuttings from shale formations to contain NORM,on-site disposal should not be permitted. Landfill disposal should be allowed when NORM levelsindicate no significant enrichment beyond background levels. State action levels for NORMrange from 5-30 picocuries per gram (pCi/g) of total radium and levels considered to represent‘uncontaminated’ materials are often set at twice the background level (NYSDEC 1999). Asdiscussed in more detail in Section P, radioactivity monitoring at landfills is recommended toavoid unintentional comingling of radioactive wastes in conventional landfills. 4-23

94. K.N. Eshleman & A.J. Elmore (2013) Chapter 4K. On site management of produced waters and wastewatersOnce drilling and completion have taken place, produced water will begin to be generated fromthe well during the flowback period, and later during the production phase. Flowback andproduced water are not distinct wastewaters. The definition of flowback water is imprecise andcan vary from well to well. It is sometimes operationally defined as water returning after the wellcompletion for the first 10-14 days. It can also be defined as water returning after the completionand before the well head is installed for production (which can be sooner than ten days aftercompletion). Because there is no consensus on when the transition from flowback to producedwater occurs, and the water quality of both can vary significantly over time, flowback andproduced water should not be treated as distinct classes of wastes. If distinction is desired, forexample, because flowback may be higher quality (lower salts) and thus have alternative disposaloptions, the distinction should be made based on the quality of the water (i.e., a specificconcentration of salts or specific chemicals such as strontium or NORM), not its classification aseither flowback or produced water.Direct discharge of drilling wastewaters at the development sites is precluded by federal law25,which requires zero discharge from onshore gas wells (see Section C). Thus, all produced watermust be collected and stored for either reuse on-site or shipment off-site for treatment ordisposal. Treatment and disposal off-site are discussed in section L below.Significant quantities of water initially return to the surface. The volumes of produced water canvary considerably in different shale gas plays and even in different wells in the same formation.Typically 10-25% of the injected water returns to the surface as flowback during the pre-production phase (Hayes and Schroeder 2009). Marcellus formation wells have reported lower(10-15%) flowback rates, however (Hoffman 2010, Mantell 2011). Thus, on-site storage ofsignificant volumes of produced water must be accommodated immediately after wellcompletion. This has typically been accomplished in open impoundments where produced wateris mixed with freshwater for makeup of the next well completion, however we stronglyrecommend that well pads sited in wet climates such as western Maryland utilize closed wastetanks for wastewater containment (with adequate secondary containment). Secondarycontainment (including dikes, liners, pads, curbs, sumps and other relevant structures) should beemployed to minimize the potential for accidental releases of production wastes from thesecontainment facilities.Despite the challenges associated with on-site management of large volumes of produced water,recycling this water for use in subsequent hydraulic fracturing operations either on the same siteor at another site is an obvious best practice. Water produced in the flowback period can bestored on-site for use in a subsequent completion without any transport costs. Minimal treatmentis necessary (e.g., settling) prior to dilution with additional freshwater for the next completion(Blauch 2010, Grottenthaler 2010). Recently Pennsylvania issued a general permit(WMGR1221) that covers treatment of produced water for subsequent reuse in hydraulicfracturing and encourages 100% recycling for water produced at well pads under development. Itis not clear how such water is tracked or reported in Pennsylvania, but most large companiesreport nearly 100% recycling of early phase produced water (i.e., flowback) (Grottenthaler 2010,25 40 CFR §435.32. 4-24

95. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandVeil 2010). Maryland should include a very strong preference for on-site recycling in permittingshale gas development.Reuse in subsequent activities is the best management practice for later produced water as well,provided active new well development is taking place within the region and shipment distancesare reasonably short. This is principally a logistical challenge, although trucking produced waterincreases both costs and the risks of spills during transit. While produced water is much saltier,its volume is quite low—often less than 200 gallons per million cubic feet (MMCF) of gas isreported in the Marcellus region (Mantell 2011). Thus, extensive dilution would occur increating the next makeup water to achieve the necessary volumes for a completion. Many drillershave current goals of 100% recycling of all produced water, however this management optionwill not be economical if newer drilling pads are sited long distances from existing producingwells. This may not be a huge problem in Maryland, especially if MSGD can be sited inclustered industrial developments as discussed in Chapter 1. Produced water can also be treatedoff-site and returned for reuse at the same pad or to other well pads (as discussed in Section Lbelow). This approach is common in Pennsylvania where centralized treatment plants offerpartial treatment of produced water (removal of everything except monovalent ions: Na+, Cl-,Br-) with return of the highly saline water to the well pads for reuse. Maryland should include astrong preference for reuse of produced waters for subsequent shale gas activities, but shouldconsider whether cross-state transfers of produced waters should be permitted for this purpose.Given western Maryland’s centralized location between two neighbor gas-producing states, thismight be an efficient option. Permit applications should definitely include plans for producedwater reuse and should specify which wells, within defined distances, will share water for reuse.L. Management of produced water (including recordkeeping, manifesting)As noted in Section K, there is no generally accepted definition of flowback and no legaldefinition that distinguishes between flowback and produced water. Best management practicesshould not attempt to distinguish based on these imprecise classifications but rather should referto water quality characteristics if distinct handling is warranted. In most cases, management ofthe two wastewaters should be similar. Flowback may contain lower concentrations of salts, buthigher concentrations of residual chemicals from the original hydraulic fracturing fluids.Produced waters generally become more concentrated in salts over production time; NORM mayalso increase with time in the produced water.Wastewaters produced during oil and gas development in the U.S. are considered non-hazardousby statutory exemption from RCRA26. As a non-hazardous waste, oil and gas production wastesare subject to different requirements for generation, transportation, treatment, storage anddisposal. Maryland has been authorized by the USEPA to operate its hazardous waste regulatoryprogram in lieu of the federal government, based in part on state regulations being at least asstringent as corresponding federal regulations. In some instances, Maryland’s regulations aremore stringent than federal regulations as is allowed by federal law; Maryland has adopted the26 “Identification and Listing of Hazardous Waste; 40 CFR 261.4(b): Exclusions: solid wastes which are not hazardous wastes”. 40 CFR 261.4(b)(5) 4-25

96. K.N. Eshleman & A.J. Elmore (2013) Chapter 4exemption for exploration and production wastes associated with oil and gas development,however27.Some testing is generally undertaken to evaluate usability for recycling at treatment plantsaccepting these wastes to determine if they can be adequately treated by the methods employedat the plants. Wastes that do not meet certain criteria will not be accepted at certain treatmentplants. Similarly, some testing is undertaken at deep well injection sites to determine if anytreatment is necessary to avoid well clogging during injection. Results of this type of testing areused internally at these facilities. They are not reported to any regulatory agency, nor are theynecessarily kept beyond the decision-making process. Trucks transporting any wastewater arerequired to carry manifests regarding their cargo, but are not required to be placarded ashazardous. Maryland should review its requirements for testing and manifesting of hazardousand non-hazardous waste to determine which regulatory structure applies to oil and gas producedwastes.Volumes of produced water from each well are generally reported on regular intervals to thestate. These reports include the well number, volume of produced water, and the name andlocation of the waste operator to whom the water was taken (e.g., waste treatment facility,underground injection well site, etc.). If produced water was reused rather than disposed of, thisis also noted, although in Pennsylvania the location of the reuse is not specified. These reportsare tabulated in Pennsylvania and released to the public (via a web site) twice a year. It is notpossible using Pennsylvania data to track water from extraction to use to reuse or ultimatedisposal. A water balance for the industry cannot be completed because of insufficient detail inthe water withdrawal plans (which withdrawals are for which wells) and insufficient detail on thereuse of flowback and produced water within multiple wells prior to ultimate disposal. In WestVirginia, additional transportation records are required that might allow tracking of water fromoriginal withdrawal to final disposition, although the state does not undertake to evaluate wateruse and wastewater generation in this way. Maryland should require reporting of produced watervolumes from every well, including the well location, the company providing transport of theproduced water and the ultimate disposition of the waste, including the location of thesubsequent well if the produced water was reused in hydraulic fracturing.In many areas of the country, road spreading of oil and gas brines is used for dust control ordeicing. Generally due to higher levels of residual fracturing chemicals, this is not permitted forearly produced water (i.e., flowback), but is commonly practiced with the low volume, high saltwater that returns during the production phase. Spreading on roads within the oil and gasdevelopment region (often dirt roads created by the drilling companies), as well as insurrounding rural areas, is not uncommon. In Pennsylvania and West Virginia this activity ispermitted for certain brines. Clearly, surface applications to roads or land will result in eventualrunoff and entry of constituents present in the brine into the surface and ground water systems.Brines from the Marcellus formation contain very high concentrations of salt and are notappropriate for open discharge to the environment, particularly given western Maryland’sdependence on drinking water obtained from private groundwater wells and surface waterobtained from headwater streams. Road spreading of the original produced water or any residualof its treatment should be prohibited. Several treatment facilities have suggested the creation of a27 COMAR 26.13.02.04-1A(5) 4-26

97. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandsalt product that would be suitable for road application through treatment of gas well brines. Thisshould only be permitted if the created salt is low in bromide and iodide and replaces aconventional salt product already being used in the region. Underground injection is the commonfinal disposal option for produced water in the U.S., and it is also appropriate for liquid residualsfrom treatment operations (e.g., highly concentrated brines). We discuss this option further insection N below, but it is unlikely to be used in Maryland.M. Treatment of produced water28Produced water can be treated using a variety of chemical and physical processes to removecontaminants. Generally, radionuclides and multivalent metal ions are relatively easy to removethrough coagulation, precipitation and filtration. Organics and oils can be removed throughskimming or sorption. Monovalent ions (Na+, Cl-, Br-) are particularly difficult to remove,requiring either membrane or thermal technologies that are energy-intensive. Concentrations ofsalt in produced water from the Marcellus formation in Pennsylvania and West Virginia are toohigh for membrane systems such as reverse osmosis; therefore, thermal technologies, includingevaporation, multi-state flash distillation, and humidification-dehumidication methods, are theonly viable treatment technology for the simple salts (Hayes 2009). Thermal desalination resultsin either an even more concentrated brine solution or a solid salt product as a residual. Theseresiduals must be managed while the desalinated water can either be reused or dischargedthrough a NPDES permit to surface water.Typically, partial treatment for recycling can be performed on-site (as described in section K),but all other treatment methods typically take place off-site. Produced water is picked up from anumber of wells in tanker trucks on a regular schedule and taken to a centralized brine treatmentplant. There are no permitted centralized brine treatment plants currently in Maryland. However,there are plants in West Virginia and Pennsylvania, and additional plants have been opening inresponse to additional produced water volumes requiring management. Two types of plants areoperating in Pennsylvania. A few brine treatment plants that were operating before regulatorychange to discharge standards on salt continue to operate without TDS limits for their discharge.These plants remove most contaminants except monovalent ions (Na+, Cl-, Br-), and dischargethe residual high salt brine to surface waters (creeks and rivers). However, these surfacedischarging plants no longer receive Marcellus formation produced water as the PA DEPrequested that drillers stop using this disposal method. In Pennsylvania, new treatment plantsmust meet a discharge limit of 500 mg/L TDS. Maryland does not have a numerical criterion forTDS (or specific conductance—a related parameter) for in-stream water quality or for dischargepermit limits. We do not recommend the use of brine treatment plants that partially treatproduced water and discharge high concentrations of salt to the environment in Maryland. Toavoid salt discharges into critical drinking water areas, prior to approving brine treatment plants,Maryland should enact a discharge permit limit for TDS and in-stream standards for TDS,chloride and bromide. Under no circumstances should Maryland allow discharge of partially-treated brine or residuals from brine treatment facilities into the waters of the state. Further,development of brine treatment plants that recycle water to drillers should be discouraged infavor of on-site treatment by mobile units and immediate reuse as this decreases truck transportand associated impacts.28 An extensive review of treatment options is provided in Hammer et al. (2012). 4-27

98. K.N. Eshleman & A.J. Elmore (2013) Chapter 4Other brine treatment plants operating in Pennsylvania and West Virginia do not dischargewastewater to the environment. They treat produced water to remove metals and organics, andreturn the brine containing only monovalent ions to drilling companies for reuse. This is verysimilar to the on-site partial treatment for reuse discussed in Section K. Because this treatmentrequires additional transportation of the water (and adds to the associated impacts and risks), off-site treatment for reuse would be an even poorer option than on-site reuse in Maryland.However, if on-site treatment were deemed infeasible or off-site treatment facilities were closerthan subsequent wells requiring water for reuse, such treatment plants could play a useful role.Materials removed from the water as sludges during treatment processes are typically dried anddisposed of at landfills. Some of these treatment plants have plans to add additional treatment todesalinate water to acceptable discharge levels; however, such second stage treatment is notoperational at most plants in Pennsylvania due to low demand for that type of extensivetreatment. If the market for partially-treated water for reuse declines, these plants will likely offerfull distillation services, but this will increase the treatment costs significantly. Distillation willproduce highly concentrated brines or solid salts that will require subsequent management, eitherat deep well injection sites or landfills. The potential to create usable salt products from thisprocess has been discussed, but technological and regulatory hurdles remain. Best managementpractices for residuals are discussed below.During the recent rapid expansion of development in the Marcellus formation in Pennsylvania,another management method for produced water was used. Produced water was sent toconventional wastewater treatment plants, classified by USEPA as publicly-owned treatmentworks (POTWs). These plants treat domestic sewage and are not designed to remove chemicalsfrom oil and gas brines. However, relatively large wastewater volumes diluted the salts, and thismethod was considered acceptable as long as brine flows were low. In 2008 and 2009, manyPOTWs in Pennsylvania were accepting higher flows than their wastewater could adequatelydilute. Concentrations of salts in the receiving waters rose unacceptably high in the fall of 2008.Concentrations of bromide, an ion with implications for drinking water treatment, rose as well.The PA DEP intervened in 2008 and 2009 to restrict use of POTWs for brine treatment. USEPAhas also provided clarifications regarding the acceptance of oil and gas wastewaters and madesubstantial changes that required permit modifications29. In 2011, PA DEP requested thatMarcellus wastewater not go to any surface-discharging POTWs without a TDS standard in theirpermit. Anecdotally, TDS and bromide levels have reportedly been lower in 2012 in somewaterways (e.g., the Monongahela), but not in others (e.g., the Allegheny) following thesechanges30. In addition to Pennsylvania, Ohio is considering use of POTWs for brine treatmentand disposal, and several lawsuits surrounding this situation are pending in that state.Following concerns regarding the use of POTWs, often without pretreatment of the producedwater, the USEPA announced plans to develop pre-treatment requirements for oil and gaswastewaters being sent to POTWs (USEPA 2011). These rules are pending at this time. Ifpromulgated, the rules would specify pre-treatment methods or water quality criteria for pre-29 Ibid., 530 To the best of our knowledge, these data have not been published in the peer-review literature, but the situation was widely covered in press accounts. 4-28

99. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandtreated wastewater that would be sent to POTWs. These rules are unlikely to address all traceconstituents of concern in produced water, as POTW treatment relies heavily on dilution tomanage low concentration pollutants. For example, it is unclear if pre-treatment rules wouldconsider bromide, a low concentration pollutant of concern only for downstream drinking waterplants.Use of POTWs for dilution of produced water from oil and gas development is not a bestmanagement practice. Disposal practices that load salts (especially those containing chloride andbromide) to surface waters that are used for drinking water sources should not be permitted. Thisactivity impedes treatment of water to provide water that is potable and safe for consumers.Higher chloride levels cause taste and odor problems in finished water. High bromide levels leadto increased formation of carcinogenic disinfectant by-products that can persist in the water tothe point of consumption. Treatment of produced water by POTWs and other conventionalwastewater treatment methods that do not remove salts should be prohibited in Maryland.A significant concern for any treatment method is the production and management of residuals.For most treatment systems, solids are removed into wet sludges, which can be disposed of inlandfills as non-hazardous wastes. Treatment residuals created from exempt oil and gas producedwaters are also exempt from federal laws related to hazardous waste, provided the exempt wasteis not mixed with a non-exempt waste prior to the treatment process31. Maryland has not objectedto the ‘derived from’ interpretation in RCRA that exempts residuals produced from exemptwastes. Treatment residuals will contain removed contaminants such as NORM, heavy metals,organic compounds, and salts, and these residuals should be evaluated for their constituents andmanaged accordingly. Since treatment residuals will generally have more concentrated levels ofcontaminants found in the original wastewater, deep well injection disposal is the preferredmanagement strategy.N. Disposal of produced water or residual treatment wastesMost produced water, as well brine residuals from treatment of produced water, in the U.S. isdisposed of through deep injection in Underground Injection Control (UIC) Class II wellsspecifically designed for disposal of brines and other fluids associated with oil and gasproduction, following requirements of the Safe Drinking Water Act of 1974 (Part C § 1421-1426). The Marcellus region does not have an extensive deep well injection infrastructure.Pennsylvania has only a few UIC Class II wells that are mostly privately owned. West Virginiaand Ohio have considerably more such facilities, including both private and commercialfacilities. In general, trucking costs can make this disposal option prohibitively expensive fordevelopment in neighboring states. Concerns in Ohio regarding earthquakes associated withunderground injection have also limited the new citing of commercial disposal wells. Marylanddoes not have a single UIC Class II disposal well. At present, disposal through deep wellinjection will require either trucking wastes to neighboring West Virginia or siting, permittingand drilling injection wells within Maryland. Maryland may prefer to develop UIC Class IIinjections wells to avoid long distance trucking of produced waters. However, these wells arecommonly sited in played out areas of gas development, which Maryland does not have, or inareas that are also suited for gas storage, which Maryland is currently using for such purpose.31 58 FR 15284, 15285 (March 22, 1993) 4-29

100. K.N. Eshleman & A.J. Elmore (2013) Chapter 4We anticipate that locations in Maryland for siting injection wells may be very limited. Reviewof injection well concerns in neighboring states and geological survey of Maryland subsurfaceformations should be undertaken prior to consideration of this activity in Maryland. Further,Maryland should review the relevant regulations surrounding development and use of UIC wellsfor produced water from shale gas development, and at the same time evaluate the capacity ofnearby states to accept produced water or residual brine from treatment of produced water beforepermitting any development in the state.O. Reclamation and closure (decommissioning)In general, once production begins, interim reclamation of the site can be achieved through re-vegetation and modifications to the stormwater management systems as discussed earlier(Section C). This should normally take place within 60 days of the initiation of gas production,and the operator should submit a site restoration report to the state. Interim reclamation of thesite after completion restores the disturbed land but allows continual access for collection ofproduced water, transport of gas (via gathering pipelines) and access for any future wellreworking, if necessary. Road access must be maintained, but other areas of the site that weredisturbed for heavy equipment needed during drilling and completion should be regraded andrevegetated. All solid wastes should be disposed of using methods appropriate for the waste type,following state regulations. Residuals from gas drilling activities should be evaluated for theirconstituents to determine their status as hazardous wastes and managed accordingly. Waterstorage impoundments should be closed and these disturbed areas reclaimed. Pit liners should beremoved and landfilled off-site. On-site disposal of residuals should not be permitted.Stormwater management for control of erosion and sedimentation should continue until the siteis fully reclaimed. In Chapter 1 we discussed the rationale behind allowing development atmulti-well pads to proceed more cautiously (especially as the first wells are drilled and theproductive capacity of these wells is ascertained) and some of the ramifications for such aprocess from the standpoint of interim reclamation. We recommend that Maryland study verycarefully how the development process at multi-well sites has taken place in other states(particularly Pennsylvania) and establish suitable regulations that balance the need to keep wellpads completely operational for extended periods of time against the goal of ensuring that partialreclamation of these sites is not unduly delayed.Once production declines, wells are sometimes shut in for possible future development (i.e.,refracking or drilling again from the same well pad location). A time limit should be establishedfor wells in this status. If no additional development takes place within 12 months, sitereclamation must begin unless an extension is issued by the state. Once no further wellproduction is expected, final reclamation of the site should take place. As discussed in Chapter 3,this requires plugging the well to ensure isolation from the surface and near surfaceenvironments—a critical process in preventing water quality impacts from movement of residualgas or brine in the formation (as well as unintended losses of the gas resource). Permanentsignage should be left in place to allow the well to be located if necessary in the future.Restoration plans should be developed in detail and submitted to the state. They should includestabilization and revegetation of all disturbed areas, including recontouring to reestablish theoriginal topographic contours, use of native plant species and use of agency-approved seeds, andremoval of all surface components of the facility (see Chapter 1). The goal of reclamation shouldnormally be to return the developed area to native vegetation (or pre-disturbance vegetation in 4-30

101. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandthe case of agricultural land returning to production) and restore the original hydrologicconditions to the maximum extent possible.Improperly closed wells have led to significant environmental impacts in oil and gas states in thepast (including Pennsylvania and West Virginia32,33) and thus should be avoided. Detailed,proscriptive methods for well closure should be developed through review of industry bestmanagement practices and other state regulations (see Chapter 3). Maryland should take aproactive approach to regulation in this area. With the implementation of best practices in wellclosure, Maryland should be able to avoid the problems that other states have experiencedthroughout their long oil and gas development histories.P. Naturally occurring radioactive materials (NORM)Shale gas formations often contain naturally-occurring radioactive materials (NORM) and drillcuttings and production brines can contain NORM. NORM should be assessed in all componentsof waste associated with gas production from shale. A 1999 study in New York State is a modelfor assessment of the industry for NORM concerns (NYSDEC 1999). Extensive on-site samplingand monitoring found NORM levels near background at most oil and gas production sites.Similar monitoring should be routine at MSGD sites to ensure adequate protection of workersand the environment.Oil and gas production wastes that exceed target levels should be reclassified as radioactivewaste (RW) and not fall under the federal exemption of oil and gas production wastes34.Similarly, when drill cuttings and production wastes are treated or disposed of, residuals fromthese processes can become enriched in NORM (then called technologically-enhanced, naturally-occurring radioactive materials or TENORM). This enrichment process should be monitoredand, if necessary, residuals should be reclassified as radioactive waste (RW) to ensure they aretracked and protective disposal practices are used. All drilling wastes should be evaluated for thepresence of NORM to ensure adequate disposal. Pennsylvania landfills have utilized radiationdetection systems to ensure that radioactive wastes are not incorrectly comingled withconventional non-hazardous solid waste (PADEP 2012). Programs to monitor for radioactivewaste at landfills should be adopted in states where shale gas drilling wastes may be sent tolandfills. Maryland should adopt monitoring at solid waste disposal landfills for radioactivity.If NORM or TENORM waste associated with oil and gas production or waste treatment containslevels of radioactivity that would result in classification as low-level radioactive waste (LLRW),these residuals should be treated in accordance with LLRW regulations. LLRW generated intreatment of produced water from MSGD in Maryland would likely be disposed of outside of thestate. The Appalachian States Low-Level Radioactive Waste Commission is an interstate agencyestablished to assure interstate cooperation for the proper management and disposal of low-level32 Methane Emissions Project, Borough of Versailles, Pennsylvania. Available on the web at: http://www.netl.doe.gov/newsroom/versailles/Versailles%20Methane%20Emissions%20Project%20- %20Final%20Report.pdf33 PA Department of Environmental Protection, Bureau of Oil and Gas Management, Stray Natural Gas Migration Associated with Oil and Gas Wells, available on the web at: http://www.dep.state.pa.us/dep/subject/advcoun/oil_gas/2009/Stray%20Gas%20Migration%20Cases.pdf34 Federal Low Level Radioactive Waste Policy Act specifically exempts NORM. 4-31

102. K.N. Eshleman & A.J. Elmore (2013) Chapter 4radioactive wastes. The Commission identified Pennsylvania as the host state to receive anddispose of radioactive waste from the party states (Delaware, Pennsylvania, West Virginia andMaryland). Maryland producers of LLRW can also contract with out of state facilities throughapproval by other interstate commissions (e.g., Texas compact).Gas operations are typically short term and thus, build up of NORM at a given site is notexpected; however, equipment is moved from site to site and could develop scale thatincorporates NORM. Oil and gas production equipment should be assessed at regular intervals.Tanks on-site holding produced water could also develop precipitates with time that containNORM. Regular inspections and cleaning of equipment and facilities that might be susceptible tothe development of TENORM are recommended as part of best management practices for theon-site management of produced water.Q. Key recommendations4-A A best practice for Maryland would be establishment in regulation of 500 ft. and 2,000 ft. setbacks (measured from the well pad, not from the individual wellbores) for private wells and public system intakes (both surface and groundwater), respectively.4-B We support Maryland Environmental Code § 14-110.1 (H.B. 1123) and recommend pre- development notification should be made to public and private drinking water well owners.4-C Pre-drilling groundwater testing should be required to be conducted by the operator and the results provided to MDE and to the well owner. Post-drilling testing is often at the discretion of the well owner, but a best management practice that would enable improved understanding of the potential for effects on groundwater would be to require post- drilling and completion testing by the operator for all wells within a pre-determined potentially affected region for a specified time period after completion of well construction activities.4-D Maryland might wish to consider ways of strengthening its anti-degradation policy to take account of the impacts of non-point source pollution that are a major threat to its high quality waters. One way that this might be accomplished would be by revising the WQS rules to require that any land development practices (e.g., forest management, MSGD, etc.) conducted in Tier II watersheds meet an anti-degradation standard.4-E Maryland needs to carefully review its stormwater regulations as they pertain to oil and gas extraction; we recommend oil and gas extraction sites be considered “hotspots.” Based on our review of stormwater management practices in other states, we recommend the use of both “active” and “passive” stormwater management: (1) the construction of properly bermed “zero-discharge” pads that effectively collect all water on a pad site and enable the reuse of this water during drilling and completion operations; and (2) construction of a below-grade lined pond adjacent to the bermed zero-discharge pad that could be used as a sump during active stormwater management phases and easily converted into a retention pond prior to a passive phase.4-F Post-construction inspections of stormwater structures should occur prior to well drilling and completion. 4-32

103. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland4-G There are very long gage records available from USGS for most of the major western Maryland rivers (Youghiogheny, Casselman, Savage, Potomac, Georges Creek) that could possibly be used to support MSGD; data for these and other gaged systems can be used to inform a quantitative analysis of acceptable water withdrawals for MSGD. This analysis is much more difficult for smaller streams and rivers due to data limitations, although we believe that such an analysis should be done. Our experience in Maryland watersheds as well as review of other areas that have completed such analysis, suggest that in western Maryland, water withdrawals for proposed MSGD would need to occur solely from the region’s large rivers (and perhaps from one or more reservoirs). Small streams (1) have significant existing withdrawals for drinking water; (2) have small catchment areas and discharges under most conditions; (3) are very unlikely to have excess flow capacity for new permitted withdrawals; and (4) can be readily dewatered. Water may need to be temporarily stored in centralized freshwater impoundments specifically constructed for this purpose, but such impoundments should never be allowed to receive or store any wastewaters.4-H To support preparations and training by first responders and well pad staff for any chemical emergencies, lists of chemicals to be used on site (plus appropriate toxicological data, chemical characterizations, MSDS, and spill clean-up procedures) should be included in permit applications.4-I Closed-loop drilling systems that sit within secondary (and perhaps tertiary) containment are preferable to open pit systems and should be considered a best practice for Maryland.4-J Maryland should include a very strong preference for on-site recycling of wastewaters in permitting of shale gas development. Under no circumstances should Maryland allow discharge of untreated brine, partially-treated brine, or residuals from brine treatment facilities, into the waters of the state. Development of brine treatment plants that recycle water to drillers should be discouraged in favor of on-site treatment by mobile units and immediate reuse as this decreases truck transport and associated impacts.4-K Maryland should review the relevant regulations surrounding development and use of underground injection wells for produced water from shale gas development and, at the same time, evaluate the capacity of nearby states to accept produced water or residual brine from treatment of produced water before permitting any development in the state.R. Literature citedBlauch, M. E. 2010. Developing effective and environmentally suitable fracturing fluids using hydraulic fracturing flowback waters. SPE Unconventional Gas Conference, Pittsburgh PA, Society of Petroleum Engineers.Boyer, E. W., et al. 2011. The Impact of Marcellus Gas Drilling on Rural Drinking Water Supplies, Pennsylvania State University.Castro, M. S., et al. 2001. Water Quality Assessment of the Piney Creek Reservoir and Watershed in Western Maryland. A. L. University of Maryland Center for Environmental Sciences, Report for Maryland Department of Natural Resources.Curtright, A. E., and K. Giglio. 2012. Coal Mine Drainage for Marcellus Shale Natural Gas Extraction. Pittsburgh PA, Rand Corporation. 4-33

104. K.N. Eshleman & A.J. Elmore (2013) Chapter 4USEPA. 2001. Industrial Surface Impoundments in the United States. Washington DC, U.S. Environmental Protection Agency.USEPA. 2010. Coalbed Methane Extraction: Detailed Study Report. Washington DC, U.S. Environmental Protection Agency.USEPA. 2011. Fact Sheet: EPA Initiates rulemaking to set discharge standards for wastewater from shale gas extraction. U.S. Environmental Protection Agency. http://water.epa.gov/scitech/wastetech/guide/upload/shalereporterfactsheet.pdf.Grottenthaler, D. 2010. Cabot gas well treated with 100% reused frac fluid. DUG 2010, Pittsburgh PA.Hammer, R., et al. 2012. In Frackings Wake: New Rules are Needed to Protect Our Health and Environment from Contaminated Wastewater. New York, New York, Natural Resources Defense Council.Hansen, E., et al. 2010. The Benefits of Acid Mine Drainage Remediation on the North Branch Potomac River. DownstreamStrategies. Morgantown WV, Maryland State Water Quality Advisory Committee.Hayes, T. 2009. Sampling and Analysis of Water Streams Associated with the Development of Marcellus Shale Gas. Prepared for Marcellus Shale Coalition.Hayes, T., and C. Schroeder. 2009. Barnett and Appalachian Shale Water Management and Reuse Technologies. DOE-NETL Oil Technology E&P. Environmental Projects Review, Morgantown, WV.Hoffman, J. 2010. Susquehanna River Basin Commisssion Natural Gas Development. Presentation to public at Marywood University.Kasraei, S. 2012. Letter to Ms. Paula Ballaraon and Mr. John Balay, Susquehanna River Basin Commission. MDE. Baltimore, MD.King, G. E. 2012. Estimating frac risk and improving frac performance in unconventional gas and oil wells (paper SPE 152596). Hydraulic Fracturing Conference, The Woodlands, TX (February 2012).Lien, A. M., and W.J. Manner. 2010. The Marcellus shale: resources for stakeholders in the Upper Delaware watershed region. Pinchot Institute for Conservation.Mantell, M. E. 2011. Produced Water Reuse and Recycling Challenges and Opportunities Across Major Shale Plays. EPA Hydraulic Fracturing Study Technical Workshop #4: Water Resources Management, Chesapeake Energy.MDE. 2000. Conservation Practice Standard for POND, USDA Natural Resources Conservation Service.MDE. 2009. Maryland Stormwater Design Manual, Volumes I and II. Baltimore MD: Maryland Department of the Environment.MDE. 2011. Maryland Standards and Specifications for Soil Erosion and Sediment Control. Baltimore, MD: Maryland Department of the Environment.MDE. 2012. Maryland Model Erosion and Sediment Control Ordinance. Baltimore, MD: Maryland Department of the Environment.Marcellus Shale Advisory Commission. 2011. [Pennsylvania] Governors Marcellus Shale Advisory Commission Report .Murphy, J., et al. 2012. Predicting Ecological Flow Regime at Ungaged Sites: A comparision of methods. River Research and Applications.NYSDEC. 1999. An Investigation of Naturally-Occurring Radioactive Materials (NORM) in Oil and Gas Wells in New York State. Albany, NY: New York State Department of Environmental Conservation.NYSDEC. 2011. Revised Draft Supplemental Generic Environmental Impact Statement on the Oil, Gas, and Solution Mining Regulatory Program.PADEP. 2003. Evitts Creek Source Water Assessment, Pennsylvania Department of Environmental Protection: 2. 4-34

105. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandPADEP. 2012. Monitoring for Radioactive Materials in Solid Waste. Retrieved November 27, 2012, from http://www.dep.state.pa.us/brp/Radiation_Control_Division/SolidWasteMonitoring/SolidWasteR adMonitoring.htm.Razavi, T. and P. Coulibaly. 2012. Streamflow prediction in ungauged basins: review of regionalization methods. Journal of Hydrologic Engineering.Shu, C., and T. B. Ouarda. 2012. Improved methods for daily streamflow estimates at ungauged sites. Water Resources Research 48 (W02523), doi: 10.1029/2011WR011501.SRBC. 2012. Natural Gas Shales and Natural Gas Well Development. Retrieved December 4, 2012, from http://www.srbc.net/programs/projreviewnaturalgas.htm.State of Oregon Department of Environmental Quality. 2010. Issue Paper: Evaluating the Antidegradation Policy as a Means to Reduce Nonpoint Sources of Toxic Pollutants to Oregon Waters (Final Draft); prepared by D. Sturdevant and J. Wigal.TNC. 2010. Ecosystem Flow Recommendations for the Susquehanna River Basin: Report to the Susquehanna River Basin Commission and the U.S. Army Corps of Engineers. The Nature Conservancy, M. DePhilip and T. Moberg. Harrisburg, PA.Veil, J. A. 2010. Water Management Technologies Used by Marcellus Shale Gas Producers. Argonne, IL, U.S. Department of Energy. National Energy Technology Laboratory.Ziemkiewicz, P., et al. 2003. Long-term performance of passive acid mine drainage treatment systems. Mine Water and the Environment 22:118-129. 4-35

106. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland5. Protecting terrestrial habitat and wildlife1Despite high levels of development in many areas of the state, western Maryland (Allegany andGarrett County) remains a largely intact landscape relative to other regions of the eastern UnitedStates2. The dominant land cover type of the Appalachian mountains of western Maryland isforest (>75%; Figure 5-1), including extensive areas of forest interior habitat (Figure 5-2). Thetotal extent of habitat loss potentially caused by Marcellus Shale gas development is anticipatedto be small relative to other forms of land conversion in the state (e.g., urban/suburbandevelopment, surface mining for bituminous coal), but not insignificant (e.g., Drohan et al.2012). We estimate that with careful planning new disturbances could be less than 1-2% of theland surface (see Chapter 1). Nevertheless, not all forests are of equal value, and rare or sensitiveforest habitat should be avoided. Beyond the direct loss of habitat, many species in westernMaryland are potentially sensitive to new construction that reduces the amount of core andconnected habitat, creates opportunities for direct exposure to toxic contaminants, providesopportunities for biological invasions, or alters the soundscapes and night-time lightscapes.Eighty-seven animal species and 117 species of plants are currently listed as rare, threatened, orendangered within Marylands Marcellus Shale region (Appendix 5A). Many of therecommendations in this chapter are aimed at minimizing the fragmentation of the forestedlandscape with special emphasis on protecting irreplaceable habitats and imperiled terrestrialbiota. Forest protection is also protective of downstream aquatic resources (Chapter 6), andpractices such as the preservation of forested riparian buffer are important to both terrestrial andaquatic resources. In general, no-net-loss of forest is a goal consistent with the states overallresource stewardship and a useful guiding principle for shale gas development in westernMaryland.A. Well pad spacing and sitingAny surface disturbance that punctures intact forest provides an impact that is greater than theamount of forest loss alone (Harris 1984). For example, a 1% net loss of total forest, as wasobserved for the conterminous United States from 2001 to 2006, can translate to net losses of asmuch as 10% of forest interior area (Ritters and Wickham 2012). A tentative pattern has beenreported for Marcellus shale sites in Pennsylvania that interior forest loss is approximately twicethat of the overall forest loss (Slonecker et al. 2012). Another recent study in Pennsylvania foundthat Marcellus well pads and associated infrastructure (roads, water impoundments, andpipelines) required approximately 9 acres per well pad with an additional 21 acres of indirectedge effects3. Loss of interior habitat, defined as areas of forest at least 100 m (328 ft) from the1 Chapter co-authors: Todd R. Lookingbill, Ph.D. (Department Geography and the Environment, University of Richmond, Richmond, VA, 23173); and Andrew J. Elmore, Ph.D. (Appalachian Laboratory, University of Maryland Center for Environmental Science, Frostburg, MD 21532)2 Saylor, K.L. 2008. Land Cover Trends Project: Central Appalachians. U.S. Department of the Interior, U.S. Geological Survey, Washington, DC, USA, available at http://landcovertrends .usgs.gov/east/eco69Report.html.3 Johnson, N. 2010. Pennsylvania energy impacts assessment, Report 1: Marcellus Shale Natural Gas and Wind, The Nature Conservancy, Pennsylvania Chapter, and Pennsylvania Audubon, http://www.nature.org/wherewework/northamerica/states/pennsylvania/news/news3511.html. 5-1

107. K.N. Eshle eman & A.J. Elmore (2013) E Ch hapter 5forest edg (Harper et al. 2005), can limit the ability of fo ge e e forests to sup pport ecologi function ical nsdue to ed effects (F dge Figure 5-2). Edge forests tend to hav more inva s ve asive species higher rate of s, esatmosphe deposition, greater levels of ligh increased predation, a altered b eric l ht, d and biochemicalcycling among other differences with forest core (Harper et al. 2005) To minimi these a c r ). izedisturban and edge impacts, Ne York Sta encourag well pads spacing to be clustered to nce e ew ate ges s dminimize the total su e urface disturb bance (NYSDEC 2011). The Pennsy . ylvania DCN Bureau o NR ofForestry requires app proval of spaacing for any leases on S y State Forest L Land and als promotes so ation of well pads (PADcentraliza l DCNR 2011). As discusse in Chapte 1, multi-w pads are ed er well ehighly enncouraged beecause they increase the regulatory e efficiency fo operators a or along withminimizi surface disturbance (NYSDEC 2011). ing d ( 2Figure 5-1 Land cover map of Allegan and Garrett County (Sourc NLCD 2001)4. Over 75% of the landsca is 1. m ny ce: apecovered by forest. Statew y wide, Maryland is dominated by agricultural land and less than 40% of th state is fores d l he sted.Urban and other lands coomprise 7.5% of the western Maryland, with urban ISA re o M h epresenting the percentage of each f30 m cell that is covered by impervious surfaces. t sThe siting of well pads also is ad ddressed in Chapter 1. In general, open lands are preferred fo C n e ordevelopm over for ment restlands. Peennsylvania emphasizes that well pa should be located in a ads emanner that reduces impacts to forested areas, preferably in already d fo y disturbed lan (PADCN nds NR2011). Upland core forests of Ap U f ppalachia pro otect warbler and other species of s rs special conceern,while cov forests of the region are highly di ve f a iverse and prroductive du to their high moisture and uesoil fertil (Wickha et al. 200 Maxwell et al. 2012). However, i should be n lity am 07, . it noted that so ome4 Througho this report the NLCD 200 is used to illustrate pattern of land-use a land-cover. At this scale, the out t 01 ns and NLCD 2001 is indisting guishable from the NLCD 20 Any subse m 006. equent mapping completed to support the g o regulatio of MSGD an on nd/or the protec ction of natura resources sho al ould use the mo up-to-date and scale- ost appropria data availab ate ble. 5-2

108. Recomm mended Best Management Practices for Marcell Shale Ga Developm s lus as ment in Mary ylandgrassland bird specie are listed as rare in Ma d es a aryland (e.g. the Upland Sandpiper) which cou ., d ), uldlimit the siting of MS SGD in some non-forest areas. Optim e mally, well p pads and asssociated facil litiesshould be sited in ind e dustrial park designed and zoned fo this type o industrial activity, and ks a or of d/orin close proximity to major inters p state highwa and exit r ays ramps desiggned to effici iently handleeround-thee-clock transsportation.Figure 5-2 Forests of we 2. estern Marylan classified as forest core (gr nd s reen), edge (ye ellow), and frag gment (red)assuming a depth of edge influence of 328 ft (Source: NLCD 2001). Of the approximately 815 s e 3 . square miles of fforest in Allegany and Ga A arrett counties, 60% is core, 38% edge, and 2% fragment by this criterio , 3 d on.Under the Maryland Forest Cons servation Act (FCA) of 1 1991, any ac ctivity requir ring anapplicatio for a subd on division, gra ading permit or sediment control per t rmit on areas 40,000 squ s uarefeet or gr oximately 1 acre, or abou ¼ the area of a typical drill pad site) requires a reater (appro a ut a lForest Co onservation Plan. Althou the FCA does not sp ugh A pecifically ca for no-ne all et-loss of forest inMaryland the require Forest Co d, ed onservation Plan does in P nclude tree prrotection speecifications,mitigatio planting, and a long-te protectio agreemen that is pla on a erm on nt aced on the re etained foresstand mitiggation areas. The Act is at least parti a ially respons sible for reve ersing the ra of forest l ate lossin the sta from a hig of 0.6%/y in 1990 to 0.1%/yea 20025. In 2008, Mary ate gh year 0 ar yland Senate Bill431 creat a task for to develo a plan to implement a no-net-loss of forests p ted rce op s policy forMaryland A key find d. ding of the ta force wa that: ask as Marylland needs to move toward a No-Net-Lo of forests as a strategic component in the effort t o d oss c to restor the health of the Chesap re o peake Bay and its watershe and promo the econom well-being of d ed ote mic rural Maryland wi strategies and policies that measurab contribute to enhanced forest land ith t bly e d retent tion and impr roved forest la stewardsh without n egatively imp and hip pacting produ uctive agricul ltural5 Maryland No Net Loss of Forest Task Force Final Report and Reco d k R ommendations January 2009 s. 9. http://ww ww.dnr.marylan nd.gov/dnrnewws/pdfs/NNLTF FFINALREPO ORT1.pdf 5-3

109. K.N. Eshleman & A.J. Elmore (2013) Chapter 5 lands. Such efforts must be conservation oriented rather than preservation oriented in nature to be 6 able to sustain native plant communities in a developed landscape.Unfortunately, Allegany and Garrett County are currently exempt from the Maryland FCAbecause of their high percentages of forest land-cover and perceived lack of threat fromresidential development relative to the rest of the state. Marcellus shale gas developmentrepresents a new land-cover change process on these landscapes. Therefore, Maryland shouldconsider adopting a no-net-loss of forest approach to Marcellus Shale gas development. Theprimary mechanism for implementing this approach would be through the requirement of forestplantings elsewhere in Maryland to mitigate any well pad or related MSGD that remove forestfrom the landscape. The requirement might be best implemented by expanding the FCA toinclude the two counties of western Maryland. Regardless of the mechanism, the siting ofmitigation plantings should consider regional conservation goals and water quality improvementpotential. Clearly such an approach would help to incentivize the siting of well pads and otherinfrastructure on non-forest lands.Following the examples of Colorado and as proposed in Pennsylvania, Maryland might considerwell-pad permitting as part of a comprehensive gas development planning process (described inChapter 1). New York State law allows the environmental impact of more than one project to beconsidered at the same time (Strong 2008). A small project may not have a negative impact onhabitat alone, but when considered in the context of nearby or related projects, the negativeimpact may be significant. With respect to terrestrial habitat, a comprehensive gas developmentplan could help channel development into areas with greater amounts of existing disturbance andavoid areas with intact forests (especially forest interior habitat and other high priorityconservation areas). Efforts in this area would greatly improve Maryland’s ability to addresscumulative impacts of MSGD which are likely to be significant without proper regulation.B. ImpoundmentsDirect exposure to contaminated water stored on-site and during transportation on- and off-sitecan come from tank leaks and spills during tank transfers, and is one of the biggest threats towildlife from hydraulic fracturing operations (Thompson 2012). Although many of the chemicalscontained in hydraulic fracturing fluids are potentially toxic to wildlife, there are few studies onthe exposure effects of gas operations to animals. One recent study of farm animals in six states(Pennsylvania, New York, Ohio, Colorado, Texas, Louisiana) suggested increased mortalityrates in livestock and companion animals (i.e., dogs and cats) living close to active gas-drillingoperations (Oswald and Bamberger 2012), with several caveats associated with the lack ofcontrols due to the case study aspect of the survey (Thompson 2012). Although chemicals can bevolatized (e.g., by impoundment aerators) and misted into the air creating an inhalation exposurepathway, the most common source of toxicity exposure was likely via contaminated water.Pathways of exposure included, for example, spills of hydraulic fracturing fluids, tears in theliners of wastewater impoundments, and spreading of wastewater on roads to reduced dust andice followed by animals licking their paws after crossing the roads (Table 5-1). Health impacts6 Maryland No Net Loss of Forest Task Force Final Report and Recommendations. January 2009. http://www.dnr.maryland.gov/dnrnews/pdfs/NNLTFFINALREPORT1.pdf 5-4

110. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandranged from neurological to sudden death with the most common effects being reproductive.Animals affected include cows, horses, goats, llamas, chickens, dogs, cats, and koi. Because themovement of farm animals is confined they may experience higher cumulative exposure thanwildlife with less restricted mobility. However, photographic evidence has been reported of deadand dying songbirds, deer, frogs, and salamanders (Oswald and Bamberger 2012). The lack ofcontrolled dose-response studies is due in part to the lack of information on the specificchemicals used in the hydraulic fracturing. Additionally, substances that occur naturally in theshale may come to the surface as part of the process. These constituents are poorly quantified,but can be equally or moretoxic than the hydraulic Table 5-1. Case survey documenting impacts to farm animalsfracturing fluid. living near gas drilling operations (Oswald and Bamberger 2012).To limit the exposure of # ofwildlife to toxic chemicals, Source of exposure casesimpoundment ponds used to Hydraulic fracturing fluid spill from holding tank 2store flowback and produced Drilling fluids overran well pad during blow out 1wastewater should not be Storm water run-off from well pad to property 3permitted in Maryland.Evaporation in wastewater Wastewater impoundment leak 1impoundments can increase the Wastewater impoundment allegedly compromised 1concentration of some toxins, Wastewater spread on road 2making them fatal traps to Wastewater dumped on property 1migratory birds and other Wastewater dumped into creek 3wildlife that may try to use the Wastewater impoundment not contained 3ponds7. In the State of New Well/spring water 17Yorks revised guidelines, Pond/creek water 8watertight tanks are thepreferred option to store Pipeline leak 1flowback and produced water Compressor station malfunction 2(NYSDEC 2011). Pennsylvania Flaring of well 3DCNR recommends thefollowing steps to protectecological resources from off-site spills: (1) use of double-wall tanks for the storage of chemicalsand liquids; (2) wherever possible, chemicals and liquids should be stored inside storage trailers(PADCNR 2011). Closed storage tanks with secondary containment should be used for allstorage of chemicals and produced waters, especially in areas with significant rainfall such aswestern Maryland (see Chapter 4). Where impoundments are used (e.g., for temporary storage offreshwater only) fencing of these water features and freeboard of several feet should bemaintained. The construction of any impoundments should consider the increases in stormintensities that are projected as a consequence of the states changing climate (Boicourt andJohnson 2010). Runoff and spill prevention, response and remediation plans should be anecessary part of the permitting process. Finally, to protect wildlife and downstream waterquality, spraying of wastewater (flowback or produced water) on roads to minimize dust, forexample, should not be permitted under any circumstances.7 Ramirez, P. 2009. Reserve Pit Management: Risks to Migratory Birds. U.S. Fish & Wildlife Service, Region 6, Environmental Contaminants Program, http://www.fws.gov/mountain-prairie/contaminants/documents/ReservePits.pdf 5-5

111. K.N. Eshleman & A.J. Elmore (2013) Chapter 5C. Riparian setbacksForested riparian buffers enhance biodiversity by (1) providing foraging, nesting, brooding,thermal, and escape cover; (2) protecting sensitive habitats; and (3) maintaining landscapeconnectivity (Bentrup 2008). These terrestrial habitats also provide valuable buffer benefits toaquatic environments, for example, by shading streams to maintain favorable temperature(Moore et al. 2005). Despite strong evidence that forested riparian buffers are an important bestmanagement practice, the scientific basis for determining a specific width for the BMP dependson the overall rational for the buffer. Much of the scientific evidence (Table 5-2) supports the useof relatively large forest buffers when the intent is to preserve biological diversity (Bearer et al.in press). For example, herptiles [i.e., amphibians (e.g., frogs, toads, and salamanders) andreptiles (e.g., snakes, turtles, and terrapins)] exhibit life stages during which individuals willmigrate great distances from streams and wetlands in search of new habitat (Grant et al. 2010). Ina study of marbled salamanders, over 200 juvenile salamanders were captured at distancesbetween 365 and 4,035 ft (median = 883 ft) from natal ponds (Gamble et al. 2006). Similarly,considerable work has been devoted to evaluating the typical dispersal distances of turtles, withmost recommendations for forest buffer widths falling in the range of 500-1,000 ft (e.g., Bodie2001) which is intended to capture 95% of all migrating individuals. There is also abundantevidence that aquatic insects utilize riparian buffers during adult life stages (Bried and Ervin2006), indicating the unique nature of riparian forests as foraging habitat for rare, threatened andendangered bats (Lookingbill et al. 2010). Table 5-2. Representative list of studies providing evidence supporting different riparian setback widths. Response tested Setback Citation recommended Neo-tropical bird activity 330 ft (Hodges and Krementz 1996) Dragonfly activity >530 ft (Bried and Ervin 2006) Turtle migration 910 ft (Burke and Gibbons 1995) Salamanders 330 ft (Crawford and Semlitsch 2007) Salamanders 890 ft (Gamble et al. 2006) Birds 660 ft (Perry et al. 2011) Freshwater reptiles 1,240 ft (Roe and Georges 2007) Salamanders 410 ft (Semlitsch 1998) Amphibians and reptiles 960 ft (Semlitsch and Bodie 2003) Wildlife 330 ft (Wanger 1999) Aquatic diversity 330 ft (Castelle et al. 1994) Amphibians and small mammals 330 ft (McComb et al. 1993)Many riparian areas in western Maryland are currently forested (e.g., 75.8% of the land areawithin 300 ft of streams is forest). Any denuded riparian zones provide opportunities forreforestation as part of our recommended no-net-loss of forest policy. Riparian setbacks wouldhelp ensure that MSGD was designed to minimize harm to this critical terrestrial habitat. Werecommend minimum setbacks of 300 ft from floodplains, wetlands, seeps, vernal pools,streams, or other surface water bodies (Figure 5-3). This distance is consistent with estimatedrequirements from the scientific literature for terrestrial species that use riparian areas asmovement corridors and amphibians, turtles and other aquatic species that use the land for atleast part of their life cycles (Table 5-2). The distance is also consistent with U.S. Department of 5-6

112. Recomm mended Best Management Practices for Marcell Shale Ga Developm s lus as ment in Mary ylandAgricultu (USDA) recommend corridor widths for te ure ded errestrial bio (Bentrup 2008). otaOperationnally, these setbacks cou be consid uld dered minimmum threshol and wou be increas lds uld sedby any addditional req quirements set by law for specific rar threatene or endang r re, ed gered species for s,example. Setbacks sh . hould be mea asured from the edge of disturbance (not the we f e ellbore) to thespecific habitat of co h oncern. In the case of floo e odplains, the Federal Em e mergency ManagementAgency (FEMA) floo ( odplain map which hav recently b ps, ve been updated for both Ga d arrett andAllegany County, could be used to delineate these habita y ats.Figure 5-3 Waterways and proposed ri 3. a iparian buffer setbacks for w estern Marylan As discusse in Chapter 6 s nd. ed 6,much of th area included in these recommended setb he d backs is unsuitaable or undesir rable for multip reasons, pleincluding designation as the 100-yr floo d odplain, riparia forest, or top an pographically ssteep (slope >1 15%) land.D. Special protect tion areas sSeveral states have im s mplemented additional setback requ d s uirements for forest focus areas. r sPennsylv vania require that a 600 ft boundary be maintain between well pad sit and state park es y ned n tesland or designated wild and natur areas on state forest l d ral land (PADC CNR 2011). N York S New Statesproposed guidelines designate fo d orest focus arreas based on Nature Co n onservancy a Landscap and peConnecti ivity Analysi These foc areas wo is. cus ould be subje to site-sp ect pecific ecological assessmment,including pre-disturb g bance biological studies and an evalu a uation of pot tential impac on forest ctsinterior birds from th project. More stringen mitigation measures m then be i b he M nt n may instituted for rthese areas (NYSDEC 2011).Fortunate Marylan has alread made sign ely, nd dy nificant prog gress in ident tifying critic areas for calspecial protection. The Maryland Departmen of Natural Resources (DNR) Natu Heritage T d nt l ural e 5-7

113. K.N. Eshleman & A.J. Elmore (2013) Chapter 5Program has developed a digital map known as BioNet (Biodiversity Conservation Network) thateffectively prioritizes areas of the state for conservation of freshwater and terrestrial plants,animals, habitats, and landscapes. The purpose of the assessment was to provide decision supportfor species and land conservation programs including the guidance of compatible land uses andland management practices. These data represent the most up-to-date understanding of the spatialdistribution of biotic resources in the state and guide the state’s overall biodiversity stewardshipstrategy. The criteria used within BioNet have a dual focus on the rarest species and habitats, aswell as the habitats that concentrate the largest numbers of rare species and habitat. Thus, BioNetconsiders areas with: (1) only known occurrences of species and habitats; (2) globally rarespecies and habitats; (3) state rare species and habitats; (4) animals of greatest conservation need;(5) watch list plants and indicators of high quality habitats; (6) animal assemblages (e.g., colonialnesting waterbirds, forest interior species); (7) hotspots for rare species and habitats; (8) intactwatersheds; and (9) wildlife corridors and concentration areas. These areas are prioritized into afive-tiered system: • Tier I – area is critically significant for biodiversity conservation • Tier II – area is extremely significant for biodiversity conservation • Tier III – area is highly significant for biodiversity conservation • Tier IV – area is moderately significant for biodiversity conservation • Tier V – area is significant for biodiversity conservationAreas identified include those that support the 204 state-listed species (Table 5A), rare and highquality plant and animal communities, high wildlife densities, and important habitats needed forwildlife migration and movements related to climate change. Tier I and Tier II sites representlocations that Maryland DNR has designated as “irreplaceable natural areas”8.Consistent with Pennsylvania state forest policy, we recommend a no-disturbance setback withinand 600 ft around any priority conservation area (Figure 5-4). Priority conservation areas shouldbe defined using the best available science and designed so that no irreplaceable areas of uniquehabitat could be impacted by MSGD. Portions of the BioNet dataset, specifically irreplaceablenatural areas (BioNet Tier I and II sites) and wildlands should initially be considered as priorityconservation areas and receive the recommended 600 ft setback. BioNet data products should bekept up to date and publically available, so that potential MSGD operators have the informationrequired to identify lands for their activities that would have the least impact on the naturalresources of Maryland. The methods used to generate BioNet products should be published inthe scientific literature.Caves were addressed in Chapter 1 while discussing potential complications arising duringdrilling, well casing and cementing; however, caves (including those subterranean voids that areman made) are also a terrestrial habitat of special concern. Western Maryland is home to twoglobally ranked species of bat that are critically imperiled in the state because of extreme rarity(Table 5-1). The Indiana bat (Myotis sodalis) is a federally listed endangered species in danger ofextinction throughout significant portions of its range. However, even common species of cave-dwelling bats are currently in a status of extreme flux due to the poorly understood white-nose8 Jonathan McKnight (DNR - Wildlife and Heritage Service, Associate Director Natural Heritage Program), personal communication 5-8

114. Recomm mended Best Management Practices for Marcell Shale Ga Developm s lus as ment in Mary ylandsyndrome that has resulted in the death of mi e illions of No America bats (Fric et al. 2010 In orth an ck 0).response, the U.S. Fish and Wild dlife Service (USFWS) c called on all activities in affected are to easbe curtail 9. Extrem caution is advisable ar led me round all Ma aryland cave that suppo obligate c es ort cave- 10dwelling species . The true extent of caves in western M T i Maryland is llikely highlyyunderrepresented by the 33 caves that are cur s rrently mapp in Garre and Allegany County ped ett(Franz an Slifer 197 Because drilling acti nd 71). e ivities have the potentia to add to th already al hesignifican levels of stress being experienced by cave-dw nt s d welling bat poopulations, ppre-drillingtesting sh hould be use to identify the location of subterr ed y ns ranean voids and a 1,000 ft setback s,should be observed around any naturally occ e a n curring cave. This recom . mmendation a assumes that all tvoids are connected to the surfac via condui sufficient in size to be accessible to bats. Wh e t ce its t hilesome of these spaces may be inac t s ccessible, hi ibernacula en ntrances can be very small and diffic n cultto detect by simple observation of the land su o urface in ma cases11. R any Regardless, t occurren the nceof voids is also indicative of pote ential habitat for other fo t orms of subt terranean life whose pres e senceis not dic ctated by con nnections to the surface.Figure 5-4 Maryland Bi 4. ioNet I and II12 sites in Allega and Garret County with a 600 ft setbac buffer. Thes 2 any tt ck sesites repres locations that are conside sent t ered irreplacea natural are by the state Tier III sites (highly signifi able eas e. icantfor biodiveersity conserva ation) are also shown. s9 http://wwww.caves.org/W WNS/USFWS-W WNS_cave_ad dvisory_news_ _rls_2009-03-226_final-1.pdf.10 A map of the density of obligate cave dwelling spec in the east o o e cies tern United Sta can be foun at: ates nd http://ww ww.karstwaterss.org/files/speciesmaps.htm11 Dan Fell (Maryland DNR - Wildlif and Heritage Service, West ler D fe e tern Region Eccologist), perso communic onal cation12 www.dn nr.maryland.govv/wildlife/Plan nts_Wildlife/digitaldata.asp 5-9

115. K.N. Eshleman & A.J. Elmore (2013) Chapter 5E. Noise and lightNoise and light pollution associated with unconventional gas development have not been studiedin depth, but increasing evidence suggests that these factors have significant effects on wildanimals (Beier 2005, Pijanowski et al. 2011). More research on the topic is needed, particularlywith respect to noise. Noise pollution can affect wildlife in at least two ways: (1) disruption ofcommunication by masking acoustic signals (primarily an issue for birds, but may also affectamphibians and terrestrial mammals); and (2) reductions in abundance, distribution and densityof species due to avoidance behavior (Patricelli and Blickley 2006, Barber et al. 2009). In a studyof gas well operations in northwestern New Mexico, noise amplitude was amplified overbaseline amplitudes up to distances of 3000 ft from a compressor (Francis et al. 2010). Nestoccupancy was depressed for bird species within this affected area by an average of 5%. In astudy of Wyoming sage-grouse, natural gas drilling and road noises were recorded and played at70 dB(f) in front of leks13 (Blickley et al. 2012). This sound pressure level is similar to what ismeasured 1300 ft from drilling rigs and main access roads in that landscape. After three years ofstudy, intermittent noise from roads decreased male attendance at the leks by 73% relative topaired controls. Drilling noise decreased attendance by 29%. Another study on mule deerindicated significant decreases in populations in areas within close proximity to well pads due toavoidance behavior (Francis et al. 2010).The use of sound barrier walls around compressors can reduce the area affected by noise by up to70% and maintain occupancy and nest success rates at levels close to those expected in alandscape without compressor noise (Francis et al. 2010). New York State (NYSDEC 2011) andAPI (API 2011) have established techniques for evaluating and mitigating noise impacts of gasoperations. Following these guidance documents and consistent with Chapter 9 we recommendthat Maryland require as part of the permitting process: (1) maintaining a maximum distancebetween well pads and BioNet irreplaceable natural areas to reduce noise effects on thesesensitive lands; (2) constructing artificial sound barriers where natural noise attenuation wouldbe inadequate; (3) equipping all motors and engines with appropriate mufflers; and (4) avoidingconstruction and drilling operations during sensitive migrating and mating seasons.All bats and many other mammals, amphibians, and birds are nocturnal. When subjected toartificial lighting at night, documented animal responses include altered forage and matingbehavior that, depending on the species and circumstances, can lead to changes in survival andfecundity (e.g., Beier 2005). Bats, for example, use their limited vision to exploit low levels oflight as they leave a roost and to avoid obstacles [such as the capture nets used by biologists thatstudy bats (Wang et al. 2004)]. Notably, bats are known to avoid disorienting bright light. Batsare farsighted, suggesting they use light more often when it is further away and dim, and switchto echolocation for objects that are closer. When migrating, bats will use vision rather thanecholocation, which helps to explain why they are known to fly into wind turbines (Johnson et al.2003). With song birds, artificial lighting has been shown to shift singing to earlier morninghours, alter mating success, and result in an earlier start of egg laying in spring (Miller 2006,Kempenaers et al. 2010). Artificial lighting alters foraging, reproductive, and defensive 13 A lek is a gathering of males, of certain animal species, for the purposes of competitive mating display. 5-10

116. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandbehaviors in amphibian species (Andrews et al. 2008). For example, frog calls are reduced underartificial lighting, which can reduce mating success and thus affect population dynamics (Bakerand Richardson 2006). Artificial lighting near aquatic habitat can also become an attractivenuisance, attracting insects that subsequently die before finding new habitat.There are two sources of light that are commonly found at drill pad sites during MSGD: (1)artificial lighting, used to illuminate the site or transportation routes to and from, or between drillpad sites, and (2) flaring of unwanted or waste gas during well completion. Although the impactof artificial light in rural settings such as western Maryland is fairly well documented, relativelylittle has been recommended as BMPs by the various states for unconventional gas development.Some states have suggested flaring is the most egregious source of light. For example,Pennsylvania requires gas operators to provide a minimum of 10 days notice to the ForestDistrict Manager when flaring activities are anticipated, but this requirement only pertains toDark Sky Areas. The Forest District Manager is required to coordinate with drilling operators toreduce conflicts during special events on state lands that require dark skies (PADCNR 2011).Proposed rules in Colorado and BMPs recommended by API also mention light pollution (API2011), but only suggest considering light as a motivation for increasing setbacks fromresidences. Proposed rules in New York are the most protective of wildlife, requiring lightingused at well pads to shine downward during bird migration periods (April 1–June 1 and August15–October 15)(NYSDEC 2011).Additional best management practices available to Maryland include requiring: (1) diffusedownward pointing lighting at all times; (2) the use of low-pressure sodium lamps (most energyefficient) instead of high-pressure sodiumor mercury lamps; (3) the use of UVfilters; (4) reduced lighting to onlylocations and intensity needed; and (5)using green light rather than red or bluewhere possible. Green light and UV Lampfiltered light has been found to be lessdisorienting to migrating birds than is redor blue light (Wiltschko et al. 1993, Pootet al. 2008). Limiting the height oflighting columns (e.g., to a height lessthan 8 m) and directing light downwardreduces the ecological impact of the light(Fure 2006). In some circumstances,outfitting sensors to lights that are Figure 5-5. The authors observed lamps at this well pad inactivated when light is needed could be Brooke County, WV (operated by Chesapeake Energy) thatan effective means of reducing light appeared to be consistent with recommended lighting BMPs.levels on average. Flaring during the Photo by A.J. Elmore.completion process should be minimizedor eliminated, which will be required by USEPA starting in 2015 (Chapter 1). When we visited adrill pad site in West Virginia operated by Chesapeake Energy, we noted that lighting wasmounted appropriately low and was covered with diffusing fabric, presumably to reduce glareand shadows. Each lamp was “capped” with a downward-reflecting shield, which might beeffective at reducing light pollution of dark skies (Figure 5-5). We did not see these lamps in 5-11

117. K.N. Eshleman & A.J. Elmore (2013) Chapter 5operation, but they appeared to be consistent with the BMPs recommended by the states and byAPI.F. Construction of roads and pipelinesRoads, pipelines, and other built linear features can have significant effects on even a largelyforested landscape such as western Maryland. Fragmentation created by infrastructuredevelopment is a major threat to biodiversity and a primary concern resulting from MSGD (e.g.,Alkemade et al. 2009). A recent meta-analysis of 49 studies on 234 species of birds andmammals found the effect of infrastructure to extend up to 1 km for bird populations, and 5 kmfor mammal populations (Francis et al. 2010). These impacts include: mortality from roadconstruction, mortality from collision with vehicles, modification of animal behavior, alterationof the physical environment, alteration of the chemical environment, spread of exotics, andincreased use of areas by humans (Trombulak and Frissell 2000). Threats to wildlife can beelevated by increased traffic on roads that have historically received little activity (Gibbs andShriver 2005) and by genetic isolation of species with poor dispersal abilities (Keller et al. 2004).These studies illustrate the importance to terrestrial wildlife of minimizing new infrastructuredevelopment in lands that are currently relatively undisturbed. The potential impacts of roadnoise have been discussed above, and the consequences of increased impervious surface onaquatic resources are addressed in Chapter 6 of this report.A relatively sparse network of roads currently fragments the Marcellus Shale region of westernMaryland, which translates to significantly lower cumulative ecological risk than the muchdenser road network of the eastern part of the state. In general, regions in which more than 60%of the total land area is within 1,250 ft of a road are at elevated risk of having those roads impactecosystem condition (Riitters and Wickham 2003). The median distance to roads in westernMaryland is currently 630 ft (Figure 5-6). The contribution of new roads to forest fragmentationis greatest in largely forested areas such as the Appalachian Mountains (Riitters and Wickham2003). An intensive network of new secondary roads and pipelines can be anticipated withMSGD. A study of the landscape changes due to natural gas extraction in the Marcellus shaleregion of Pennsylvania found an increase in forest edge of 380 miles in Bradford County and 721miles in Washington County between 2004 and 2010, with the largest amount (55%) attributableto road and pipeline construction (Slonecker et al. 2012). For both counties, pipeline constructionwas the major contributor to overall forest loss, increase in patchiness, and decrease in meanforest patch size.We recommend minimizing the amount and impact of new road and pipeline construction asmuch as practicable by following the guidance proposed by New York State: (1) limiting thelinear distance of new roads through strategic siting of operations; and (2) co-locating projectinfrastructure with current roads, power lines, and pipelines (NYSDEC 2011). Centralization andco-location of infrastructure also offset air pollution by decreasing truck traffic (PADCNR2011). As the extent of road effects is thought to be at least as far-reaching as drillingoperations, it would be most protective if all setbacks described in this chapter be applied to roadconstruction as well as well-pad development: 300 ft from riparian areas, 600 ft fromirreplaceable natural areas and wildlands, and 1,000 ft from caves. All new infrastructure 5-12

118. Recomm mended Best Management Practices for Marcell Shale Ga Developm s lus as ment in Mary ylandconstruct c f d ghts-of-way1 . When pra tion should consolidate facilities and pipeline rig 14 actical, newconstruct tion could ut tilize pre-exi isting disturb bance and be scheduled using season restrictio to e nal onsavoid migratory and mating seasons (e.g., pe breeding season appr eak g roximately M 15-July 15) May y(NYSDE 2011). ECFigure 5-6 Map of the ro network of Garrett and Allegany Count showing dis 6. oad f A ty stance to the clo osest road for e each30-m grid cell. Locations close to roads would be pref s s ferred for siting well pads to reduce the am ng mount of newinfrastructu and associa disturbanc required. Th histogram (in ure ated ce he nset) provides the distribution of landscape area eand foreste landscape ar at specified distances from roads (for al l land only, the median = 0.12 miles and m ed rea d m e mean =0.28 miles are shown).G. Invas sive plants and wild s dlifeGas deve elopment can disrupt nat plant com n tive mmunities, p providing op pportunities for invasion of nsexotic sp pecies. Most states with active shale gas operatio have curr or propo a ons rent osedrecomme ended practic for pre-c ces construction inventory, pprevention, m management control, an t, ndmonitorin of invasiv species. New York St has prop ng ve N tate posed baselin surveys a the ne anddevelopm of a com ment mprehensive managemen plan for a taxa of inv e nt all vasive speci in the stat ies tethat emph hasizes early detection and rapid res y a sponse (NYS SDEC 2011) Pennsylvan DCNR h ). nia hasoutlined BMPs for st forests th includes pre-constru tate hat uction inventtories, cleani procedur ing resfor equippment, annua surveys, an species sp al nd pecific contr measures for post-disturbance rol sinfestatio (PADCN 2011). Co ons NR olorado requ uires all heav equipmen hand tool boots and vy nt, ls, d14 Universi of Colorado Intermountai Oil and Gas BMP Project: http://www.oi ity o. in ilandgasbmps.o org/ 5-13

119. K.N. Eshleman & A.J. Elmore (2013) Chapter 5other equipment to be disinfected prior to moving the equipment for use at a new site15.Additional BMPs advocated by Pennsylvania DCNR include (PADCNR 2011): • A pre-construction inventory should be performed within the anticipated areas of disturbance to determine the appropriate prevention methods, predict control needs and assess its level of responsibility for management. • Soil disturbance should be minimized to decrease introduction. Consider co-location within previously disturbed areas and/or alternative construction methods. • The operator should clean equipment in an appropriate manner prior to bringing equipment into un-invaded areas or ecologically sensitive areas. • It is recommended that the operator use weed-free seed, soil, gravel, and mulch. Failure to use of weed-free material increases the potential to introduce invasive plant species and requires stringent monitoring. • Pre-treat invasive plant species infestations that reproduce prolifically from rhizome/root segments prior to disturbance. Pre-treatment may limit the spread of the invasive plant infestations upon completion of disturbance activities. • Disturbed areas should be surveyed annually at the appropriate time of year to detect early infestations. • Management and control of post-disturbance infestations of invasive plant populations should be species specific. In some situations, it may be best to wait another growing season to assess the spread before moving forward with management techniques.We recommend that permitting require specific plans for: (1) flora and fauna inventory surveysof sites prior to operations; (2) interim reclamation following construction and drilling to reduceopportunities for invasion; (3) annual monitoring and treatment of new invasive plantpopulations as long as the lease is active; and (4) post-activity restoration to pre-treatmentcommunity structure and composition using seed that is certified free of noxious weeds.H. ReclamationReclamation and site decommissioning were discussed in Chapter 4 in the context of water andwater quality management, but reclamation is also important in terms of providing forage andcover habitat for terrestrial species and minimizing opportunities for non-native plant invasions.Consistent with our recommendations in Chapter 4, reclamation planning should be conducted intwo separate phases. Within an established period of time following the construction and drillingphases of development and bringing a well into production, any portion of a well pad site that isnot needed for gas production should be revegetated. New York and Pennsylvania providespecific recommendations for this interim phase including creating soft edges around newclearings by maintaining current understory shrubbery or planting native plants (NYSDEC2011); revegetation should avoid wet seasons and wet periods outside of wet seasons tominimize impacts on soils, water, and vegetation (PADCNR 2011). Monitoring of native andinvasive species could occur on-site throughout the length of the lease. For example, proposedprotocols in New York call for monitoring of forest interior or grassland birds during theconstruction phase of the project and for a minimum of two years following well completion(NYSDEC 2011). A second phase of terrestrial habitat reclamation occurs after the wells havebeen plugged and gas production activities have ceased at a site. Permanent site restorationshould remove the built infrastructure, restore the disturbed soil, re-contour the site, and provide15 Ibid. 5-14

120. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylanda permanent vegetative cover. Clear objectives for final site restoration should be specified aspart of the permitting process. Options include reverting the site back to its original land cover orrestoring the landscape to its native condition, which in this area of the country is forest cover. Inall cases, species planted should be in their natural geographic range and local stock should bepreferred (PADCNR 2011).I. Key recommendations5-A Minimize well pad size, cluster multiple well pads, and drill multiple wells from each pad to minimize the overall extent of disturbance and reduce fragmentation and associated edge effects. 5-A.1 Concentrate operations including roads on disturbed and open lands, ideally in locations zoned for industrial activity and/or close proximity to major roads. 5-A.2 Adopt a no-net-loss of forest policy requiring any activities that remove forest to be offset by plantings elsewhere in the region. 5-A.3 Implement comprehensive planning process to address the cumulative impact of multiple projects, to channel development into areas with greater amounts of existing disturbance, and to avoid areas with intact forests (especially forest interior habitat).5-B Allow for freshwater impoundments only. Impoundments should not be used for flowback or produced wastewater. 5-B.1 Require watertight, closed metal tanks with secondary containment for all storage of chemicals and wastewater. 5-B.2 Include runoff and spill prevention, response, and remediation plans as part of the permitting process5-C Establish and enforce setbacks to conserve terrestrial and aquatic biodiversity. 5-C.1 Enforce 300 ft minimum setbacks from all floodplains, wetlands, seeps, vernal pools, streams, or other surface water bodies. 5-C.2 Exclude all development activities from priority conservation areas (BioNet Tier I and Tier II sites and wildlands). Enforce a 600 ft setback from these areas. 5-C.3 Enforce 1,000 ft setback from any cave to reduce stress to bats and other obligate subterranean species.5-D Review local noise ordinances to ensure they are sufficiently protective. Artificial sound barriers and mufflers should be considered where natural noise attenuation would be inadequate, especially in proximity to priority conservation areas. 5-D.1 Avoid construction and drilling operations during sensitive migratory and mating seasons.5-E Reduce the amount of light pollution at drill pad sites by restricting night lighting to only when necessary and to only the amount of lighting required, direct light downward, instead of horizontally, use fixtures that control light directionality well, minimize glare, and use low pressure sodium (LPS) light sources whenever possible. 5-15

121. K.N. Eshleman & A.J. Elmore (2013) Chapter 5 5-E.1 When drill pads are located within 1,000ft of aquatic habitat, vegetative screens and additional lighting restrictions could be required to reduce light pollution into these sensitive areas.5-F Co-locate linear infrastructure as practicable with current roads, pipelines and power lines to avoid new disturbance. 5-F.1 Avoid stream crossings and any disturbances to wetlands and riparian habitat.5-G Submit an invasive species plan as part of permit application for preventing the introduction of invasive species and controlling any invasive that is introduced. 5-G.1 The invasive species management plan should emphasize early detection and rapid response and include baseline flora and fauna inventory surveys of site prior to operations and long-term monitoring plans for areas that could become problematic after gas development occurs.5-H Develop a two-phased reclamation strategy comprised of (1) interim reclamation following construction and drilling to reduce opportunities for invasion and (2) post- activity restoration using species native to the geographic range and seed that is certified free of noxious weeds.J. Literature citedAlkemade, R. M., M. van Oorschot, L. Miles, C. Nelleman, M. Bakkenes, and B. ten Brink. 2009. GLOBIO3: a framework to investigate options for reducing global terrestrial biodiversity loss. Ecosystems 12:374-390.Andrews, K. M., J. W. Gibbons, and D. M. Jochimsen. 2008. Ecological Effects of Roads on Amphibians and Reptiles: A Literature Review. Herpetological Conservation 3:121-143.API. 2011. Practices for Mitigating Surface Impacts Associated with Hydraulic Fracturing. API Guidance Document HF3. American Petroleum Institute, Washington, DC.Baker, B. J., and J. M. L. Richardson. 2006. The effect of artificial light on male breeding-season behaviour in green frogs, Rana clamitans melanota. Canadian Journal of Zoology-Revue Canadienne De Zoologie 84:1528-1532.Barber, J. R., K. R. Crooks, and K. M. Fristrup. 2009. The costs of chronic noise exposure for terrestrial organisms. Trends in Ecology & Evolution 25:180-189.Bearer, S., E. Nicholas, T. Gagnolet, M. DePhilip, T. Moberg, and N. Johnson. in press. Evaluating the scientific support of conservation best management practices for shale gas extraction in the Appalachian basin. Environmental Reviews.Beier, P. 2005. Effects of Artifical Night Lighting on Terrestrial Mammals. Pages 19-42 in C. Rich and T. Longcore, editors. Ecological consequences of artificial night lighting. Island Press, Covelo, CA.Bentrup, G. 2008. Conservation buffers: design guidlines for buffers, corridors, and greenways. Gen. Tech. Rep. SRS-109. Department of Agriculture, Forest Service, Southern Research Station, Asheville, NC.Blickley, J. L., D. Blackwood, and G. L. Patricelli. 2012. Experimental evidence for the effects of chronic anthropogenic noise on abundance of greater sage-grouse at leks. Conservation biology 26:461- 471.Bodie, J. R. 2001. Stream and riparian management for freshwater turtles. Journal of Environmental Management 62:443-455. 5-16

122. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandBoicourt, K., and Z. P. Johnson, editors. 2010. Comprehensive strategy for reducing marylands vulnerability to climate change. University of Maryland Center for Environmental Science and Maryland Department of Natural Resources, Cambridge, MD and Annapolis, MD.Burke, V. J., and J. W. Gibbons. 1995. Terrestrial Buffer Zones and Wetland Conservation: A Case Study of Freshwater Turtles in a Carolina Bay. Conservation Biology 9:1365-1369.Castelle, A. J., A. W. Johnson, and C. Conolly. 1994. Wetland and Stream Buffer Size Requirements - A Review. Journal of Environmental Quality 23:878-882.Crawford, J. A., and R. D. Semlitsch. 2007. Estimation of Core Terrestrial Habitat for Stream-Breeding Salamanders and Delineation of Riparian Buffers for Protection of Biodiversity. Conservation Biology 21:152-158.Drohan, P. J., M. Brittingham, J. Bishop, and K. Yoder. 2012. Early trends in landcover change and forest fragmentation due to shale-gas development in Pennsylvania: A potential outcome for the northcentral Appalachians. Environmental Assessment 49:1061-1075.Francis, C. D., J. Paritsis, C. P. Ortega, and A. Cruz. 2010. Landscape patterns of avian habitat use and nest success are affected by chronic gas well compressor noise. Landscape Ecology 26:1269- 1280.Franz, R., and D. Slifer. 1971. Caves of Maryland. Educational Series No. 3. Maryland Geological Survey.Frick, W. F., J. F. Pollock, A. C. Hicks, K. E. Langwig, D. S. Reynolds, G. G. Turner, C. M. Butchkoski, and T. H. Kunz. 2010. An emerging disease causes regional population collapse of a common North American bat species. Science 329:679-682.Fure, A. 2006. Bats and lighting. The London Naturalist 85.Gibbs, J. P., and W. G. Shriver. 2005. Can road mortality limit populations of pool-breeding amphibians? Wetlands Ecology and Management 13:281-289.Grant, E. H. C., J. D. Nichols, W. H. Lowe, and W. F. Fagan. 2010. Use of multiple dispersal pathways facilitates amphibian persistence in stream networks. Proceedings of the National Academy of Sciences of the United States of America 107:6936-6940.Harper, K. A., S. E. Macdonald, P. J. Burton, C. Jiquan, K. D. Brosofske, S. C. Saunders, E. N. S. Euskirchen, D. A. R. Roberts, M. S. Jaiteh, and P.-A. Esseen. 2005. Edge Influence on Forest Structure and Composition in Fragmented Landscapes. Conservation Biology 19:768-782.Harris, L. D. 1984. The fragmented forest. University of Chicago Press, Chicago, USA.Hodges, M. F., Jr., and D. G. Krementz. 1996. Neotropical Migratory Breeding Bird Communities in Riparian Forests of Different Widths along the Altamaha River, Georgia. The Wilson Bulletin 108:496-506.Johnson, G. D., W. P. Erickson, M. D. Strickland, M. F. Shepherd, D. A. Shepherd, and S. A. Sarappo. 2003. Mortality of bats at a large-scale wind power development at Buffalo Ridge, Minnesota. American Midland Naturalist 150:332-342.Keller, I., W. Nentwig, and C. R. Largiader. 2004. Recent habitat fragmentation due to roads can lead to significant genetic differentiation in an abundant flightless ground beetle. Molecular Ecology 13:2983-2994.Kempenaers, B., P. Borgstroem, P. Loes, E. Schlicht, and M. Valcu. 2010. Artificial Night Lighting Affects Dawn Song, Extra-Pair Siring Success, and Lay Date in Songbirds. Current Biology 20:1735-1739.Lookingbill, T. R., A. J. Elmore, K. A. M. Engelhardt, J. B. Churchill, J. E. Gates, and J. B. Johnson. 2010. Influence of wetland networks on bat activity in mixed-use landscapes. Biological Conservation 143:974-983.McComb, W. C., K. McGarigal, and R. G. Anthony. 1993. Small mammal and amphibian abundance in streamside and upslope habitats of mature Douglas-fir stands, western Oregon. SNorthwest Scientist 67.Miller, M. W. 2006. Apparent effects of light pollution on singing behavior of American robins. Condor 108:130-139. 5-17

123. K.N. Eshleman & A.J. Elmore (2013) Chapter 5Moore, R. D., D. L. Spittlehouse, and A. Story. 2005. Riparian microclimate and stream temperature response to forest harvesting: A review. Journal of the American Water Resources Association 41:813-834.NYSDEC. 2011. Revised draft supplemental generic environmental impact statement on the oil, gas, and solution mining regulatory program. Albany, NY.Oswald, R., and M. Bamberger. 2012. Impacts of gas drilling on human and animal health. New Solutions 22:51-77.PADCNR. 2011. Guidelines for administrating oil and gas activity on state forest lands. Commonwealth of Pennsylvania Department of Conservation and Natural Resources Bureau of Forestry.Patricelli, G. L., and J. L. Blickley. 2006. Avian communication in urban noise: Causes and consequences of vocal adjustment. Auk 123:639-649.Perry, R. W., T. B. Wigley, M. A. Melchiors, R. E. Thill, P. A. Tappe, and D. A. Miller. 2011. Width of riparian buffer and structure of adjacent plantations influence occupancy of conservation priority birds. Biodiversity Conservation 20:625-642.Pijanowski, B. C., L. J. Villanueva-Rivera, S. Dumyahn, A. Farina, B. L. Krause, B. M. Napoletano, S. H. Gage, and H. Pieretti. 2011. Soundscape ecology: The science of sound in the landscape. BioScience 61:203-216.Poot, H., B. J. Ens, H. D. Vries, M. A. H. Donners, M. R. Wernand, and J. M. Marquenie. 2008. Green Light for Nocturnally Migrating Birds. Ecology and Society 13:14.Riitters, K. H., and J. D. Wickham. 2003. How far to the nearest road? Frontiers in Ecology and Environment 1:125-129.Ritters, K. H., and J. D. Wickham. 2012. Decline of forest interior conditions in the conterminous United States. Scientific Reports 2:1-4.Semlitsch, R. D. 1998. Biological Delineation of Terrestrial Buffer Zones for Pond-Breeding Salamanders. Conservation Biology 12:1113-1119.Semlitsch, R. D., and J. R. Bodie. 2003. Biological Criteria for Buffer Zones around Wetlands and Riparian Habitats for Amphibians and Reptiles. Conservation Biology 17:1219.Slonecker, E. T., L. E. Milheim, C. M. Roig-Silva, A. R. Maliza, D. A. Marr, and G. B. Fisher. 2012. Landscape consequences of natural gas extraction in Bradford and Washington Counties, Pennsylvania, 2004-2010. U. S. Geological Survey, Open-File Report 2012-1154.Strong, K. 2008. Conserving natural areas and wildlife in your community: smart growth strategies for protecting the biological diversity of New Yorks Hudson River Valley. New York Cooperative Fish and Wildlife Research Unit, Cornell University, and New York State Department of Environmental Conservation, Hudson River Estuary Program, Ithaca, NY.Thompson, H. 2012. Fracking boom spurs environmental audit. Nature 485:556-557.Trombulak, S. C., and C. A. Frissell. 2000. Review of Ecological Effects of Roads on Terrestrial and Aquatic Communities. Conservation Biology 14:18-30.Wang, D., T. Oakly, J. Mower, L. Shimmin, S. Yim, R. Huoneycutt, H. Tsao, and W.-h. Li. 2004. Molecular evolution of bat color vision genes. Molecular Biology and Evolution 21:295-302.Wanger, S. 1999. A review of the scientific literature on riparian buffer width, extent and vegetation. Office of Public Service & Outreach Institute of Ecology, University of Georgia.Wiltschko, W., U. Munro, H. Ford, and R. Wiltschko. 1993. Red light disrupts magnetic orientation of migratory birds. Nature 364:535-537. 5-18

124. Recomm mended Best Management Practices for Marcell Shale Gas Development in Maryland s lus a m yAppend 5A dix 5-19

125. K.N. Eshle eman & A.J. Elmore (2013) Chapter 5 h 5-20

126. Recomm mended Best Management Practices for Marcell Shale Gas Development in Maryland s lus a m y 5-21

131. K.N. Eshle eman & A.J. Elmore (2013) E Ch hapter 5(Hodges and Krementz 1996), (Bried and Ervin 2006), (Burke and Gibbons 1995), (C ( 2 Crawford and Semmlitsch 2007), (Gam et al. 2006), (Perry mbleet al. 2011), (Roe and Georges 2007), (Semlitsch 1998), (Semlitsch and Bodie 2003) , (Wanger 1999), ( ( h (Castelle et al. 199 (McComb et a 94), al.1993) 5-26

132. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland6. Protecting aquatic habitat, wildlife, and biodiversity1Freshwater aquatic habitats (defined as all streams, rivers, seeps, springs, wetlands, lakes, ponds,reservoirs, and floodplains) are unique and critical components of watersheds, providing hotspotsof biodiversity and transfers of energy, nutrients, and matter to coastal ecosystems such asChesapeake Bay (Likens and Bormann 1974, Lowe and Likens 2005, Alexander et al. 2007,Meyer et al. 2007). Risks to aquatic habitat and wildlife from Marcellus shale gas developmentactivities are numerous and include both direct impacts to the aquatic environment through streamdewatering, runoff generation, non-point and point source pollution, elevated thermal regimes, andindirect impacts through riparian habitat degradation. In Chapter 4 we discussed the possibleadverse impacts of temporary stream dewatering on surface water supplies, but it must beemphasized that streamwater itself is a critical physical habitat characteristic that must bemaintained to support aquatic biota, including species that contribute to Maryland’s bioticdiversity and recreational opportunities (e.g., trout). Land surface and channel erosion can greatlyincrease suspended sediment concentrations and turbidity levels—particularly during runoffevents—resulting in decreases in water clarity and increased sedimentation of fine materials.Turbidity and sedimentation reduce water column and benthic light availability, influencing fishforaging success and the quality of substrate for habitat. Excessive sedimentation fills the porespaces in which fish lay their eggs, adversely impacts benthic organisms such as freshwatermussels, and reduces the production of submersed aquatic vegetation (SAV). Increases in watertemperature due to forest clearing, riparian vegetation disturbances, or the inadvertent discharge ofpreviously impounded water into streams could negatively affect habitat quality and imposeadditional stresses on trout and other cold-water fish populations.Conservation of aquatic habitat and wildlife requires more extensive analysis than a simpleevaluation of the closest wetland or stream landscape feature. The numbers of tributaries, and theirrespective size and location within the stream network (Palmer et al. 2000, Benda et al. 2004) arecritical to understanding population dynamics and ecosystem function (Rice et al. 2001, Rice et al.2006). Organisms certainly vary in their susceptibility to disruptions in aquatic habitat quality,area, and connectivity due to variation in degree of habitat specificity (e.g., restricted movementwithin a small range of stream size) and ability or inability to travel over land between wetlands orstream reaches (Fagan 2002). While less common than dispersal within aquatic habitat, overlanddispersal plays a critical role in the exchange of individuals and genetic material between distantpopulations (Bunn et al. 1999, Bilton 2001). Human alteration of land cover proximate to andbetween aquatic habitat has the potential to adversely impact the fitness, survival, and matingsuccess of a wide variety of organisms (Oke et al. 1989, Sweeney 1993, Urban et al. 2006),leading to population declines or localized extinctions by restricting overland dispersal. As part ofa comprehensive plan for conserving aquatic habitat and wildlife, Maryland should consider thelarger landscape context of aquatic habitat. This should include considering: (1) how habitat isconnected through physical transport and biological dispersal; and (2) cumulative impacts towatersheds from the combined effects of agriculture, urbanization, and MSGD. Becauseconservation of aquatic habitat, wildlife, and diversity is related to land use within the entirewatershed, much of the discussion of terrestrial habitat, wildlife, and biodiversity is relevant to this1 Chapter co-authors: Andrew J. Elmore, Ph.D., and Keith N. Eshleman, Ph.D. (both at: Appalachian Laboratory, University of Maryland Center for Environmental Science, Frostburg, MD 21532) 6-1

133. K.N. Eshleman & A.J. Elmore (2013) Chapter 6chapter and many of the key recommendations (e.g., riparian buffer setback distances) werediscussed in Chapter 5.Figure 6-1. Stream density in western Maryland ranges from approximately 1.5 km/km2 (stream length divided bywatershed area) in the west to nearly 4.0 km/km2 in the east (Julian et al. 2012, Elmore et al. in review). High streamdensity has the potential to complicate the siting of well pads, particularly in Stronghold watersheds (cyan) or wherereproducing populations of native brook trout2 have been identified (purple)3. Many streams contain uniqueassemblages of rare, threatened and endangered species and all streams are essential hotspots for biotic life,supporting a wide range of visitors from the terrestrial landscape. Maryland DNR as part of their Maryland BiologicalStream Survey measures stream community composition annually4. Recent work to synthesize these data hasadvanced in many areas resulting in detailed maps (inset) delineating classes of stream reaches with similarcommunity composition (represented by different colors in the map)5.During Marcellus shale gas development, BMPs could be employed that reduce non-point sourcepollution, maintain habitat of sufficient quality for rare, threatened, and endangered species as wellas for species of significant commercial or recreational value (e.g., trout fisheries), and limit theintroduction of non-native species. Fortunately, Maryland’s aquatic environment has been underintensive study for many decades, including detailed stream survey work (MDNR 2010) andsynthetic analyses aimed at establishing robust descriptions of both reference and impacted aquaticpopulations (Utz et al. 2009). There are many important considerations to be made whenevaluating potential BMPs, but perhaps the most important is to generate and use accurate maps ofwetlands and the stream network (Julian et al. 2012, Elmore et al. in review). Such data ideally2 Data on stronghold watersheds and brook trout were acquired from Maryland Department of Natural Resources3 Additional complications can arise when MSGD is proposed in Tier II streams and watersheds. We view this as awater quality concern and therefore address this topic in Chapter 4; however, clearly the available BMPs haveconsiderable overlap.4 Maryland Department of Natural Resources. 2010. Retrieved from http://www.dnr.state.md.us/streams/MBSS.asp5 Unpublished model results, Matthew Fitzpatrick, UMCES, Appalachian Laboratory. 6-2

134. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandprovide a detailed spatial representation of the aquatic resources at risk and their proximity to theproposed development. BMP selection should also be sensitive to existing conservation efforts,which are currently used in Maryland to identify watersheds and wetlands of particular value tothe overall biodiversity of Maryland’s aquatic habitat and to identify current threats to thisbiodiversity (e.g., climate change). As has been the case in Maryland’s neighboring states thathave active shale gas development, a variety of BMPs have been developed that cover activities inboth upland and riparian environments. To best protect aquatic, wetland, and riparian habitats andwildlife in Maryland, the choice of BMPs should be based on the best available science anddetailed site analyses.A. Buffers and setbacksA primary BMP that has been widely employed to mitigate against adverse impacts on aquaticsystems is the use of a forest riparian buffer or, where forest is not present, a minimum setbackdistance from aquatic habitat. Upland forest buffers provide benefits to aquatic environments thatcan be classified as chemical, physical, or biological. The scientific basis for imposing a bufferwith a specific width depends on the overall rationale for the buffer. Many favorable chemical andphysical characteristics (e.g., stream nitrate concentration, sediment concentration, watertemperature, benthic light availability) can be achieved by imposing a relatively small forest bufferthat might not be much wider than the average canopy height (~100 ft). Consequently, a broadarray of literature that focuses on aquatic habitat condition and biological diversity supports theadoption of a 100 ft buffer from aquatic habitat (see Wanger 1999 and references therein). Widerbuffers are required to protect herptiles (reptiles and amphibians) that use forest riparian buffers(particularly on floodplains) for forage or dispersal. Therefore, the most appropriate forest riparianbuffer width is generally larger than what would be determined if only the aquatic environmentwere considered. For this reason, forest riparian buffer widths described in Chapter 5 (terrestrialbiodiversity conservation) generally supersede setbacks required for aquatic habitat, wildlife, andbiodiversity wherever forest is present in the proposed setback.Where forest riparian buffers are not present, either because agricultural activity or developed landcover extends all the way or most of the way to aquatic habitat, the benefit realized from setbackrestrictions can take different forms. Setbacks from aquatic habitat in agricultural lands can bejustified wherever MSGD would produce stormwater runoff, potentially transporting sediment-laden or nutrient-rich water to streams, rivers, and wetlands. For example, productive pastureshave the potential to abate the impact of such stormwater before it enters aquatic habitat.Similarly, in urban settings setbacks provide needed space for stormwater management, includingretention and diversion of stormwater, thus reducing the chance that any spills and leaks wouldlead to contamination of aquatic habitat (see Chapter 4). Finally, in all settings (forest, agricultural,and urban), setbacks from aquatic habitat provide benefits to recreational resources. Manystreams, some located close to urban and suburban communities, are used frequently for birding,fishing, boating, and swimming. To maintain the quality of these locations and reduce conflictswith these other uses, MSGD infrastructure should always be set back from aquatic habitat.Wherever possible, this setback should be forested (i.e., a riparian forest buffer) and Marylandshould use MSGD as an impetus to continue it’s overall efforts to increase the coverage of riparianforest buffers throughout western Maryland (e.g., mitigation plantings under Maryland’s no-net-loss of forest program). However, as stated above, the lack of forest in the riparian zone should notbe used as justification to reduce the setback distance. 6-3

135. K.N. Eshleman & A.J. Elmore (2013) Chapter 6As in many other states, no direct disturbance of any aquatic habitat for shale gas developmentshould be permitted in Maryland. We specifically define direct disturbance as any site preparation,earth-moving, well pad construction, grading, well drilling, equipment storage, or otherdevelopment activity on the land surface anywhere aquatic habitat is present (exceptions would bedisturbances associated with any necessary access road, utility and pipeline corridor constructionactivities that are specifically addressed in a later section of this chapter). In addition, consistentwith the most stringent setback requirements that we identified in our reviews (Lien and Manner2010), we recommend that a minimum 300 ft aquatic habitat setback be applied in westernMaryland (with the distance measured from the edge of any land disturbance, not from thelocation of a particular wellbore, to the edge of a particular habitat). Because setbacks are intendedin part to reduce the chance that spills of contaminated or sediment laden water reach aquatichabitat, the 300 ft setback should not be decreased in situations where the setback space is alreadycleared of forest (e.g., in agricultural settings). In all cases the intent is to protect the aquatichabitat, therefore while slope, land-cover, and soil condition likely influence the calculation of themost appropriate setback distance, a 300 ft setback should never be reduced. The 300 ft setback isconsistent with, but slightly more protective, than what is being proposed or used in other states.Pennsylvania DCNR uses a 300 ft no-disturbance buffer for situations where a body of watercontains threatened or endangered species or is considered either a high quality or exceptionalvalue stream or body of water6. Similarly, West Virginia also enforces a 300 ft buffer fordevelopment near streams with naturally reproducing trout populations7. The 300 ft buffer exceedsthe 100 ft requirement used on non-DCNR lands in Pennsylvania, as well as the 150 ftrequirement used by New York for conventional gas development to protect permanent surfacebodies of water and springs that provide domestic water (NYSDEC, 2011).Due to the nature of horizontal drilling and thelinear shape of many aquatic habitats (e.g.,streams and rivers), the requirement of a 300 ftminimum setback will not significantly restrictthe placement of MSGD infrastructure. Throughthe use of compact industrial parks for MSGD,we are confident that MSGD operators will beable to effectively access the majority ofMarcellus shale gas reserves without disturbinglands within 300 ft of streams. Further, the Figure 6-2. A plot of mean slope vs. distance to a streamtopography of western Maryland is highly reveals that aside from the floodplain (occurring on averagedissected, with the majority of streams located in within the first 100 ft of streams), the mean slope of landdeep, narrow ravines. Once out of these ravines, within 300 ft of streams is too steep for MSGD (i.e., >the Appalachian Plateau affords considerable 15%). Note that the mean slope includes many very steeparea of relatively flat (<15% slope) land that, if ravines, which skews the mean towards higher than expected values. In a separate calculation we found thatother conditions are met, could be available for only 40% of lands within the 300 ft stream setback wereMSGD. Excluding floodplains8, which we do not sufficiently flat to accommodate MSGD.6 However, PADEP enforces a 150 ft buffer for high quality streams and a 100 ft buffer for Class A trout streams outside of Pennsylvania state forests.7 West Virginia Horizontal Well Act; http://www.legis.state.wv.us/Bill_Status/bills_text.cfm?billdoc=hb401%20enr.htm&yr=2011&sesstype=4X&i=401.8 Floodplains can be adequately described by FEMA floodplain maps, which have recently been updated for western Maryland. 6-4

136. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandrecommend for MSGD, only 40% of the area within 300 ft of streams exhibits a slope <15%,making it appropriate for MSGD. This statistic, combined with the fact that most of the areawithin 300 ft of streams is forested (and no-net-loss of forest is recommended), suggests that thereare multiple interacting characteristics of lands within 300 ft of streams that make these landsunsuitable for MSGD.Figure 6-3. Wetlands in western Maryland are defined for regulatory purposes by the National Wetlands Inventory9(NWI) and by additional mapping efforts conducted by the Maryland DNR10. Total wetland area as defined by theunion of these data sets is 20,000 acres, 7000 acres of which is made up of small wetlands with an area less than 10acres. Certain wetlands with rare, threatened, endangered species or unique habitat receive special attention. Code ofMaryland Regulations (COMAR) Title 26, Subtitle 23, Chapter 06, Sections 01 & 02 identify these Wetlands ofSpecial State Concern (WSSC) and affords them additional protections; MDE is responsible for identifying andregulating these wetlands.B. Special protection of high-value assetsWe have identified many examples of specific BMPs that are being used by other states to provideadditional protection of high-value or highly sensitive assets such as 100-year floodplains,wetlands, high quality streams, natural trout streams, and rare, threatened and endangered speciesbeyond that provided through minimal setbacks. It has been strongly recommended that stateswith actual or proposed unconventional gas development undertake efforts to identify critical areaswith known endangered species, unique habitats, significant migration and breeding areas forbirds, mammals and aquatic organisms, and significant riparian areas (Lien and Manner 2010).9 U.S. Fish and Wildlife Service. 2012. National Wetlands Inventory website. U.S. Department of the Interior, Fish and Wildlife Service, Washington, D.C. http://www.fws.gov/wetlands/10 All DNR data for this map was acquired from Maryland Department of Natural Resources at http://dnrweb.dnr.state.md.us/gis/data/index.asp 6-5

137. K.N. Eshleman & A.J. Elmore (2013) Chapter 6The goal would be to develop maps that would allow identification and consideration of highvalue assets before initial site selection, thus reducing the chance of selecting sites that turn out tobe unsuitable or unfeasible for MSGD. Pennsylvania has already developed a PennsylvaniaNatural Diversity Inventory (PNDI) and associated Environmental Review Tool (ERT) that allowsthe public, a consultant, property owner, or project planner to perform on-line searches to identifypotential impacts to threatened, endangered, special concern species and special concern resourcesin the Commonwealth. The ERT can now accommodate linear projects up to 10 miles in lengthand area projects up to 1,200 acres in size; projects that exceed these limits can be submitted forenvironmental review as “large projects”. The County Natural Heritage Inventory (CNHI) effort inPennsylvania is another example of a cooperative program undertaken by the PennsylvaniaNatural Heritage Program (PNHP) partnership. The CNHI performs systematic studies of criticalbiological resources of the state on a county-by-county basis that form the basis for the PNDIpermit review data.As discussed in Chapter 1, prior to submitting a drilling application or comprehensive drilling planfor review and approval, a prospective shale gas developer could be required to consult availabledata on high-valued biological and water quality assets (e.g., Tier II streams and watersheds, seeChapter 4) within the western Maryland region. Similar to activities in other states, such anexercise would ideally allow a prospective operator to quickly determine the applicable BMPsgoverning MSGD at a particular site—thus saving considerable time and money during theplanning stages of a project. To support this effort, Maryland will likely need to continue itsefforts to identify high-value assets and publish in the scientific literature the methods used tomake this designation. Maryland has made considerable progress in this area with the creation ofits BioNet classification of irreplaceable natural areas (Chapter 5). Addressing aquatic biodiversityspecifically, Maryland has taken the BioNet approach one step further by identifying those“stronghold watersheds” (Figure 6-1) that are: (1) the most important areas for the protection ofMaryland’s aquatic biodiversity; (2) where rare, threatened, or endangered freshwater fish,amphibians, reptiles, or mussel species occur in the highest densities; and (3) where specialprotection is deemed necessary to ensure the persistence of imperiled fauna11. When documentedproperly, such data could be effectively used to channel MSGD into watersheds or sites withinwatersheds where it will have the least impact on aquatic habitat and biodiversity.The use of “stronghold watersheds” implicitly assumes that aquatic biodiversity conservationshould take a “watershed approach” since some particularly sensitive species may cease to persistif even relatively small portions of these watersheds become degraded. Such is the case withremaining populations of brook trout (Salvelinus fontinalis), which is the only salmonid native toMaryland. Brook trout were once widely distributed throughout the central Appalachians,including western Maryland. Brook trout are sensitive to increased stream temperatures(McCormic et al. 1972, Eaton and Scheller 1996), sediment and habitat alteration, and alteredstream chemistry (Leivestad 1982, Mount et al. 1988, Ingersoll et al. 1990). Altered land use isassociated with disappearance of brook trout populations, with sensitivity to agriculture west ofthe Blue Ridge in Maryland (Utz et al. 2010) and near universal extirpation from watershedsexceeding 4% impervious surfaces (Stranko et al. 2008). Brook trout are now dramaticallyreduced throughout their historic range (Hudy et al. 2008). Although populations occasionallyoccur in highly modified watersheds, these are the exceptions rather than the rule, as is evident by11 http://www.streamhealth.maryland.gov/stronghold.asp 6-6

138. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandtheir continued disappearance from watersheds. Many of Maryland’s populations currently inhabitonly portions of streams, are disconnected from other streams, and are present at low abundances.Thus, many of the existing populations do not have sufficient space or numbers for long termviability (Hilderbrand and Kershner 2000, Hilderbrand 2003) and many existing populations willlikely be extirpated within 20 years if the previous 20 years is a guide (Stranko et al. 2008)).Activities that decrease abundance or fragment existing populations will further increaseextirpation risk in even the stronghold watersheds.Our research into BMPs proposed or in use to protect aquatic habitat in other states suggest thatstream setbacks represent the primary instrument used to provide protection of aquaticbiodiversity, which in many cases leaves the door open to cumulative impacts from the linearcombination of many disturbances, regardless of their distance to aquatic habitat. Increasingsetbacks might offer some additional protection, but primarily for species with terrestrial lifestages (see Table 5-3). Therefore, some states have included language attempting to addresscumulative impacts in other ways. For example, in Colorado the responsibility is put on drillingoperators to minimize land disturbance, consolidate facilities, and co-locate infrastructurewhereever possible. Similarly, in Pennsylvania, operators are required to provide PADEP with adescription of their efforts to avoid, minimize, or mitigate for impacts to high-valued biologicalassets (e.g., co-location and centralization of infrastructure, use of specialized BMPs, well padspacing and density adjustments, working with other companies holding leases in this area toreduce cumulative impacts, etc.). In particular, there is recognition that minimizing the number ofwell pads through coordinated planning, consultation, and utilization of existing rights of way, canmitigate the cumulative impact on forests. Our opinion of these efforts is that, while well intended,they generally lack teeth, and will do little to address cumulative impacts in watersheds highlysensitive to even low levels of development.In select high-value watersheds12, Maryland should consider novel ways of establishing areallimits on surface development of all kinds (e.g., residential, commercial, wind power,unconventional shale gas, etc.) to address cumulative impacts. There is substantial scientificevidence that aquatic habitat and biodiversity respond to cumulative land disturbances or land usechanges (e.g., urbanization), either linearly or non-linearly (Booth et al. 2002, Walsh et al. 2005,Petty et al. 2010, Merriam et al. 2011). While establishing fixed response thresholds for aquaticsystems has proven difficult, there is considerable empirical evidence from the urban stormwaterliterature that cumulative surface imperviousness causes declines in aquatic biodiversity orecological condition beginning at impervious cover well below 10% (Walsh et al. 2005, Petty etal. 2010). In Maryland’s Piedmont watersheds, over half the aquatic insect species have becomeextinct from watersheds with 10% impervious cover (Utz et al. 2009). Among the most sensitiveaquatic species is brook trout, which is almost entirely restricted to watersheds with less than 4%impervious surface (Stranko et al. 2008).To provide an adequate margin of safety, we recommend that cumulative impervious cover(including all well pads, access roads, public roads, etc.) be maintained at less than 2% of thewatershed area in select high-valued watersheds. In some cases, stronghold and Tier II watersheds(e.g., the Savage River watershed) might be excellent candidates for such additional protections,but also possibly many or all of the watersheds containing brook trout. However, additional12 For example, stronghold watersheds and Tier II watersheds, the second of which have anti-degradation protection under MDE’s Clean Water Act regulatory authority (see Chapter 4). 6-7

139. K.N. Eshle eman & A.J. Elmore (2013) E Ch hapter 6analysis is warranted given the im i d mportant ram mifications o any approa limiting developmen We of ach nt.believe th use of multi-well pad to access relatively la hat m ds arge (~2 mi2) resources o shale gas ofthrougho western Maryland wi help to ma out M ill aintain these recommend low leve of surface e ded els edevelopm ment, thus prrotecting aquuatic systems. In addition to recomm n mending relat tively large (~2mi2) drill ling units, Maryland cou also cons M uld sider employ ying strategic land purch c hases (andsubseque incorpora ent ation into the existing sta forest sys e ate stem), as we as strategi conservat ell ic tioneasement programs, to maintain surface impe t t erviousness at current le evels in selec high-value ct ewatersheds. An overa policy of no net loss of forest (Ch all f hapters 1 and 5) will pro d ovide additioonalprotection to aquatic diversity wh other, more specific restriction fall short. MSGD oper here m c, ns ratorsshould be required to follow state e o e-of-the-art land reclama l ation proced dures (Chapters 1 and 5) thateffectively return soi permeabili to pre-dis il ity sturbance va alues before aattempting to permit addditiondrill pads in the same watershed. Finally, it should be rec s e s cognized tha most stream segments in at mwestern Maryland cu M urrently have very low (< e <1%) cumul ative imperv vious cover wwithin theircontributting basins (FFigure 6-4), suggesting that it may b possible to site MSGD in many ar t be o D reaswithout exceeding th recommen e he nded 2% imp pervious cov threshold ver d.Figure 6-4 New mappin technologies (Tarboton and Baker 2008) make possible the calculatio of percent 4. ng s d e onimpervious cover within the contributin basin of eve stream segm s ng ery ment in a lands scape. The resu ulting map (usi 2001 ingNLCD data on imperviou cover) show that many western Marylan watersheds have less than 1% imperviou cover. us ws nd usThis low im mpervious cover enables the continued exis stence of brook trout populati k ions and gener rally elevated a aquaticbiodiversit relative to ot ty ther areas of th state. he 6-8

140. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandC. Design and construction of well pads, access roads, pits, utility corridors, andpipelinesImplementation of BMPs in design and construction of well pads, access roads, and other ancillaryinfrastructure associated with Marcellus shale gas development can also minimize degradation orlosses of aquatic habitat and aquatic biodiversity. In addition to utilizing appropriate no-disturbance setbacks to protect aquatic habitat, proper implementation of appropriate erosion andsediment control measures (see Chapter 4) and stormwater pollution prevention techniques areimportant ways that aquatic habitat can be conserved.Well pads. We have recommended “zero-discharge” well pads as a BMP for western Marylandprimarily to control stormwater (and associated sediment pollution; see Chapter 4), as well asspills/leakage of chemicals from the surface to ground and surface water systems. We expectstormwater impacts to be most significant during well pad construction when the system is mostsusceptible to failure resulting from heavy rainfall events and after well drilling and completionactivities have ended and active collection, treatment, and disposal of stormwater runoff hasceased. During well drilling when well pads are being regularly monitored, operation of passiveand active stormwater collection should be able to minimize downstream impacts from the pad(i.e., this is the period in which stormwater can be actively collected in vacuum trucks, treated,disposed of, or used on site). Vacuum trucks should be kept on site throughout the period in whichactive stormwater collection is needed. However, only passive structural stormwater BMPs wouldbe operational after well development is completed (or possibly during periods between individualwell drilling events). Under these conditions, and without additional BMPs, the well pad wouldlikely be functioning largely as an impervious surface, thus increasing stormwater discharge,channel erosion in small receiving streams, and downstream sedimentation. Stormwater could alsobecome contaminated with salts or other pollutants through leaks from produced water storagetanks or liquid lines on-site, or from tanker trucks used to transport the produced water off-site.Therefore, other (passive) urban stormwater BMPs would still be needed. Since there exists thevery real possibility that runoff from these pads could carry pollutants off-site, we do notrecommend use of any BMPs that would promote infiltration due to the concern for groundwaterpollution. The best solution for addressing both quality (i.e., suspended solids) and quantity (i.e.,peak discharge) issues might be through construction of a below-grade lined pond adjacent to thebermed zero-discharge pad that could be used as a sump during active stormwater managementphases and easily converted into a retention pond prior to any passive phases. Regular (annual)maintenance of the pond would also be needed to ensure that the system is functioning correctly atall times. Additional water quality treatment could be obtained through operation of a constructedwetland sited downstream of the pond outlet.As discussed in Chapter 4, these recommendations are at least in partial conflict with two ofMaryland’s performance standards for controlling stormwater pollution13: standard no. 1: sitedesigns shall minimize the generation of stormwater and maximize pervious areas for stormwatertreatment; and standard no. 2: annual groundwater recharge rates shall be maintained bypromoting infiltration through the use of structural and non-structural methods. Clearly,13 Maryland Stormwater Design Manual , Volumes I & II (effective October 2000, revised May 2009); http://www.mde.state.md.us/programs/Water/StormwaterManagementProgram/MarylandStormwaterDesignManual/ Pages/programs/waterprograms/sedimentandstormwater/stormwater_design/index.aspx 6-9

141. K.N. Eshleman & A.J. Elmore (2013) Chapter 6Maryland’s stormwater designs originate from experience mostly with urban and suburbandevelopment and thus emphasize the use of BMPs that tend to maximize infiltration. Since theseare not recommended, there will need to be some significant attention given to alternativeapproaches such as the ones we have identified to address these problems during periods whenactive stormwater management would not be a viable option.Access roads. Wherever feasible, use of existing roads is the preferred option for facilitatingtransport of materials and personnel to well sites. However, we anticipate that there will be manycases where existing roads are nonexistent or inadequate and new roads will be needed. To protectaquatic habitat and minimize associated stormwater runoff, the design, routing, construction, andmaintenance of any access roads to a well pad should be done in a manner that can safely supportconsiderable heavy truck traffic, minimizes the clearing of forests, avoids steep slopes, avoidswetland and stream crossings, utilizes bridges or arched culverts for all stream crossings (leavingthe stream bed relatively undisturbed), and promotes sheet flow runoff from the road surface ontosurrounding soils wherever possible. Roads should not be located in or parallel to perennial orintermittent stream channels (i.e., no stream fords). Consistent with Pennsylvania DEP proposedrules, all wetland crossings should be avoided. Pennsylvania DCNR describes particularly goodpractices for constructing and maintaining such gravel roads to facilitate Marcellus shale gasdevelopment in Pennsylvania state forests that would also be highly applicable to westernMaryland (PADCNR 2011). For road construction, Pennsylvania DCNR: (1) recommendsutilizing materials and designs (e.g., crowning, elimination of ditches, etc.) that encourage sheetflow as the preferred drainage method for any new construction or upgrade of existing gravelroadways; (2) provides specific recommendations about aggregate depth, type, and placement; and(3) promotes the use of geotextiles as a way of reducing rutting and maintaining sub-base stability(PADCNR 2011). In Pennsylvania (as in western Maryland), it is typical for water to be a seasonalproblem on dirt and gravel roads and one of the best ways to minimize the risk of road failures isto selectively schedule hauling operations to avoid or minimize traffic during the spring thaw andother wet weather periods.Pits. We do not recommend the use of any open pits on-site for collecting and storing drillingwastes, flowback, or produced water due to concerns about surface water quality (see Chapter 4).We strongly recommend closed drilling systems in which all drilling and hydraulic fracturingfluids, chemicals, and liquid wastes are collected and stored in steel tanks that provide superiorprimary containment. Secondary containment can be provided by berms and liners placedstrategically under tanks and areas where liquid transfers take place. Tertiary containment can beprovided by construction of zero-discharge pads.Utility corridors and pipelines. In addition to providing vehicular access to sites, road corridorscan also be designed and constructed to facilitate below-ground transmission of gas, water, andAC power (if desired) to each well pad (as in the case of Pennsylvania state forests). Gathering (orfeeder) pipelines provide a way of transmitting the gas to compressor stations and to largertransmission pipelines that would also need to be co-located along major roads and highways.Flexible (e.g., HDPE) pipelines could be used to transmit water to each well pad to supporthydraulic fracturing operations. AC power could be used to power drilling equipment, lights, andother equipment on-site (in lieu of diesel generators). At Tiadaghton State Forest, for example, itwas possible to co-locate such infrastructure within a ~35-foot wide corridor immediately adjacentto the access road. Co-location of this ancillary infrastructure along the road corridor helps 6-10

142. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandminimize the extent of surface disturbance. Another viable alternative that would further minimizesurface disturbance is transfer of freshwater in flexible pipes above ground (King 2012)—althoughthis practice might be problematic in western Maryland where winter temperatures could causethese pipes to freeze and burst.Wherever possible, any belowground transmission of gas, water, and AC power should be co-located with road infrastructure to minimize impacts on aquatic habitat. Aquatic habitat crossings,where necessary, should be accomplished with appropriate use of bridges or arched culverts toensure free flow of water, particularly during flood stages (API 2011). The ecological effects,particularly on fish populations such as brook trout, of in-stream disturbance and semi-permanentbarriers to dispersal (e.g., culverts) are well documented, and therefore should be avoidedwherever possible (Burns 1972, Barton 1977, Meyer et al. 1999, Poplar-Jeffers et al. 2009). As inNew York, when utility lines are to be buried beneath streams, minimum burial depths should beenforced (NYSDEC 2011). Due to documented impacts of road crossings on fish spawningsuccess (Lachance et al. 2008), a general trout spawning substrate evaluation by DNR biologistsshould be required if any portion of a stream is expected to sustain a temporary or permanentblockage to fish passage. Alternatively, Maryland could ban the practice of diverting streams forany purpose and require the building of bridges or arched culverts to accomplish stream crossings.Likewise, open trenches within streams should be avoided in favor of using directional boringtechniques for installation of pipelines. Directional boring is a trenchless construction techniqueby which an operator can drill down next to a stream, bore horizontally under a stream, and thenbore up to the surface on the other side. The technique is highly advantageous over streamtrenching because it leaves the stream banks and streambed intact and the need for temporarilydewatering the stream is eliminated. Risks associated with directional boring are related to thepossibility of encountering unexpected subsurface voids, which have the potential to releasedrilling fluids and cuttings into stream waters. While such an event would be unfortunate, webelieve the benefits of directional boring outweigh the risks, which can generally be mitigated forby maintaining a depth of at least 10 ft below the streambed and avoiding drilling through highlyfractured substrate. In addition, efforts should be taken to avoid surface and subsurface spills orleaking of drilling fluids.Surface impoundments. There are currently no construction standards for the kind of small (< 15MG) freshwater impoundments that are being used throughout the state to temporarily store waterprior to its use for hydraulic fracturing in Pennsylvania (Marcellus Shale Advisory Commission,2011), although larger impoundments require dam construction and operation permits). Thefacility that one of us (KNE) visited in Tiadaghton State Forest was a shallow (3-4 ft deep), linedpond equipped with automated water level monitors that is capable of being continuously-monitored from either a remote or centralized office location (Figure 1-4). Adequate freeboard canbe maintained by pumping out water as needed, and a series of standpipes provide a means ofsafely refilling the pond from water tankers. A buried pipeline enables transfer of water from theimpoundment to nearby well pads. This seemed to be a particularly well designed facility thateffectively isolated the stored water from the stream network (i.e., there is not an obviousmechanism other than overflow of the structure by which non-native species could be introducedinto a nearby waterway). At a minimum, planning for one of these facilities should includeprecautionary measures to identify invasive species at water sources and avoid transporting thesespecies to impoundments located in watersheds where these species are not present. Further, thedischarge of any impounded water back into a natural water body should be prohibited to avoid 6-11

143. K.N. Eshleman & A.J. Elmore (2013) Chapter 6increasing water temperatures in groundwater fed streams and to avoid inadvertent non-nativeintroductions.Surface water intake structures. Intake structures should be designed to avoid entrainment ofaquatic organisms (Lien and Manner 2010) and invasive species management plans should includeprocedures for effectively washing intake equipment before leaving the site (see below).D. Erosion and sediment controlsHigh quality headwater streams—tributary streams, intermittent streams, and spring seeps — areessential to the health of stream and river ecosystems. Headwaters, when functioning properly,help to reduce sediment in the lower reaches of the stream network. Forested buffer zones slowerosion during peak stream discharge and help maintain low stream water temperatures, a criticalfactor in streams that support trout and other cold-water species (e.g., Koehn and Hairston-Strang2009, Henley et al. 2010). When forested buffers are removed or when headwater streams aredirectly disturbed, these channels become conduits for sediment and pollution that leads directly tolarger streams and coastal receiving waters (Kaplan et al. 2008). The direct effects of sediments onfish will vary with the concentration of suspended matter, duration and timing of exposure, degreeof sediment deposition, particle size distribution and type of sediment, and fish species and lifestages at which the fish is exposed (Kemp et al. 2011). Known impacts from sediments in streamsinclude: (1) reduced photosynthesis throughout the water column leading to reduced primaryproductivity and, therefore, reduced forage for higher trophic levels; (2) reduced periphytonattachment (a mixture of algae, cyanobacteria, heterotrophic microbes, and detritus found in mostaquatic habitats) and macrophyte growth, leading to reduced animal and plant abundance, speciesrichness, and diversity; and (3) increased sediment deposition and the loss of physical habitat(Kemp et al. 2011).Best management practices for sediment and erosion controls are covered in detail in Chapter 4.Aside from forest buffer disturbance (which we recommend protecting against with a 300 ftriparian forest buffer), much of the risk to aquatic habitat, and headwater stream ecosystemsspecifically, comes from ungraded roads on steep slopes or erodible soils, and stream crossings.Research has shown that 90% of the sediment that ends up in our nation’s waters from forestedlands is associated with improperly designed and maintained roads (Daniels et al. 2004).Unsurfaced roads, even with only moderate levels of light vehicle traffic, produce the greatestamount of sediment per unit of rainfall. Gravel roads with a maintained driving surface ofsufficient aggregate can be built to produce significantly less sediment (Sheridan and Noske2007). In Chapter 4 and elsewhere in this report we discussed the importance of regulating roadconstruction during MSGD and have recommended the use of gravel road design principlesrecommended by Pennsylvania DCNR (PADCNR 2011). Many BMPs related to road constructionhave proven effective, including elevating the road profile, building grade breaks and additionaldrainage features, removing berms, etc. (e.g., Scheetz and Bloser 2008). Other possible BMPsinclude the use of silt fencing, sedimentation ponds, mulches, and grass seeding, which have beenshown to be effective at sediment removal during periods of little rain, but inadequate duringperiods of flashy flows typical of mountain streams (Hedrick et al. 2010). Therefore, using acombination of BMPs and recognizing that additional protective measures might be necessaryduring certain times of the year (primarily late winter and early spring), is itself a BMP forMaryland. 6-12

144. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandE. Invasive species controlsEquipment used in MSGD is often transported great distances and used in relatively pristinewatersheds. In particular, water withdrawals from large rivers and reservoirs (where permitted)have the potential to introduce non-native and invasive species that can become a risk to nativeaquatic habitat and biodiversity. Maryland should take precautions to reduce the transmission ofinvasive plant and animal species by requiring an invasive species management plan of industryprior to any drilling operations. Of particular concern is the potential for harmful algal blooms(HAB), such as those produced by the non-cyanobacterial taxa, Prymnesium parvum (commonlyknown as “golden algae”), which is likely the most problematic HAB taxa in U.S. waters. P.parvum has caused large fish kills worldwide since as early as the 1930’s, and was first suspectedof fish kills in Texas in 1982 and confirmed in 1985 (Lopez et al. 2008). P. parvum blooms canspan many miles, across entire lakes, and can even propagate hundreds of miles downriver. Thisalgae has been implicated in the largest fish-kill associated with MSGD in PA and WV (DunkardCreek), suggesting it is an emerging threat to freshwater systems throughout the region (Brooks etal. 2001, Renner 2009).To protect aquatic habitat, each operator should be required to submit a site-specific invasivespecies management plan prior to any drilling operations. Such a plan should describe proceduresto be used during any water withdrawal from a local water source. At the very minimum,equipment should be power-washed and rinsed with clean water before leaving the withdrawalsite. Loose plant and soil material (potentially containing seeds, roots, or other viable plant parts)and unfiltered water, that has been removed from clothing, boots and equipment, or generatedfrom cleaning operations, should be disposed of in appropriate containers for disposal. Duringpower washing, wash water (including spray) should not discharge within 100 ft of any stream,existing or proposed wetland, or stormwater conveyance (e.g., ditch, storm drain, etc.). In nocircumstances should water that has been transferred between watersheds or moved upstreamabove confluences be discharged into aquatic habitat. This would include water that has beenstored in tanks or impoundments, but was not subsequently used in the drilling or completionprocess.F. Key recommendations6-A Direct disturbance of any aquatic habitat for shale gas development should not be permitted.6-B A minimum 300 ft aquatic habitat setback should be applied, with the distance measured from the edge of any land disturbance, not from the location of a particular wellbore, to the edge of a particular habitat.6-C Data that describe the biological resources of western Maryland should be developed and made available to MSGD applicants. These data should be used to effectively channel development away from high-value biological resources and into industrial zones accessible via existing roads and highways.6-D The use of multi-well pads to access relatively large (~2 mi2) resources of shale gas would enable the maintenance of reasonably low levels of surface development. 6-13

145. K.N. Eshleman & A.J. Elmore (2013) Chapter 66-E Cumulative surface development (including all well pads, access roads, public roads, etc.) could be maintained at less than 2% of the watershed area in high-value watersheds.6-F Initially, all MSGD could be excluded from areas of high-value assets (e.g., BioNet sites, stronghold watersheds, Tier II watersheds, etc.)6-G Closed drilling systems on zero-discharge drilling pads on which all drilling and hydraulic fracturing fluids, chemicals, and liquid wastes are collected and stored in steel tanks that provide superior primary containment to holding ponds are a best management practice. Vacuum trucks could be used to handle on-site runoff during drilling and well completion (see Chapter 4).6-H Maryland should require an invasive species management plan of industry prior to any drilling operations. Such a plan should include, at the minimum: 6-H.1 A description of water sources to be used to fill any impoundment, including analysis of any invasive species that might be present at the withdrawal site but absent from the watershed where the impoundment will be located. 6-H.2 Water withdrawal equipment should be power-washed and rinsed with clean water before leaving the withdrawal site.6-I Maryland should prohibit the discharging of any previously impounded water back into a natural water body, thus reducing the chance for the introduction of invasive species and short-term elevated thermal regimes in streams.6-J Wherever possible, existing roads should be used in MSGD. Where new roads are required, PA DCNR recommendations could be adopted: 6-J.1 Use materials and designs (e.g., crowning, elimination of ditches, etc.) that encourage sheet flow as the preferred drainage method for any new construction or upgrade of existing gravel roadways. 6-J.2 Where stream crossings are unavoidable, use bridges or arched culverts to minimize disturbance of streambeds. 6-J.3 Promote the use of geotextiles as a way of reducing rutting and maintaining sub- base stability. 6-J.4 Open trenches within streams should be avoided in favor of using directional boring techniques.6-K In general, during road and pad construction a combination of BMPs should be used to reduce sediment and erosion, recognizing that additional protective measures might be necessary during wet times of the year (primarily late winter and early spring).G. Literature citedAlexander, R. B., E. W. Boyer, R. A. Smith, G. E. Schwarz, and R. B. Moore. 2007. The role of headwater streams in downstream water quality. Journal of the American Water Resources Association 43:41- 59.API. 2011. Practices for Mitigating Surface Impacts Associated with Hydraulic Fracturing. API Guidance Document HF3. American Petroleum Institute, Washington, DC. 6-14

146. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandBenda, L., K. Andras, D. Miller, and P. Bigelow. 2004. Confluence effects in rivers: Interactions of basin scale, network geometry, and disturbance regimes. Water Resources Research 40.Bilton, D. T. 2001. Dispersal in Freshwater Invertebrates. Annual Review of Ecology and Systematics 32:159-181.Booth, D. B., D. Hartley, and R. Jackson. 2002. Forest cover, impervious-surface area, and the mitigation of stormwater impacts. Journal of the American Water Resources Association 38:835-845.Brooks, B. W., J. P. Grover, and D. L. Roelke. 2001. An emerging threat to inland waters. Environmental Toxicology and Chemistry 30:1955-1964.Bunn, S. E., P. M. Davies, and T. D. Mosisch. 1999. Ecosystem measures of river health and their response to riparian and catchment degradation. Freshwater Biology 41:333-345.Eaton, J. G., and R. M. Scheller. 1996. Effects of climate warming on fish thermal habitat in streams of the United States. Limnology and Oceanography 41:1109-1115.Elmore, A. J., J. P. Julian, S. M. Guinn, and M. C. Fitzpatrick. in review. Potential stream density in mid- Atlantic U.S. watersheds. Plos One.Fagan, W. F. 2002. Connectivity, fragmentation, and extinction risk in dendritic metapopulations. Ecology 83:3243-3249.Hedrick, L., S. Welsh, J. Anderson, L.-S. Lin, Y. Chen, and X. Wei. 2010. Response of benthic macroinvertebrate communities to highway construction in an Appalachian watershed. Hydrobiologia 641:115-131.Henley, W. F., M. A. Patterson, R. J. Neves, and A. D. Lemly. 2010. Effects of Sedimentation and Turbidity on Lotic Food Webs: A Concise Review for Natural Resource Managers. Reviews in Fisheries Science 8.Hilderbrand, R. H. 2003. The roles of carrying capacity, immigration, and population synchrony on persistence of stream-resident cutthroat trout. Biological Conservation 110:257-266.Hilderbrand, R. H., and J. L. Kershner. 2000. Conserving inland cutthroat trout in small streams: how much stream is enough? North American Journal of Fisheries Management 20:513-520.Hudy, M., T. M. Thieling, N. Gillespie, and E. P. Smith. 2008. Distribution, status, and land use characteristics of subwatersheds within the native range of brook trout in the Eastern United States. North American Journal of Fisheries Management 28:1069-1085.Ingersoll, C. G., D. R. Mount, D. D. Gulley, T. W. Lapoint, and H. L. Bergman. 1990. Effects of PH, Aluminum, and Calcium on survival and growth of eggs and fry of brook trout (Salvelinus fontinalis). Canadian Journal of Fisheries and Aquatic Sciences 47:1580-1592.Julian, J. P., A. J. Elmore, and S. M. Guinn. 2012. Channel head locations in forested watersheds across the mid-Atlantic United States: A physiographic analysis. Geomorphology 177-178:194-203.Kemp, P., D. Sear, A. Collins, P. Naden, and I. Jones. 2011. The impacts of fine sediment on riverine fish. Hydrological Processes 25:1800-1821.King, G. E. 2012. Estimating frac risk and improving frac performance in unconventional gas and oil wells.in Society of Petroleum Engineers Hydraulic Fracturing Conference, The Woodlands, TX.Koehn, S. W., and A. Hairston-Strang. 2009. Forestry Best Management Practices in Maryland: Implementation and Effectiveness for Protection of Water Resources. Maryland Department of Natural Resources Forest Service.Leivestad, G. 1982. Physiological effects of acid stress in fish. Pages 157-164 in R. E. Johnson, editor. Acid rain/fisheries. American Fisheries Society, Bethesda, MD.Lien, A. M., and W. J. Manner. 2010. The Marcellus shale: resources for stakeholders in the Upper Delaware watershed region. Pinchot Institute for Conservation.Likens, G. E., and F. H. Bormann. 1974. Linkages between terrestrial and aquatic ecosystems. BioScience 24:447-456.Lopez, C. B., E. B. Jewett, Q. Dortch, B. T. Walton, and H. K. Hudnell. 2008. Scientific assessment of freshwater harmful algal blooms. Joint Subcommittee on Ocean Science and Technology, Washington, D. C.Lowe, W. H., and G. E. Likens. 2005. Moving headwater streams to the head of the class. BioScience 55:196-197. 6-15

147. K.N. Eshleman & A.J. Elmore (2013) Chapter 6McCormic, J. H., B. R. Jones, and K. E. Hokanson. 1972. Effects of temperature on growth and survival of young brook trout, Salvelinus fontinalis. Journal of the Fisheries Research Board of Canada 29:1107-&.MDNR. 2010. Maryland Biological Stream Survey.Merriam, E. R., J. T. Petty, J. G. T. Merovich, J. B. Fulton, and M. P. Stranger. 2011. Additive effects of mining and residential development on stream conditions in a central Appalachian watershed. Journal of the North American Benthological Society 30:399-418.Meyer, J. L., D. L. Strayer, J. B. Wallace, S. L. Eggert, G. S. Helfman, and N. E. Leonard. 2007. The contribution of headwater streams to biodiversity in river networks. Journal of the American Water Resources Association 43:86-103.Mount, D. R., C. G. Ingersoll, D. D. Gulley, J. D. Fernandez, T. W. Lapoint, and H. L. Bergman. 1988. Effect of long-term exposure to acid, aluminum, and low calcium on adult brook trout (Salvelinus fontinalis) .1. Survival, growth, fecundity, and progeny survival. Canadian Journal of Fisheries and Aquatic Sciences 45:1623-1632.NYSDEC. 2011. Revised draft supplemental generic environmental impact statement on the oil, gas, and solution mining regulatory program. Albany, NY.Oke, T. R., J. M. Crowther, K. G. McNaughton, J. L. Monteith, and B. Gardiner. 1989. The Micrometeorology of the Urban Forest [and Discussion]. Philosophical Transactions of the Royal Society of London 324:335-349.Palmer, M. A., C. M. Swan, K. Nelson, P. Silver, and R. Alvestad. 2000. Streambed landscapes: evidence that stream invertebrates respond to the type and spatial arrangement of patches. landscape ecology 15:563-576.PADCNR. 2011. Guidelines for administrating oil and gas activity on state forest lands. Commonwealth of Pennsylvania Department of Conservation and Natural Resources Bureau of Forestry.Petty, J. T., J. B. Fulton, M. P. Strager, J. G. T. Merovich, J. M. Stiles, and P. F. Ziemkiewicz. 2010. Landscape indicators and thresholds of stream ecological impairment in an intensively mined Appalachian watershed. Journal of the North American Benthological Society 29:1292-1309.Renner, R. 2009. Salt-loving algae wipe out fish in Appalachian stream. Environmental Science & Technology 10.1021/es903354w.Rice, S. P., R. I. Ferguson, and T. B. Hoey. 2006. Tributary control of physical heterogeneity and biological diversity at river confluences. Canadian Journal of Fisheries and Aquatic Sciences 63:2553-2566.Rice, S. P., M. T. Greenwood, and C. B. Joyce. 2001. Tributaries, sediment sources, and the longitudinal organisation of macroinvertebrate fauna along river systems. Canadian Journal of Fisheries and Aquatic Sciences 58:824-840.Scheetz, B. E., and S. M. Bloser. 2008. Environmentally Sensitive Maintenance Practices for Unpaved Roads: Sediment Reduction Study. Center for Dirt and Gravel Road Studies, Penn State University.Stranko, S. A., R. H. Hilderbrand, R. P. Morgan, M. W. Staley, A. J. Becker, A. Roseberry-Lincoln, E. S. Perry, and P. T. Jacobson. 2008. Brook trout declines with land cover and temperature changes in Maryland. North American Journal of Fisheries Management 28:1223-1232.Sweeney, B. W. 1993. Effects of streamside vegetation on macroinvertebrate communities of white clay creek in eastern North America. Proceedings of the Academy of Natural Sciences of Philadelphia 144:291-340.Tarboton, D. G., and M. E. Baker. 2008. Towards an algebra for Terrain-Based Flow Analysis.in N. J. Mount, G. L. Harvey, P. Alplin, and G. Priestnall, editors. Representing, Modeling and Visualizing the Natural Environment: Innovations in GIS. CRC Press, Florida.Urban, M. C., D. K. Skelly, D. Burchsted, W. Price, and S. Lowry. 2006. Stream communities across a rural-urban landscape gradient. Drivers Distrib 12:337-350.Utz, R. M., R. H. Hilderbrand, and D. M. Boward. 2009. Identifying regional differences in threshold responses of aquatic invertebrates to land cover gradients. Ecological Indicators 9:556-567. 6-16

148. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandUtz, R. M., R. H. Hilderbrand, and R. L. Raesly. 2010. Regional differences in patterns of fish species loss with changing land use. Biological Conservation 143:688-699.Walsh, C. J., A. H. Roy, J. W. Feminella, P. D. Cottingham, P. M. Groffman, and R. P. Morgan. 2005. The urban stream syndrome: current knowledge and the search for a cure. Journal of the North American Benthological Society 24:706-723. 6-17

149. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland7. Protecting public safety1Modern shale gas development is an industrial activity that involves handling of very largequantities of hazardous or toxic chemicals, hydraulic fracturing fluids, and wastewaters (“brines”)at outdoor sites often located in remote or rural areas. It is very common for most of thesematerials to be transported by trucks for considerable distances on public roads to the drillingsites. Further, like any other outdoor activities, these drilling sites are exposed to extreme weatherand environmental conditions (e.g., snowstorms, rainstorms, floods, windstorms, freezingconditions, etc.) that not only make working at such sites difficult, but also elevate the risk ofaccidents, spills, or leakages away from a particular site. Unless such spills are prevented and/orquickly contained, surface water or groundwater contamination may result, which can exposehumans or ecosystems to toxic chemicals. For this reason, New York State has concluded thatshale gas well pads and all associated on-site infrastructure should be treated like other industrialfacilities. The first step in protecting public safety from some of the primary hazards associatedwith industrial facilities is siting such facilities as far away as possible from homes, businesses,public buildings, or places with high levels of recreational activity (e.g., hiking trails, parks, picnicareas, etc.). This can be achieved through the use of setbacks and careful permitting in the vicinityof parks and other recreational resources (see Chapter 8). Secondly, employing best managementpractices in well construction (e.g., casing and cementing) in order to ensure wellbore integrityand isolation are important steps that must be used to control migration of hydrocarbons, brines, orhydraulic fracturing fluids into groundwater, causing pollution of underground drinking watersupplies (see Chapter 3). As discussed below, security measures such as adequate signage,lighting, fencing and supervision that are appropriate to other industrial facilities should berequired to ensure that shale gas development is conducted in as safe a manner as possible(NYSDEC 2011).A. Spill prevention and emergency responseThe prevention and containment of spills involving hazardous or toxic chemicals used in thecompletion process, hydraulic fracturing fluids, or wastewaters at a well site—or during transit toor from a well site—is a very important component of providing protection of public safety, aswell as the surrounding environment. As noted previously in Chapters 1 and 4, a best practice inspill prevention and protection of public safety in general is the development of a site-specific,emergency response plan (ERP) that describes specifically in writing how a particular operatorwill respond to different emergencies (e.g., spills) that may occur during each phase of shale gasdevelopment at a particular site (or off-site) and across the operators’ many related facilities (e.g.,multiple wells and water impoundments). The procedures outlined in an ERP are intended toprovide for the protection of lives (workers and the public at large), property (both on-site and off-site), and the environment, through appropriate advance planning, safety training, and coordinateddeployment of company and community assets. In addition to addressing spill prevention andclean-up procedures, an ERP would also logically include procedures for protecting the publicfrom fires, explosions, or blow-outs that could occur on a well pad2. While the names of such1 Chapter co-authors: Keith N. Eshleman, Ph.D. and Andrew J. Elmore, Ph.D. (both at: Appalachian Laboratory, University of Maryland Center for Environmental Science, Frostburg, MD 21532)2 West Virginia Horizontal Well Act. http://www.legis.state.wv.us/Bill_Status/bills_text.cfm?billdoc=hb401%20enr.htm&yr=2011&sesstype=4X&i=401 7-1

150. K.N. Eshleman & A.J. Elmore Chapter 7plans vary from state to state, documents describing emergency preparedness are required orproposed in all states in our review (and are advocated by API). The ERP proposed for shale gasdevelopment in New York State would have, at a minimum, the following elements (NYSDEC2011): • Identity of a knowledgeable and qualified individual with the authority to respond to emergency situations and implement the ERP • Site name, type, location (including a copy of 7½ minute USGS map), and operator information • Emergency notification and reporting (including a list of emergency contact numbers for the area in which the well site is located; and appropriate regulatory office), equipment, key personnel, first responders, hospitals, and evacuation plan • Identification and evaluation of potential release, fire and explosion hazards • Description of release, fire, and explosion prevention procedures and equipment • Implementation plans for shut down, containment, and disposal • Site training, exercises, drills, and meeting logsIn addition, as required by PADEP, a Prevention, Preparedness, and Contingency (PPC) Plan inPennsylvania (similar to an ERP in New York) must include a list of all chemicals or additivesused and the different wastes generated by hydraulic fracturing (and approximate quantities ofeach material and the method of storage on-site), as well as MSDS data, toxicological data, andwaste chemical properties. A more comprehensive and standardized PPC Plan would also include:(1) assigning 9-1-1 addresses to sites to aid in emergency responses; (2) providing geographicpositioning system (GPS) coordinates for access roads and well pad sites; and (3) distributing PPCPlans to the appropriate county emergency management coordinator so that emergency responderswould have immediate access to MSDS information in the event of an actual emergency(Marcellus Shale Advisory Commission 2011). Another best practice related to implementation ofan ERP is that inspectors be given 24-hour notice before any major operation occurs at a particularwell site (i.e., cementing, hydraulic fracturing, drilling, flaring) (STRONGER 2011).As is the policy in most states that we reviewed, an ERP should be developed in Maryland andsubmitted to the appropriate state regulatory authority as part of the well permit applicationprocess. The ERP could also be part of an overall site-specific safety plan developed by anoperator to address the full gamut of public safety issues involving such topics as site security andoff-site transportation of materials. In addition to development and implementation of an ERP,there are many other BMPs that are critical for spill prevention and containment. Most of these areprimarily used for protecting water resources (both surface water and groundwater) and werediscussed in more detail in Chapter 4.Effective implementation of an ERP in the event of an actual emergency requires specializedteams of emergency responders, appropriately trained in specific well pad emergencies, publicsafety, and methods isolating and securing an incident site. Each county should have at least onespecialized team of emergency responders available at all times to respond to an emergency. Theemergency responders should leave control of well blow-outs, fires or contaminant releases toprofessional, operator-trained experts utilizing equipment staged in a manner to provide a timelyresponse to emergencies. It is important that each well operator maintain all necessary equipmentto respond to various types of emergencies in a satisfactory operating condition and on-sitethroughout the drilling and completing phases of the operation. 7-2

151. Recomm mended Best Manageme Practice for Marcel Shale Ga Developm t ent es llus as ment in Mary rylandA county or regional task force of public/ind y o dustry partner should be formed to f rs e facilitatecoordinat tion, knowleedge sharing and refinem of eme g, ment ergency respoonse protoco (Marcellu ols usShale Ad dvisory Com mmission 201 The Com 11). mmission als recommen so nded design andimplemenntation of a unified com mmand system for addres m ssing well pa incidents— ad —with the FederalEmergen Managem Agency (FEMA) IC 300- and 400-level tr ncy ment y CS d raining prog grams serving as gappropria models for state use (Marcellus Shale Adviso Commis ate f S ory ssion 2011).Figure 7-1 The road net 1. twork in Garret and Allegany Counties con tt y nsists of primar and seconda roads and ry arynumerous bridges, some with specific weight restrictions. A detailed transportation plan would c b w d n consider the ex xisting tion network within a formal network analytransportat w ysis, designed to reduce conflicts between M o MSGD and exis stingroad uses. The road data shown are curr as of 2011 and were acqu rent 1 quired from the Maryland Stat Highway e teAdministraation (SHA)3. Allegany and Garrett County provided bridg locations an weight limit A G y dge nd ts.B. Site security sDuring si preparation, drilling, and completing of shale gas wells, t presence and operati of ite e the e ionheavy eq quipment suc as drill rig and the sto ch gs orage and us of large q se quantities of chemicals pr resentsafety haazards that ar comparable to those present at ma other ind re p any dustrial facil lities. For th hisreason, it is importan that well sites and asso t nt ociated infra astructure be treated like other indust trialsites—inncluding secu uring these facilities so they can be o f t operated in a safe manne Once dril er. llingand comp pletion equip pment and ch hemicals hav been rem ve moved and we are prod ells ducing, other rsecurity measures (or a reduced level of secu m r l urity) may be more appro e opriate than during the e earlierdrilling and completi phases of an operatio Specific best practice to be implemented by an a ion on. es yoperator would certainly include: (1) adequat perimeter fencing (at least a 6 ft h te r high chained link dor equivaalent), gates (with keyed locks), and signage in p d place around drill rigs, engines, d3 http://ww ww.sha.marylan nd.gov/Index.a aspx?PageId=282 7-3

152. K.N. Eshleman & A.J. Elmore Chapter 7compressors, tanks, impoundments, and separators, to restrict public access; and (2) use of safetyor security guards to further control access (particularly important during active drilling andcompletion phases of an operation). As in Ohio, duplicate keys to all locks should be provided tothe regulatory agency and to local emergency responders upon request.C. Transportation planningTransportation planning is an important consideration in shale gas development due to the needfor moving large quantities of heavy equipment, chemicals, water, and wastewater either to orfrom various sites distributed throughout a particular region. We discuss transportation planning inChapter 9 in the context of protecting quality of life and aesthetics in predominantly rural westernMaryland, but it should be kept in mind that transportation planning must also address risks topublic safety—especially those specifically posed by frequent truck transport of materials on ruralpublic roadways and bridges that in many cases were neither designed nor constructed for suchpurposes.The natural gas industry faces significant logistical challenges associated with transporting andstoring the tremendous volumes of sand, pipe, water, and other materials that are necessary to drilland complete a Marcellus shale gas well. According to the Governor’s Marcellus Shale AdvisoryCommission (2011), the maximum distance to effectively serve a well head in Pennsylvania is 75miles, especially due to the steep terrain found in many of Pennsylvania’s drilling locations. Thecloser a drilling company can get to areas where it can store the vast quantities of materialsrequired for drilling a well, the better the efficiency of the drilling operation. Railroads, already inplace and operational, could provide an alternative system for effectively and efficiently receiving,storing, and trans-loading commodities to well heads in Pennsylvania and throughout theMarcellus shale region. The railroads (and rail terminals) provide an added benefit of reducing theneed to develop a staging area on forest land or other vacant land within a 50- to 75-mile radius.Nevertheless, even with the benefit of rail transportation, there would still exist the need for trucksto move material and equipment from a rail terminal to the well pads (Marcellus Shale AdvisoryCommission 2011), so the impact of using rail transportation on truck transportation may not bethat great. Further, while western Maryland is obviously home to a major CSX rail yard atCumberland, the closest actual rail terminal to the region is in Hagerstown—60-100 miles away—perhaps too far away for railroads to play a major role in staging MSGD operations. The proximityof the Cumberland rail yard to a major interstate highway (I-68) that bisects western Marylandwould seemingly make this an ideal location for an MSGD staging area, however. If Marylanddecides to move ahead with MSGD, the state might consider investing public funds in newterminal facilities in western Maryland to support the activity, provide incentives for privatefinancing, or both (as has been done in Pennsylvania through Rail Freight and Rail TransportationAssistance Programs).With respect to truck transportation, there are obvious risks to public safety (injury, death)associated with accidents involving additional traffic onto roads and bridges, plus additional risksassociated with exposure to spilled hazardous chemicals, fires, or explosions resulting from suchaccidents. As recommended for Pennsylvania, we believe it is reasonable to expect the appropriatestate transportation authorities to calculate, evaluate, and address the major impacts of additionaltruck traffic on the road and highway system prior to shale gas development occurring in anundeveloped part of the state (Marcellus Shale Advisory Commission 2011). At the same time,counties and municipalities should also undertake an inventory and structural evaluation of 7-4

153. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandlocally-owned bridges currently exempt from federally mandated inspections (typically 8 ft to 20ft) to ensure that these structures are capable of safely handling the additional traffic (and loads)associated with shale gas development. While this recommendation was made for Pennsylvania,we believe that the same type of analysis could and should be done for western Maryland. Wherethe road network or bridges are deemed inadequate for supporting the additional traffic, the roadsystem (including inadequate bridges) should be upgraded to support such traffic prior to shale gasdevelopment occurring or such traffic should not be permitted on these roads. With respect tomovement of heavy equipment on state highways, we also agree with the recommendation that thestate should be responsible for establishing a protocol to allow for emergency transport of suchequipment during off-hour periods (evenings, nights, and weekends) in cases where there is animmediate need of the equipment (Marcellus Shale Advisory Commission 2011). The protocolwould be similar to how ‘wide loads’ are presently transported in the state of Maryland and wouldthus require cooperation and coordination with the state police who assume primary responsibilityfor the highway system.D. Key recommendations7-A The first line of defense in protecting public safety is designing MSGD operations in a way that maintains separation between MSGD infrastructure (including transportation routes) and the public. 7-A.1 Facilities should be sited as far away as possible from homes, businesses, public buildings, or places with high levels of recreational activity (e.g., hiking trails, parks, picnic areas, etc.) (see Chapter 9 also). 7-A.2 Best management practices in well construction (e.g., casing and cementing) should be followed to ensure wellbore integrity and isolation (see Chapter 3). 7-A.3 Proper monitoring and pre-development assessment are important steps to limit the migration of hydrocarbons, brines, or hydraulic fracturing fluids into groundwater, causing pollution of underground drinking water supplies and to enable rapid detection in the event of migration (see Chapters 1 and 4).7-B MSGD applicants should be required to develop site-specific, emergency response plans (ERP) that describes in detail how a particular operator will respond to different emergencies that may occur during each phase of shale gas development at sites, or transportation routes between sites, permitted for MSGD. 7-B.1 The ERP must include many types of standard information, including the names and contact information for first responders, and location (including GPS coordinates) of MSGD sites. 7-B.2 The ERP must include variations on standard responses demonstrating sensitivity to weather, time of day, time of year, and the particular geography of sites (e.g., topographic and soil conditions). 7-B.3 The ERP must also include a list of all chemicals or additives used, expected wastes generated by hydraulic fracturing, approximate quantities of each material, the method of storage on-site, MSDS for each substance, toxicological data, and waste chemical properties.7-C Best management practices implemented to avoid emergencies should include: 7-5

154. K.N. Eshleman & A.J. Elmore Chapter 7 7-C.1 Adequate perimeter fencing (at least a 6 ft high chained link or equivalent), gates (with keyed locks), and signage in place around drill rigs, engines, compressors, tanks, impoundments, and separators, to restrict public access. 7-C.2 Use of safety or security guards to further control access (particularly important during active drilling and completion phases of an operation). 7-C.3 Duplicate keys to all locks should be provided to the regulatory agency and to local emergency responders.7-D Maryland’s Department of Transportation should calculate, evaluate, and address the major impacts of additional truck traffic on the road and highway system prior to the state permitting MSGD. 7-D.1 Counties and municipalities should also undertake an inventory and structural evaluation of locally-owned bridges currently exempt from federally mandated inspections to ensure that these structures are capable of safely handling the additional traffic (and loads) associated with MSGD. 7-D.2 The state should establish a protocol to allow for emergency transport of heavy or oversized equipment during off-hour periods (evenings, nights, and weekends).E. Literature citedMarcellus Shale Advisory Commission. 2011. [Pennsylvania] Governors Marcellus Shale Advisory Commission Report.NYSDEC. 2011. Revised Draft Supplemental Generic Environmental Impact Statement on the Oil, Gas, and Solution Mining Regulatory Program.STRONGER. 2011. Ohio Hydraulic Fracturing State Review. Oklahoma City, OK: State Review of Oil and Natural Gas Environmental Regulations, Inc. 7-6

155. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland8. Protecting cultural, historical, and recreational resources1Western Maryland (Allegany and Garrett County) contains a plethora of cultural, historical andrecreational resources. Many types of sites (e.g., national and state historic sites, heritage areas,local historic districts, state parks, wildlife management areas, wildlands, etc.) would be at risk ofimpairment, either through physical, visual, auditory, or olfactory degradation. In addition to theirintrinsic value, some of these resources, such as historical landmarks or unique natural landscapefeatures (lakes, waterfalls, etc.), draw considerable tourism, generating revenue for localcommunities and the state of Maryland. In 2011, nearly $6M in sales tax was collected throughcombined tourism-related sales in these two counties.2 The local job market depends in part ontourism, accounting for over $50M in tourism-related salaries for the two-county area in 20083,4.A decline in the quality or quantity of resources that attract tourism would potentially limit furthereconomic development in this area. Disturbance associated with site preparation work, welldrilling activities, truck traffic, and operation of heavy equipment—unless successfully avoided ormitigated for—could negatively impact the enjoyment of natural areas for hunting, fishing, hiking,boating, and other recreational activities. Finally, natural areas might also be impacted throughinadvertent introductions of invasive species or losses of natural biological/landscape biodiversity(see Chapters 5 and 6).A. Identification of sitesProtection of cultural, historical, and recreational resources must begin with identification of sitesthat would be adversely affected by Marcellus shale gas development. However, many of theseresources (e.g., national and state historic properties) are virtually unknown and are typicallyunmapped. In cases like this, New York State requires identification of all sites that are eligible forinclusion on state and national registers of historic properties, or are included on the stateinventory, to ensure that they receive special consideration, protecting them from disturbance orimpairment. Many state and federal databases exist to provide such information, but theinventoried data are typically not transmitted to users in the form of a digital map. New York hasactually mapped out its visually sensitive resource areas and has proposed that applicants submit avisual resource mitigation plan as part of the permit application process (NYSDEC 2011). Inwestern Maryland, there are six listed items on the National Trust for Historic Preservation, plus52 sites in Allegany County and 23 sites in Garrett County listed on the National Register ofHistoric Properties.5 Moreover, there are literally hundreds of sites in these counties that are listedon the state inventory including historic properties, local historic districts (Cumberland andFrostburg), historic cemeteries and monuments, roadside historical markers, and a state heritage1 Chapter co-authors: Keith N. Eshleman, Ph.D. and Andrew J. Elmore, Ph.D. (both at: Appalachian Laboratory, University of Maryland Center for Environmental Science, Frostburg, MD 21532)2 Maryland Tourism Development Board. 2008. Annual report, available online at: http://www.visitmaryland.org/AboutMDTourism/Pages/TourismNewsAndReports.aspx.3 Department of Business and Economic Development. 2011. Maryland Tourism Fastfacts. Available online at: http://visitmaryland.org/AboutMDTourism/Documents/Annual_Report_2011.pdf.4 Additional information on the economic impact of Maryland State Parks can be found in the 2010 Maryland State Parks Economic impact & Visitor Study, accessed at the following URL on February 13, 2013: http://www.dnr.maryland.gov/publiclands/pdfs/economicimpactstudy2010.pdf5 http://www.mht.maryland.gov/nr/NRPickCounty.html. 8-1

156. K.N. Eshleman & A.J. Elmore (2013) Chapter 8area (Mountain Maryland Gateway to the West Heritage Area that includes the towns of Accident,Deer Park, Friendsville, Grantsville, Kitzmiller, Loch Lynn Heights, Mountain Lake Park,Oakland, McHenry, and Bloomington plus scenic byways that connect these towns in GarrettCounty) (not shown). Given the large number of sites that could be impacted in these twocounties, best practice would be for operators to consult with the Maryland Historical Trust(MHT) within the Maryland Department of Planning and other county and local historicpreservation offices during the planning and permit application process to ensure that no eligibleor existing cultural or historical sites would be potentially disturbed or impaired by any aspect ofshale gas development.Western Maryland also contains extensive public recreational resources that will requireidentification and mapping, including: a national historical park, a national scenic trail, stateparks, state forests, state forest trails, state wildlife management areas, natural areas, wildlands, awild river, a national byway, and two state scenic byways. Most of these state and federalrecreational resources and important natural areas are reasonably well known and mapped inMaryland. On an areal basis, state forest land (118,099 acres) is by far the largest publicrecreational space in the two-county region, covering 21% of the total land area. Wildlifemanagement areas cover 17,809 acres (2.6%) and state parks cover 10,203 acres (1.5%).Confronted with a similar level of diversity in the types of resources that exist, PennsylvaniaDCNR Bureau of Forestry recommends that any constraints mapping done by gas drillingcompanies operating in the Pennsylvania state forests should be done in close consultation withlocal stakeholders who typically have the best knowledge of these resources. In Maryland,regardless of whether or not a proposed operation would be located on state or federal land, bestpractice would require close consultation with local governments, state park and forest officials,national park managers, and wildlife managers who are familiar with the resources that could beimpaired by shale gas development. To facilitate this planning activity, we have provided a list ofthe major public recreational and natural resource areas that could be impacted by shale gasdevelopment in western Maryland (Table 8-1, Figure 8-1, Figure 8-2).B. Setback requirements and mitigationTo avoid disturbances or impairment of major cultural and historical resources, New York Statehandles proposed oil and gas drilling near these sites on a case-by-case basis. A variety ofmitigative actions can be required including: (1) visual screening of drilling operations; (2)setback requirements greater than minimums for private homes (100 ft) and public buildings orareas (150 ft); (3) restriction on times of operation (e.g., avoid tourist season, museum hours,whitewater release dates, opening days for hunting and fishing, etc.); and (4) landscapingreclamation requirements. In New York, many of these mitigative actions are presently added asconditions to drilling permits (NYSDEC 2011). Other mitigative BMPs that have been proposedin New York include: relocation of MSGD infrastructure found damaging by local residents orresource managers, use of camouflage or disguise to reduce the impact of MSGD infrastructure,maintaining low facility profiles, downsizing the scale of a project, using alternative technologies,using non-reflective materials, and controlling off-site migration of lighting (NYSDEC 2011).With respect to state forest recreational areas, Pennsylvania DCNR Bureau of Forestry requires a300 ft. setback from any state forest picnic area, trail, road of historic value, tree plantation,overlook, vista, fire tower site, or existing right of way; this setback also affords additionalprotection of public safety through conflict avoidance. The Bureau of Forestry also relies on local 8-2

157. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandknowledge of cultural sites, recreational trails, trailheads, high use areas, viewsheds, vistas, andhigh aesthetic areas during the permitting process with a goal of avoiding areas of (or providingincreased setbacks from) concentrated recreational activity and developed recreational sites whenpermitting gas related infrastructure (PADCNR 2011). In the Pennsylvania state forests, animportant criterion in site selection for drilling pads is the degree to which locations can providenatural vegetative or topographic screening (PADCNR 2011). Additionally, API recommends thatsetbacks be increased to take into account prevailing winds and topography; in New York, sitesare assessed for their archeological importance (NYSDEC 2011).Table 8-1. Public recreational resources in Allegany and Garrett County, Maryland. Name of Resource Administered by County Chesapeake and Ohio Canal National Historical Park National Park Service Allegany Potomac Heritage National Scenic Trail National Park Service Allegany Historic National Road U.S. Dept. of Transportation Allegany, Garrett Dan’s Mountain State Park Maryland Dept. of Natural Resources Allegany Rocky Gap State Park Maryland Dept. of Natural Resources Allegany Big Run State Park Maryland Dept. of Natural Resources Garrett Casselman River Bridge State Park Maryland Dept. of Natural Resources Garrett Deep Creek Lake State Park Maryland Dept. of Natural Resources Garrett Deep Creek Lake Natural Resources Management Maryland Dept. of Natural Resources Garrett Area Herrington Manor State Park Maryland Dept. of Natural Resources Garrett New Germany State Park Maryland Dept. of Natural Resources Garrett Swallow Falls State Park Maryland Dept. of Natural Resources Garrett Youghiogheny River State Park Maryland Dept. of Natural Resources Garrett Green Ridge State Forest Maryland Dept. of Natural Resources Allegany Garrett State Forest Maryland Dept. of Natural Resources Garrett Potomac State Forest Maryland Dept. of Natural Resources Garrett Savage River State Forest Maryland Dept. of Natural Resources Garrett Bell Grove Wildlife Management Area Maryland Dept. of Natural Resources Allegany Billmeyer Wildlife Management Area Maryland Dept. of Natural Resources Allegany Dan’s Mountain Wildlife Management Area Maryland Dept. of Natural Resources Allegany Sideling Hill Wildlife Management Area Maryland Dept. of Natural Resources Allegany Warrior Mountain Wildlife Management Area Maryland Dept. of Natural Resources Allegany Mt. Nebo Wildlife Management Area Maryland Dept. of Natural Resources Garrett Deep Run Wildland Maryland Dept. of Natural Resources Allegany Maple Run Wildland Maryland Dept. of Natural Resources Allegany Potomac Bends Wildland Maryland Dept. of Natural Resources Allegany Rocky Gap Wildland Maryland Dept. of Natural Resources Allegany Bear Pen Wildland Maryland Dept. of Natural Resources Garrett High Rock Wildland Maryland Dept. of Natural Resources Garrett Middle Fork Wildland Maryland Dept. of Natural Resources Garrett Savage Mountain Wildland Maryland Dept. of Natural Resources Garrett Savage Ravines Wildland Maryland Dept. of Natural Resources Garrett South Savage Wildland Maryland Dept. of Natural Resources Garrett Youghiogheny State Wild River Maryland Dept. of Natural Resources Garrett Mountain Maryland Scenic Byway State Highway Administration Garrett Chesapeake and Ohio Canal Scenic Byway State Highway Administration Allegany 8-3

158. K.N. Eshle eman & A.J. Elmore (2013) E Chap 8 pterFigure 8-1 Cultural and historical reso 1. ources in Garre and Allegany Counties con ett ny nsist primarily of historical d districtsand registe ered historic bu uildings. Sceni byways, His ic storic National Road, C&O CCanal National Historical Park and k,major transsportation rout are also sho tes own. In the case of the Maryla Inventory of Historic Pro e and operties, polyggonsrepresentin properties sm ng maller than 1 acre have been replaced with a point symbo so that they a visible at th a ol are hisscale. Data provided by the Maryland Historical Trust a t H t.Figure 8-2 Public recrea 2. ational resource in Garrett an Allegany County are plen es nd ntiful and contribute to the eco onomicand cultura vitality of the region. With the exception of the Potoma Heritage Nat al ac tional Scenic T Trail, Mountain nMaryland Gateway to the West Heritag Area, and Wildlands (no da available), all resources li G e ge W ata isted in Table 8 are 8.1shown here (or in Figure 8-1). e 8-4

159. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandWith respect to state forest recreational areas, Pennsylvania DCNR Bureau of Forestry requires a300 ft setback from any state forest picnic area, trail, road of historic value, tree plantation,overlook, vista, fire tower site, or existing right of way; this setback also affords additionalprotection of public safety through conflict avoidance. The Bureau of Forestry also relies on localknowledge of cultural sites, recreational trails, trailheads, high use areas, viewsheds, vistas, andhigh aesthetic areas during the permitting process with a goal of avoiding areas of (or providingincreased setbacks from) concentrated recreational activity and developed recreational sites whenpermitting gas related infrastructure (PADCNR 2011). In the Pennsylvania state forests, animportant criterion in site selection for drilling pads is the degree to which locations can providenatural vegetative or topographic screening (PADCNR 2011). Additionally, API recommends thatsetbacks be increased to take into account prevailing winds and topography; in New York, sitesare assessed for their archeological importance (NYSDEC 2011).Figure 8-3. Many western Maryland recreational sites, such as this section of the C&O canal southeast ofCumberland, draw tourism from eastern portions of the state as well as throughout the Midwest.Knowledge of potential conflicts with cultural, historical and recreational resources is presumablyafforded through good communication between MSGD operators, local governments, and stateregulatory and management agencies. Certainly a first step is the identification of the location ofcultural and recreational resources potentially impacted by MSGD. However, this should befollowed up with an in-depth analysis of the ways in which the local and visiting population usesthese resources. API recommends that operators communicate with land owners and/or surfaceusers concerning activities planned for a particular site and provide information on the measures tobe taken for safety, protection of the environment, and minimization of impacts to surface uses.The goals of any interactions should be for transparency and increasing the flow of timely andrelevant information to surface owners, users, and other stakeholders. As recommended inPennsylvania (Ubinger et al. 2010), Maryland might consider developing a standardizedstakeholder process that could be implemented as part of comprehensive planning strategy; thegoal of such a process would be to engage stakeholders and the community in the most effectiveways possible, while allowing the permit review process to be expedited.C. Key recommendations8-A Applicants for drilling permits should be required to consult with Maryland Historical Trust during the planning and permit application process to identify all eligible or existing cultural or historical sites in the vicinity of proposed MSGD activity (including all drill pad sites, gas pipelines, roads, and transportation routes to and from MSGD facilities).8-B Regardless of whether or not a proposed operation would be located on state or federal land, best practice would require close consultation with local governments, state park and 8-5

160. K.N. Eshleman & A.J. Elmore (2013) Chapter 8 forest officials, national park managers, and wildlife managers who are familiar with the resources that could be impaired by shale gas development.8-C Applicants should be required to submit a visual resource mitigation plan as part of the permit application process based on site-specific assessment (i.e., viewshed analysis).8-D Site selection for drilling pads in Maryland should be locations that can provide natural vegetative or topographic screening.8-E Siting of well pads, or the routing of MSGD-related truck traffic, near high use recreation areas should be avoided if possible.8-F Maryland should impose a minimum 300 ft setback from all cultural and historical sites, state and federal parks, trails, wildlife management areas, natural areas, wildlands, scenic and wild rivers, and scenic byways to protect the region’s most important cultural, historical, recreational, and ecological resources. Setback considerations should include high use areas, noise and visual impacts, and public safety concerns.8-G The calculation of setback distances should consider prevailing winds, topography, and viewsheds, and repeatable formulas for calculating setbacks should be established.8-H Mitigative techniques, such as the use of visual screens, sound barriers, camouflage, and landscaping near cultural and historical sites, as well as restricting the times of gas development operations, should be required to minimize disturbances and conflicts with recreational activities in areas adjacent to gas development zones.8-I Any permitted shale gas development activities in the vicinity of public recreational sites—including state forests—should be timed so as to avoid periods of peak recreational activity (e.g., holiday weekends, first day of trout season, spring and fall hunting seasons, whitewater release dates, etc.). Maryland DNR should collect and provide data to help inform peak activity times.D. Literature citedMarcellus Shale Advisory Commission. 2011. [Pennsylvania] Governors Marcellus Shale Advisory Commission Report.NYSDEC. 2011. Revised Draft Supplemental Generic Environmental Impact Statement on the Oil, Gas, and Solution Mining Regulatory Program.PA DCNR. 2011. Guidelines for Administering Oil and Gas Activity on State Forest Lands. Commonwealth of Pennsylvania Department of Conservation and Natural Resources Bureau of Forestry.Ubinger, J. W., J. J. Walliser, C. Hall, and R. Oltmanns. 2010. Developing the Marcellus Shale: Environmental Policy and Planning Recommendations for the Development of the Marcellus Shale Play in Pennsylvania. Harrisburg, PA: Pennsylvania Environmental Council. 8-6

161. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland9. Protecting quality of life and aesthetic values1 The overall quality of life and aesthetic values in the two western Maryland counties derives inlarge measure from the mostly undeveloped rural mountainous landscape dominated by forestsinterspersed with agricultural lands and relatively small towns. The few cities in westernMaryland have changed relatively little over recent decades despite explosive population andexurban growth to other parts of the state. The relatively slow-paced way of life, minimalautomobile traffic, and associated amenities are attractive features of the area for residents andvisitors alike (Wainger and Price 2004, Chancellor et al. 2011). As discussed in Chapter 8, thequality of life is also significantly enhanced by the recreational opportunities afforded by theextensive state and federal public lands that exist in both Allegany and Garrett County. Hiking,biking, hunting, fishing, swimming, and boating are just some of the recreational activities thatare afforded through access to the state parks, forests, wildlife management areas, and wildlandsin the region (Boller et al. 2010). While providing economic benefits to the region, shale gasdevelopment in western Maryland clearly has the potential to negatively impact the area’squality of life and aesthetic character through altered land use, increased traffic (particularlyheavy truck traffic), noise pollution, visual and light pollution, and by creating conflicts withordinary community activities that would not exist in its absence. The purpose of this chapter isto provide recommendations of best practices that could significantly mitigate for these negativeimpacts. Note that many other practices that address quality of life and aesthetic issues (e.g.,constraints mapping for well siting; setback requirements for protecting cultural, historical, andrecreational resources, and public safety) were previously addressed in Chapters 1, 8, and 7,respectively.A. Hours of operationNew York State has proposed that shale gas development activities be conducted in a way thatavoids peak traffic hours, school bus hours, museum hours, community events, tourist periods,and overnight quiet periods (NYSDEC 2011). Similarly, as discussed in Chapter 8, PennsylvaniaDCNR Bureau of Forestry (PA DCNR 2011) mandates that any permitted shale gas developmentactivities in Pennsylvania state forests be timed to avoid periods of peak recreational activity(e.g., holiday weekends, first day of trout season, whitewater release dates, spring and fallhunting seasons, etc.). Considering the potential intensity of truck transport during the drillingand completion process, a comprehensive plan to protect the quality of life and aesthetic valuesin western Maryland should include multiple synergistic strategies to limit gas-developmentrelated disturbance. Similar to best practices proposed by New York State and employed byPennsylvania state forests, Maryland could restrict hours and times of operation to avoid orminimize the greatest conflicts, but this practice by itself is unlikely to be sufficient. Bestmanagement would employ thoughtful siting and visual screens, and rely on restrictions on hoursof operation to mediate the most disruptive activities (e.g., during well completion). As discussedin Chapter 1, practices that would be generally effective at reducing conflicts would be: (1) sitingwell pads away from populated areas (especially those with schools and other regularly-visited1 Chapter co-authors: Andrew J. Elmore, Ph.D., and Keith N. Eshleman, Ph.D. (both at: Appalachian Laboratory, University of Maryland Center for Environmental Science, Frostburg, MD 21532). 9-1

162. K.N. Eshleman & A.J. Elmore (2013) Chapter 9public buildings); (2) siting well pads and associated facilities in industrial parks designed andzoned for this type of industrial activity; (3) siting well pads in close proximity to majorinterstate highways and exit ramps designed to efficiently handle round-the-clock transportation;and (4) reducing truck traffic associated with water, chemical, and wastewater hauling (e.g.,through use of temporary pipelines). Used in combination with such siting criteria, restrictionson hours and times of operation (based on input from the public) would likely provide significantadditional mitigation of the most problematic conflicts.B. Noise controlMany studies illustrate a link between exposure to noise and negative effects on public health.Noise may severely impair quality of life (disrupt sleep, interfere with speech intelligibility), orpossibly give rise to both social and psychological problems (Bodin et al. 2008). Excessive noisealso has a broader environmental impact, for instance it can reduce optimal habitat area forcritical species or alter their behavior (Yong 2008) (also discussed in Chapter 5). Several statesand API provide specific best management practices to deal with issues of noise control.Colorado has established maximum permissible noise levels for oil and gas operations at wellsites and gas production facilities. In Colorado, operations involving a pipeline or gas facilityinstallation or maintenance, the use of a drilling rig, completion rig, workover rig, or wellstimulation are all subject to the maximum permissible noise levels for industrial zones. In thehours between 7:00 a.m. and the 7:00 p.m., the noise levels of different land uses surrounding anindustrial zone may be increased 10 db(A) for a period not to exceed 15 minutes in any one-hourperiod.2 New York also has established techniques for assessing, mitigating, and evaluating noiseimpacts and specific sound levels and characteristics of proposed or existing facilities (NYSDEC2011). API (API 2011) and these two states have also identified specific BMPs that can beemployed for mitigating noise impacts through: (1) careful siting of facilities—distance,direction, timing, and topography are the primary considerations in mitigating noise impactsfrom hydraulic fracturing and trucking operations (API 2011); (2) requirement for ambient noiselevel determination prior to operations; (3) placement of walls, artificial sound barriers, orevergreen buffers between sources and receptors (i.e., especially around well pads andcompressor stations) (API 2011); (4) use of noise reducing equipment (e.g., mufflers) on flares,drill rig engines, compressor motors, and other equipment (API 2011); and (5) use of electricmotors in place of diesel-powered equipment if feasible. We recommend that Maryland requireas part of the permitting process: (1) the enforcement of minimum distances between well padsand surrounding homes, businesses, and heavily-used recreational facilities to reduce noise asmuch as possible; (2) require ambient noise level determination prior to operations; (3)construction of artificial sound barriers where natural noise attenuation would be inadequate; (4)equipping all motors and engines with appropriate mufflers; and (5) requiring electric motors inplace of diesel-powered equipment for any operations within 3,000 ft of any occupied building.No drilling or compressor stations should be permitted within 1,000 ft of a residence.C. Road impacts and transportation planningAssessing the environmental impact of gas development activities should include an assessmentof the impact of vehicle traffic moving into, through, and out of sensitive areas via the existingroad network. Such an analysis must determine: (1) which segments of the network are2 COGCC Rule 802 Noise Abatement 9-2

163. Recomm mended Best Manageme Practice for Marcel Shale Ga Developm t ent es llus as ment in Mary rylandaccessibl to vehicular traffic; (2) when they are accessib and (3) t what class of vehicl le ble; to ses lesthey are accessible. The road netw a T work in Gar rrett and Alle egany Count consists o several ty ofprimary roads suppor by a den network of secondar roads (Fig r rted nse ry gure 9-1). Hoowever,limitation and barrie to vehicle traffic are many, includ ns ers e ding one-wa streets, sc ay cenic byways s,one-lane and restricte weight br ed ridges, gates, speed bum low-pass bridges, rai mps, s ilroad crossi ings,etc. In a formal traffi network an f ic nalysis, thes features ar used to co se re onstrain the f flow of traffficresulting in the determmination of a least-cost path betwee any two lo en ocations. The advantages ofsuch an approach are that it can incorporate multiple type of inform a e i m es mation, gather from sta red ate,county, and local sou a urces, and it synthesizes the many co onstraints tha should be considered in atattempts to reduce th severity an number of conflicts b he nd o between gas- -developmen related tra nt afficand existting uses of roads. Consi r istent with pr ractices reco ommended b API and t states of by theColorado and New York, all perm applican should de o Y mit nts evelop and su ubmit a detaailedtransporttation plan fo approval by the regula or b atory authori prior to c rity conducting a site anydevelopm ment, drilling well work g, kover, or wel completion activities; the approval process sho ll n l ouldalso allow for adequa comment by the publ state tran w ate t lic, nsportation a agencies, and county roa d adsdepartme to specif ents fically identi potential road use co ify l onflicts or iss sues unbekn nownst to the eapplicant or the prim t mary state reggulatory agen (since al road issues are really l ncy ll local issues).Figure 9- The road network of Al -1. n llegany and Garrett County is relatively sparse. Beca G y y ause roads are eoften conffined to areas with reduced topographic slope, many stream valley contain roa This, s d c y ys ads.combined with other re d estrictions meentioned in th text, sugges that increas truck traff on this roa he st sed ffic adnetwork will have mult w tiple impacts on the quality of life and a y aesthetic valu in the regi ues ion. Road dattawere acquuired from the Maryland St Highway Administrati (SHA). e tate y ion uired transportation plan would be a detailed, comThe requ d mprehensive document i e includingnetwork analysis and plans for ot d ther transpor rtation-relate activities. Such a plan would be ed . n 9-3

164. K.N. Eshleman & A.J. Elmore (2013) Chapter 9designed to achieve the following: (1) maximize efficient driving; (2) route vehicles onto roadsand across bridges specifically designed to carry heavy truck loads on a repetitive basis; (3)ensure public safety; (4) avoid peak and sensitive traffic hours; (5) ensure that all trucks are DOTcompliant; (6) coordinate with highway departments and emergency responders; (7)upgrade/improve roads as needed; (8) inform the public of any necessary detours; (9) utilizeflowlines to reduce truck traffic (if feasible and cost effective); and (10) assure adequate off-roadparking at well site and delivery areas (API 2011). Further, as proposed in New York, anydeviation from the plan, detours or closures, must be done with advanced public notice. Road-use agreements should also be established between operators and municipalities to ensure publicsafety and provide a mechanism for addressing road damages attributable to shale gasdevelopment in a timely way.Existing roads should be utilized wherever feasible; if new roads are needed, however, potentialimpacts should be considered along with landowner recommendations, consideration forhistorical and cultural resources, and a mitigation strategy to prevent erosion and protectenvironmentally-sensitive areas. Both API (API 2009) and Pennsylvania DCNR (PADCNR2011) provide specific recommendations for the design and construction of new roads in rurallandscapes. While roads should be designed and constructed in ways appropriate for theirintended use, it is recommended that construction crews consider using the PA DNCR forconstruction of permanent non-paved roads to address potential environmental impacts, controlerosion, and avoid damage to environmentally sensitive areas (PADCNR 2011).D. Visual pollution/viewshedsAs discussed in Chapter 8 in thecontext of protecting cultural,historical and recreational resources,there are two types of mitigativetechniques that are appropriate foraddressing visual pollution andminimizing degradation of visuallysensitive resources in general. Thefirst type of technique involves theuse of viewshed analysis to helpcarefully site well pads andassociated infrastructure at locationsthat are least visible from heavilyused roads, overlooks, or publicrecreational facilities. The second Figure 9-2: Drill rigs must be lit at night to facilitate 24-hr operations,type of mitigation involves the use of however, during the production phase artificial lighting might be eliminated.visual screens, camouflages, paintschemes, evergreen buffers, andlandscaping techniques to obscure drilling equipment and shale gas development activities fromview as much as possible. We recommend use of both types of mitigative techniques to minimizedegradation of western Maryland viewsheds by shale gas development activities as much aspossible. It should be emphasized that because well drilling and completion operations thatemploy large amounts of heavy (and, in some cases, three stories tall) equipment on-site are 9-4

165. Recommended Best Management Practices for Marcellus Shale Gas Development in Marylandtemporary in nature, the most severe degradation of visually-sensitive resources occurs duringperiods of maximum development activity. We believe that removal of major equipment alonewould in some cases contribute significantly to restoring these natural viewsheds. In other cases,careful land reclamation practices (e.g., revegetation of well pads, planting of evergreen screensaround permanent gas infrastructure, etc.) would provide additional benefits (see Chapter 4).Light pollution has the added potential (above and beyond general impacts to viewsheds) to posesignificant direct and indirect effects on the quality of life and aesthetic values in westernMaryland (Figure 9-2). Indirect effects of light pollution were covered in Chapter 5, and take theform of the different ways in which artificial lighting can influence wildlife and biologicaldiversity more generally. Artificial lighting causes direct effects on the quality of life andaesthetic values by being a distraction while driving on primary and secondary roads, obscuringdark night skies, and reducing the rural aesthetic qualities of the region. Many visitors to westernMaryland frequent campgrounds and other state facilities, and expect dark night skies as part oftheir experience. Maryland could put an emphasis on preserving these conditions. Similar towhat was discussed in Chapter 5, the primary BMPs for reducing the impact of artificial lightaim to reduce the amount of lighting used, keep lights low and directed down on the work site asmuch as possible, and increase the use of low-pressure sodium lights relative to other types oflighting. Most polluting are lamps with a strong blue emission, like metal halide and white light-emitting diodes (LEDs) (Falchi et al. 2011). Following these guidelines, Maryland could takesteps to reduce the amount and nature (color) of artificial lighting used during MSGD. It shouldalso be noted that the light required at different stages of MSGD can vary substantially; whilehigh light levels might be required during drilling and well completion, during productionartificial lighting could be reduced or eliminated altogether (after addressing security concerns.)E. Key recommendations9-A Well-pad siting should consider the multiple factors that influence the quality of life and aesthetics of rural life in western Maryland (e.g., location of existing infrastructure, traffic loads on existing roads, etc.) 9-A.1 Site well pads away from occupied buildings (e.g., dwellings, churches, businesses, schools, hospitals, and recreational facilities) 9-A.2 Site well pads and associated facilities in industrial parks (either new or existing) designed and zoned for this type of industrial activity 9-A.3 Site well pads in close proximity to major interstate highways and exit ramps designed to efficiently handle round-the-clock transportation 9-A.4 Reduce truck traffic associated with water hauling through use of temporary pipelines where possible.9-B Each of the counties in western Maryland should revisit noise regulations and enforcement policies and confirm they are appropriate for this industrial activity.9-C No drilling or compressor stations should be permitted within 1,000 ft of an occupied building.9-D Require electric motors (in place of diesel-powered equipment) for any operations within 3,000 ft. of any occupied building 9-5

166. K.N. Eshleman & A.J. Elmore (2013) Chapter 9 9-D.1 Encourage electric motors in place of diesel-powered equipment wherever possible. 9-D.2 Restrict hours and times of operation to avoid or minimize the greatest conflicts between the public and MSGD. 9-D.3 Require ambient noise level determination prior to operations. 9-D.4 Require construction of artificial sound barriers where natural noise attenuation would be inadequate. 9-D.5 Equip all motors and engines with appropriate mufflers.9-E All permit applicants should develop and submit a detailed transportation plan for approval by the regulatory authority prior to conducting any site development, drilling, well work over, or well completion activities 9-E.1 The approval process for the transportation plan should allow for adequate comment by the public, state transportation agencies, and county roads departments.9-F It is recommended that new road construction follows PADCNR guidelines for construction of permanent non-paved roads to address potential environmental impacts, offset erosion, and avoid damage to environmentally sensitive areas.9-G We recommend the use of viewshed analysis to help determine the best location for MSGD-related infrastructure as well as to determine what mitigative techniques would be appropriate.9-H We recommend use of mitigative techniques (e.g., the use of visual screens, camouflages, paint schemes, evergreen buffers, and landscaping techniques) to minimize degradation of western Maryland viewsheds by MSGD.F. Literature citedAPI. 2009. Environmental Protection for Onshore Oil and Gas Production Operations and Leases (API RP-51R, First Edition). Washington, DC: American Petroleum Institute Publishing Services.API. 2011. Practices for mitigating surface impacts associated with hydraulic fracturing (API Guidance Document HF3, First Edition). Washington, DC: American Petroleum Institute Publishing Services.Bodin, T., M. Albin, J. Ardo, E. Stroh, P. Ostergren, and J. Bjork. 2008. Road traffic noise and hypertension: results from a cross-sectional public health survey in southern Sweden. New Scientist 197:33-35.Boller, F., M. Hunziker, M. Conedera, H. Elsasser, and P. Krebs. 2010. Fascinating Remoteness: The Dilemma of Hiking Tourism Development in Peripheral Mountain Areas Results of a Case Study in Southern Switzerland. Mountain Research and Development 30:320-331.Chancellor, C., C. P. S. Yu, and S. T. Cole. 2011. Exploring Quality of Life Perceptions in Rural Midwestern (USA) Communities: an Application of the Core-Periphery Concept in a Tourism Development Context. International Journal of Tourism Research 13:496-507.NYSDEC. 2011. Revised Draft Supplemental Generic Environmental Impact Statement on the Oil, Gas, and Solution Mining Regulatory Program. 9-6

167. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandPADCNR. 2011. Guidelines for Administering Oil and Gas Activity on State Forest Lands. Commonwealth of Pennsylvania Department of Conservation and Natural Resources Bureau of Forestry.Wainger, L. A., and E. W. Price. 2004. Evaluating quality of life, economic vulnerabilities, and drivers of ecosystem change. Environmental Monitoring and Assessment 94:69-84.Yong, E. 2008. City songbirds change their tune. New Scientist 197:33-35. 9-7

168. Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland10. Protecting agriculture and grazing1After forested land, agricultural land is the second largest contributing land cover in Allegany andGarrett Counties on an areal basis, covering 15.6% (108,420 acres) of these two counties. In 2007,the most recent year for which data are published, it is estimated that Allegany County had 302farms covering a land area of 36,343 acres2, while Garrett County had 677 farms covering 95,514acres.3 The value of all agricultural product sales in the two counties was estimated as $3.16M and$27.73M, respectively. Farms in these two counties are typically small family operations, with theaverage farm covering 121 acres in Allegany and 141 acres in Garrett County4. While many typesof county-level data from the National Agricultural Statistics Service are not published to protecteconomically sensitive information, major crops produced in these two counties include corn forgrain, corn for silage, soybeans, winter wheat, other hay, barley, and vegetables, in addition tomilk, cattle and calves, sheep, hogs, and poultry. In addition to the economic value of crops andother agricultural production, agriculture also contributes aesthetically to the quality of life andcultural fabric of rural western Maryland as discussed in Chapter 9. While important in this regard,it must be noted that agricultural production in both western Maryland counties is dwarfed byproduction in many other counties in the state. For example, Carroll County which has about142,000 acres in farmland as of 2007 (roughly the same acreage as Allegany and Garrett Countiescombined) had agricultural product sales that were about a factor of three greater ($87.4M).5 Thisdifference is consistent with the generally low fertility soils and cool climate of western Maryland.Nonetheless, as an important economic activity in the region and as a component of Maryland’sgeneral effort to maintain farming activities throughout the state, contributing to overall quality oflife, we believe that there are some best practices that can protect both cropland and grazing landfrom negative impacts of shale gas development.One recent study of farm animals in six states (Pennsylvania, New York, Ohio, Colorado, Texas,Louisiana) suggested increased mortality rates in livestock and companion animals (i.e., dogs andcats) living close to active gas-drilling operations (Oswald and Bamberger 2012), with severalcaveats associated with the lack of controls due to the case study aspect of the survey (Thompson2012). Although chemicals can be volatized (e.g., by impoundment aerators) and misted into theair creating an inhalation exposure pathway, the most common source of toxicity exposure waslikely via contaminated water. Pathways of exposure included, for example, spills of hydraulicfracturing fluids, tears in the liners of wastewater impoundments (which we do not recommend buthave been used in PA), and spreading of wastewater on roads to reduced dust and ice followed byanimals licking their paws after crossing the roads (again, the spreading of wastewater on roads isnot recommended for Maryland). Health impacts ranged from neurological to sudden death withthe most common effects being reproductive. Animals affected include cows, horses, goats,1 Chapter co-authors: Keith N. Eshleman, Ph.D. and Andrew J. Elmore, Ph.D. (both at: Appalachian Laboratory, University of Maryland Center for Environmental Science, Frostburg, MD 21532).2 Allegany County 2010 Agricultural Profile, Maryland Department of Agriculture. http://www.mda.state.md.us/on_web/ag_links/countyag.php#allegany.3 Garrett County 2010 Agricultural Profile, Maryland Department of Agriculture. http://www.mda.state.md.us/on_web/ag_links/countyag.php#garrett.4 The difference between agricultural land area and farm area in these counties is likely due to the presence of woodlots on many of these farms. In other words, some of the farm area is actually mapped as forest cover.5 Carroll County 2010 Agricultural Profile, Maryland Department of Agriculture. http://www.mda.state.md.us/on_web/ag_links/countyag.php#garrett. 10-1

169. K.N. Eshleman & A.J. Elmore Chapter 10llamas, chickens, dogs, cats, and koi. Because the movement of farm animals is confined they mayexperience higher cumulative exposure than wildlife with less restricted mobility. However,photographic evidence has been reported of dead and dying songbirds, deer, frogs, andsalamanders (Oswald and Bamberger 2012).A. Protection of prime farmlandPrime farmland is an official designation used by U.S. Department of Agriculture to define landthat has the best combination of physical and chemical characteristics for producing importantagricultural crops. Prime soils have the following inherent characteristics: a minimum amount ofsurface rocks, low susceptibility to erosion and have not had been excessively eroded in the past, afavorable pH, an acceptable level of content of salt and sodium, water and air permeability, andare not subject to prolonged saturation. They also have the following related qualities: have nearlylevel to gently sloping topography, and rarely or never flood during the growing season.6 Fewerthan 2% of western Maryland soils are considered prime soils7, and 71% are considered class VIor VII, designating them as suitable for planting of permanent pasture, trees, or reserved forwildlife management and recreation8. As recommended by Lien and Manner (2010), we agree thatsoil conditions at sites being considered for shale gas development be evaluated as part of theplanning process; prime agricultural soils and prime farmland should generally not be disturbedfor well pad siting, road construction, or any ancillary gas development activities. Further, highlyerodible soils should also be identified as part of the planning process and appropriate bestpractices should be employed to prevent erosion and sedimentation problems in developing theseareas (see Chapter 4).Some agricultural lands in western Maryland are already protected to some extent by the MarylandAgricultural Land Preservation Foundation (MALPF). MALPF—which protects agricultural landin Maryland through the use of perpetual easements—was created by the Maryland GeneralAssembly in 1977 and is housed within the Maryland Department of Agriculture (MDA). Prior to2007, applications for easements were only accepted from landowners in designated AgriculturalPreservation Districts. Easements may be donated or purchased with a goal of providing for theperpetual production of local food and fiber. Agricultural land easements in western Maryland aredisplayed in Figure 10-1 and comprise about 4.3% of the agricultural land in the two counties.The text in the current standard deed of easement found at the MALPF website reads as follows:“No rights-of-way, easements, oil, gas or mineral leases, or other similar servitude may beconveyed, or permitted to be established on the land for any commercial, industrial or residentialuse, without the Grantees express written permission.”9 Thus, any surface uses of the land forshale gas development without the Grantee’s (i.e., state of Maryland’s) written permission wouldappear to expressly violate the protective status granted under MALPF.6 http://www.md.nrcs.usda.gov/technical/nritext.html.7 Ibid.8 Land Capability Classification. Agriculture Handbook No. 210. Soil Conservation Service. U.S. Department of Agriculture. Issued September 1961. Reprinted January, 19739 Current Standard Deed of Easement, para. II.A.2, pp. 3-4, Maryland Agricultural Land Preservation Foundation; http://www.malpf.info/laws.html 10-2

170. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandFigure 10-1. Agricultural land is generally confined to floodplains (Allegany County) and the Appalachian Plateau(Garrett County). Maryland Agricultural Land Preservation Foundation (MALPF) agricultural easements (sometimeslocated within agricultural districts) protect land from non-agricultural uses and are administered by MarylandDepartment of Agriculture. Rural Legacy Area Properties have multiple goals, including to protect economies basedon farming and forestry, and are administered by the Department of Natural Resources.A second mechanism through which agricultural land has been protected is Maryland’s RuralLegacy Program, which was enacted by the General Assembly in 1997, and has dedicated over$210 million to preserve 68,675 acres of valuable farmland, forests, and natural areas throughoutthe state10. In western Maryland, 62,747 acres have been identified as Rural Legacy Areas, asubset of this area has been protected through conservation easements (Figure 10-1). The RuralLegacy Programs goals are to establish greenbelts of forests and farms around rural communitiesto preserve their cultural heritage and sense of place, and critical habitat for native plant andwildlife species. Relevant to prime farmland, the Rural Legacy Program also aims to supportnatural resource economies such as farming, forestry, tourism and outdoor recreation. Similar toMALPF conservation easements, subsurface activities on Rural Legacy properties are prohibitedwithout the Grantees’ approval and require the Grantee to consider whether the impact would bedestructive of the conservation attributes the easements were designed to protect. In GarretCounty, the Bear Creek Rural Legacy Area overlays the Accident gas storage dome, and wasestablished by the Maryland Department of Natural Resources in collaboration with GarrettCounty to protect farms with severed or leased mineral rights with the understanding that gasstorage activities would continue to take place on lands encumbered with Rural Legacy easements,provided that such storage activities do not unduly compromise the natural and working resourcesthe Area were established to protect. In Allegany County, the Mountain Ridge Rural Legacy Areais delineated around 10,163 acres of existing protected lands that may be further connected and10 http://www.dnr.state.md.us/land/rurallegacy/12thAnniversary.asp 10-3

171. K.N. Eshleman & A.J. Elmore Chapter 10consolidated, forming a greenway potentially linking ridgetops in West Virginia withPennsylvania, as well as westward into the Allegheny Plateau.Protected lands throughout western Maryland have clearly been established with consideration ofagriculture and prime soils as an objective. Therefore, restricting MSGD so as to preserve primesoils and agricultural lands could be achieved by enforcing MALPF and Rural Legacy Areaeasements throughout western Maryland. From our reading of the MALPF and Rural LegacyProgram websites, it appears this would just require that Maryland not approve any MSGD inthese areas. With respect to other non-protected agricultural lands where shale gas developmentmight be permitted, some of the best practices proposed by New York State would provide anappropriate level of protection, of agriculture and grazing, namely (NYSDEC 2011): • Well pads, infrastructure, roads, and utility corridors should generally be sited along field edges, thus avoiding bisection of fields. • Topsoil should be stockpiled during site development activities, covered during storage, and redistributed back onto agricultural land as part of the land reclamation process. • In active agricultural areas, operators must: (1) keep drill cuttings and topsoil separate; (2) remove any drilling muds from fields; (3) avoid soil compaction; and (4) fence in active pasture areas (alternately fence livestock out of gas development areas).B. Key recommendations10-A Soil conditions at sites being considered for shale gas development should be evaluated as part of the planning process.10-B Prime agricultural soils and prime farmland protected by Maryland’s existing land easement programs should not be disturbed for well pad siting, road construction, or any ancillary gas development activities.10-C Highly erodible soils should also be identified as part of the planning process and appropriate best practices employed to prevent erosion and sedimentation problems in developing these areas (see Chapter 4).10-D Well pads, infrastructure, roads, and utility corridors should generally be sited along field edges, thus avoiding bisection of fields.10-E Topsoil should be stockpiled during site development activities, covered during storage, redistributed back onto agricultural land as part of the land reclamation process, and soil compaction should be avoided at all times.10-F Operators must fence livestock out of gas development areas.C. Literature citedLien, A. M., and W. J. Manner. 2010. The Marcellus shale: resources for stakeholders in the Upper Delaware watershed region. Pinchot Institute for Conservation.NYSDEC. 2011. Revised draft supplemental generic environmental impact statement on the oil, gas, and solution mining regulatory program. Albany, NY.Oswald, R., and M. Bamberger. 2012. Impacts of gas drilling on human and animal health. New Solutions 22:51-77. 10-4

172. Recommended Best Management Practices for Marcellus Shale Gas Development in MarylandThompson, H. 2012. Fracking boom spurs environmental audit. Nature 485:556-557. 10-5

SlideShare for iOS

Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland

by Marcellus Drilling News

on Mar 12, 2013

Statistics

Views

Actions

1 Embed 78

Accessibility

Categories

Upload Details

Usage Rights

Report content

Recommended Best Management Practices for Marcellus Shale Gas Development in Maryland Document Transcript