Hosted by the National Climatic Data Center
This page last updated 29 October 2003

Background:

The Drought Monitor Forum discussed user, administrative, and technical issues related to the weekly U.S. Drought Monitor (USDM) product. The Drought Monitor consists of a map depicting areas of drought and an associated narrative. The latest edition can be viewed at the following web site: http://www.drought.unl.edu/dm/monitor.html

To accomplish these goals, the October workshop agenda was divided into three parts:

1. Wednesday, October 1, was devoted to a discussion of defining and monitoring drought in an arid/semiarid environment.
2. Thursday, October 2, focused on USDM user discussions and input and began addressing USDM technical issues.
3. Friday, October 3, finished discussing technical issues and covered USDM administrative issues.

The workshop presentations, decisions, and recommendations are summarized below. They consist of reports written by workshop organizer/USDM author Richard Heim and authors Michael Hayes and Mark Svoboda.

Workshop Summary and Recommendations:

• Defining and Monitoring Drought in an Arid/Semiarid Environment - Heim
  
• USDM Forum Summary - Heim
  
• Workshop Summary - Hayes
  
• Workshop Summary - Svoboda

The following are key points made by the speakers, at the Cedar City workshop, and workshop conclusions regarding defining and monitoring drought in an arid/semiarid environment such as the western U.S.:

1. How to define drought in a desert (arid/semiarid environment):
   1. it is basically a supply and demand issue (water balance concept is central to drought). If you have enough moisture to meet demand, it doesn't matter how much it normally rains. Insufficient water to meet needs equals drought.
   2. drought is defined by its impacts: no impacts equals no drought regardless of the weather that is occurring.

2. Useful monitoring tools include:
   1. precipitation departures during key seasons
   2. energy release components
   3. monitoring wells
   4. streamflow
   5. reservoir levels (levels of Clark Canyon Reservoir, in Montana, were the lowest ever observed this summer)
   6. SWSI
   7. SPI
   8. PDSI
   9. soil moisture (this is being addressed in Montana through the deployment of the Montana Counties Soil Climate Network)
   10. snow water equivalent
   11. crop water use (varies with crop)

3. An SPI-type map for streamflow would be helpful. However, the period of record varies by streamflow station, thus making the resulting percentiles difficult to compare between stations.

4. Some state agencies use multiple maps for drought monitoring. Three maps are used in Texas and soon to be used in Arizona (a climatological/meteorological index, an agricultural index, and a water availability/hydrological index). Two are used in New Mexico (a meteorological index and a hydrological index). Four are used in Alabama (Lawn and Garden Moisture Index, Weekly Crop Moisture Index, River Flow Index, Ground Water Index).

5. During the Wednesday afternoon session, Mark Svoboda summarized important points raised by the day's speakers. For drought in an arid/semiarid region:
   1. the SWSI is a very important tool, as it integrates several factors relevant to the West. The USDM authors would like to see the SWSI issued westwide on an operational basis for all 12 months.
   2. timing of precipitation and elevation are important
   3. a paleoclimatic perspective gives us a better idea of drought recurrence frequencies and relative severity of current droughts
   4. GAPS - station locations and parameters needed
   5. better information is needed on rangeland and livestock conditions
   6. model-based indices are a potential tool
   7. groundwater is an important indicator that could be monitored better. We should locate baseline stations where there are no human drawdown impacts.
   8. focus on the impacts of the drought, specifically the socioeconomic impacts, not just the meteorological indicators. We should look at the autocorrelation between the indicators (what the meteorologists have access to) and the impacts (what affects peoples' lives).
   9. the longer timeframes are more important than the shorter timeframes
   10. the management/mismanagement of water has important impacts on the user. In this respect, we need to determine the difference between what is a drought and what is a water supply problem.
   11. it's very important to know who we are making the product for (who are the customers)
   12. need a dry spell indicator as opposed to a drought indicator (actually, the USDM already makes this distinction with D0 for dry spell vs. D1-D4 for drought)

6. In his "Assessment of Drought Assessment", Kelly Redmond noted:
   1. data on the water demand side of the equation are not as well developed; impacts have shorter histories; more information is needed on impacts at the state and local levels
   2. not all aspects of drought are adequately depicted on maps, i.e., long distance (spatial) relationships and lagged (time delay) relationships
   3. there is uneven input from the state level
   4. drought monitoring efforts are almost completely unfunded; funding is needed

The following are discussion points and decisions made, at the Cedar City workshop, concerning U.S. Drought Monitor (USDM) technical, administrative, and user feedback issues:

1. the short-term and long-term aspects of drought need to be better delineated on the USDM map

2. the USDM is a weekly product. In the West, drought is largely a hydrological phenomenon. Wait a week to see if the effects/impacts of a precipitation event are significant enough to warrant changing the map depiction (don't make major changes on the USDM map right away, or at most make just a one category change). These precipitation events could be addressed in the narrative immediately, and later changed due to prior events can also be addressed in the subsequent narratives.

3. the one map vs. two map issue was discussed by several participants. A weekly map, plus a monthly hydrological drought map, was considered to be the ideal arrangement. Mark Svoboda presented a talk for a group led by Harry Lins (the partnership consisting of NDMC-Mark, USGS-Harry, NOAA- Frank Richards, USDA-Phil Pasteris) which addresses this issue. The group proposes that this partnership prepare a hydro-oriented companion map product to the USDM. This monthly Water Resources Monitor (USWRM) would complement -- not replace -- the weekly USDM. The USWRM would be a consolidation of indices/indicators onto one comprehensive national drought map by basin or hydrological unit and would be based on the following observational data: i) primary: reservoirs (most important), surface water, ground water
   ii) secondary: snowpack (most important), soil moisture, water supply forecasts The Army Corps of Engineers and Bureau of Reclamation are potential partners who could help with reservoir data. The USWRM would be supported by a separate web site and open list server (like the USDM) through which value added information, subjective assessment, and interpretation of the objective factors are provided by local field experts. Mark identified issues of concern, which include:
   1. resources
   2. who does the work
   3. lack of adequate current and historical observations
   4. better approach (use PRISM) to account for elevation
   5. address how significant rains affect short vs. long-term recovery
   The next step is to create a prototype product and web site by January 2004. The workshop participants are eagerly looking forward to this new product.

4. additional useful potential tools include:
   1. 14-day precipitation
   2. estimated (crop) water use
   3. a differential map (water use minus precipitation)
   4. a satellite-based percent of average seasonal greenness index (discussed by Jess Brown)

5. have the USDM weekly analysis period go through 0800 EST Monday instead of 12Z Tuesday

6. for areas that have water restrictions or governor-declared drought emergencies, keep them in D0(H) on the USDM map until the restrictions/declarations are lifted. But there is the danger of a feedback loop if we use local water restrictions/drought declarations as input: localities may continue their restrictions because we say they are in a drought category on the USDM, but we are showing they are in a drought because they have drought restrictions.

7. suggestions for Impacts Labeling on the USDM:
   1. use a priority tag on the impacts (Aw, aW, AW), where upper case denoted primary impact and lower case denotes secondary impact
   2. what are Agricultural Impacts (A) in the West? They are soil moisture or short-term conditions. April-May-June are the important months for Wyoming and Colorado rangeland.
   3. instead of A, use V (for all Vegetation) for the short-term impacts. Or say, "A stands for vegetation (non-irrigated crops and natural vegetation)", and "H stands for hydrological (water) and irrigated crops".
   4. use cross hatching for the A and H impacts

8. issue the USDM on a bi-weekly or monthly schedule during the winter

9. I (Richard Heim) noted, to the western local experts, that the SWSI is a very useful tool that we could utilize more readily if it were available plotted as a map instead of published as a table

10. the LDAS Model Soil Moisture product described by Andy Wood may be a useful input indicator for the USDM

11. consider bringing in somebody from way outside the USDM user community (e.g., a sociologist) to help guide us on where to go regarding human impacts in the West

12. as the USDM moves into the ArcGIS environment, it would be helpful if impacts (as well as all of the input indicators) information could be provided to the USDM authors in a gridded or point format, operationally, on a weekly (or at least monthly) basis, nationwide, in a format that could be objectively integrated (i.e., by computer) into the USDM. The useful impacts include:
    1. counties/localities having drought declarations/emergencies/restrictions
    2. USDA topsoil short/very short
    3. USDA pastures/rangeland short/very short
    4. status of crops

13. Jay Breidenbach identified USDM strengths as:
    1. concisely conveys a summation of everything we know about drought in one picture
    2. rates drought intensity on a familiar scale (0-4)
    3. easy to understand
    He categorized the Idaho USDM customers as:
    1. government (federal, state, local)
    2. hydroelectric companies
    3. news media
    4. river recreational interests
    5. ordinary citizens
    The bulk of these customers are interested in longer-term drought as indicated by streamflow, reservoir storage, and mountain snowpack. Short-term dryness would be of interest to such customers as ordinary citizens (lawns, trees) and wildfire managers. Jay summarized the following points concerning USDM Idaho customer feedback:
    1. prefer county overlays on the USDM map
    2. it is important to include local impacts without being influenced by politics
    3. it's okay to have some subjectivity in producing the product, but it should not be swayed by politics
    4. most customers like the USDM
    5. most like the archive and animation features on the NDMC web page
    6. the exploder list server feedback is working well
    7. want a more detailed comparison of the current drought to past droughts (e.g., time series of the Dx values by division would be nice)

14. Joe Abraham identified two classes of Arizona USDM stakeholders:
    1. water, fire, land managers (state and federal)
    2. tourism, energy, cooperative extension, ranching, media
    The general consensus of the stakeholders:
    1. prefer location information on the map (cities, counties, rivers, etc.)
    2. the drought depiction detail is generally just right
    3. the USDM information is useful
    4. two separate maps would be better
    5. the county or climate division scale is most adequate
    6. prefer interactive data layers (ArcIMS, or access to the ArcGIS shape files) to make their own maps

15. there was a feeling that the USDM is plagued by a "horse race" attitude of daily/weekly drought monitoring

16. in his Thursday morning discussion session, Mark Svoboda addressed the "What's In It For Me?" issue. Some drought stakeholders may not participate in the USDM process because they don't see any advantage to it. A long-range goal for the USDM is to serve as a portal to a suite of drought monitoring products which address drought on a finer scale. This would work by, on the USDM map, clicking on a state and that action leads to a state/local web page where the stakeholder can derive some attention and credit. With the USDM author team partnering with state/local offices, the fine scale (i.e., county level) detail can be addressed. This would be a good item to bring up with the WGA/NOAA NIDIS group. In the meantime, we could put on the NDMC USDM web page a skeletal map with the states highlighted, where clicking on a highlighted state will bring the user to that state's drought page(s).

17. if we want to depict county-level drought conditions on the USDM, then we need station-based instead of climate division- based input

18. at last year's USDM Forum, the discrepancy between CPC's weekly PDI and NCDC's monthly PDI products was discussed (this discrepancy is due to the different time steps). The solution was to have NCDC compute the monthly PDI on a weekly basis using weekly updates to the current month data provided by CPC. Progress has not been made on this due to CPC resource issues, but Rich Tinker explained that he should be getting prototype CPC updates to NCDC within the next one to two months.

19. Phil Pasteris offered to provide the USDM authors with shape files that give full annotation of the western U.S. river basins (HUCs). Harlan Shannon can provide updated instructions on how to utilize this information in the production of the USDM map. We will need to get shape files of the eastern U.S. HUCs from Harry Lins or Frank Richards.

20. we should send to the drought exploder list server a difference map of the USDM categories from last week's final map to this week's draft maps. This is something Harlan Shannon can document for the authors in the instruction manual.

21. It is unavoidable that some politics enters into the mix during the exploder list discussions and local expert feedback process. The influence of politics should be minimized. We should use as much objective and quantitative input as possible.

22. Socioeconomic Impacts:
    1. socioeconomic drought impacts were considered to be an important factor that we need to do a better job incorporating into the USDM. One suggestion was to focus on them in the narrative or as a deeper (ArcGIS) data level. Another suggestion was to use NDMC's U.S. Drought Impact Map to depict the impacts for the USDM.
    2. in the Possible Impacts column on the Drought Severity Classification Table, we should attempt to quantify and test these descriptions (re crop losses, etc.). This is a good item to bring before the WGA/NOAA NIDIS group.

23. another unfinished item from last year's USDM Forum was the discrepancy in statistical distributions used for computing the SPI. Simple lack of resources has delayed WRCC's conversion to using the Pearson III distribution. Mark Svoboda said the Nebraska group can also convert to the Pearson III for the station-based weekly SPI.

24. as far as the Dx boundaries having a 50 km thick "fuzzy border", this is something ArcGIS can do with buffer, but we should keep this on the back burner

25. NCDC will not organize and run the next USDM Forum workshop. There was consensus among the authors that the JAWF/CPC folks will hold the next Forum in the Washington, D.C.-Rockville area, but that we can probably get by with just biennial Forums instead of holding them every year. Since Ted O'Brien was also at the workshop, he offered to hold a future North America Drought Monitor and Climate Extremes (NADM) workshop in Canada. There was discussion of alternating the USDM and NADM workshops, with the following tentative schedule: 2004 - NADM workshop in Canada
    2005 - USDM workshop at JAWF/CPC
    2006 - next NADM workshop
    2007 - next USDM workshop I (Richard Heim) said I wanted to run this idea past NCDC management to get their feedback before commenting, but it sounds reasonable to me.

26. Jan Curtis is very willing to provide us with Wyoming data -- digital values in an ArcGIS importable format (SWSI, basin shape files, etc.) and daily data from other (non-COOP) stations

27. having the USDM show the worst of the short- and long-term conditions, instead of the average of them, is the more logical/reasonable approach

28. Advanced Feedback:
    1. the workshop participants like the idea of the author putting out on Friday a "Looking Ahead Thoughts" email
    2. also favorably received was the idea of the local experts sending feedback a week or two ahead of time on gradual changes they'd like to see, especially relevant in the West

29. Authoring:
    1. it was agreed that someone from the West should be added as an author. Potential candidates include USGS, the RCCs (e.g., WRCC or MRCC-Mike Palecki/Jim Angel), the CDC in Colorado, the NRCS in Portland, OR. An NRCS person was desired, but the Portland office work schedule would allow authoring during only July-December. Phil Pasteris asked that we revisit this in half a year.
    2. Ted O'Brien suggested that an authoring team consist of three people each week
    3. another idea was two authors each week, where one would do the eastern U.S., the other the western U.S., and the two sub-maps then be integrated into the USDM
    4. Barbara Mayes suggested NWS/CSD could provide operational people as USDM authors
    5. IRI was also suggested as a potential source for authors (talk to Tony Barnston or Chet Ropelewski)
    6. UCAR/NCAR was also suggested as a possible source for authors

30. Phil Pasteris noted that NRCS would like to focus in the next six months or so on getting the westwide SWSI and the Basin Water Index input indicators up

31. Written Agreement/Guidelines:
    1. there was no consensus about whether an MOU/MOA between the authors/agencies was required, nor whether a set of written guidelines was required
    2. but there was agreement that, if written guidelines were to be prepared, they be recommendations, not dictations, to the authors on what they can and can't do. Such a document could serve as a white paper to justify getting funding support (possibly through WGA/NOAA NIDIS) and might consist of such things as who can be authors, what products those authors can or should consider, what they should not look at, etc. As noted on the drought list server by a local expert, "this would serve as proof that the authors are qualified, and acting within the bounds of a standard base of data and related information. It may also help to explain the potential problems of going from a National level assessment to that of a county."
    3. Barbara Mayes will investigate writing an MOU for among the authors

32. Advisory Board:
    1. general response was this is a good idea
    2. could serve as an auditing board, as a mechanism to get funding and/or agency support and/or to deflect politics, as a lobbying group; more than one board may be required for these diverse activities. As noted on the drought list server by a local expert, the board "could dissuade critics from accusing the authors of being swayed by outside pressures, while showing that we are working with the public to improve the product."
    3. board composition could be federal chiefs, local/state/regional and user group representatives
    4. it was noted that the exploder list server group local experts are unofficially serving as an auditing group already. But we also would want members on the board who aren't on the exploder list.
    5. possibly use the Oklahoma Mesonet Board as an example/template

33. Funding:
    1. Ted O'Brien noted that there are resourced international agreements which we probably aren't accessing, but which we could/should be accessing for drought monitoring. More likely to succeed in tapping these sources for NADM than USDM.
    2. Ted also suggested someone senior write a letter, and send to someone with resourcing authority, explicitly identifying what is needed to author the USDM (personnel, how it is done, time it would take, computer equipment/software and funding needed, etc.). The following were suggested: NRCS Chief Bruce Knight, NWS HQ or HPC, Office of Hydrology- Andrea Bair-Frank Richards' boss-Greg Mann-corporate board.
    3. it was suggested that the WGA/NOAA NIDIS effort could provide graduate students to do research on correlating impacts to the various drought indices

34. there was some concern about unauthorized early release of the USDM product. We don't want anyone (private industry, etc.) on the exploder list server who would have a financial stake in the USDM. There was discussion about signing a pre- release form and about having this issue be part of the written guidelines.

35. making the USDM ArcGIS shape files available to the public was clouded by concern about users changing the Dx depictions:
    1. make them available in the future, but not now
    2. instead of zipping the entire file directory structure, zip just the shape files and make just those available to the public
    3. protect the shape files -- make them read-only -- so the public can't change them

36. there was a desire to have the entire USDM archive from 1999 onward converted to an ArcGIS format. We could georeference the gif images and resolve them to the climate divisions, but limited resources dictates that this be a project for later.

37. in a month, Dave Miskus will be making short-term and long- term maps, with separate Dx categories, and producing ArcGIS shape files to put online as a prototype. He will coordinate with Harry Lins on the long-term map.

38. Rich Tinker will be tweaking the long-term blend for the West, will send out to the exploder list server any changes for comment and verification before finalizing. Possible changes include giving the longer term (60-month, 12-month, etc.) precipitation percentiles higher weights, and including a hydrologic-year-to-date (October to present) precipitation term with a large weight.

39. Roy Kaiser/Gina Loss indicated Montana has ArcGIS SWSI data back ten years which they would be happy to provide to us as input

1. One of the key points that Don Jensen brought up in his presentation is that in the West it is important to calculate various indices by calculating index values (his example was ET) at local stations first before spatially interpolating, rather than interpolating the individual parameters (temperature, precipitation, etc...) And then calculating an index value for the region.

2. To emphasize one of your points (Item 2b), I thought the time series plots of energy release components presented by Gina Loss were extremely valuable and to get some of those parameters would be helpful in monitoring conditions in the West...even though we don't technically do "fire". I think they would help us put conditions into a "return period" perspective that is important for us as authors.

3. Gina Loss also emphasized that there is more behind the USGS streamflow numbers than meets the eye. Her examples showed how the normal range for USGS (25-75%) is very large, and oftentimes reflects the very short period of record for some gauges.

4. I wanted to re-emphasize Montana Counties Soil Climate Network (MCSCN). It is a prototype of a system that could provide a lot of beneficial information. The network includes counties that MUST have 5 stations within the county, with one station co- located with an NWS office or COOP site.

5. Ron Abramovich emphasize the importance of "cumulative drought impacts" from multiple-year droughts. That is an issue that is not well capture by any objective indicator. Others also talked about the "hysteresis" effect of drought severity and impacts. This has been mentioned at previous drought forums as well.

6. Ron A. also talked about waiting for making changes to the map after precipitation events. This suggestion was echoed by Al Dutcher. It is a suggestion made a previous forums as well. On a related note, Al Dutcher suggested making the Drought Monitor current through Monday rather than Tuesday. That MIGHT help the issue of waiting in that at least changes to the Drought Monitor would not be made between drafts 1 and 2.

7. Gregg Garfin emphasized the need to understand the relationship between drought severity (as measured by the variety of indicators and the DM) and the impacts! This is a huge research need.

8. Another useful drought monitoring tool is the Crop Specific Drought Index (CSDI) discussed by Al Dutcher.

9. Tom Pagano mentioned that the SWSI values, by their nature, are generally conservative.

10. Tree ring analyses are critical because of the reasons Connie Woodhouse mentioned, but they also provide an opportunity to better understand the relationships between various oceanic and atmospheric parameters that may give us insight to the future predictability of droughts.

11. There is still a very critical need for water supply/demand information for livestock, as well as better indicators of pasture conditions.

12. Jay Breidenbach gave quite a list of DM "successes" and customer feedback. One of the items on the list that caught my attention was the fact that the customers he has interacted with were okay with some subjectivity within the DM map.

13. I loved Joe Abraham's perspective "Leave No Stakeholder Behind". It is one we should remember with the Drought Monitor map.

14. I loved how your notes describe the 1 map vs. 2 map perspective: "A weekly map, plus a monthly hydrological drought map, was considered to be the ideal arrangement." You continued "The group proposes that this partnership prepare a hydro-oriented companion map product to the USDM. This monthly Water Resources Monitor (USWRM) would complement not replace the weekly USDM." This is a very key point...do all the authors agree?

The key themes, notes and recommendations I took from the meetings are already spelled out quite nicely by Mike and Richard. Let me just add or reemphasize the following:

1. It was agreed upon to hold the DM Forum bi-annually, perhaps alternating w/ the North American DM Forum.

2. It can be hard to get a feel for impacts and their accumulative effect during on- going multi-year droughts...........MT, UT, ID and WY for example. The DM perhaps doesn't take this into account enough, especially when there is a short- term improvement period. Determining drought classification is perhaps tougher to define coming out vs. going in.

3. Akin to real estate, in the drought monitoring world of the West: elevation, elevation, elevation! Should we be looking at incorporating PRISM?

4. Get off of Climate Division analysis and move to gridded and/or station based calculations

5. Tap into local/regional Mesonet for soil moisture observations and groundwater well readings (NE, OK, MT, etc.)

6. Do I need to hit on this issue again??.............probably, 1 vs. 2 vs. 3 maps. I think it was agreed upon to keep the map as we do it now, but propose moving to separate water resources assessment and ag maps now that we are moving into ARC GIS. Concern about resources and commitment to getting this done!

7. A real need for SWSI info for the entire West. These could be a real aid to the Drought Monitor, the water resources assessment (Water Monitor) product development, and possibly the blends (a watershed basin approach)

8. Utilize ACIS in the future to include crop-specific drought indices as well as PET calculations

9. Along w/ elevation, elevation, elevation...........we have to be aware of timing, timing, timing, especially on the Ag side.

10. People expressed a real need for a detached impacts drought map. This would help to account for "lingering" impacts as well as other social-economic drought impacts. As the forefather to the Drought Monitor, the NDMC maintains a Drought Impacts map at: http://www.drought.unl.edu/risk/us/usimpacts.htm Archives here go back to the summer of 1998. The NDMC will investigate how to further enhance this product in the future.

11. The Paleo-Perspective: more public awareness/education needs to be conducted w/ regards to planning for drought and determining worst case scenarios based on more paleo data.

12. As part of the project "National Integrated Drought Information System" underway through the Western Governor's Association (WGA) and NOAA/CSD, Phil Pasteris talked to the group about "gap analysis" w/ regards to what parameters we are lacking and where monitoring is needed at the local level. Input is needed from the DM exploder group to help in this effort.

13. Obviously, drought doesn't stop at geo-political borders, so the talk from Canada was appropriate. Every effort should be given to maintaining the North American Drought Monitor effort between Canada, Mexico and the U.S. This is currently an experimental monthly product, which can be found at: http://www.ncdc.noaa.gov/oa/climate/monitoring/drought/nadm/index.html

14. An area really lacking for monitoring in the U.S. (besides soil moisture obs, groundwater levels and more upper elevation data) falls into range/pasture lands

15. Discussions were had about looking into the future of the blends to account for gridded vs. climate divisions plus adjusting the makeup/weights w/in them as well. The blends are doing a better job of picking up short-term drought but there are some areas where they may not really be picking up and holding onto long- term droughts..............this is not just a problem in the West. Rich Tinker will look into this when he can get to it.

16. "Off-season" drought DO matter. Economic impacts can be huge for the snow droughts for tourism and winter recreation activities. A better snow depth climatology is needed as well as maintaining the snow observers themselves.

17. Could we better use model-based indices as another tool? Some alternative simulation based soil moisture monitoring methods were demonstrated through the North American Land Data Assimilation System.

18. Much talk about the role of the DM as a "trigger" in USDA's disaster relief programs. Concerns about spatial resolution of the DM (approximately 10 county) being used at a "county" level. There was talk about how our resolution will improve in ARC GIS once we start incorporating more station or cell-based inputs on a grid. There were also discussions about potentially creating "buffers" or transition boundaries between the categories when it comes down to where the line cuts through a county. States and USDA need things at a county level.

19. There is a need for archiving the ARC GIS shape files for each week's map and for going back and redrawing the maps back to 1999.

20. Need for more authors and resources commitments to the DM