Wetland in the Muscatatuck National Wildlife Refuge. USFWS photo by Thomas Simon.

South Fork Patoka River showing headwater restoration efforts. USFWS photo by Thomas Simon.

Conservation of imperiled species and the protection of biological integrity of National Wildlife Refuges is an important policy mandate of the U.S. Fish and Wildlife Service.  The policy 601 FW 3 is a directive for refuge managers to follow while achieving refuge purpose(s) and refuge ystem mission. The policy provides for the consideration and protection of the broad spectrum of fish, wildlife, and habitat resources found on refuges and associated ecosystems. Further, it provides refuge managers with an evaluation process to analyze their refuge and recommend the best management direction to prevent further degradation of environmental conditions; and where appropriate and in concert with refuge purposes and system mission, restore lost or severely degraded components.  The National Wildlife Refuge System Administration Act of 1966 as amended by the National Wildlife Refuge System Improvement Act of 1997, 16 U.S.C. 668dd-668ee (Refuge Administration Act), Section 4(a)(4)(B) of this law states that "In administering the System, the Secretary shall . . . ensure that the biological integrity, diversity, and environmental health of the System are maintained for the benefit of present and future generations of Americans . . . ." This is one of 14 directives to the Secretary contained within the Refuge Administration Act. 

Biological integrity, diversity, and environmental health can be described at various landscape scales from refuge to ecosystem, national, and international. Each landscape scale has a measure of biological integrity, diversity, and environmental health dependent on how the existing habitats, ecosystem processes, and wildlife populations have been altered in comparison to historic conditions. Levels of biological integrity, diversity, and environmental health vary among refuges, and often within refuges over time. Individual refuges contribute to biological integrity, diversity, and environmental health at larger landscape scales, especially when they support populations and habitats that have been lost at an ecosystem, national, or even international scale. In pursuit of refuge purposes, individual refuges may at times compromise elements of biological integrity, diversity, and environmental health at the refuge scale in support of those components at larger landscape scales. When evaluating the appropriate management direction, refuge managers will consider their refuges' contribution to biological integrity, diversity, and environmental health at multiple landscape scales.

This study evaluated the influence of contaminants on fish and wildlife resources of National Wildlife Refuges in Indiana.  The principal goals of this project was to determine the current fish, macroinvertebrate, and crayfish assemblage condition of National Wildlife Refuges in Indiana using a variety of scales including refuge and watershed approaches based on a spatially intensive sampling design.  A second goal was to determine those areas that contaminants caused aquatic assemblage degradation and document the magnitude and extent of problems using biological indicators such as fish, macroinvertebrate, and crayfish assemblage structure and associated physical and chemical sampling. 

Fish, macroinvertebrate and crayfish community composition, structural and functional attributes, feeding groups, and spatial distribution was studied at three refuges in Indiana.  The Index of Biotic Integrity (Simon 1990; Simon and Dufour 1998; Simon, unpublished data) provided a method that was a direct measure of impacts to biological assemblages and water quality and habitat data that were impairing biodiversity and biological integrity.  A stressor identification model was developed that evaluated the spatial extent that contaminants, habitat, and land use was impairing biological integrity on National Wildlife Refuges.

Physical habitat, land use, and chemical variables were measured to evaluate influences on aquatic community expectations.  Aquatic community data were evaluated using multivariate analysis to determine if specific physical and chemical variables were predictive or correlative with the degraded biological integrity of the system. A new analysis technique (Morris et al. 2005) termed the “Hot spot” analysis was used to identify the magnitude and extent of the impact from the identified contaminant using a spline smoothing kringing technique within a local refuge and watershed context.

The design strategy used was a probability, stratified sampling design that was used to select locations from each of the refuges both within and upstream.  Biological, chemical, physical, and land use information was acquired from each location so that patterns in indicator response could be determined.  Fish assemblage information was measured using an index of biotic integrity (IBI) calibrated for the Eastern Corn Belt Plain and Interior River Lowland Ecoregions.  Macroinvertebrates and crayfish assemblages were collected and data analyzed based on species richness measures since no index of biological integrity has been calibrated for Indiana.  Chemical samples were collected on the date of the fish collection, preserved in the field, and transported on ice to the State of Indiana Board of Health for analysis.  Physical habitat based on the qualitative habitat evaluation index (QHEI) and land use was determined for each location.  This information was the basis for determining the species distribution, conservation status, and IBI scores for each location. 

The second objective of this study was to evaluate the status of the aquatic assemblages on the three Indiana refuges.  The project basically consisted of two parts, the evaluation at local or refuge scale and the evaluation at a larger watershed scale.  During 2006, fish, crayfish, and macroinvertebrate assemblages were collected from the Big Oaks National Wildlife Refuge.  Macroinvertebrates were collected at all three refuges during the spring of 2007, while  fish and crayfish were collected from the proper index periods during the summer and fall of 2007.  Fish sampling methodology included back pack and long-line electrofishing in wadeable streams for distances of 50-150 m, while boat electrofishing was conducted in non-wadeable large river, lake, and pond habitats within a 500 m sampling distance.    Data was analyzed with consideration of drainage area and the index of biotic integrity (IBI) biological criteria used by the State of Indiana.  Significant differences were observed between sites on refuge compared to sites upstream of the refuges. 

Additional data analysis to identify watershed stressors was based on data collected by the Indiana Department of Environmental Management Biological Studies Section.  This information was collected during 2006 and 2007, to connect interpretive ability between the refuges in the Vernon Fork Muscatatuck River watershed.  An additional 63 locations were assessed including duplication of some sites collected during the refuge scale assessment conducted by the Service.  During 2007, the State of Indiana completed surveys during drought conditions in the Vernon Fork Muscatatuck River.  The 2006 sampling seasonal was typical and considered a normal water year, while the 2007 season was a dry, hot year that resulted in non-flowing conditions in many of the streams on the refuge.  Some sites became isolated pools that restricted the migration of species.  This provided an excellent opportunity for best and worst case opportunities to evaluate the refuges.  For most of the refuges, samples collected during 2007 provided higher IBI scores than 2006; however, surprisingly most of the refuges possessed fish assemblages that included many sensitive species, including a few species considered to be extirpated from Indiana. 

The highest species richness was collected from the Patoka River National Wildlife Refuge followed by Big Oaks and Muscatatuck.  Several important distribution records included the presence of flier, redspotted sunfish, eastern sand darter, harlequin darter, and popeye shiner.  The flier and redspotted sunfish records are the furthest north and eastern records for these species, while the popeye shiner is the first record of the species in Indiana since the species was described from the White River near Indianapolis in the late 1800s.   New distribution records for the Indiana crayfish and Sloan’s crayfish were discovered during this study.  The Indiana crayfish is a former federal candidate species, while Sloan’s crayfish has been identified as imperiled by the Natural Heritage database.  The Indiana crayfish was collected from sixteen new locations in the Patoka River watershed in the vicinity of the National Wildlife Refuge.  Despite the report in the Natural Heritage database, Sloan’s crayfish has not declined in either abundance or status in the vicinity of the Big Oaks National Wildlife Refuge.  Improvements in the presence of Sloan’s crayfish have been observed near the Muscatatuck National Wildlife Refuge with increased abundance of Sloan’s crayfish and declines in rusty crayfish abundance.  The rusty crayfish was observed in nutrient impacted areas east of the Big Oaks National Wildlife Refuge and downstream near the West Perimeter Road during 2007.  No other specimens were collected from within the refuge.  Rusty crayfish were not collected from either the Muscatatuck National Wildlife Refuge or the Patoka River National Wildlife Refuge during either 2006 or 2007 collections. 

Based on refuge scale assessment, it is clear that the three biological indicators are measuring different environmental components in the environment.  Fish assemblage structure reflected large scale land use, habitat, and surface water contaminants, while macroinvertebrate assemblage structure reflected groundwater contaminant constituents and to a lesser extent habitat and land use attributes.  Crayfish assemblages provided the clearest indication of measurement at the Big Oaks National Wildlife Refuge, but did not provide an important measure at the Patoka River National Wildlife Refuge.  Further monitoring efforts would be needed to evaluate the practicality of crayfish as an indicator of stressor response.  Little overlap in contaminant, habitat, or land use variables were observed between the three biological indicators. 

The Patoka River National Wildlife Refuge was being impaired by legacy acid mine drainage leachate.  The acidity and associated heavy metals emanating from the South Fork Patoka River was the primary area responsible for the contaminants identified by the stressor model.  In addition, a heated water impact from the Black Beauty Coal Mine was detected during March sampling.  The stressor model also detected poor performance by the Oakland City WWTP, which showed degraded conditions downstream of the discharge. 

Significant differences between refuge and upstream land use and habitat caused most of the non-contaminant variables to be significant in the stressor model.  Big Oaks National Wildlife Refuge was primarily impacted by nutrient impacts from the headwaters of Little Graham Creek and from metal contamination from within the refuge and upstream on Otter Creek.  The metal contamination from within the refuge appears to be contained to areas previously known and no migration of heavy metals was observed downstream from the refuge.

The refuge scale stressor identification model identified commercial areas in Sandy Branch as contributing to the loss of biological integrity on the Muscatatuck National Wildlife Refuge.  In addition, a legacy train derailment from the Baltimore and Ohio Railroad near milepost 81.4 near Hayden was being detected in Storm Ditch.  A processing facility upstream of the northeast corner of the refuge is causing high biological oxygen demand and depleted  dissolved oxygen conditions in the tributary of Richart Lake. 

The watershed scale assessment found different stressor responsible for explaining the variability in fish assemblage structure in the Patoka River watershed.  The local scale stressor model identified fish assemblage structure was explained by zinc, chloride, and grass-pasture land use.  The watershed indicator identified twelve chemical contaminants to explain variation.  Both zinc and chlorides were identified by both models, but the local scale identified grass-pasture land use as a variable, while the watershed scale identified four land use variables not including grass-pasture.  The watershed scale stressor model identified two habitat variables including pool-glide score and drainage area, while habitat did not significantly explain local scale model variance in the Patoka River National Wildlife Refuge. 

At the watershed scale, variables able to explain the variation in the fish assemblage structure were primarily habitat and land use in the Vernon Fork Muscatatuck River.  Chloride, chromium, sodium, and iron were identified as significant by the stressor identification model, while chloride, reactive silica, and sulfate were significant at the local scale.  The differences in significantly predictive chemical stressors is a result of several factors including the percent-of-range concentrations and the density of sites.  As site density increases across a watershed, greater ability to detect significant patterns occurs.  In addition, if sites are too close together, despite high levels of potential contaminants being present there is no significance because levels become background.  It is important in designing stressor assessment studies to select assessment scales that would capture the greatest variety of potential contaminants and include sufficient site density that chemical contamination rather than land use variables become the significant predictive variables.