An Official Website of the Commonwealth of Kentucky
Volume IV identifies spatially explicit geographic areas for the purpose of focusing conservation efforts that benefit the largest number of Species with Greatest Conservation Need (SGCN). These areas represent Priority Conservation Areas (PCA’s) within Kentucky. Volume IV addresses part of the requirements of Element 2 by identifying the location of key habitat areas for each taxonomic group and all SGCN combined. The location and condition of key habitat areas was also identified for each species in Volume II under the Species Accounts. Locations listed in Volume II were site specific in that counties, United States Geological Survey (USGS) quadrangles, stream reaches, etc. were listed. This approach differs from Volume II by locating landscape level areas that harbor large numbers of SGCN. The species occurrence records used to generate maps in Volume II formed the primary data source for identifying conservation areas in Volume IV. The following sections explain the methodology used, sources of information used, conservation areas for each taxonomic group, and finally the PCA’s for all taxonomic groups combined.
The KY-CWCS Team examined several methods and data sources for locating PCA’s within the state. Relational database management systems (SQL Server and Access; Microsoft Corporation; Redmond, WA) and Geographic Information Systems (GIS; ArcView 3.2 and ARCGIS 9.0; ESRI, Redlands, CA) were used to organize and analyze these data for all species and groups of species. A wide variety of data sources across species groups at various spatial scales were available for analysis. Initial examination of all these sources quickly revealed that simply ‘throwing’ all data into a GIS produced more noise than useful information. A methodology was developed that allowed these data layers to be utilized in a systematic manner but also allowed for significant input from species experts during the process (Appendix 4.1). This ‘bottom-up’ approach uses species occurrence data from Volume II as the base to build the analysis. Only occurrence records dated 1984 or later were used for determining PCA’s. This approach allowed flexibility in selection of a species or species group, level of detail for species occurrence records (i.e., point, quad, or county level), spatial unit (e.g., watershed or Ecoregion), and a numeric index to quantify the results (species richness and/or rarity).
Species data were organized taxonomically by Class and categorically into aquatic and terrestrial groups. The spatial units for examining occurrence data differed between aquatic and terrestrial groups. For aquatic species 8-digit (larger watersheds) and 14 digit (smaller watersheds) USGS hydrologic units were used (United States Geological Survey, 2001a; United States Geological Survey 2001b). Level IV Ecoregions (Woods et al. 2002) formed the spatial basis for terrestrial species. Some hand digitized areas were included for mammals and birds (see following sections).
Richness of SGCN was used as the primary index for all taxonomic groups regardless of the spatial unit. Rarity of individual SGCN was also used by the Fishes and Lamprey’s Team to select some watersheds. Maps that depicted SGCN richness were generated by combining species occurrence records with spatial units. Taxonomic experts examined these maps to aid in selecting important conservation areas for each group. After all taxonomic conservation areas were selected they were analyzed to identify areas within the state where overlap occurred between groups. Regions of highest overlap were identified as Tier I Priority Conservation Areas. Specific methodology and results are covered for each taxonomic group in the following sections.
Relational database and GIS tools were created to give biologists the ability to examine other data sets such as predicted habitat suitability from the Kentucky GAP Analysis Project (Wethington et al. 2003) and large forest block data developed by the Kentucky State Nature Preserves Commission (KSNPC). These tools varied slightly between taxonomic groups to accommodate biologic differences between groups, data availability, data quality, and existing prioritization systems (e.g., Partners in Flight focus areas). Predictive vertebrate models developed by KY-GAP were used to examine where significant overlap of suitable habitat for SGCN occurred. This approach provided a valuable tool for terrestrial taxonomic groups but aquatic species were not modeled by KY-GAP. Also, some terrestrial models were lacking important recent information. For example, KY-GAP coded reclaimed surface mines as unsuitable for golden-wing warblers yet recent surveys have shown this type of habitat to be valuable for these birds (Shawchyi Vorisek, pers.comm). We also included priority or focus areas already determined by partners in Kentucky.
Species occurrence data were used to generate SGCN richness maps at the 14-digit and 8-digit hydrologic unit (watershed) levels for both the fish and mussel taxonomic groups. The USGS hydrologic units at the 8-digit (large watershed units) and 14-digit (small watershed unit) were selected because they are watershed based. Since aquatic species are influenced by any action upstream, we chose the 8-digit units as the first priority boundary. These are the appropriate size to allow conservation actions to be implemented in a broader area that includes areas both upstream and downstream of important sites. The larger 8-digit watersheds often cross political boundaries and allow managers from multiple agencies to coordinate conservation actions. Smaller 14-digit watersheds were selected to represent the best areas within the larger 8-digit watershed. These 14-digit units are areas of high species diversity where managers can focus conservation actions and conduct site-specific research, monitoring, and management. Both the 8-and 14-digit watersheds were prioritized based on the SGCN richness. The 8-digit watershed with the highest richness was examined first. The remaining hydrologic units were included until at least 80% of all aquatic SGCN were included.
Mussel SGCN richness maps were inspected to determine the location of important watersheds (Appendix 4.2, Appendix 4.3). Nine 8-digit watersheds were identified as Mussel Conservation Areas (Appendix 4.4). A total of 36 mussel SGCN occurred in these 9 watersheds (Appendix 4.5). These included the following watersheds in order of priority by species diversity: Upper Green, Barren River, Upper Cumberland-Lake Cumberland, South Fork Cumberland, Rockcastle River, Licking River, Ohio Brush-Whiteoak, Rolling Fork, and Lower Tennessee and Lower Ohio Rivers. The Upper Green and Barren Rivers contain 23 bivalve SGCN (52% of all SGCN). The Green River has been identified nationally as an important area for freshwater mussel conservation and the most biologically rich branch of the Ohio River system (Master et al. 1998). Several federally endangered species are found throughout many of these areas and some of the best populations are also contained within these watersheds.
Fishes and Lampreys SGCN richness maps were inspected to determine the location of important watersheds (Appendix 4.6, Appendix 4.7). A total of 10 8-digit watersheds that contained at least 5 SGCN were identified as Fish and Lamprey Conservation Areas (Appendix 4.8). A total of 50 (85%) SGCN occurred in these watersheds (Appendix 4.9). The top 5 of these watersheds included: Obion Creek (15 SGCN); Upper Green River (10 SGCN); South Fork Cumberland River (9 SGCN); Lower Mississippi Memphis (8 total SGCN); and Barren River (8 SGCN).
Fourteen 8-digit watersheds were identified for the aquatic group; 5 of these overlap for both mussels and fishes, 5 were fish only and 4 were mussel only (Appendix 4.10). Two 8-digit hydrologic units were combined (Lower Ohio and Lower Tennessee) where the fauna was similar enough to be considered a single management unit. The highest priority watersheds included 5 overlapping 8-digit watersheds and one non-overlapping 8-digit watersheds (Appendix 4.11). The non-overlapping watershed (Obion Creek) was included as high priority because it contained the highest fish diversity. These included from West to East; Obion Creek (non-overlapping), Lower Tennessee and Lower Ohio Rivers, Barren River, Upper Green River, South Fork Cumberland, and Rockcastle River. Lower priority watersheds included the 8 remaining non-overlapping 8-digit watersheds identified in the bivalve and fish priority sections.
Species occurrence data was used to determine 8-digit hydrologic units (watersheds) where there were endemic or multiple species of SGCN crayfish. Eleven 8-digit watersheds were identified as Crayfish Conservation Areas (Appendix 4.35), in alphabetic order: Barren, Bayou De Chien, Kentucky Lake, Little Kentucky, Lower Ohio-bay, Middle Green, South Fork Cumberland, Upper Cumberland, Upper Cumberland-Lake Cumberland, Upper Green, and Upper Levisa.
Determining conservation areas for terrestrial species presented several challenges to the CWCS Team. The most significant issues were selecting a consistent spatial unit and the lack of occurrence data for most terrestrial species (Appendix 2.7). A watershed approach was not used as with aquatic species because terrestrial species range across the landscape and are not confined to waterways (Eastern Hellbender is an exception). SGCN Richness was examined at several spatial scales that included county, USGS quadrangle, USGS ¼ quadrangle, and Level IV Ecoregion. After considerable deliberation, the Terrestrial CWCS Team used Level IV Ecoregions as the basic mapping unit. This allowed implementation of landscape level management practices and accommodated the lack of quality point occurrence records.
The reptile and amphibian groups had particularly low numbers of occurrence data. Many reptile and amphibian SGCN were not tracked by the Heritage Program (Appendix 2.2). Most data for these species are still in paper form. A large number of occurrence records exist for birds (Appendix 2.7) but selecting conservation areas remained difficult. Filtering breeding records from transient and/or accidental records was not always possible. Also, concentration of bird occurrence records was directly related to the location of existing monitoring stations, not necessarily the location of important conservation areas. Given these were the best available data, combined with species expert knowledge, the Terrestrial Team examined SGCN richness maps for Level IV Ecoregions and modified these as needed to make the best possible representation statewide conservation areas.
Species occurrence data for all amphibian SGCN were extremely sparse (Appendix 2.7). This was particularly true for amphibians not tracked by the Heritage Program at Kentucky State Nature Preserves Commission (KSNPC). For this reason only occurrence records for species monitored by KSNPC were used to generate richness values. The map of these richness values by Level IV Ecoregion (Appendix 4.12) revealed that diversity within this group was concentrated in the western 1/3 of Kentucky resulting in the selection of 10 ecoregions. In order to include species not monitored by KSNPC 2 additional areas were added. These were the Cumberland Mountain Thrust Block Level IV Ecoregion and a small, but important, area in northern Kentucky. The final Amphibian Conservation Area map (Appendix 4.13) included occurrence records for all amphibian SGCN (Appendix 4.14).
More species occurrence records were available for birds than any other taxonomic group (Appendix 2.7). As previously discussed, interpreting these data made selecting conservation areas difficult. The Level IV Ecoregion richness map (Appendix 4.15) generated using all bird SGCN did not match well with biologist knowledge of important areas. We therefore divided the entire bird group into three sub-groups: Forest Birds, Grassland Birds, and Wetland Birds (Appendix 4.16). This classification scheme closely mimics the Central Hardwoods Joint Venture Concept Plan (Fitzgerald et al. 2003). Conservation areas were identified for each of these sub-groups and then combined to generate the overall bird conservation area. The process for identifying conservation areas for each sub-group of birds is described below.
Conservation areas for Forest Birds were created using a myriad of techniques and data layers. Occurrence records for this sub-group were inspected in relation to Level IV Ecoregion, KSNPC Large Forest Block data, focus areas from the Central Hardwoods Joint Venture Concept Plan (Fitzgerald et al. 2003), and KY-GAP predicted distributions. The resulting Forestland Bird conservation areas map (Appendix 4.17) was based almost entirely on Ecoregion boundaries (with slight modifications to include contiguous forest blocks) and the addition of the Licking River watershed.
Data used to generate conservation areas for Grassland Birds included focus areas for grassland birds identified in the Central Hardwoods Joint Venture Concept Plan (Fitzgerald et al. 2003), KY-GAP vegetation and predicted distribution, and maps created by KSNPC indicating areas where native grasslands most likely occurred in 1793 (KSNPC). The Bird conservation areas for Grasslands (Appendix 4.18) was a mixture of ecoregional boundaries and hand digitized polygons to include reclaimed surface mine areas, concentrations of species occurrence records, and other important grassland areas known to biologists.
Conservation areas for Wetland Birds were determined by inspecting KY-GAP vegetation and predicted distribution, focus areas identified in the Central Hardwoods Joint Venture Concept Plan (Fitzgerald et al. 2003), KDFWR survey data, and wetland inventory data (Kentucky Environmental and Public Protection Cabinet, 2002). The final Bird conservation areas for Wetlands map (Appendix 4.19) consisted primarily of Ecoregion boundaries in western Kentucky, the Ohio River corridor, the Licking River watershed, and, polygons around important wetland areas and public waters.
To determine areas in Kentucky most important to all avian SGCN, we combined the three bird sub-group conservation area maps to find areas of overlap (Appendix 4.20). These overlap areas represented the highest priority, as projects to benefit all three groups could be implemented. Species occurrence records for 71 of the 78 (91%) extant bird SGCN fell within these areas (Appendix 4.14). The overlap areas for the bird sub-groups were used in later analysis to determine conservation areas for terrestrial groups and all CWCS taxonomic groups. (See Terrestrial Conservation Areas and Combined Priority Areas for Terrestrial and Aquatic Taxonomic Groups)
Examination of SGCN richness by Level IV Ecoregion for mammals (Appendix 4.21) revealed distinct differences in diversity between eastern and western portions of the state. High-elevation species and numerous species of bats accounted for the importance of areas in the east while 2 species (i.e., cotton mouse and swamp rabbit) were restricted to western portions of Kentucky. The ecologic importance of karst topography present in south-central Kentucky was not reflected by the SGCN richness map in Appendix 4.21. Portions of 6 different ecoregions converge here making the choice of important ecoregions very difficult. We therefore hand digitized a polygon based on all mammalian SGCN occurrence data and known locations of hibernacula (i.e., caves) for bats identified as SGCN. This karst region includes Mammoth Cave National Park and a significant portion of the Middle Green River identified by the Aquatic CWCS Team as highly diverse. This ecologically unique landscape is important to numerous other invertebrates not addressed in this CWCS version. For example, there are at least 15 crayfish with distribution records very near or within our Karst Conservation Area (Taylor and Schuster 2004), and the federally endangered Kentucky Cave Shrimp (Palaemonias ganteri) is endemic to the Green River within this area. Given the diversity of the Green River, the abundance of caves and sinkholes, and the overall uniqueness of habitats within this karst region, it is very likely that other invertebrates found here will warrant consideration as SGCN in future versions of CWCS. Overall, the conservation areas for mammalian SGCN (Appendix 4.22) contained occurrence records for all species in this taxonomic group (Appendix 4.14).
Selection of reptile conservation areas was similar to that of amphibians. Again, only occurrence records for KSNPC monitored species were used to generate a Level IV Ecoregion richness map (Appendix 4.23). As with amphibians, the western 1/3 of Kentucky was most important to these species (Appendix 4.12). The remaining reptilian SGCN were included by adding the Southwestern Appalachians Plateau Escarpment Ecoregion in southeastern Kentucky. Overall, the conservation areas for reptilian SGCN (Appendix 4.24) contained occurrence records for all species in this taxonomic group (Appendix 4.14).
Conservation Areas for each terrestrial taxonomic group were combined to determine areas of overlap (Appendix 4.25). All taxonomic group conservation areas intersected in far western Kentucky and in the south central karst region of Kentucky. Other areas containing conservation areas for at least 3 groups included the lower Ohio River, portions of the Tradewater River, and Lake Cumberland.
The 6 Conservation Area maps created for each taxonomic class provided landscape level information about important areas for each group. This approach is beneficial in that it helps ensure that our strategy addresses the geographic diversity of SGCN. The difficulty is that a simple overlap of all taxonomic conservation areas does little to prioritize or focus where actions are implemented. Ideally, specific geographic areas could be identified where conservation efforts could benefit the maximum number of taxonomic groups and SGCN. This subset of all conservation areas would then represent our overall CWCS Priority Conservation Areas (PCA’s).
Our first step in locating PCA’s was to overlap conservation areas for all taxonomic groups. Each taxonomic conservation area (6 total) was assigned a value of 1 and then all were summed together. A map was created to illustrate the overlapping of taxonomic conservation areas (Appendix 4.26). Regions with values of 6 indicate that at least some species from all 6 taxonomic groups are likely to occur there. The portions of Kentucky with at least 1 taxonomic conservation area totaled 21,249,578 acres or approximately 82% of the state (Appendix 4.27). Nearly 68% of the state was a conservation area for 1, 2, or 3 groups while only 14% showed overlap for 4 or more groups. To narrow our focus we inspected areas where values of 4, 5, and 6 occurred. We observed that these higher values tended to cluster together into 3 distinct geographic regions. These 3 areas became our Tier I PCA’s. We created 3 different polygons around each clustering of values 4, 5, and 6 and tentatively named them (Appendix 4.26). From west to east they are; 1) Mississippi-Ohio Valley Plains (MOVP) Conservation Area or Purchase Region, Interior Low Plateau Karst (ILPK) Conservation Area or Green and Barren Rivers, and the Southwestern Appalachian Plateau (SWAP) Conservation Area or Upper Cumberland River.
We generated a list of known SGCN occurrences from the KFWIS database that fell within the boundaries of each Tier I PCA (Appendix 4.28). With these data we calculated an overall SGCN richness for each Tier I PCA and richness by taxonomic class (Appendix 4.29). With this information we hoped to prioritize within the Tier I PCA’s themselves. The MOVP was the most SGCN rich (149), followed by the ILPK (133), and the SWAP (104). Richness of SGCN indicates the total number of species that can be addressed in each area but does not account for species that may occur in more than one area. Logically, we would designate the MOVP as the most important because the largest numbers of SGCN were found there. We employed a repetitive algorithm to select the next PCA (Williams, 1998). This type of ‘greedy’ algorithm emphasizes stepwise selection of new conservation areas that are richest in targets or features not already represented. In other words, which PCA would harbor the most species not found within the MOVP? Beginning with the richest area (MOVP) we compared SGCN composition between all areas. We found the greatest SGCN overlap between the MOVP and ILPK. More new species were found (49 species; (Appendix 4.29) in the SWAP even though its SGCN richness was lower. Combining unique SGCN from the MOVP and SWAP generated a list of 198 SGCN total found in these 2 areas. This procedure was repeated adding the ILPK and 27 new species were added for a total of 225 SGCN found in the 3 Conservation Areas.
All Tier I PCA’s combined comprised only 14.4% of Kentucky’s total land area (Appendix 4.30) but contained occurrence records for 91% of all SGCN (Appendix 4.29). These figures are in agreement with United Nations guidelines for reserve design (Bruntland, 1987) that gives a target of 12% land area encompassing 85% of targeted species. This ’12 % fixation’ target has also been adopted by Canada (Hummel, 1995). We feel this provides evidence for the soundness of our methodology and the validity of the PCA’s selected by the CWCS Team. However, we must also be mindful that additional research is needed to confirm the persistence of these occurrence records (i.e., one time sighting or persistent population?) and the quality of the records (were there data entry errors or misidentifications?). While these Tier I PCA’s provide a valuable guide for determining the most important locations to implement conservation actions they do not represent the only areas in the state in need of conservation. Many Tier II areas harbor large numbers of SGCN and contribute to securing the biodiversity of Kentucky. Special consideration will be given to those areas supporting SGCN that were not included in Tier I PCA’s. For example, the Cumberland Mountain Thrust Block ecoregion in southeast Kentucky contains several SGCN (e.g., Canada Warbler and Kentucky Red-backed Vole) not found in Tier I PCA’s.
Additional habitat and species mapping projects are needed that focus on these areas to determine if and where ‘hot spots’ may occur. This also will allow better coordination with partners already working in these areas. A first step towards coordination was to identify the public lands within each area (Appendix 4.31, MOVP; Appendix 4.32, ILPK; and Appendix 4.33, SWAP). Nearly 38% of the SWAP PCA is in public ownership (Appendix 4.30). The Daniel Boone National Forest (DBNF) comprises most of this public land although the Big South Fork National Recreation Area and KDFWR wildlife management areas also contribute significant area. Inspection of the map, however, reveals that in-holdings within the DBNF may need to be addressed in that region. Public ownership is much less significant in either the MOVP or the ILPK (3.7% public land for each). Habitat composition of each area is another important consideration for future research. The MOVP is approximately 20% pasture and 47% agriculture (Appendix 4.34). This indicates that working closely with Farm Bill programs are likely important actions. The ILPK also has significant portions of land use in pasture and agriculture, 33% and 31% respectively but is more heavily forested than the MOVP. Conversely, the SWAP is nearly 80% forested. Differences in land ownership patterns and land use patterns between areas will obviously play an important factor when selecting appropriate conservation actions within each Tier I PCA.
Bruntland, G. Our common future: The World Commission on Environment and Development. 1987. Oxford, England, Oxford University Press.
Fitzgerald, J.A., C. D. True, D. D. Diamond, T. Ettel, L. Moore, T. A. Nigh, S. Vorisek, G. Wathen. Delineating focus areas for bird conservation in the Central Hardwoods Bird Conservation Region. C.J. Ralph and T.D. Rich. Proceedings of the Third International Partners in Flight Conference. GTR-PSW-191, Pgs. 192-202 . Albany, CA., U.S.D.A. Forest Service. 2002.
Fitzgerald, J.A., Wathen, G., Howery, M.D., Lisowsky, W.P., McKenzie, D.F., Pashley, D.N.. The Central Hardwoods Joint Venture Concept Plan. 2003. Tennessee Wildlife Resources Agency.
Hummel, M., 1995. Protecting Canada's Endanagered Spaces. Key Porter Books, Toronoto.
Kentucky Environmental and Public Protection . Kentucky's Wetland Polygon Features (GIS Portal Single zone). 2002.
Master, L.L., S.R. Flack, and B.A. Stein, 1998. Rivers of Life: Critical watersheds for protecting freshwater biodiversity. The Nature Conservancy, Arlington, Virginia.
Taylor, C. A. and G. A. Schuster. The Crayfishes of Kentucky. 2004. Illinois Natural History Survey Special Publication .
U.S. Geological Survey. 14 Digit Watershed (USGS 14-Digit Hydrologic Units) shapefiles . 2001.
U.S. Geological Survey. 8 Digit Watershed (USGS 8-Digit Hydrologic Units) shapefiles. 2001.
Wethington, K., Derting, T., Kind, T., Whiteman, H., Cole, M., Drew, M., Frederick, D., Ghitter G., Smith, A., Soto, M.. The Kentucky Gap Analysis Project Final Report. 2003. Frankfort, KY., United States Geological Survey, Biological Resources Division.
Williams, P.H. 1998. Key sites for conservation: area-selection methods for biodiversity. Pages pp. 211-250 in .M. Mace, A. Balmford, J.R. Ginsberg editors. Conservation in a Changing World. Cambridge University Press, Cambridge, UK.
Woods, A.J., Omernik, J.M., Martin, W.H., Pond, G.J., Andrews, W.M., Call, S.M., Comstock, J.A., Taylor, D.D.. Ecoregions of Kentucky (color poster with map, descriptive text, summary tables, and photographs): (map scale 1:1,000,000). 2002. Reston, VA, U.S. Geological Survey.