Volume No. 12, 2003/2004
FINAL REPORT SUMMARIES
The Iowa Gap Analysis Project (IA-GAP) began in 1997 to identify areas in the state where vertebrate species richness lacked adequate protection under existing land ownership and management regimes.
To accomplish this goal, the IA-GAP team prepared an assortment of data sets that led to three main pieces of information:
When the project began, there were few statewide data sets available that provided the type of data needed for this project. Consequently, much effort was devoted to building the previously mentioned key data layers at a sufficiently fine scale and resolution for subsequent analysis. At the completion of the project, these data became freely available, with the intent that they will be used by those responsible for managing the state’s valuable natural resources and by the public, so that everyone can be better informed. With this in mind, we emphasize that these data are dynamic and, in some places, already out-of date. Nonetheless, the data and analyses that constitute IA-GAP represent an important first step toward understanding the status of vertebrates and land cover in Iowa and planning for the conservation of their biodiversity.
The land cover of Iowa was mapped by a two-phase, digital classification procedure, which was applied independently to 12 Landsat 5 Thematic Mapper (TM) images covering the state. All TM images were from mid-April to early October between 1990 and 1994. In the first phase, classification of the satellite TM imagery was done by the Iowa Department of Natural Resources, Geological Survey Bureau between 1997 and 1998. In this phase the land cover was separated into six cover types: cropland, grassland, trees, artificial, barren, and water, using unsupervised classification. The resolution of the satellite imagery and resulting classified image was a 30 m pixel, and all subsequent classifications were done at that same resolution. Phase two of the image analysis further differentiated certain land cover classes generated in phase one into one of 29 vegetation alliance aggregations. These alliance aggregations are part of an Iowa vegetation alliance list developed for IA-GAP by experts within the state.
Ground-reference data were used in an unsupervised classification to label each mapping unit according to its land cover type. A total of 29 different land cover types were mapped across the state. Digital National Wetland Inventory (NWI) data was used to provide the wetland data for Iowa GAP. An NWI aggregating model lumped the many classes of wetlands into five general groups: temporary, seasonal, semipermanent, permanent, and water.
The single most extensive cover type was cropland, which comprised almost 60% of the state. As a group, grasslands covered about 28% of the state. Six and a half percent of the state was forested, the upland deciduous forest type making up 5.7% of that total.
Several factors influenced our decision to use existing data sets for accuracy assessment and forego the implementation of a statewide data collection effort. Two data sets were available for use from the same time period as the land cover map: a partial statewide coverage of land cover from the Iowa county offices of the Natural Resource Conservation Service (NRCS) and the joint USGS and EPA generated National Land Cover Dataset (NLCD). Overall accuracies were 77% and 75%, respectively. Both data sets were aggregated to the Andersen Level 1, seven classes, and then compared to Iowa GAP land cover aggregated to the same seven classes.
Predicted Vertebrate Distributions
Distributions of 288 terrestrial vertebrate species were predicted, including 21 amphibians, 44 reptiles, 170 birds, and 53 mammals. The modeling process involved five steps. First, hexagon-based range limits for each species were determined, based on the location of species records or breeding bird survey blocks. This step included input from experts in the field. Next, associations between each species and its habitat, features such as land cover, soil types, and distance to water, were researched and summarized in a Wildlife-Habitat Relationship Model (WHRM) database. After preparing the necessary GIS layers to represent these habitat features, a raster-based modeling approach was used to determine the predicted distribution; the distribution grid was clipped to the extent of the range map. The range maps and WHRM for each species were reviewed by various experts within the state. After review, any necessary changes were made to the range limits and model rules. No accuracy assessment was done for vertebrate species. It is hoped that this report will form the basis for future accuracy assessment studies.
Geographic patterns of species richness generally suggest higher diversity in the northeastern and southeastern portions of the state, with the lowest diversity found in regions where farming is predominately intense row cropping in north central and northwest Iowa. Greater species diversity occurred in the most heavily forested counties in northeastern Iowa and along the major streams and rivers associated with the Mississippi River system. Grasslands in south central Iowa and in the Loess Hills Region also showed greater diversity across taxa.
Considering the issue of scale, we feel confident that our models performed reasonably well for Iowa land cover types. With this coarse-scale model approach, errors of commission will be more common than errors of omission. In other words, overestimation of a species distribution is more likely. Failure to predict a species’ presence in an area where it actually occurs may cause inadvertent harm if land-use decisions are made without that species in mind. If, however, a species is predicted to occur where it has never been recorded, it is more likely that the species will be targeted in future surveys and also considered in subsequent land-use decisions.
Land Stewardship and Management
The term “stewardship” is used in place of “ownership” because legal ownership, especially in the case of public lands, does not necessarily identify the entity responsible for management of the land resource. At the same time, it is necessary to distinguish between stewardship and management status, because a single land steward may manage portions of its lands differently.
The digital land stewardship layer was created by incorporating various administrative boundaries into a base layer of land ownership obtained from various sources. State lands were obtained from the Iowa Department of Natural Resources as an Arc/Info coverage. County lands were done by conducting an extensive mail survey through the Iowa Association of County Conservation Boards (IACCB). Individual counties submitted data on paper maps or as ArcView shapefiles if they possessed GIS capabilities. Each map feature in the stewardship layer was assigned a management status code and other required National GAP attributes. Status codes were determined by consulting management plans if they existed, talking with agency personnel, or looking at legislation that pertained to a particular land designation such as the State Preserves System.
Lands were assigned to one of four management classes based on the relative degree to which land stewards were responsible for maintaining biodiversity values. Status 1 lands reflected the highest, most permanent level of restrictive management; such lands included National Monuments, lands designated as a State Preserve, Nature Conservancy Preserves, and some National Wildlife Refuges where multiple uses were not permitted. Management could be changed more easily on Status 2 lands, such as wildlife management areas and National Wildlife Refuges where multiple uses were permitted, but it was still more restrictive than the remaining multiple-use public lands or private lands, which were assigned to Status 3. Status 4 included lands with no irrevocable easement or mandate to preserve biodiversity values or where the status otherwise could not be determined.
Private land makes up approximately 98% of land in Iowa. Public lands administered by federal, state, and county agencies consist of less than 2% of the state. Other than a few exceptions, most of Iowa’s public land consists of relatively small, disjunct areas within a vast amount of private land. Exceptions are areas along the Mississippi and Missouri Rivers, reservoirs along the Des Moines, Cedar, and Iowa rivers, and a scattering of larger complexes managed by many agencies and private individuals. Status 1 and 2 lands occupy less than 0.5 %. Status 3 and 4 lands, which actually actively contribute to the state conservation system, occupy less than 2% of the state; half of this is managed by the Iowa Department of Natural Resources.
Once the requisite statewide data were assembled, the actual gap analysis involved intersecting the GIS layers of land cover and predicted vertebrate distributions with land stewardship. These results form the basis of GAP’s mission to provide landowners and managers with the information necessary to conduct informed policy development, planning, and management for the long-term maintenance of biodiversity. A practical solution to the problem of defining adequate representation for vegetation or vertebrate species is to report both percentages and absolute area of each element in management areas and allow the user to determine which types are adequately represented in areas under active management.
Land Cover
Being an agricultural state, most land in Iowa is privately owned, and it was expected that the gap analysis results would reflect this situation. Cropland is 99% privately owned as is 98% of the grassland types, 90% of the forest types, and 86% of the herbaceous wetlands. Open water had the lowest private ownership at 53%. All 29 land cover types have less than 10% of their managed areas in Status 1 and 2. Actually, all 29 land cover types have less than 0.5% of their managed areas in Status 1 and 2. The area-weighted average percentage for all status 1 and 2 land in Iowa is 0.05%. Herbaceous wetlands as a group fared the best with 0.22% of their total area in status 1 or 2 land. Forest types follow with 0.17% of their total area in status 1 or 2 land.
Predicted Vertebrate Distributions
Greater than 90% (95.75%) of the predicted habitat for all species modeled in Iowa were on private lands followed by state lands (2.00%), and then federal lands (1.03%). The total amount of land falling into the status 1 and 2 categories was very small (< 0.5 % or 6,678 ha) and reflected in the amount of predicted habitat within these categories. For almost all species (98.26%, 283), the amount of predicted habitat within status 1 and 2 areas was less than 1.0%. The remaining five species (1.74%) modeled were found in the category of 1-<10% of predicted distribution in status 1 and 2 lands.
Intensive agriculture, urban development, drainage, soil erosion, deforestation, channelization of streams and rivers, and an extensive grid of transportation corridors have reshaped Iowa’s landscapes since the beginning of European settlement more than a century ago. The tallgrass prairies that covered the state’s highly productive soils have been reduced by more than 99%, and about 95% of the once abundant prairie potholes have been drained. Over half of the original forest has been lost, and the remainder has been severely fragmented and disturbed. Most of the natural areas that remain have experienced some kind of disturbance by grazing, fire suppression, or drainage.
Only a tiny proportion (2%) of the land area of Iowa is in public ownership, and only a few tracts are larger than a few thousand acres. Scattered remnants of prairies, forests, and wetlands have been preserved in state and county parks, preserves, wildlife management areas, state forests, and a few privately owned areas. Most public lands are managed for multiple uses, and few areas are managed for biodiversity conservation.
Much of Iowa’s biodiversity occurs along stream corridors where the land is less suitable for agriculture. Bluffs and bottomlands along the Mississippi River on the eastern border of the state, and the Loess Hills and Missouri River on the western border represent some of the best of the remaining natural habitats. These major rivers together with smaller rivers and stream corridors are important for species movement for both terrestrial species and migratory birds. The Des Moines river corridor, the Loess Hills, grassland areas in the northwest and south central sections of the state, the Iowa Great Lakes, and the northeast paleozoic plateau are also important centers of biodiversity and have potential for restoration and management.
Because of Iowa’s fertile soils and favorable climate, it is likely that the land will remain in agriculture and private ownership in the foreseeable future. Gap analysis can assist natural resource planners with identifying existing centers of biodiversity so that conservation efforts can be directed where they will do the most good. Large tracts of land for biodiversity management are seldom available; therefore, ways must be found to protect biodiversity on private lands such as through long-term conservation easements and other voluntary initiatives.
A 48-class land cover map of Kentucky was developed as one of the primary inputs for vertebrate distribution mapping. The classification was developed based on the National Vegetation Classification System (NVCS): International Classification of Ecological Communities: Terrestrial Vegetation of the Southeastern United States (Weakley et al. 1998) but used alliance aggregations as final map units. The map units were derived after consideration of the availability of statewide ancillary data sets for modeling, success achieved by other states with similar vegetation types, and expert review by state professionals.
The map was created using various combinations of input data, including
The vegetation classification used 29 Landsat Thematic Mapper (TM) images and a hybrid of supervised and unsupervised classification routines and preclassification image stratification to produce the primary modeling input. A DEM of Kentucky was created by mosaicking 721 USGS DEM quadrangles comprising Kentucky and a 10 km buffer zone surrounding the state. Derivatives of the DEM including slope, aspect, and landform were used extensively during the vegetation modeling procedures. Other data were used where appropriate and available. The final computer models were developed independently for each ecoregion in the state with the specific input dependent on data availability and the differing physiographic and geomorphic properties of the region.
Final map accuracy on a pixel-by-pixel, per-class basis was 51% with user’s and producer’s accuracy ranging from 8-100%. A “fuzzy” accuracy assessment was also conducted. The accuracy using this method was 75% with user’s and producer’s accuracy ranging between 25-100%.
The distributions of 365 native terrestrial vertebrate species, including 52 amphibians, 52 reptiles, 63 mammals, 152 breeding birds, and 111 wintering birds (65 bird species occurred in both the breeding and wintering groups) were predicted. Several steps were required to complete the modeling process. Experts from around the state reviewed the products from each step before the next step was taken. First, we determined range limits for each species based on current information about species’ presence or absence within counties and quads, or the Environmental Protection Agency’s (EPA) hexagon grid system. Second, the association of each species with habitat features such as land cover, water, edge, and elevation was researched and compiled in a Wildlife-Habitat Relationships (WHR) database. Third, the necessary GIS layers to represent these habitat features were prepared. Fourth, a raster-based modeling approach was used to combine the species’ ranges and WHR databases into maps of predicted distributions for each species at a resolution of 30 m grid cells. The final step was to determine the accuracy of the predicted species distributions.
Accuracy assessment was conducted at three spatial scales. In the first level, records of species occurrences were obtained within specific areas of the state. The data were based on species checklists from 51 validation areas, including 47 Breeding Bird Survey routes, one national park, one national forest, one national recreation area, and one state park. Predicted species’ presence and absence were then compared with those indicated on the species checklists. In the second level, an accuracy assessment of physiographic provinces was conducted. A predicted species occurrence list was created for each province and compared with the occurrences documented in the species checklists. Finally, an accuracy assessment at the state level was conducted by comparing the predicted species occurrences with observational data compiled from databases around the state. The latter two assessment processes only evaluated species’ presence and not species’ absence because of our inability to distinguish between observational errors and true absence of species. Thus, these two assessments were not complete.
Geographic patterns of richness of terrestrial vertebrate species indicated that biodiversity (i.e., predicted species richness) was generally higher in the western portion of the state. Highest mean biodiversity was associated with land cover of wet forested habitats (i.e., riparian, bottomland, and floodplain forests) and drier deciduous/coniferous forest habitats. L owest biodiversity was predicted in the Cumberland Plateau and the northernmost region of the state. Not surprisingly, lowest biodiversity among the habitat types was predicted for areas with little or intermittent vegetative cover (i.e., nonforested mine lands, agricultural lands, and high-density urban areas) across the state.
Comparisons between predicted and observed species presence/absence at 51 validation areas indicated high agreement rates overall. Low rates of omission errors (failure to predict a species at a location in which it has been recorded) occurred, averaging 4% for all taxa combined. Commission error rates (prediction of a species in a location in which it has not been recorded) were also relatively low (< 10%) for all groups except breeding birds. High rates of commission errors for breeding birds (32%) showed that the models were more likely to overpredict bird distributions than to underpredict them. High commission errors are preferable to high omission errors in the context of management decisions. Failure to predict a species in an area in which it actually occurs (omission errors) can lead to management decisions that inadvertently harm the state’s biodiversity. In contrast, if a species is predicted to occur where it has never been recorded (commission errors), then that species can be targeted for future surveys and can be considered in land use decisions.
The Gap Analysis Program (GAP) uses a scale of 1 through 4 to denote the relative degree of management for biodiversity maintenance for each tract of land, with “1” being the highest, most permanent and comprehensive level of maintenance and “4” being the lowest, or unknown, status. Status codes were assigned to land parcels by Kentucky Department of Fish and Wildlife Resources staff based on conversations with managing entities and field staff regarding management goals and practices. A flow chart adapted from the Gap Analysis Handbook was used to make final status determinations. The gap analysis of Kentucky consisted of intersecting stewardship lands with vegetation and vertebrate distribution. V ertebrate species and natural vegetation types were considered as underrepresented (i.e., “gaps”) if < 1% of their statewide distribution fell within Status 1 and 2 lands.
Less than 2% of Kentucky was classified as Status 1 and 2 lands. None of the 35 natural vegetation map units were considered well (> 50%) protected. Two (5.7%) of the 35 units had < 1% of their area in Status 1 and 2 lands. One of these, dry-oak forest in the Upper East Gulf Coastal Plain, is relatively common in other physiographic regions. The other underrepresented vegetation unit ( Cumberland highlands forest) is a good candidate for future protection efforts. As a whole, vegetation units exhibited moderate to low levels of protection; 21 of 35 (60% of vegetation units) had < 10% of their area in Status 1 and 2 lands. Other candidates for further protection include floodplain, riparian, and bottomland forest types, because they were consistently associated with higher vertebrate diversity.
Of the 428 native terrestrial vertebrate species assessed, 27 (6.6%) had >10% of their predicted distributions within lands assigned management Status 1 or 2, while 57 (13.3%) species had < 1%. Only 1 (0.2%) species had > 50% of predicted distribution in Status 1 and 2 lands. The distribution of this species (red-breasted nuthatch) was restricted to a single 7.5’ USGS quadrangle. Similarly to vegetation units, vertebrate species on the whole showed low to moderate levels of protection.
The GAP analysis of Kentucky confirms that few vertebrate species or vegetation types have high levels of long-term protection. More land under public ownership is warranted, but budgets require land purchases to be strategically located. Data products from this report provide the basic tools needed to examine and compare the merits of potential land purchases. More involvement from private landholders is also needed to improve protection where land for purchase is not available. Finally, continued research that builds upon this present work is needed to monitor long-term trends in vertebrate species and habitat protection.
The Maryland, Delaware, and New Jersey Gap Analysis Project (MDN-GAP) was initiated as a cooperative effort between the U.S. Geological Survey (USGS) Biological Resources Division, the Maryland Department of Natural Resources, and the U.S. Fish and Wildlife Service (USFWS). Late in 1999, the New Jersey Department of Environmental Protection (NJDEP) was added to facilitate work in that state. Various state, federal, and private natural resources entities in the three states were embraced as cooperators throughout the project.
The objectives were to (1) develop databases which describe current land cover, management of lands managed for conservation of biological diversity, and land stewards; (2) identify elements of biological diversity that are underrepresented or not represented in the existing network of lands managed for long-term maintenance of biological diversity; and (3) provide the databases in a format accessible to land managers and stewards, natural resources groups, and others interested in long-term maintenance of biological diversity.
As is true throughout the eastern United States, beginning over 200 years ago with the settlement by western Europeans, the landscape has changed dramatically from heavily forested regions to the current configuration. The middle Atlantic is highly fragmented and continues to be a rapidly urbanizing corridor. This complex matrix of elements made landscape characterization a challenging undertaking.
The 5,051,578 hectares (12,482,449 acres) in the study area were mapped into 62 classes (30 wetland and beach classes, 23 upland forest classes, 7 urban/disturbed/agricultural classes, and 2 water classes). The classification scheme follows the physiognomic hierarchical structure of the National Vegetation Classification System (NVCS). Out of necessity because of the resolution of the satellite imagery and ancillary data, most of the classes ended as aggregates of alliances based on ecological similarities. Landsat TM imagery from 1991-1993 was used as the basis for the classification along with aerial videography flown in 1996-1997.
Stewardship data sets were generated to allow a management status class to be assigned to all public and private conservation properties in the project area. The characteristics used to determine status were
This data layer combined with the land cover layer allowed the gap analysis to be performed. As the purpose of the analysis was to identify gaps in current protection of biological diversity, MDN-GAP took a conservative approach in assignment of status categories. Every parcel was considered on the merits of the actual use and management activities for that parcel without an initial bias of ownership. A result of the extensive multiple use policies and practices in these areas is that properties may have received a status characterization different from that which the land manager would have assigned.
It was not surprising, given the preponderance of fragmented and urbanized landscapes in the eastern U.S., to see that 87% of the area in Maryland, Delaware, and New Jersey was held in private ownership. Federal and state lands made up 10% of the area, while regional and local lands were less than 2%. Private conservation organizations owned less than 1% of the land area.
There are a number of public and privately held lands with some degree of biological diversity management activities not included in this data set. These include state university, correctional facilities, and hospital lands, and local land trusts and easements. These data were just being developed during the timeframe encompassing this project, and subsequent updates should make an effort to include these properties in the assessment.
The results of the gap analysis of the land cover are generally described here for MDN-GAP. These are arranged by major land cover class for simplicity; specific details are found within the report.
Forest classes: For the evergreen forest category, there were 10 mapped classes included. Status 1 or 2 lands were approximately 5,299 ha (13,094 ac), which comprised 2.8% of those classes’ total area. There were 16 deciduous forest mapped classes with an area of 70,270 ha (173,637 ac) listed as Status 1 or 2 lands, representing 5.6% of those classes’ total area. The mixed forest category contained six mapped classes and covered an area of 20,235 ha (50,001 ac) in Status 1 or 2 lands. This was approximately 4.2% of those classes’ total area. For forest classes as a whole, there were 32 mapped classes with 95,804 ha (236,732 ac) in Status 1 or 2 classification, which represents 5% of those classes’ total area.
Shrubs: The shrub category contained eight mapped classes covering an area of 1,946 ha (4,809 ac) in Status 1 or 2 categories. This represents 6.8% of those classes’ total area.
Herbaceous: The herbaceous category contained 11 mapped classes covering an area of 14,544 ha (35,938 ac) in Status 1 or 2 categories. This represents 7.9% of those classes’ total area.
Sparsely vegetated: There was only one mapped class included here, the Sparsely Vegetated Beach Alliances (SVBA). Other classes that may have been included were determined to be more transitional than permanent and therefore counted as anthropogenic and not included in the analyses directly. All other classes may be found in the data sets which accompany this report. The SVBA mapped class found in Status 1 or 2 lands had an area of 451 ha (1,114 ac), which represented 22.4% of that class’ total area.
The vertebrate predictive distribution modeling aspect of MDN-GAP is under development by a separate GAP project. Expected completion of these models is 2004. As a result, the gap analysis conducted and reported on in this report is limited to the land cover mapping. MDN-GAP provides a baseline upon which more detailed studies and efforts needed for the long-term conservation of biodiversity in the area may be placed. It is hoped that the data sets provided with this report benefit a wide variety of uses.
s
Seventy vegetation alliances were identified in South Dakota using the National Vegetation Classification System. SD-GAP used a regional approach to mapping land cover, separating the state by biogeographic regions: eastern South Dakota (e.g., prairie potholes and Northern Great Plains), western South Dakota (e.g., Northern Great Plains, Badlands, Sand Hills, minus the Black Hills), and the Black Hills. Landsat 5 Thematic Mapper satellite imagery with a resolution of approximately 30 m was obtained through the Multi-Resolution Land Characteristics Consortium. Vegetation was mapped at a scale of 1:100,000 with a minimum mapping unit (MMU) of 2 ha for vegetation. Special features, such as temporary wetlands, also were preserved with a MMU of 2 ha. Satellite images covering eastern South Dakota were interpreted using training data from Farm Service Agency aerial photo section maps (2.56 km 2), which identified agricultural and perennial vegetation categories. A GIS database containing four wetland classes was overlaid onto the classified image. Remaining land cover classes and individual vegetation alliances were mapped by reclustering perennial vegetation and subsequently interpreting the clusters into vegetation types, as well as by onscreen digitizing. Satellite images of the Black Hills were interpreted using training data obtained from a USDA Forest Service GIS that included land cover categories for the Black Hills National Forest. Vegetation was mapped using a GIS summary that contained the percentage of land cover types comprising each spectral cluster, and a regression tree analysis incorporating ancillary databases. Western South Dakota, however, lacked training data over the majority of the land area, and therefore this region was interpreted using known data from only two locations, northwestern South Dakota and Wind Cave National Park (TNC Nature Mapping Program).
Once all regions were classified, they were mosaicked, and population information from CIESEN ( University of Missouri, Columbia) was overlaid to represent towns, cities, and industrial areas. The SD-GAP land cover classification identified 35 categories, including 9 grassland, 3 shrubland, 1 dwarf-shrubland, 2 woodland, 5 forest, 6 water and wetland, 3 barren or badland, and 6 disturbance categories. Combined grassland categories dominated the South Dakota landscape, accounting for 56.1% of the land area. Agriculture comprised 31.2% of the entire land area of South Dakota. Creeping juniper woodland, bur oak forest, cottonwood woodland, and shale barren slope sparse vegetation were among the smallest categories. Six water categories comprised 4.5% of the land area; eastern water categories were primarily prairie potholes. Badlands comprised less than 1.3% of the land area, while the forest categories of the Black Hills made up only 2.7%. A total of 931 locations were assessed for accuracy in eastern South Dakota. Of those, 797 were correctly classified for an overall accuracy of 85.6%.
Environmental Protection Agency (EPA) Environmental Monitoring and Assessment Program (EMAP) hexagons were used to create distribution maps for terrestrial vertebrate species throughout the state of South Dakota. Each hexagon in the state was coded with a 0, 1, 2, or 3 using ARC/INFO or ArcView software. Species richness was determined by overlaying all distribution maps using EMAP hexagons. Species richness was determined for all taxa of species and with an overall richness for terrestrial and aquatic species. Habitat models were created by researching habitat use described in papers published with habitat information on each species. Over 1,200 papers were reviewed for both terrestrial vertebrates and aquatic species. Each species’ habitat type was evaluated for the most appropriate GIS method. Land cover types were selected from the South Dakota land cover map. Each map was clipped along the species’ range boundary, if needed. Land cover was limited by appropriate elevation boundaries, wetland or woodland buffers, soils, or precipitation boundaries for each species by clipping the land cover with the boundary, if necessary. Range maps and models were created for 78 mammals, 215 birds, 30 reptiles, and 15 amphibians occurring in South Dakota. Overall species richness ranged from 177 to 249 of the 338 species present in South Dakota.
After initial processing was completed, habitat variables were added to the reach files. Variables used were temperature, stream size, flow, geology, groundwater potential, relative gradient, size discrepancy, floodplain reach, and dam or stream reach. These are essential variables believed to affect the community structure of the stream system. Each variable was combined to form a valley segment code. This unique valley segment code identified the properties of each individual stream. Fish distribution predictions were based on this code. When a known location was plotted on a given reach, fish species found in that reach were assumed to be in suitable habitat. If all 10 habitat variables matched another stream reach, that stream also was assumed to have suitable habitat. Additional habitat information was located with pertinent literature related to each species. When specific habitat information was found, the attributes were also considered to be suitable habitat and used to predict additional stream types where the species may be found. The literature was incorporated into the species reach models by substituting values in the known valley segment types that were different from data found in the literature. Concatenating habitat variables for each river reach resulted in greater than 6,200 unique valley segment codes. For the analysis, temperature, stream size, flow, gradient, and groundwater potential were used to create a reduced number of concatenated valley segment types. This resulted in 127 unique types in South Dakota. Range maps were created for 116 fish species using South Dakota’s 11-digit hydrologic units. After predicted river reaches were identified for 97 viable species, range maps were updated to identify watersheds with predicted reaches present. Species richness ranged from 0 to 87. Highest richness was in watersheds found along the main channel of the large rivers present in the state. Few fish species were predicted to occur in headwater streams, possibly indicating a data gap. Due to the river continuum concept, which states that as stream order increases the number of fish species increases, large rivers are not necessarily in need of protection through public lands. Headwater streams likely need higher levels of protection through public lands to maintain biological diversity.
Digital coverages of public land boundaries were obtained from the respective agencies. When coverages were not available, land ownership boundaries were obtained in paper format and digitized to create coverages. All coverages were error-checked, attributed, and combined to create a statewide stewardship map for gap analysis. The stewardship layer identified 20 land ownership categories. Over 70% of the land area of South Dakota was privately owned and managed. Federal and state entities owned approximately 9.5% and 2.1% of the land area in South Dakota, respectively. Status codes also were assigned and evaluated. Greater than 85% of the land area in South Dakota was classified as Status 4 land, meaning that conversion to unnatural land cover types is likely. Just over 1% of the land was considered “highly conserved” (Status 1 & 2 combined).
Gap analysis was conducted by intersecting land cover types with conservation status and ownership codes and exporting resulting tables. Square kilometers and percentages of each land cover type were reported within each status and ownership code. Each land cover type or species protection status was reported as 0<1%, 1<10%, 10<20%, 20<50%, or >50% of the species total land area found in Status Code 1 or 2. Three South Dakota land cover types (burned pine, vegetated badlands, and unvegetated badlands) had greater than 10% but less than 20% of land area in Status 1 and 2 management. No valley segment code was protected in at least 10% of its range. No mammal species was protected in at least 10% of its range. Five bird species (bufflehead, great egret, rednecked grebe, white-throated swift, and wood duck) were protected in at least 10% but less than 20% of their ranges. One amphibian species (mudpuppy) was protected in 15% of its range. Twelve fish species were protected in greater than 10% but less than 20% of their ranges, including American eel, lake herring, logperch, longnose gar, mottled sculpin, rainbow smelt, silverband shiner, silver chub, silver lamprey, spottail shiner, suckermouth minnow, and yellow perch. One fish species (blue catfish) was protected in 100% of its range. Protection from conversion currently is lacking in most areas in the state (<1% of South Dakota’s land is classified in Status 1 or 2 conservation lands), especially in eastern South Dakota where most land (49%) has been converted to agriculture. None of 127 revised valley segment codes were protected in at least 10% of their ranges. This includes 17 rare valley segment types that received no Status 1 or 2 protection on any portion of the streams. Of these 17, 11 (65%) were present entirely on Status 4 (private) land. Of 55 Black Hills cold-water stream types, 44 (80%) had no Status 1 or 2 protection. Of 72 warm-water stream types, 27 (38%) had no Status 1 or 2 protection. Based on this analysis, South Dakota’s biological stream conditions were not considered protected. Overall, 20 of 435 (5%) vertebrate species in South Dakota were protected in 10% or more of their ranges. Of the terrestrial vertebrates, only 6 of 338 (2%) had 10% or more of their range in Status 1 and 2 lands, while 14 of 97 (14%) of aquatic vertebrates were protected at this level.
The National Gap Analysis Program is a nationwide effort to assess and document the spatial distribution of biodiversity elements and evaluate its conservation status across the U. S. The Texas Gap Analysis Project (TX-GAP) has been conducted at the Texas Cooperative Fish and Wildlife Research Unit (TX-CFWRU) at Texas Tech University (TTU) in Lubbock, Texas. The objectives of the project were (a) to develop a map of current land cover of Texas from recent Landsat TM images that was at least 80% accurate in predicting the true vegetation types; (b) to estimate the potential distribution of Texas wildlife vertebrate species with at least 80% accuracy; (c) to depict and map land stewardship categorized by level of conservation management; and (d) to combine the above data layers in a GIS and perform analyses of species richness patterns relative to known levels of land conservation and management.
Due to the size of Texas and variability in environments, its vegetation is both diverse and complex. In addition, more than 90% of the land is privately owned, and access to the land for field verification procedures is limited. The land cover map was generated through digital classification of satellite imagery supported by field surveys and ancillary information.
Accuracy assessment involved a statistical comparison of subset samples from the classified scene to ground observations and using airborne videography. Eighty-three land cover/land use types were mapped, four of which are not natural vegetation (water, bare soil, cropland, and urban). Vertebrate distribution predictions were modeled from known location data and information on species-habitat associations. The distribution maps produced were verified through expert review, by comparing predictions with specific areas where detailed inventories exist, and by conducting field surveys. Land stewardship and management were inventoried and mapped. This information was used as reference for the analysis of vegetative communities and wildlife species conservation status.
For the final analysis, all these data were combined in the GIS to evaluate how vegetation communities, sites with maximum number of species overlap (richness), or even single species distributions were represented in existing managed areas. This allowed the identification and delineation of potential “gaps” in conservation and their potential risks.
The gap analysis for land cover types showed that the most protected land cover type is the consolidated rock sparse vegetation which overlaps 80% with conservation lands. If we isolate that type, we find that it occurs as rocky cliffs almost exclusively in Big Bend National Park and the Guadalupe Mountains. Although it would appear to be very well protected, it has a limited distribution and only accounts for a very small portion of the total land cover of Texas (0.02%). Other examples of types with restricted areas and very low amounts of protection (i.e., <1%) are juniper woodlands, playa lakes, wetlands, Harvard oak shrublands, and water oak forests. Riparian areas, hardwood bottomlands, and wetlands have experienced some of the most serious reductions in occurrence since human settlement. The grassland prairie types, while they may represent a much larger percentage of the total area (approximately 20% combined) have little overlap with conservation lands (generally <10%). In some cases, they have been overgrazed to the point of creating a type conversion to cactus shrubland types. Despite the widespread occurrence of the prairie grasslands in Texas, their low representation on conservation lands is a concern.
The gap analysis of the vertebrate species shows that most species fall into the 1-10% level of protection. Ninety percent of the 628 species are less than 10% protected. Even more sobering, we do not know what the current distribution of these animals represents in habitat quality or what portion this is of their historic distributions. The majority of the animals in the category that is greater than 50% are species whose predicted distributions are confined to the Big Bend Park area. Three of the species have endangered or threatened listings: the reticulated gecko, the greater long-nosed bat, and the spotted bat. Their level of protection does not necessarily indicate that the populations are viable. However, it does suggest how very important a natural area such as Big Bend is to the continued conservation of these species of interest. On the other hand, some endangered species have very little overlap on protected lands, such as the Comal blind salamander (0.1%) and the Texas kangaroo rat (basically 0%). Texas has a very small amount of land that is set aside for protection of biodiversity. Consequently, 85% or more of each taxon’s species are less than 10% protected. As a group, amphibians are least represented.
Because Texas is so dominated by private lands, we clearly need to consider more solutions that offer economic and social incentives to landowners and encourage them to participate in creative management activities that would benefit our natural resources. Education and outreach programs should be developed and promoted throughout the state, and partnerships between agencies, corporations, and the private sector should be encouraged.
Return to Table of Contents