Ecological Land Cover Classification For a Natural Resources Inventory in the Kansas City Region, USA

Transcription

1 Ecological Land Cover Classification For a Natural Resources Inventory in the Kansas City Region, USA by Applied Ecological Services, Inc. In cooperation with the Mid-America Regional Council 600 Broadway, Suite 300, Kansas City, MO Phone: Steven I. Apfelbaum 1, Kim Alan Chapman 2, Jason Carlson 1, Joshua D. Lippold 3, Frank J. Norman 3, Neil Thomas 4, and Theodore Hartsig 3 November Applied Ecological Services, Inc., Smith Road, Brodhead, WI Applied Ecological Services, Inc., Mushtown Road, Prior Lake, MN Applied Ecological Services, Inc., 201 Main Street, Suite 201, Kansas City, MO Resource Data, Inc., 305 Westover Drive, Asheville, NC Applied Ecological Services, Inc.

2 Executive Summary The Kansas City metropolitan region historically was the gateway to the Great Plains, and to this day it harbors high quality natural resources and vegetative species reminiscent of those seen by early settlers. In the 150 years since settlement, communities comprising this 3,000 square mile region have grown and changed the environment of these natural resources, and these once abundant resources have diminished greatly. Now in the 21 st Century, there is wide recognition that a coherent natural resource inventory is needed in the region. This natural resource inventory will be a wealth of information that can be used to plan for the future growth of our region and preserve valuable natural assets which benefit the people who live here. This recognition is timely. The region s many streams are threatened by unmanaged growth and land use change, yet the many municipalities that cut across watersheds weaken unified action. Widespread changes in land use must be understood to fully address the challenge of habitat fragmentation and degradation. Developing effective policies to protect these natural resources requires knowing where, how abundant, and in what state of ecological health these resources are. Recognizing these needs, Mid-America Regional Council (MARC) initiated the development of a regional map to depict the region s natural resources assets and ecological land features. MARC intends for this map and associated data to be used as a tool in conservation planning and ecological preservation by local governments and planning agencies throughout the Kansas City region. The ecological land cover map includes an extensive Geographic Information System (GIS) and associated database that is the primary basis of this tool for use in conservation planning. This initiative is funded by the U.S. Environmental Protection Agency and managed by MARC. With its completion, MARC will begin to educate local planning agencies and government agencies the public in using the regional ecological land cover map and GIS data as the conservation planning process moves ahead. With this project, Kansas City joins other national leaders such as Chicago and Milwaukee in using conservation planning tools for environmental quality. Thanks to similar projects, these cities increasingly enjoy better air and water quality, reduced flood damage, ecosystem and biodiversity conservation, habitat and wildlife protection, Ecological Land Cover Map of the Kansas City Metropolitan Area i

3 stream course stabilization, the creation of neighborhood and development amenities, better opportunities for outdoor recreation, and models of sustainable urban development. To achieve these goals, MARC teamed with Applied Ecological Services, Inc. (AES) and Patti Banks & Associates (PBA), two Kansas City firms with expertise in conservation planning and ecological services. AES created a natural resources classification and inventory method, conducted field surveys, and developed descriptions of ecological land cover types in the Kansas City region. The AES classification and inventory method used previous work by state and federal agencies, or data gathered by MARC and the region s counties and cities. AES intended the classification and inventory to be understandable to lay people and amenable to future revisions and updates. During the field inventory, hundreds of locations were visited by staff of AES and PBA to assess the condition of and collect new information about the region s natural resources, and verify existing data about vegetative communities and their occurrences. Results of the inventory found that 22 percent of the metropolitan region retains areas of high quality vegetative communities with numerous native species that are worthy of conservation. These areas are limited in number and in their extent, but they present many opportunities for conservation of ecosystems, habitats, and other natural resources that will benefit the overall Kansas City region. Results of the inventory also revealed a tremendous need for ecological restoration work so that stream water quality and fish habitat can be improved, and that wildlife habitat in forests, savannas, and grasslands can again sustain high numbers of native species. When incorporated into conservation plans, woodlands, grasslands, and shrublands can result in better stormwater management, cleaner water, healthier riparian areas and reduced sediment loadings in streams and other water bodies. Native species are vital for providing habitat for maintaining healthy wildlife populations. These lands also improve the aesthetic character of the landscape, provide scenic views and open space, and provide for recreational activities for our communities. The Kansas City region has several areas where natural resources are similar to those present 150 years ago, but they are rapidly being converted to non-native uses. ii

4 The urban areas in the Kansas City region continue to support large tracts of woodland and restorable savannas. Urban areas are included in this 22 percent of land area that retain high quality ecological conditions, and largely consist of mature deciduous forests, lowland hardwood forests in stream valleys and river bottoms, marshes and other wetlands, and the native grasslands and savannas that were the dominant natural features of the region 150 years ago. These natural resources are concentrated along rivers and streams, near open water, and on steeply-sloping land. The woodlands and restorable savannas in urban settings play important roles in providing recreational opportunities, serving existing and future parkland, creating buffers for streams, and increasing awareness of the region s environmental heritage. Rural and semi-rural areas on both Kansas and Missouri present large tracks of both forest land and native prairie for conservation and/or restoration. The current data demonstrates numerous areas in Leavenworth and Johnson Counties in Kansas that have large tracts of natural vegetative communities that contain quality native vegetation and provide great environmental quality. Wyandotte County, Kansas, while largely developed for urban utilization, also contains several large tracts of deciduous forest. Platte, Ray, Clay, Jackson, and Cass Counties in Missouri exhibit extensive areas dominated by land used for agriculture and/or urban development, however, the original classification of the data from the United States Geological Service (USGS) Gap Analysis Program (GAP), because it was collected several years ago, may not reflect many areas and tracts that contain native vegetation that provide opportunities for conservation. The results of the natural resources inventory and development of an ecological cover map for the Kansas City region provide the tools that will catalyze the next phase of MARC s initiative to address conservation and planning needs in the Kansas City region. Conservation and planning not only ensure that plant and animal species continue to thrive in their native habitats, they are vital to the future success of the region s communities. A focus on natural resource conservation and restoration benefits a resident s quality of life by preserving natural breaks in development for nearby recreation and relaxation; by maintaining and raising land values; by improving stormwater management and reducing flood risk; and by enhancing environmental quality. Along the way, conservation and planning will maintain or improve the stream flow and water quality, wildlife habitat and populations, and the rich diversity of life forms in the region. If successful, conservation planning for sustaining our natural resources will bequeath the heritage of the region s natural ecosystems and its many benefits to the future generations in the Kansas City metropolitan area. iii

5 TABLE OF CONTENTS Section Page 1. Project Overview Ecological Land Cover Assessment and Natural Resource Inventory Methods Data Assembly and Base Mapping Regional Crosswalk Digital Database Field Reconnaissance First Field Reconnaissance Second Field Reconnaissance Field Reconnaissance Data Evaluation Initial ELC Classification and Natural Resources Inventory (Jackson County) Aquatic Communities Natural Communities Forests Lowland Communities Grassland-Savanna-Woodland Communities Cultural or Sparsely Vegetated Land Developed Land Urban Forest Other Cultural Types Extrapolation to Eight-County Kansas City Region Final ELC Classification and Natural Resources Inventory (Kansas City Region) Ecological Land Classification and Natural Resource Inventory Outcomes ELC/NRI Results Ecological Land Cover Conservation Establishing Regional Conservation Priorities Public Outreach and Education Project Evaluation Summary Existing Data Gaps and Method Development Needs Future Data Acquisition and Analysis Acknowledgments...30 i

6 TABLE OF CONTENTS Tables Section Page Table 2-1 Crosswalk of Kansas and Missouri GAP Classifications... 4 Table 2-2 Ecological Condition Ranks... 8 Table 2-3 Ecological Land Cover Types for the Kansas City Region...15 Table 3-1 Ecological Land Cover in the Kansas City Region...18 Figures Figure 2-1 Ecological Land Cover Classification of Natural Resources in the Kansas City Region...16 Figure 3-1 Distribution of Ecological Land Cover in the Kansas City Region...17 Figure 3-2 Example of Natural Resources Conservation Planning Map...22 Appendices Appendix A A-1: Detailed Data Collection and Conversion Methodology A-2: Natural Resource Inventory (NRI) Metadata A-3: Shape Files Used for the Kansas City NRI GIS Program Appendix B Soil Types in the Kansas City Region ii

7 1. Project Overview The Kansas City metropolitan region has historically served as the gateway to the Great Plains of the United States, and to this day the region continues to exhibit the rich natural resources and high quality vegetative communities. The extent of our native natural resources, however, has largely disappeared with development of the cities and communities that make up the metropolitan region, with relatively small remnants remaining in areas along the major drainage ways and in small pockets throughout the area. These remnant prairies, wetlands, and forests are worth protecting for the benefit of the people that live in the region, and for sustaining and promoting native wildlife and vegetation. The Kansas City region exists at the confluence of the Kansas and Missouri Rivers, natural resources that dominate the geographical features of the area. This area is situated in the western glaciated plains and the Osage Plains on the southern portions, consisting of gently rolling topography dominated by thin soils on uplands, rock outcrops, and rich, deep soils in bottomlands. Prior to its settlement, vegetation in the Kansas City region consisted of upland and bottomland prairies, although relatively large tracts of forests were present in the bottom lands. The majority of the area, however, was tallgrass prairie on the gently rolling plains that were dissected by the tributaries of the Kansas and Missouri Rivers. Extensive aquatic and wetland communities were present, particularly in the floodplains of the larger rivers and, to a lesser extent, their tributaries. In 2003, Mid-America Regional Council (MARC) initiated the development of a regional map depicting natural resource assets and ecological land features oriented towards conservation planning and ecological preservation. The objective of this program is to develop a comprehensive database and GIS program that features multiple mapping layers that depict the region s vegetation resources, natural resource features (including rivers, streams, and other water features), and regional infrastructural features, such as streets and roads, political boundaries, parks, and features of MetroGreen, the regional network of trails and greenways. These layers can be used to evaluate regional natural resources and how they can be conserved and integrated into community planning. This project, inclusive of the regional map and the fieldwork it incorporates, represents a natural resource inventory across a broad region that is part of a larger initiative to produce the tools for communities in the Kansas City region to develop conservation plans. These tools are intended to be used as the basis for identifying important conservation goals and strategies to protect the region s natural resources. With this project MARC will implement the region s first multi-phase, collaborative, community-based initiative to document, map, and ultimately conserve natural resources within the Kansas City metropolitan area. The development of a natural resources inventory is a critical first step toward solid environmental planning at the local level utilizing a systems-based framework for watershed management, resource conservation, and restoration at the regional level. This report presents the methodology used to develop the ecological classification system and natural resources inventory for the Kansas City region, as well as the results of initial data collection and the creation of a regional ecological land cover database and map. The report describes how the project creates new digital maps of the region s natural resources - useful to local governments, planners, engineers, developers, ecologists, and citizens - to ensure that critical natural areas such as floodplains, wetlands, and quality upland areas are conserved as 1

8 development continues. Over the long term, the intent of this regional natural resource inventory and the associated map series is to provide a base that will lead to the creation of a regional framework for an interconnected, landscape-scale conservation and restoration plan. This inventory and map can be utilized to form a key component of all local and regional planning efforts related to such factors as land use, economic development, transportation, water resources, and air quality. To assure that this project is as complete as possible, the ecological land cover map and GIS builds on past inventory and conservation work for critical ecosystems and valuable natural resources in the region. The results provide an up-to-date assessment of the condition of ecosystems and natural resources while compiling consistent baseline information in a usable Geographic Information System (GIS) for use in local planning processes (e.g., watershed plans and greenway plans); support the identification and implementation of high priority MetroGreen trail segments around the region; and offer opportunities to educate Kansas City audiences about the value of local natural resources and the value of sound environmental planning and stewardship. This is an on-going process however, and this report describes the first steps in a program that will be added to from data collected through new studies and planning processes in the Kansas City region. 2

9 2. Ecological Land Cover Assessment and Natural Resource Inventory Methods This section describes the general approach and methodology used to complete the ecological classification and inventory for the Kansas City region. Applied Ecological Services, Inc. (AES) used the following tasks to complete the ecological survey and inventory, which are described in detail in the following sub-sections: 1. Data Assembly and Mapping: digital information from several government sources was used to establish baseline information about land cover in the region. 2. Field Reconnaissance: The digital information was validated and/or refined through field inspections and verifications. 3. Ecological Land Cover Classification Development: Using data from the data assembly and subsequent field reconnaissance, AES created an ecological classification representing existing natural resources in the region, a GIS-based information database, and a regional map of ecological land cover. 4. Data Extrapolation and Second Field Verification: The ecological classification involved an iterative process in which initial data were assembled, evaluated in the field, revised, and then re-evaluated in the field a second time. Final data were assembled after the second field reconnaissance, evaluated, and incorporated into the GIS program and the regional land cover map. Details of the methodology of this process are provided in Appendix A. This program was completed between June 2003 and June Data Assembly and Base Mapping The initial phase of the ecological classification and inventory work involved data identification and assembly; the synthesis of data and creation of GIS base maps and graphics; and solicitation of input from local experts on the type and condition of natural resources in the Kansas City metropolitan area. This initial phase included two primary components: a regional crosswalk and development of the digital database. These components are described below Regional Crosswalk Early assessment of information about vegetative cover and natural resources in the Kansas City area indicated a lack of a comprehensive, up-to-date data set from which to create an ecological land cover classification and complete a natural resources inventory. Classifications from Missouri and Kansas used different criteria to map natural resources, and classifications by different agencies varied in their mapping results. To overcome this challenge, AES examined existing land cover classifications and tested the mapping results against recent digital orthophotos of the Kansas City metropolitan area. A classification was considered acceptable based on its ability to correctly describe a uniform patch of land surface on digital orthophotos (e.g., forest, grassland, and urban) and correctly 3

10 match the edges of the digitized polygon to that land surface patch. Based on this examination, AES decided to use United States Geological Survey (USGS) Gap Analysis Program (GAP) data rather than other land cover data systems. The GAP project is a nationwide inventory of land cover and habitat suitability whose purpose is to identify conservation opportunities especially as it relates to the needs of wildlife. While available for both Missouri and Kansas, each state used a different GAP classification. In order to join these classifications and create a regional ecological land cover classification, AES created a two-level classification linked to the two GAP classifications (Table 2-1). This approach to combining different classifications is termed a crosswalk. The crosswalk contains two types of classifications. The first classification, herein called AES Type 1, is broader than the second classification, called AES Type 2. For example, Cultural or Sparsely Vegetated Land (AES Type 1) in Kansas is subdivided into Cultivated Land, Developed Land, and Cultural Grassland (AES Type 2). The AES Type 1 names are used for initial mapping work and the AES Type 2 names are used for field inventory. Table 2-1. Crosswalk of Kansas and Missouri GAP Classifications. Kansas GAP Classification AES Type 1 AES Type 2 Missouri GAP Classification Water Aquatic Communities Open Water Open Water Cultural or Sparsely Barren or Sparsely Vegetated Vegetated Land Land Barren or Sparsely Vegetated Cultivated Land Cultural or Sparsely Vegetated Land Cultivated Land Row and Close-grown Crops Cultural or Sparsely Vegetated Land Developed Land Urban Impervious Cultural or Sparsely Vegetated Land Developed Land Urban Vegetated Non-native Grassland Cultural or Sparsely Vegetated Land Cultural Grassland Cool-season Grassland CRP (Conservation Reserve Cultural or Sparsely Program) Vegetated Land Cultural Grassland Maple-Basswood Forest Natural Communities - Forest Deciduous Forest Deciduous Forest Oak-Hickory Forest Natural Communities - Forest Deciduous Forest Deciduous Forest Post Oak-Blackjack Oak Forest Mixed Oak Ravine Woodland Tall Grass Prairie Mixed Prairie Pecan Floodplain Forest Natural Communities - Grassland and Transitional Natural Communities - Grassland and Transitional Natural Communities - Grassland and Transitional Natural Communities Grassland and Transitional Natural Communities - Forest Natural Communities - Forest Natural Communities - Forest Natural Communities - Grassland and Transitional Natural Communities - Grassland and Transitional Natural Communities - Lowland Deciduous Woodland Deciduous Woodland Deciduous Woodland Deciduous Woodland Mixed Evergreen-Deciduous Forest Mixed Evergreen-Deciduous Forest Mixed Evergreen-Deciduous Forest Grassland Grassland Lowland Hardwood Forest and Woodland Deciduous Woodland Glade Complex Eastern Red Cedar and Red Cedar-Deciduous Forest and Woodland Shortleaf Pine-Oak Forest and Woodland Shortleaf Pine Forest and Woodland Warm-season Grassland Warm-season Grassland 4

11 Kansas GAP Classification AES Type 1 AES Type 2 Missouri GAP Classification Ash-Elm-Hackberry Floodplain Forest Cottonwood Floodplain Forest Mixed Oak Floodplain Forest Bur Oak Floodplain Woodland Cottonwood Floodplain Woodland Low or Wet Prairie Freshwater Marsh Cattail Marsh Natural Communities - Lowland Natural Communities - Lowland Natural Communities - Lowland Natural Communities - Lowland Natural Communities - Lowland Natural Community Lowland Natural Communities - Lowland Natural Communities - Lowland Natural Communities - Lowland Lowland Hardwood Forest and Woodland Lowland Hardwood Forest and Woodland Lowland Hardwood Forest and Woodland Lowland Hardwood Forest and Woodland Lowland Hardwood Forest and Woodland Lowland Hardwood Forest and Woodland Marsh and Wet Herbaceous Vegetation Marsh and Wet Herbaceous Vegetation Marsh and Wet Herbaceous Vegetation Bottomland Hardwood Forest and Woodland Bottomland Hardwood Forest and Woodland Bottomland Hardwood Forest and Woodland Swamp Marsh and Wet Herbaceous Vegetation Marsh and Wet Herbaceous Vegetation Digital Database Data collected for use in the ecological classification and natural resource inventory were assembled in a functional spatial database. These data came from a variety of sources, including the US Environmental Protection Agency (EPA), Missouri Department of Conservation (MDC), Kansas Biological Survey (KBS), MARC, Missouri Spatial Data Information Service (MSDIS), Kansas Data Access and Support Center (DASC), Natural Resources Conservation Service (NRCS), and local governments) and were obtained in digital form, or digitized by AES if the information was critical (e.g., Missouri Natural Features Inventory reports). Field data gathered by AES also were incorporated into this spatial database. All data were rectified to a base projection. In addition to field information, orthorectified aerial photography (2001), and land cover data, the following data were assembled: MARC planning area boundaries Municipal, state, and federal jurisdictional boundaries Floodplains and flood-prone areas Streams and water resources (including National Wetland Inventory (NWI) wetlands and ponds) MARC s files of roads and similar information Contours and other topography features Natural resource inventories and rare natural features locations Soils, including hydric soils In addition, a subset of information was also collected in Jackson County where stream assessments were completed for three demonstration watersheds. Three separate reports, currently being finalized, contain information from Jackson County stream assessments. Information gathered for these assessments included the sources cited above, as well as field verification of problem areas; opportunity areas for the treatment of stream erosion, flooding, and poor water quality; and stream-related physical data necessary for modeling. A summary of data conversion methods, data applicability, and related information is provided in Appendix A. 5

12 2.2. Field Reconnaissance The objective of field reconnaissance was to obtain information to verify land cover and species and, if necessary, revising the ecological land cover classification and resulting mapping. By evaluating the condition of selected natural resources, an understanding of the conservation needs in the Kansas City metropolitan area would be achieved. The field reconnaissance was completed in two phases: an initial field reconnaissance, and a second, followup field reconnaissance. The first field reconnaissance was performed in Jackson County, with the intent of extrapolating the findings to the rest of the region. The second field reconnaissance was conducted throughout the Kansas City region at plant communities where classification questions remained. The field reconnaissance involved visiting locations that represented the entire county and region, as well as to locate significant and rare natural resources, such as remnant prairies and undisturbed forests. The results of the field reconnaissance were used to create an initial ecological land cover classification and natural resources inventory for the region First Field Reconnaissance The first field reconnaissance was completed September 15-18, Five AES staff and one employee of Patti Banks and Associates (PBA) formed two teams and visited over 150 locations in Jackson County. The following data were obtained at locations that were used as references for revising the ecological land cover classification: County Site Number AES Type 1 AES Type 2 Canopy Dominants (trees collectively comprising 90% of highest vegetation layer) Subcanopy Dominants (trees/shrubs collectively comprising 90% of layer below canopy) Groundlayer Dominants (trees/shrubs/herbs collectively comprising 90% of <1m vegetation) Other Species (common species, species indicative of site diversity, rare species, etc.) Condition Rank (based on level of disturbance and quality of community) Notes (remarks that further describe the site) Soil Series (from soils map) Kansas/Missouri GAP Type (from GAP land cover map) Stewardship Problems (invasive species, fire suppression, over-grazing, erosion, etc.) Locations of visited sites were digitized and the field information entered into a database. Prior to visiting sites in the field, field maps were prepared. Separate maps showed AES Type 2 ecological land cover, GAP land cover, and soils data on a base of streams, roads, municipal boundaries, and contours. These maps initially were produced on paper at a 1:2,000 scale. At each site in Jackson County the GAP type was checked and the edge of the GAP polygon was examined relative to the aerial photographs. This provided more detailed understanding of the limitations of the GAP classification and the changes necessary to improve its quality. 6

13 2.2.2 Second Field Reconnaissance After evaluation of the data compiled from the first field reconnaissance, two AES teams convened in Kansas City during the period of November 18-21, 2003 to complete the second field reconnaissance. Staff sought to represent all habitats, slopes, aspects, and geographical locations in the Kansas City region for those natural communities where classification questions remained. During the second field reconnaissance, all data were placed in digital form on laptop computers carried in the field. Digital information could be displayed on orthophotographs from which field staff selected sites and navigated to them. With each site inspected, staff digitized the location in the GIS program, and entered data into the database linked to that digital location. Other data were written on field forms for later data entry into the GIS program. In addition, AES searched for natural communities in good ecological condition, although project constraints did not permit region-wide ground inventory work. AES planned to add these better condition natural communities to the reference sites already assembled from previous inventories in order to later contribute to a conservation plan for the Kansas City region. AES teams checked over 150 locations outside Jackson County. AES digitized these locations and entered field data in the spatial database. These field data were combined with previous field data, Heritage program field data, and Natural Resources Inventory data collected by others. These More than 300 sites in the Kansas City region were evaluated during the first and second field reconnaissances. digitized field data established a set of reference stands to aid future classification and inventory work, and help to establish conservation priorities in the Kansas City region Field Reconnaissance Data Evaluation Correlation of the field information on ecological conditions of natural resources to the aerial photography available for the region is essential for successful data evaluation. To accomplish this, ecological conditions were ranked at many locations using a pre-determined ranking system (Table 2-2). Aerial photography was then examined to determine if the ecological conditions seen in the field were readily apparent on the aerial photography. In general, it was found during the two field reconnaissances that the condition of natural resources in the field was not readily apparent on aerial photography, except for the more mature forests and remnant prairies. The more mature forests were generally classified as forests in GAP, while less mature forests were classified by GAP as woodlands. It was also determined, using soils information, that most of the GAP woodlands in the Kansas City metropolitan area have developed where prairie grew 150 years ago. As such, they represent forests of poorer ecological conditions compared to forests which have existed at the same location for a more substantial period of time, as is the case for the more mature forests of the 7

14 region. Forests on historical forest and savanna soils are significant and deserving of conservation because they represent the original forest conditions of the region. Table 2-2. Ecological Condition Ranks (letters may be combined to suggest a range of conditions) A Excellent. The plant community is intact and fully functional. Its soils, vegetation structure, plant diversity, and ecosystem functions have not changed substantially for decades, if not centuries. B Good. The plant community is intact and functional, but soils, vegetation structure, and plant diversity were slightly modified by land use during the past century or more. C Fair. The plant community is damaged and has lost species and ecosystem functions as a result of incompatible land use during the past century or more. Its vegetation structure can be quite different from that existing in the plant community over a century ago. D Poor. The plant community is highly damaged by long-term incompatible land use, but is still recognizable as a plant community of the type. Soils are usually intact, but vegetation structure has changed greatly, and many species have been lost and replaced by non-native invasive species. NR Not Ranked. The location is culturally created (e.g., residences, cropland, orchard, pasture, and impoundment). Generally, the poor-quality forests became established on soils that were transitional between typical grassland and typical forest soils. It was also found that some GAP forests located on these soils were recovering from severe cutting decades ago. These forests contained large honey-locust (Gleditsia tricanthos), an indicator of disturbance, rather than large oak (Quercus) trees encountered in the more mature forests on forest soils of the region. Generally, a predominance of oak in a forest of the Kansas City region indicates better ecological conditions. Remnant prairies were also evident on the aerial photography by the color and the texture of the photograph. The GAP classification and inventory work did not accurately locate remnant prairies. Remnant prairies visited during AES field reconnaissance were discovered by AES or already known through previous inventories. While AES staff identified potential remnant prairies by their color-texture signature, project constraints limited opportunities to locate and visit all remnant prairies in the Kansas City metropolitan area. To assure that field data collected for the natural resource inventory were consistent with accepted standards, a meeting of technical experts and other stakeholders from the region was held October 23-24, 2003 in Kansas City. Meeting participants evaluated the ecological land cover classification and natural resource inventory methods of this project, and contributed information on important natural resource areas known to them. The stream assessments performed in Jackson County were discussed, and suggested modifications were provided at this meeting and in separate meetings with federal and state regulatory agency personnel and Jackson County staff. The outcomes of the first field reconnaissance and this meeting provided an initial ecological land cover classification as described below. 8

15 2.3 Initial Ecological Land Cover Classification and Natural Resources Inventory (Jackson County) The initial ecological land cover (ELC) classification was created and the natural resources of Jackson County were mapped in November This map and classification formed the basis for developing the eight-county Kansas City Region map used in the second field reconnaissance. Detailed information on the creation of this classification and map are provided in Appendix A, but a brief summary follows Aquatic Communities Aquatic Communities include areas of open water for a substantial part of the year. The information was derived from NWI data that was incorporated into the project database. Polygons coded as open water or aquatic bed were selected from the NWI data set and mapped. These polygons were verified and corrected against the digital orthorectified aerial photography. In addition to using digital mapping information, stream assessments were completed in Jackson County for the Bur Oak, Little Cedar, and Round Grove Creek watersheds. While this information was not directly used for the creation of the regional ELC classification, it was used to establish the methodology that could be used for completing future stream assessments in the region. Also, the reports provided a template for completing similar work in the region. Stream assets are numerous throughout the Kansas City metropolitan region These assessments identified the critical issues which related to stormwater and natural resource management and simultaneously developed strategies and actions to address those critical issues. The reports will be available from Jackson County when completed Natural Communities Natural communities are recognizable as examples of plant communities that existed in the Kansas City metropolitan area for several thousand years. Many of these plant communities have experienced some alteration due to land use practices, but still retain many of the same native species, exhibit much of the same vegetation Natural communities of the Kansas City region have existed for several thousand years, and they still retain many of the same native species and provide similar wildlife habitat and ecological functions as they would have 150 years ago. structures, and provide similar wildlife habitat and ecological functions (such as water retention and infiltration) as they would have 150 years ago. Natural communities are the focus of inventory work because, due to their rarity in the Kansas City region and the great expense and difficulty of replacing them, they will become the chief object of conservation and protection work. Natural communities in the Kansas City 9

16 metropolitan region are used by the Missouri and Kansas state governments, and local nonprofit conservation groups (e.g., The Nature Conservancy) in their own natural resource inventories. Using natural communities is a way to create a common language among parties interested in natural resource conservation, protection, and use. The following categories of natural communities were mapped in this report Forests An accurate forest layer was created for Jackson and Johnson Counties from recent aerial photography (circa 2001). Forests in the other six Kansas City region counties were classified using the AES classification and GAP data. Forest boundaries created from GAP data do not match recent aerial photography and date from 1991 in Missouri and the late 1990 s in Kansas. In this study forest classification and natural resource inventory work for Jackson and Johnson Counties are more accurate than the forest work in other counties. Jackson and Johnson Counties provide a model and goal for future classification and inventory work in the other six counties. Oaks are often dominant in upland settings, with the addition of other tree species in lowland settings. Layers of tree saplings and shrubs grow beneath the uppermost tree canopy. Forests in good ecological condition have a groundcover comprised of tree seedlings, shrubs, and native wildflowers, grasses, and sedges. Forests in poor ecological Forest community on rolling terrain in Leavenworth County, Kansas condition have few if any oaks in the tree canopy, few native plant species in the ground layer, and an abundance of non-native or highly invasive woody plants beneath the tree canopy [e.g., buckbrush (Symphoricarpos orbiculatus), Tartarian and Japanese honeysuckles (Lonicera tatarica and L. japonica). Forests identified by the GAP classification work ranged from good to poor ecological condition, and up to 50% of the forest boundaries for some polygons locations did not match the orthorectified aerial photography. In addition, the GAP data were developed from 1991 imagery, and therefore were out-of-date. Development has since destroyed or altered some of the forests. To rectify this, AES completed an unsupervised classification of 2001 aerial photography in Jackson County. A detailed explanation of this process is provided in Appendix A. AES then edited this unsupervised classification to create a forest land cover layer. Three sources of error were encountered. First, the unsupervised classification selected tree shadows adjacent to forests and included them in the forest layer, resulting in slightly unmatched forest polygon boundaries. Second, old fields being invaded by brush and young trees were also selected as forest areas. Third, areas of moist soils that appeared dark on aerial photographs were also selected. AES staff removed these errors by hand-editing the unsupervised classification. 10

17 While this method created a very accurate and up-to-date forest layer in Jackson County, it proved too expensive to repeat in the remaining counties. Johnson County recently completed a land cover classification and inventory which resulted in a forest layer with accurate polygon boundaries. AES classified the Johnson County forest layer polygons using GAP data. For example, a Johnson County forest polygon that was classified as deciduous forest by GAP was called deciduous forest. The forest classification and inventory was further refined in Jackson and Johnson Counties using soils data from the USDA Natural Resources Conservation Service (NRCS) soil surveys (Appendix B). Soil conditions have a direct effect on the types and quality of vegetation that will grow in different areas. Different NRCS soil units correspond to different AES soil types, which in turn are typically associated with certain natural communities. Association of soils to vegetative communities was tested and employed in Jackson and Johnson Counties, but was not used in the other six counties. Forest soils (technically hapludalfs, soils that are transitional between grassland and forest) correspond to upland forests in the Kansas City region. Grassland soils were upland prairies 150 years ago, and today support a variety of natural communities. Present-day forests that established on grassland soils typically are of less ecological quality than forests that have grown on soils that favor high quality trees. Similarly, existing prairie remnants on grassland soils are usually in better ecological condition than those not on grassland soils. As an example, soils that are shallow, such as soils of the Oska series, once supported glades, savannas, prairies, and other rare natural communities. Today many of these soils support poor-quality forest or disturbed pastureland. Soils in the Kansas City region that are Lack of regular burning and/or haying of wet occasionally flooded are historically capable of prairie vegetation has resulted in succession supporting wet prairie, a valuable vegetative to lowland hardwood forest and woodland. resource. The lack of regular burning and/or haying of wet prairie vegetation over the past several decades, however, has resulted in succession of these areas to lowland hardwood forest and woodland. Some soils that are occasionally or continuously flooded are classified as hydric. Hydric soils support wetlands, typically marsh and wet herbaceous vegetation, including shrubby wetlands. In Jackson and Johnson Counties, the digitized forest layers were subdivided into deciduous forest, lowland hardwood forest and woodland, and deciduous woodland/immature forest based on these soils data and GAP data. One GAP forest type (mixed evergreen-deciduous forest) corresponded to former savanna areas and was identified in the other six counties, but not in Jackson or Johnson Counties Lowland Communities The trees of lowland hardwood forests are primarily non-oak species, although bur (Quercus macrocarpa) and swamp white oak (Quercus bicolor) both occur, and even occasionally chinkapin oak (Quercus muhlenbergii). In their condition 150 years ago, these forests resembled savanna, with scattered large bur, swamp white, and chinquapin oaks, cottonwood (Populus deltoides) and sycamore (Platanus occidentalis). In the absence of fire, other trees [e.g., American elm (Ulmus americanus), hackberry (Celtis occidentalis), honey-locust, and green ash (Fraxinus pennsylvanicus)] have filled canopy gaps and overtaken the wet meadows formerly located in these areas as well. Remnant wet meadows are dominated by grasses, sedges, and various wildflowers. Poor ecological conditions are indicated by an abundance of reed canary-grass (Phalaris arundinacea), invasive tree species, and poor groundlayer diversity. 11

18 These natural communities are found on occasionally flooded and hydric soils. As described above, the lowland hardwood forest was mapped where recently flooded or hydric soils overlapped with up-to-date forest layers in Jackson and Johnson Counties and with GAP forest data in the remaining six counties. Marsh and wet herbaceous vegetation polygons were derived from NWI data and GAP data. Included are NWI polygons coded as emergent wetland, wet meadow, and shrub swamp. Marsh and wet herbaceous vegetation were required to be located on occasionally flooded or hydric soils Grassland-Savanna-Woodland Communities These rare natural communities proved the most difficult to locate using this methodology. Included in this group are limestone glades, sparsely vegetated natural communities with scattered oaks [chinquapin, post (Quercus stellata), and blackjack (Q. marilandica)] and eastern red cedar (Juniperus virginiana). A variety of native wildflowers, grasses and sedges grow in limestone glades. This plant community often contains species that are rare in the Kansas City region. Upland prairie is well known, but also rare in the Kansas City region, along with the species, which inhabit it. In this study, AES uses the term grassland to indicate natural communities recognizable as native prairie, but in addition former prairies that have been interseeded with tall fescue (Festuca arundinacea) or planted to smooth brome (Bromus inermis), yet still may support scattered native prairie plants and function as hayfield or pasture. Prairie in the best ecological condition would be located on grassland soils. Savannas are comprised of a mixture of trees, shrubs and groundcover, but require fire, light grazing, or haying to persist in recognizable form. Prior to 1850, prairie and savanna blanketed the Prior to 1850, prairie and savanna blanketed the majority of the Kansas City region. Today they are among the region s rarest natural communities. majority of the Kansas City region. Today they are among the region s rarest natural communities. Conversion to cropland, heavy grazing by livestock, and the lack of fire have altered over 95% of the region s original prairies and savannas so that today they are not recognizable as these natural community types. Because of the difficulty of locating these communities, additional, intensive inventory work is necessary to locate remnant prairies, savannas, and limestone glades in the Kansas City region. Locations for these natural communities reported by previous inventories were digitized or added to the project s spatial database, and then used by AES to specify the site conditions where these communities would mostly likely be found. AES also located remnant glades, prairies, and savannas for inclusion in this spatial database. Limestone glades were most likely to be found on shallow soils (e.g., Oska) identified as deciduous forest or woodland in GAP data. On aerial photographs, patches of light-colored 12

19 limestone and sparse vegetation were visible among small clumps of trees and shrubs. Prairie was found on grassland or thin soils and had a characteristic dried-blood color and finely-dotted texture. Winter wheat fields, old fields with abundant goldenrod (e.g., Canadian goldenrod [Solidago canadensis) and tall goldenrod (S. gigantea)], and some recently mowed areas had a similarly-colored signature, but a smoother or coarser texture than native prairies. CRP fields resembled native prairie on aerial photographs and required a site visit to determine the true origin of the grassland. AES developed a method to identify potential savannas in the Kansas City metropolitan area. Potential savannas are located on forest soils or thin soils. In addition, they are located on slopes (>6%), which face southerly and westerly (from 135 to 315 degrees). They are often classified in GAP as deciduous forest or mixed evergreen-deciduous forest. Even if current aerial photography indicates a potential savanna site is forested, those sites often are tinted reddish due to the underlying groundlayer vegetation. These locations have the greatest potential of any location in the Kansas City region for restoration to ecological conditions resembling those of 150 years ago. Restoration would involve careful thinning of selected trees and prescribed burning. Site visits would be necessary to determine the potential of the groundlayer to recover its native sedges, grasses, and wildflowers. Sites that have been heavily grazed for decades have the least potential to recover a native groundlayer. Restoring extensive Restoring extensive savannas would improve runoff infiltration, provide habitat, and create beautiful open woodlands. savannas would improve runoff infiltration above stream valleys, provide habitat for uncommon species, and create beautiful open woodlands for recreation and aesthetic enjoyment Cultural or Sparsely Vegetated Land Cultural or sparsely vegetated lands do not typically include natural communities, but rather combinations of plant life and built surfaces that are used intensively by people. They include cropland, cool-season grasslands, urban and suburban areas, and highly disturbed lands Developed Land Because development of land has progressed at a rapid rate in the Kansas City metropolitan area in the decade since the Missouri GAP mapping was completed (Kansas GAP data is more recent), AES decided to update information on developed lands in the Kansas City metropolitan area. To do this, cadastral data from municipal tax rolls were obtained and modified as Developed land includes land that has been taken out of it s natural condition and is continuously used for the benefit of the human population, including residence, commerce, transportation and industry, and other cultural uses. described in Appendix A. Where cadastral data were not available, 911 emergency response road centerline data were used to identify developed areas (Appendix A). This assumes that, outside the developed areas in municipalities, these small parcels resulted from subdivision of agricultural land and therefore contained or would soon contain a built structure. There is an ecological justification for this decision. Over the span of a decade or more, the presence of a home, cabin, commercial building, or even a staging area or parking lot within or adjacent to natural communities causes those natural communities to become less viable for certain wildlife species, as well as for native plants. The long-term survival of many wildlife and 13

20 plant species in the Kansas City region requires large, continuous blocks of natural communities. Without careful planning and long-term stewardship, the cultural and built environment leads to the introduction of non-native aggressive plants [e.g., Japanese honeysuckle, garlic mustard (Alliaria petiolata), sericea lespedeza (Lespedeza cuneata)]; shifts the balance of habitat to favor small predators (e.g., house cats, raccoons, skunks, and crows), which eat birds, eggs, reptiles, amphibians, and other wildlife; and prevents the execution of prescribed burning and other land management due to the proximity of real estate and development Urban Forest In Jackson and Johnson Counties small parcels of undeveloped forest land were mapped as urban forest. These forests have value to local residents, despite having lower ecological value than large blocks of forests in rural areas. AES split these forests from other forests using parcel data in order to elevate the value of these forests in development areas Other Cultural Types These include three types of land cultural grassland, cultivated land, and agricultural lands. Cultural grassland is cool-season, non-native grassland that largely lacks native species, and also cropland that has been abandoned and has been colonized by weedy plants, shrubs, and trees. Cultivated land is currently cultivated to crops that are harvested each year. Agricultural lands are primarily used for a mixture of agricultural uses and also small patches of natural communities, and they may contain buildings. AES derived these layers from GAP data, parcel data, and as the negative of the natural communities already mapped Extrapolation to Eight-County Kansas City Region Using the classification described above, AES mapped the eight-county Kansas City region prior to conducting the second field reconnaissance. This classification and ecological land cover map used in the field tested the accuracy of data that AES staff collected during the second field reconnaissance. Based on this classification, and following the second field reconnaissance of the entire region, a final classification was produced Final Ecological Land Cover Classification and Natural Resources Inventory (Kansas City Region) Results from the second field reconnaissance were used to refine the initial ELC classification (Table 2-3). Using the final ELC classification, the natural resources of the eight-county Kansas City region were mapped as shown in Figure 2-1, and are discussed in the next section. Table 2-3. Ecological Land Cover Types for the Kansas City Region Aquatic Communities Natural Communities Lowland Open Water Standing water for a significant part of the year Lowland Hardwood Forest & Woodland Open to closed forest canopy in flooded or wetland areas Marsh & Wet Herbaceous Vegetation Wetlands without a closed forest canopy; includes brush and scattered trees 14

21 Table 2-3. Ecological Land Cover Types for the Kansas City Region Natural Communities Forest Natural Communities Grassland-Savanna-Woodland Cultural or Sparsely Vegetated Land Unclassified Insufficient data to classify Deciduous Forest Mostly closed canopy of deciduous trees, often mature; includes former savannas on south to west slopes. Mixed Evergreen Deciduous Forest Open to mostly closed canopy of junipers and deciduous trees; often oaks; may include former savannas Deciduous Woodland/Immature Forest Open canopy of deciduous trees; often immature; may contain former savannas, or glades on soils with bedrock close to surface. Grassland Grassland, often containing native wild plants; may include CRP plantings. Urban Forest Deciduous canopy cover within an urbanized location (specific to Jackson and Johnson Counties) Cultural Grassland Grassland of planted domesticated grasses, or formerly cultivated land reverting to grassland and sometimes brush. Agricultural Land Used as farmland (specific to Jackson and Johnson Counties). Cultivated Land Used as cropland. Developed Land Urban and suburban land uses, including homes, businesses, roads 15

22 Figure 2-1 Ecological Land Cover Map Kansas City Natural Resource Inventory 1

23 3. Ecological Land Cover Classification and Natural Resource Inventory Outcomes The ELC classification and the NRI have provided an updated source of data for the Kansas City region that can be used by cities and counties within the metropolitan region for planning of community growth and development while also planning for resource conservation. The results of this study and the use of the outcomes of this project, primarily the Ecological Land Cover maps (including individual maps of the numerous natural resources, parks, green spaces, and infrastructure systems) are presented in this section. 3.1 ELC/NRI Results The distribution of ecological land cover is illustrated in Figure 3-1 below, the Ecological Land Cover Map, and is summarized in Table 3-1 on the following page. Mixed Evergreen Deciduous, 0.42% Marsh/ Wet Herbaceous Vegetation, 1.70% Low land Hardw ood Forest Woodland, 3.06% Grassland, 1.80% Developed Land, 20.42% Open Water, 1.61% Unclassified, 0.12% Urban Forest, 2.16% Agriculture, 0.55% Cultivated Land, 27.43% Agriculture Cultivated Land Cultural Grassland Deciduous Forest Dec. Woodland Immature Forest Developed Land Grassland Lowland Hardwood Forest Woodland Dec. Woodland Immature Forest, 4.45% Deciduous Forest, 8.88% Cultural Grassland, 27.41% Marsh/ Wet Herbaceous Vegetation Mixed Evergreen Deciduous Open Water Figure 3-1 Distribution of Ecological Land Cover in the Kansas City Region Unclassified Urban Forest In general, the data show that the region s undeveloped land is dominated by land used for agricultural purposes, a reflection of the region s history, and for human dwellings and businesses. About 75% of the region s land area is dedicated to crops (cultivated land), non-native pastures and hay meadows (cultural grassland), and residences and businesses (developed land). Land utilized for agriculture is well distributed in the region; however, a significant amount dominates the land cover in Cass, Platte, and Ray Counties in Missouri, and Leavenworth County in Kansas. While agricultural lands make up most of the ecological land cover in the region, and urban or developed lands also occupy much of the land in the core urban areas, 22 percent of the area, or approximately 538,000 acres (840 square miles), have been identified as areas that may retain natural or near-natural ecological conditions and present opportunities for conservation and restoration. 17

24 Table 3-1 Ecological Land Cover in the Kansas City Region Leavenworth Percent Ecological Jackson MO Clay MO Cass MO Platte MO Ray MO Wyandotte KS Johnson KS Classifi- KS Total of Acres Total cation acres % acres % acres % acres % acres % acre % acres % acres % Area Open Water 8,730 2% 7,702 3% 6,463 1% 5,677 2% 5,413 1% 1,125 1% 2,240 1% 2,147 1% 39, % Lowland Hardwood Forest & Woodland 22,866 6% 6,039 2% 13,375 3% 10,845 4% 5,753 2% 1,618 2% 8,766 3% 5,870 2% 75, % Marsh/ Wet Herbaceous Vegetation 2,175 1% 4,978 2% 2,333 1% 8,955 3% 6,870 2% 3,469 3% 8,305 3% 4,562 2% 41, % Deciduous Forest 15,294. 4% 17,078 6% 25,737 6% 27,034 10% 37,211 10% 16,164 16% 67,927 23% 11,643 4% 218, % Mixed Evergreen- Deciduous 4, % 0 0% 368 0% 0 0% 0 0% 0 0% 0 0% 5,138 2% 10, % Deciduous Woodland/ Immature Forest 31,953 8% 9,351 4% 19,591 4% 11,156 4% 19,947 5% % 5,714 2% 10,9010 4% 109, % Grassland Urban Forest 900 Cultural Grassland 60,407 0% 0 0% 7, % 0 0% 0 0% 1,825 2% 19,527 7% 13,985 5% 44, % 43, % NA NA NA NA NA NA NA NA NA NA NA NA 9,776 3% 53, % 15% 94,332 36% 188,165 42% 82,698 30% 132,690 36% 7,939 8% 70,183 24% 37,011 13% 673, % Agriculture NA NA NA NA NA NA NA NA NA NA NA NA NA NA 13,619 5% 13, % Cultivated Land 78,107 20% 47,878 18% 128,874 29% 104,227 38% 154,996 42% 11,263 11% 88,460 30% 60,008 20% 673, % Developed Land 125,037 32% 75,594 29% 56,733 13% 26,778 10% 8,337 2% 55,529 56% 23,551 8% 130,154 44% 501, % Unclassified Total Acreage % NA NA NA NA NA NA NA NA NA NA NA NA 2,696 1% 2, % 393, % 262, % 449, % 277, % 371, % 99, % 294, % 293, % 2,456, %

25 Many opportunities for conservation exist for ecological land cover areas as defined in Section 2 and in Table 3-1, including deciduous mature and immature woodland/forest, lowland hardwood forest with herbaceous vegetation, mixed evergreen and deciduous woodland, and, particularly characteristic of this area, native grasslands, marshes, and aquatic systems that contain wetland vegetation. Many of these areas are currently located in rural or semi-rural areas along river and stream corridors, near open water, and in non-forested upland areas. These features present opportunities not only for parkland and recreational use, but also for promoting the conservation of important ecological systems that both represent our environmental heritage and provide valuable resource benefits to our communities. Review of the ELC map indicates that several natural communities exist throughout the region. The current data shows numerous areas in Leavenworth and Johnson Counties in Kansas that have large tracts of native communities. Wyandotte County, Kansas, while largely developed for urban utilization, also contains several large tracts of deciduous forest. While the Missouri counties (Platte, Ray, Clay, Jackson, and Cass) show large areas dominated by land used for agriculture and/or urban development, the original classification of the data from the USGS GAP, may not reflect many areas and tracts that contain native vegetation that provide opportunities for conservation because the data collected several years ago. In fact, the Missouri GAP data don t list native grasslands when, in fact, many areas of native prairies exist within the four counties that are of high enough quality to be considered for conservation and restoration. An example of how this data may be used is provided in the next section. By county, conservation areas may include the following: 1. Cass County, Missouri: The land cover data show that most of the county is in cultivated or cultural grasslands (71 percent). The data show, however, that 16 percent of the land cover is lowland hardwood and upland deciduous forest, with another 7,868 acres, or 2 percent of the land area, in native grassland. These resources may be considered for conservation, particularly near many of the growing communities and along stream and rivers in the county. Several areas with native vegetation exist throughout the Kansas City metropolitan area that are worthy of conservation. 19

26 2. Clay County, Missouri: The ecological land cover data indicates that 14 percent or more of the land may have conservation value, particularly large areas of lowland hardwood and upland deciduous forests. The data show that the majority of the county is either developed or in cultural grassland, but areas of native grassland probably exist throughout the county that may be considered for conservation. 3. Jackson County, Missouri: The data show that extensive areas of this county are developed. However, large areas of forested land covering 18 percent of the county provide many conservation opportunities, some of which have already been undertaken by the County in their planning process. The data also show large amounts of urban forest that are valued by citizens of the county in parkland and in border areas, and 2,175 acres of wetland vegetation that can be considered in the planning process. 4. Johnson County, Kansas: Many opportunities for conservation of woodlands and prairies exist throughout Johnson County, especially in the fringes of suburban development. An estimated 19 percent of the land cover in Johnson County is listed as lowland hardwood and upland forests and native grasslands. Many of these areas are near streams and rivers that drain the county and present opportunities for parkland, buffer areas, and environmental quality. 5. Leavenworth County, Kansas: The ecological land cover data show that Leavenworth County has extensive areas with native vegetation that can be considered for conservation through the planning processes as this county grows in population. The data indicate that 38 percent of the land area in Leavenworth County, or 52,099 acres of land are mapped as lowland hardwood forest, deciduous forest (mature and immature), marsh/wetland vegetation, and as native grassland. In particular, 23 percent of the land is shown as deciduous forest, a resource that provides opportunities for maintaining environmental quality, water quality, stream buffers, and parkland throughout the county. 6. Platte County, Missouri: Twenty-one percent of Platte County is identified as lowland hardwood, deciduous (mature and immature) forest, and marsh/wetland vegetation. The nature of the land cover in the county would suggest, however, that many areas of native prairie also exist and should also be considered for conservation. Many of these areas (woodland forest and prairie) may be along the Missouri River and its tributaries that are in Platte County. 7. Ray County, Missouri: Ray County is dominantly rural (only 2 percent developed land) with 78 percent of the land cover listed as cultural grassland and cultivated land. Nineteen percent of the land cover, however, is identified as forested or marsh/wetland vegetation. Like the other Missouri counties in the Kansas City metropolitan region, the rural, undeveloped nature of the county suggests that many areas of native prairie exist, and together with forested land, there are many opportunities to conserve these native areas and utilize them for enhanced environmental quality, parkland and recreation, and wildlife/game management. 8. Wyandotte County, Kansas: The majority of Wyandotte County is developed land (56 percent), and 19 percent is identified as cultural grassland or cultivated. This smallest of the metropolitan area counties (with 99, 620 acres total land area) also contains nearly 24,000 acres of land identified as woodland/forest (19 percent), marsh/wetland vegetation (3 percent), and native grassland (2 percent) with conservation opportunities particularly in the western and northwestern portions of the county along streams that drain to the Missouri River. Regionally, conservation opportunities often extend beyond county and community boundaries, suggesting that to optimally conserve resources and provide enhance environmental quality, 20

27 local governments and conservations groups may seek to work together in formulating long term plans that include conservation and restoration measures within watersheds in the Kansas City area. 3.2 Ecological Land Cover Conservation The importance and value of conserving natural resources and the ecological systems that remain in the Kansas City region is vital to the future success of our communities. One of the most important aspects of ecological conservation is protecting our natural resources, including native vegetation and wildlife. Protected habitat assures that plant and animal species will continue to live in their native habitats and persist in the Kansas City region. Conservation of native vegetation and animal species also protects our environment and communities from unwanted invasive vegetation. Conserving natural resources is not limited to retaining spaces where native plant and animal communities can thrive. Conserving the existing ecological land cover provides substantial value and benefits for quality and protection of land planning and utilization. Such benefits include: Development of buffer regions around and between communities in order to provide natural breaks in land use. Buffers may include parks, riparian and native area green corridors, conservation areas, and residential conservation development. Utilization for parks and green space which serve recreational and education purposes. Parks can incorporate native landscape features, including aquatic, geological, and vegetation features that can provide a living monument to the original ecosystems of the Kansas City region. Within the parks, pathways and open areas can allow residents to observe vegetation and wildlife. Improving land value by increasing aesthetic and land use desirability. Properties adjacent to conservation areas hold higher value because of their desirability for working, playing, and living near. Providing buffers for stormwater management and flood prevention. Opening areas up to provide natural attenuation of stormwater allows the natural hydrologic cycle to be more balanced, lessening the risk of damaging floods while preserving the flow and aquatic wildlife of streams and rivers. Improving environmental quality by filtering pollutants from waters and air. By conserving native vegetation, a natural filtering system that captures sediments and pollutants before they enter streams, thereby improving water quality. Similarly, conservation of vegetation provides traps of airborne particulates and pollutants while also reducing the volume of noise. Reducing water and energy consumption. Conservation of native vegetation reduces water consumption in multiple ways, including: Native vegetation increases infiltration of stormwater where it is then stored in the soil, minimizing the need to continually water open areas. 21

28 Native vegetation is adapted to Kansas City s climate, and requires less water to sustain a healthy plant community. Native vegetation reduces evapotranspiration because its use of water is more efficient than non-native vegetation. Conservation areas also reduce energy consumption by lowering the ambient temperature of surrounding areas and requiring less maintenance. Of these benefits, the last three may provide the most distinct and important values to our communities in terms of protection of property and our environment, reduction in maintenance costs, and sustaining the quality and quantity of water. 3.3 Establishing Regional Conservation Priorities The ecological land cover map produced and described in this technical report will be used in combination with previous inventories, AES field data, feedback from local experts, analyses (e.g., stream buffer requirements), connectivity information, public lands, and other information from local agencies, to produce a map of conservation priorities in the Kansas City region. The ecological land cover map reflects priorities to some degree. For example, all forest communities classified as Deciduous Forests in Jackson and Johnson Counties are likely to be more mature than those classified as Deciduous Woodland/i\Immature Forests due to their origination on native forest soils as opposed to the latter s more recent origination on grassland soils. Furthermore, these immature woods are likely to be of lower ecological quality as they occur on soils not suitable for forest growth. AES learned that previous inventories, which identified high quality sites, have become outdated. This is to be expected in a rapidly developing region. For example, the Missouri Natural Resources Inventory work completed in the 1970 s and 1980 s now contains sites, which have been developed, or have deteriorated due to incompatible land use or neglect. The results of the NRI are available for conservation planning throughout the Kansas City metropolitan area. The information, including data for use in GIS, includes maps that show conservation elements for the entire metro area and for each county. GIS layers that show the various natural resource features of each county have been developed to clearly show where natural resources and various ecological features exist in relation to urban and agricultural areas, as well as to current Kansas City green corridor features. Other layers show existing parks and trails to demonstrate how these areas can be maximized for planning decisions regarding buffer areas, park development, and other conservation measures, as well as for determining how communities can develop around these areas. Figure 3-2, on the following page, demonstrates an example of how using the database and GIS can illustrate the ecological features of a county in conjunction with community boundaries, roadways, parks, and Kansas City MetroGreen corridors that can be used for planning purposes. Using these features, planners can map the locations of valuable conservation areas and prioritize their conservation based on their proximity to other features such as parks, MetroGreen corridors, and surrounding geographical features, including roads and surrounding communities. Planners can utilize this information to maximize existing natural resources in the planning process, including how natural resources can be used for stormwater management, 22

29 Figure 3-2 Ecological Land Cover and Natural Resources Inventory for the Kansas City Region Example Natural Resource Conservation Planning Map (Johnson County, KS) 20

30 where future or existing parks can be planned, and how development can take place with conservation of native vegetation, aquatic features, and soils to retain the Kansas City region s native heritage. 3.4 Public Outreach and Education This program includes public outreach to provide an outlet for the Kansas City Natural Resources Inventory information and database. The public outreach program consists of educational sessions to describe the data collection and compilation process, including development and utilization of the GIS program for communities throughout the region to use in the planning process. In addition, presentation of brochures and posters will enable public agencies to view and utilize the NRI information as an integral component of their community planning process. The public outreach program includes the following: 1. GIS Tutorial: A tutorial that provides users information about the data compiled for this project access to the GIS features that can and should be explored for its full utility and application for planning processes. 2. Informational Brochures: Brochures and supporting text on the topics of stormwater management, big rivers conservation, greenways and parks, and community planning. These brochures can be used by MARC and local communities for explaining the ELC and NRI process and use. 3. Power Point Program: A power point presentation summarizing the natural resource inventory work can be reviewed that presents ideas on how interested parties can become involved in protection, restoration, and stewardship of these resources. 4. Case Studies: Project-specific examples of conservation, stormwater management, restoration, and management activities in the Kansas City region include a brief summaries of the work involved and cost for various projects, and will demonstrate how conservation measures are already working within the region. These case studies will be presented during public meetings and education opportunities to demonstrate how conservation measures are already working. 5. Meetings: AES and PBA will assist MARC in meetings for stakeholders in the Kansas City region. The intent is to summarize the results of this project and to seek input from experts, interested parties, and others. All outreach materials will be presented and provided at these meetings. 25

31 4. Summary and Project Evaluation 4.1 Summary Historically, vegetation in the Kansas City region consisted largely of upland and bottomland prairies, with relatively large tracts of forests were present in the bottomlands. The majority of the area was tallgrass prairie on the gently rolling plains that were dissected by deep drainages of the tributaries of the Kansas and Missouri Rivers. Extensive aquatic and wetland communities were present, particularly in the floodplains of the larger rivers and, to a lesser extent, their tributaries. In 2003, MARC initiated the development of a regional map depicting natural resource assets and ecological land features that now exist in the Kansas City metropolitan area, with the objective of developing a comprehensive database and GIS program for use by local governments and planning agencies to incorporate ecological and natural resource features in future development and growth. This project, inclusive of the GIS program and its regional map, represents a natural resource inventory across a broad region that is part of a larger initiative to produce the tools for communities in the Kansas City region to develop conservation plans. The multiple mapping layers depict the region s vegetation resources, natural resource features (including rivers, streams, and other water features), and regional infrastructural features, such as streets and roads, political boundaries, parklands, and the MetroGreen Program, with pathways that connect the regions parks and greenways. These layers, either individually or combined, can be used to evaluate regional natural resources and how they can be conserved and integrated into the development of community plans. To complete field surveys of the region s natural resources, AES created a natural resources classification and inventory method, completed the field surveys, and developed descriptions of ecological land cover types in the Kansas City region. The AES classification and inventory method used previous work by state and federal agencies, or data gathered by MARC and the region s counties and cities. During the field inventory, hundreds of locations were visited by staff of AES and PBA. The inventory revealed a lot about the region s natural resources. For example, the inventory identified 22 percent of the metropolitan region as potentially supporting high quality vegetative communities and natural resources, and that many of these areas support native vegetative communities similar to those found 150 years ago. These areas of high quality vegetative communities present many opportunities for conservation of ecosystems, habitats, and other natural resources that will benefit the overall Kansas City region. The inventory also revealed a tremendous need for ecological restoration work so that stream water quality and fish habitat can be improved, and that wildlife habitat in forests, savannas, and grasslands can again sustain high numbers of native species. The urban areas in the Kansas City region continue to support extensive forests and restorable savannas. These areas largely consist of mature deciduous forests, lowland hardwood forests in stream valleys and river bottoms, marshes and other wetlands, and the native grasslands and savannas that were the dominant natural features of the region 150 years ago. These natural resources are concentrated along rivers and streams, near open water, on steeply-sloping land, and in rural and semi-rural areas. The extensive forests and restorable savannas in urban settings play important roles in providing recreational 26

32 opportunities, serving existing and future parkland, creating buffers for streams, and increasing awareness of the region s environmental heritage. Rural and semi-rural areas in both Kansas and Missouri present large tracks of both forest land and native prairie for conservation and/or restoration. The current data demonstrates numerous areas in Leavenworth and Johnson Counties in Kansas that have large tracts of native communities. Wyandotte County, Kansas, while largely developed for urban utilization, also contains several large tracts of deciduous forest. Platte, Ray, Clay, Jackson, and Cass Counties in Missouri exhibit extensive areas dominated by land used for agriculture and/or urban development, however, the original classification of the data from the USGS GAP, because it was collected several years ago, may not reflect many areas and tracts that contain native vegetation that provide opportunities for conservation. The NRI project has identified and mapped many areas throughout the region that merit conservation or restoration. When managing our environment properly, the effort can result in cleaner water, healthier riparian corridors, and reduced erosion and sedimentation in creeks and other water bodies. Our forests, prairies, wetlands, and shrublands are essential for providing habitat for healthy and viable wildlife populations. These plant communities also improve the aesthetic appeal of the landscape, provide scenic views and green space, provide for recreational opportunities, and protect the soil from erosion. The results of the natural resources inventory and development of an ecological cover map for the Kansas City region provide the tools that will catalyze the next phase of MARC s initiative to address conservation and planning needs in the Kansas City region. Conservation planning not only ensures that plant and animals species continue to thrive in their native habitats, they are vital to the future success of the region s communities. A focus on natural resource conservation and restoration benefits a resident s quality of life by preserving natural breaks in development for nearby recreation and relaxation; by maintaining and raising land values; by improving stormwater management and reducing flood risk; and by enhancing environmental quality. Along the way, conservation and planning will maintain or improve stream flow and water quality, wildlife habitat and populations, and the rich diversity of life forms in the region. If successful, conservation and planning will bequeath the landscape remaining in a near-natural state to the future generations in the Kansas City metropolitan area. This report presents the first steps in developing a comprehensive conservation planning process. Additional data acquisition, mapping, and planning will contribute to a more complete and accurate database that will provide essential tools for the communities in the Kansas City region to use as they plan for the future. Many of these data needs are described below. The base program has been established with this project, and Kansas City can join other major cities in the United States in successfully using natural resources for the enjoyment and benefit of its citizens. 4.2 Existing Data Gaps and Method Development Needs Three significant issues were recognized and addressed by this ELC classification and natural resources inventory for the Kansas City region: 1. A lack of a region-wide ecological land cover classification for natural resource inventory work. 27

33 2. A lack of a recent region-wide inventory of natural resources, and 3. Previous classification and inventory work contained imprecise boundaries of natural resource types, and in their identification. These issues were resolved largely by the creation of this ELC classification and completion of natural resource mapping. Nevertheless, data gaps remain that should be addressed for the future success of this program and for planning in the Kansas City region as a whole. Spatial data in Jackson and Johnson Counties are more recent and accurate than the other six counties. This is due to the special procedures used to refine the existing GAP data. The method provided great detail and accuracy, but required additional resources. For example, the unsupervised forest classification followed by hand-editing in Jackson County required approximately 40 hours of labor. The high labor requirement made this method too costly to employ in the other counties, but provided a model that could be followed in other counties. As explained in Sections 2.2 and 2.3, rare communities--prairies, limestone glades, oak savannas-- were not accurately identified in existing GAP inventories, and while AES staff located remnants of these communities, not all remnants were identified during this inventory. Using parcel data helped to improve the currency of the spatial data, but also introduced errors to the data. This is because the parcel criteria AES used (<10 acres) to identify potentially developable lands included small undeveloped tracts used as recreation land or simply lying idle. The 1990 s GAP classification identified many of these as forest or grassland. These parcels, nevertheless, have a great potential to be developed in the future because they have been subdivided from larger agricultural ownership blocks or are inside highly developed lands. Because of the lack of detailed inventories in counties outside Jackson and Johnson, the information in those counties must be interpreted with care. For example, in Leavenworth County it appears that there are many locations of Grassland. As defined in the ecological land use classification and used in the natural resource inventory work for this project, Grassland is a nearly treeless expanse that likely contains native plant species, and may also include CRP lands. However, in Leavenworth County these areas are not pristine prairies, and in many cases they may be significantly altered by over-seeding with nonnative tall fescue, and applying herbicides to kill broad-leaved prairie plants. Nevertheless, these locations have a greater chance of containing warm-season prairie grasses and native prairie plants than those locations identified as Cultural Grassland. Therefore Grassland is used despite the potential for inaccuracy. As already mentioned, identifying all the remnant native prairies in the Kansas City region would require special inventory work outside the scope of this project. Some data gathered for use in Jackson County stream and watershed assessment work were not used in the development of the ELC classification. These data include floodplain boundaries, stream stability models, and stream buffers. These data will be used in future work to establish conservation priorities in the Kansas City region. AES used its field data and that of the Kansas and Missouri Heritage Programs and Biological Surveys in order to test the accuracy with which the classification and inventory 28

34 could predict ground conditions. These field data represent a small fraction of the surface area of the Kansas City region, and therefore provide only a sample of what can be expected from the classification and inventory work when it is used throughout the region. In general, deciduous forest and woodland ELC data most accurately predict conditions on the ground. AES estimates that in Jackson and Johnson Counties, the classification and natural resources inventory is greater than 90% accurate in correctly identifying land as forested or wooded, and in correctly depicting the boundary of those forests. In the other counties, because GAP was used as a basis for mapping, the accuracy of the boundary is reduced, and the accuracy of the identity of more mature Deciduous Forest was less than 90%. The classification and inventory are less accurate in predicting Lowland Hardwood Forest and Woodland in Jackson and Johnson Counties, and lower still in the other counties. This lower performance is due to the fact that the trees, which comprise lowland hardwood forest, are also capable of invading upland draws and ravines throughout the region. On the ground it may appear that a site is lowland hardwood forest and woodland, but often these sites were prairie or savanna prior to In addition, in many locations the damming of stream segments saturates upland soils, creating lowland hardwood forest conditions where upland conditions existed previously. Detecting prairies, limestone glades, and oak savannas in good condition will require additional detailed inventory work outside this project s scope. This classification and inventory did not predict good examples of these communities with great certainty. However, in Jackson and Johnson Counties, the classification and inventory data identified areas with shallow soil and southern forested slopes, which have the greatest potential to support these communities. Good examples of prairies, glades, and savannas could be located with additional inventory work in the appropriate ecological land cover type and soil type. The classification and inventory predicts wetland occurrence in the field with a reasonable accuracy due to the use of NWI mapping. However, these data are generally more than a decade old, and consequently, the classification does not match wetland boundaries to field conditions well. It is better at identifying the existence of of a wetland. However, many farm ponds and gravel bars in streams were identified as wetlands by the NWI. These were not mapped as wetlands in the ELC cover map produced here. 4.3 Future Data Acquisition and Analysis Opportunities abound to improve and expand on the work completed for this project. To create a more detailed and comprehensive source of natural resources information that can be used throughout the Kansas City region, the following recommendations are made to accomplish the following: Complete an inventory for remnant prairies and limestone glades using color imagery supplemented by field assessment and verification. This inventory will be facilitated by the predictions inherent in the classification and inventory map, as well as the previous and new field inventory data assembled by AES and available with the digital data. Consultation with the KBS on the results of its 2004 tallgrass prairie inventory in Leavenworth and Johnson Counties is also suggested. 29

35 Complete an inventory for the most restorable oak savannas in the region, again, using color imagery and field assessment. This inventory will also be supported by the predictions inherent in the classification and inventory map, as well as the previous and new field inventory data assembled and available with the digital data. Update the existing natural resource inventory data for the six remaining counties using current cadastral data and 911 data. The inventory has been completed for Jackson and Johnson Counties. In conjunction with the cadastral and 911 updates to the six counties, the NRI data can be enhanced with the creation of forest boundaries and the addition of soil and aspect information. Important ecological assumptions, found in this project, that relate forest types to landscape position could then be extrapolated to the remaining counties. While this study has gone great lengths to map out natural resources through the utilization of existing GIS data, the project can also be enhanced with a high resolution land cover or land use data set. The creation of such a data set can efficiently be accomplished for the Kansas City Region using two methods: 1) Utilizing cadastral information or 2) Use of high resolution multi-spectral imagery. This land cover or land use data should be created and updated every 5 years ultimately allowing planners and land managers the ability to track land use change. To optimally utilize data across community boundaries, planners and land managers will benefit if the NRI data is summarized by watershed, subwatershed, and catchment area. Watershed delineation data has been included as part of the hard drive delivery making this step a quick and easy GIS exercise. By summarizing the NRI data by various watershed parameters much can be learned in terms of the vulnerability and sensitivity of various drainage basins. The spatial and other data assembled for this project have created an opportunity for accomplishing the next level of classification and inventory work when that time arrives. These data provide a ground-truthed basis for creating a new classification with an up-todate satellite image of the Kansas City region. Should the opportunity arise to purchase such an image, it is recommended that the compiled old and new field data (including Natural Heritage Inventory and Biological Survey data), knowledge gained about image signatures, and the current natural resource inventory be used to create a spectral-based classification of the satellite image. Given the existing data, the greatest cost of such an effort would be obtaining the imagery and completing the spectral classification. 30

36 5.0 Acknowledgements The staff of AES wishes to thank MARC, and in particular Tom Jacobs, Scott Paszkiewicz, and Andrea Repinsky, for their inspiration and vision, which set the course for this project. Walt Foster and others at the EPA were also very helpful in establishing the technical foundation for this project. Data came from many sources, but in particular, AES wishes to thank the following due to the large commitment they made in providing assistance: EPA, MDC, KBS, MARC, MSDIS, DASC, NRCS, U.S. Army Corps of Engineers, and Johnson County Automated Information Mapping System (AIMS) and other local governments. PBA, especially Patti Banks and Laurie Brown, shared expertise, time, and office space to bring this project to completion. 24

37 APPENDIX A A-1: Detailed Data Collection and Conversion Methodology A-2: Natural Resource Inventory (NRI) Metadata A-3: Shape Files Used for the Kansas City NRI GIS Program

38 Appendix A-1: Detailed Data Collection and Conversion Methodology To complete data collection for the ecological classification and inventory for the Kansas City region, AES completed the following tasks: A. Data Assembly and Mapping: digital information from several government sources was used to establish baseline information about land cover in the region. B. Field Reconnaissance: The digital information was validated and/or refined through field inspections and verifications. C. Ecological Land Cover Classification Development: Using data from the data assembly and subsequent field reconnaissance, AES created an ecological classification representing existing natural resources in the region, a GIS-based information database, and a regional map of ecological land cover. D. Data Extrapolation and Second Field Verification: The ecological classification involved an iterative process in which initial data were assembled, evaluated in the field, revised, and then re-evaluated in the field a second time. Final data were assembled after the second field reconnaissance, evaluated, and incorporated into the GIS program and the regional land cover map. Details of how data was compiled and utilized for this project are described below. Data Assembly and Mapping One of the primary objectives of the Kansas City Natural Resources Inventory was the compilation of extensive amounts of data from a variety of source that could be used as a baseline for the natural resources inventory and from which to build a comprehensive GIS program. As the project began, and data sources were accessed and compiled into the program database, it was found that a lack of comprehensive, up-to-date data sets existed. These data sets were needed to create a consistent and complete region wide ecological land cover classification and natural resources inventory. While both Missouri and Kansas and their many different agencies have extensive geographical and land use data sets that provide data of different resolutions, accuracies, and formats, AES determined that the information and data available from the states would require close examination and testing to assess its quality for use in the Kansas City Natural Resource Inventory. To overcome potential data quality issues, AES examined existing land cover classifications and existing natural resource data and tested the mapping results against recent digital orthophotos of the Kansas City Metropolitan Area. From this process, it was discovered that a Triage approach was necessary for project success. This approach would map different counties to different resolutions dependent upon existing data availability and coherency. Our examination of data determined that Johnson County, Kansas and Jackson County, Missouri had the most detailed and consistent data. Cass, Clay and Platte Counties, Missouri had less detailed information than Johnson and Jackson Counties, and Ray County in Missouri and Leavenworth and Wyandotte Counties in Kansas were found to have the lowest resolution data and were mapped similarly.

39 Land cover data was considered acceptable based on its ability to correctly distinguish a uniform patch of land surface on digital orthophotos (e.g., forest, grassland, urban) and correctly match the edges of the digitized polygon to that land surface patch. Based on this examination, AES found GAP data to be the most accurate and consistent when compared at a multi-county, multi-state level. While available for both Missouri and Kansas, each state used a different GAP classification. In order to join these classifications and create a regional ecological land cover classification, AES created a two-level classification linked to the two GAP classifications (Table 1 of the Ecological Land Cover Classification for a Natural Resources Inventory in the Kansas City Region, USA report). This approach to combining different classifications is termed a crosswalk. The final Crosswalk Classification for this study (CLASS_URBF) included the following: Aquatic Communities Open Water: Standing water for a significant part of the year. Natural Communities Lowland Lowland Hardwood Forest and Woodland: Open to closed forest canopy in flooded or wetland areas. Marsh and Wet Herbaceous Vegetation: Wetlands without a closed canopy; includes brush and scattered trees. Natural Communities Forest Deciduous Forest: Mostly closed canopy of deciduous trees, often mature; includes former savannas on south to west slopes. Mixed Evergreen-Deciduous Forest: Open to mostly closed canopy of junipers and deciduous trees, often oaks; may include form savannas Natural Communities Grassland-Savanna-Woodland Deciduous Woodland/Immature Forest: More open canopy of deciduous tress, often mature. Grassland: Often containing native wild plants; may include CRP plantings. Cultural or Sparsely Vegetated Land Urban Forest: Deciduous canopy cover within an urbanized location (Specific to Jackson/Johnson Counties) Cultural Grassland: Grassland of planted domesticated grasses, or formerly cultivated land reverting to grassland and sometimes brush. Agricultural Land: Used as farmland (Specific to Jackson/Johnson Counties) Cultivated Land: Used as cropland

40 Developed Land: Urban and suburban land uses, including homes, businesses and roads. Ares mapped may contain natural vegetation. Unclassified: Insufficient data to classify. In order to classify data into the crosswalk classification the following methods were used. 1. Developed or Undeveloped Classification Developed lands were the first class assigned and delineated within the NRI classification. This class is the most accurate spatially and temporally because of the source data availability. Because of its accuracy and consistency within the project area the developed land class was set to override all other classes. Once developed and undeveloped lands were distinguished the undeveloped lands class were further dissected Developed Classification defined: Urban and built-up areas A Land Cover/Use category consisting of residential, industrial, commercial, and institutional land; construction sites; public administrative sites; railroad yards; cemeteries; airports; golf courses; sanitary landfills; sewage treatment plants; water control structures and spillways; other land used for such purposes; small parks (less than 10 acres) within urban and built-up areas; and highways, railroads, and other transportation facilities if they are surrounded by urban areas. Also included are tracts of less than 10 acres that do not meet the above definition but are completely surrounded by urban and Built-up land. Two size categories are recognized in the NRI: (i) areas 0.25 to 10 acres, and (ii) areas greater than 10 acres. [NRI-97] 2. Two Methods for Defining Developed Lands Two methods were used in the classification of developed and undeveloped lands. (1) Parcel size and parcel use (cadastral data) provide the required information necessary to classify urban area according to NRCS s definition (below). (2) The average distance from a road provides the second approach taken to define developed land. 2.1 Cadastral Method Of the two methods, the cadastral data (parcel size and tax use values) is perhaps the most accurate and highly resolved method of classification. Because cadastral data is used for tax assessment purposes the data is developed to a set of minimum standards which are very similar across the study area. The spatial accuracy of any parcel may vary according to the data and survey methods used. The older methods are generally the least accurate. Even with this cadastral mapping variability the overall spatial accuracy of the cadastral data is significantly greater than any other land use source data available for this project. Because taxation is an annual event, parcel data is updated each time the parcels are sold and subdivided. The temporal accuracy of the data is generally accurate to within 1 year of the acquisition data (often accurate to the data of purchase) unless earlier versions are requested Redefining Urban Areas for Use with Cadastral Data Johnson, Clay and Cass Counties were the only counties to provide workable tax use values (Further work and discussion with the various counties will be required to extract these codes

41 and their descriptions.). The tax use values can be very valuable for projects like this one because they describe the primary use of the property for tax purposes. Thus, smaller parcels used for higher intensity uses such as residential homes, commercial properties, industrial properties, etc., are more than likely developed. As parcels increase in size the reliability of the use class decreases as some parcels may have multiple uses or certain portions of a parcel may have a secondary or multiple different uses not defined by the use class value. Larger parcels should be reviewed and reclassified accordingly. Parcels less than or equal to 10 acres were defined as developed. All parcels greater than 10 acres that were not specifically defined as commercial, industrial or some other developed category were defined as undeveloped Cadastral Data Availability Cadastral data was not available for all of the counties in the study area. As the map development began only one county had provided cadastral information. Currently all counties but one (Ray) have provided some form of cadastral data. Jackson County data has only provided geographical subsets of its attribute data (including tax use). Other counties have not provided the appropriate attribute data to perform this analysis. As a result, other methods for defining developed areas were required for this study. It may be useful to develop a GIS users group of key county officials and GIS staff in order to support the efforts of the various county mapping departments and their staff. Regular meeting would help MARC explain and demonstrate the value of cooperation in meeting the mapping needs of the region. The publication of annual map and data products would help support the efforts of individual GIS departments to develop and enhance their data development and distribution efforts Updating NRI Developed Lands We strongly recommend that you pursue the acquisition of the cadastral data and the related attribute information required to map urban and built-up areas using this method. The use of this data is the least expensive and most accurate method to update the NRI annually. Changes in developed lands represent the continuing encroachment of development on the natural resource inventory Urban Area/non-Urban Area Map An Urban Area/Non-Urban area map, a subset of the NRI map, can easily be generated from cadastral data. Change in urban area (using parcel size metric) can be measured accurately, reliably and regularly. Annual updates of this map will provide a visual presentation and data that decision-makers and the public can understand. Increases in urban area reflect reductions in the various natural resource classes. 2.2 Roadless Areas The roadless area method is the second method used to classify developed land. This method is less accurate than the cadastral method and is used when cadastral data is not available and when the accuracy of the road centerline data is reasonable, i.e. when there are few omissions in the road centerline network and when road centerlines accurately overlay current orthorectified aerial photos as shown in Figure 1 below.

42 Figure Road centerlines overlay 2001 aerial photos with reasonable accuracy. In a few areas the centerlines were not as accurately depicted as here. Road metrics are ideal indicators of development which includes impervious surface area and hydrologic segmentation. Roads are built to provide access. On forested land, temporary roads are needed for timber extraction and management. Agricultural land requires more permanent roads for intensive cultivation and livestock management. Residential and commercial areas require greater road densities to service smaller properties that support larger populations. Greater overall road density in a study area indicates greater intensity use of the land. In addition to their contribution to impervious surface area, the abundance, longevity, and spatial characteristics of roads affects the natural drainage system. They modify natural drainage patterns within a watershed, creating artificial, segmented hydrologic systems. They function as dikes and dams, containing and directing water into and out of these artificial basins. Overland flows enter and exit these basins through breeches in the system, e.g. culverts and bridge openings Roadless Areas Defined Roadless area polygons are bounded by roads and railroads generated from road and railroad GIS centerline data. The road network was converted from a vector line file into a polygon theme in which the roads and railroads serve as the boundary lines of the polygons. Only centerlines that create a closed boundary around unroaded areas are converted into polygons. Cul-de-sac and other dead-end streets and roads are not incorporated into the resulting metrics. These disconnected or dangling line segments do represent important access into roadless areas. Distance from a road grid metric was used to better characterize the roadless polygon areas. Below in figure #2 you see a distance grid with the roadless polygon overlaid. Red represents a least distance from a road while blue represents a furthest distance. The black lines are the roadless polygon boundaries.

43 Figure 2. The use of grid distance from roads incorporates the access issues presented by culde-sacs and dead end roads that are not addressed with roadless polygon metrics. Blue areas identify the greatest distances from roads. Using the zonal statistics function within the spatial analysis extension we were able to extract summary grid distance statistics for each roadless polygon. The average distance from a road and the maximum distance from a road were added to the roadless polygon file. The resulting roadless polygon mapped below identifies the roadless areas with the maximum distance from a road of less than 300 meters in yellow. These polygons are most often associated with developed land. Polygons with a maximum distance from a road greater than 300 meters are less often associated with developed areas. Figure 3. Summary statistics for each roadless polygon such as the maximum distance to a road within a polygon are linked and coded to the roadless polygon record.

44 Figure 4. Roadless polygon techniques identifies more developed areas (shown in yellow). The areas in yellow have a maximum distance to a road of 300 meters. This measurement roughly corresponds to the NRCS metric defining urban areas of parcel clusters of 10 acres or less. Road centerline data is developed and maintained by Mid-America Region of Governments and is available internally to MARC Roadless Area Maps An Urban Area/Non-Urban area map, a subset of the NRI map, can easily be generated from road centerline data. Change in urban area (using parcel size metric) can be measured accurately, reliably and regularly. The resolution and accuracy of the road centerline data is less than from cadastral data. However, updates of this map will provide a visual presentation and data that decision-makers and the public can understand. As with the cadastral version of this map, increases in urban area reflect reductions in the various natural resource classes. 3. High Resolution Forest Cover In Jackson and Johnson County we have very accurate forest cover GIS files. In Jackson County the data was generated from the aerial photo interpretation process. In Johnson County the forest cover data was provided by the County. 4. Land Cover Data Various sources of land use land cover data was inventoried and assessed for accuracy, precision, usability and consistency from a spatial and temporal perspective. It was determined that GAP Land Cover data was the most appropriate for this project Land Use and Land Cover (LULC) Land Use and Land Cover (LULC) data from the USGS was one of the data layers inspected. LULC data consisted of historical land use and land cover classification data based primarily on the manual interpretation of 1970 s and 1980 s aerial photography

45 ( This information could be used as a data source for change analysis. The resolution of the data fits comfortably well with NLCD or GAP data described below. Because the coverage is more than 20 years old it was not used in the development of the current natural resource maps The National Land Cover Characterization (NLCD) Also reviewed was the National Land Cover Characterization (NLCD) project. The project was created in 1995 to support the original Multi-Resolution Land Characterization (MRLC) initiative and fulfill the requirement to develop a nationally consistent land cover data set from MRLC data called National Land Cover Data 1992 (NLCD 92). Additional information about the project can be found at: According to product description on the USGS NLCD web site the land cover data results from the early to mid-1990s Landsat Thematic Mapper satellite data with a 30 meter resolution. The unsupervised classification was labeled using relatively small numbers of aerial photography for ground truth. Leaf-off and leaf-on imagery were combined in the classification process. Classification was performed at the state level often with a mosaic up to 18 TM scenes. The reliability of the data is greatest at the state or multi-state level Gap Analysis Program (GAP) Gap Analysis is a scientific means for assessing the extent native animal and plant species are being protected at a state, local, regional, or national level. One of the primary by-products of the GAP effort is a land use, land cover mapping data. The data was derive from satellite imagery similar to the NLDC data but has been refined specifically to identify natural community types. The National Gap Analysis Program or GAP function, according to its website at is to provide broad geographic information on the status of ordinary species (those not threatened with extinction or naturally rare) and their habitats in order to provide land managers, planners, scientists, and policy makers with the information they need to make better-informed decisions Gap analysis is a method used to identify the degree to which native animal species and natural communities are represented in our present-day mix of conservation lands. Those species and communities not adequately represented in the existing network of conservation lands constitute conservation gaps. The national effort brings state and national agencies together in an effort to map the following: Existing natural vegetation to the level of dominant or co-dominant plant species; predicted distribution of native vertebrate species; Public land ownership and private conservation lands The current network of conservation lands; and Distributions of any native vertebrate species, group of species, or vegetation communities of interest with the network of conservation lands. Kansas Gap data was taken from TM data since 1995 to 2000 while the Missouri Gap data was taken from TM data from the 1991 to The vegetation mapping uses the National Vegetation Classification System (FGDC 1996). Several agencies are involved in the development of the GAP including the Biological Resources Division of the U.S. Geological

46 Survey, the Department of Defense, the Environmental Protection Agency and various state agencies. Other non-government partners include The Nature Conservancy. The GAP work was conducted in Kansas by the State Biological Survey using a classification scheme prepared previously for the State (Lauver, Kindscher, Faber-Langendoen, and Schneider 1999). In Missouri land cover mapping and classification was prepared by an interagency team using the National Vegetation Classification introduced initially by The Nature Conservancy and later modified by a working group with The Ecological Society of America. Other classifications completed included those available on the DASC website in Kansas, and the MSDIS web site for Missouri. Further information is available on the Kansas GAP efforts at: Further information is available on the Missouri GAP efforts at Summary of Existing Land Cover Mapping Data All of the data described above was developed for regional planning for use at a scale of 1:100,000 or smaller. None of these data fit the specific needs of this project. None of the data was of the desired resolution for more detailed planning in the MARC study area. The GAP data provided the best and most current cover data of these land cover sources. GAP data was used and refined where possible in the development of the NRI maps. 5. Johnson County Forest Cover A detailed forest cover GIS boundary of Johnson County was provided by the county. The boundaries fit very well over the 2001 natural color aerial photography produced by AerialExpress. A description of the process used by the County to define forest cover was not provided but assumed to be an aerial photo interpretation process. The digital feature code (FEA_Code) 82 was used to represent forest cover in this project. According to additional codes the photography used in the forest cover classification was dated Figure 5. The red boundary lines represent the forest cover boundaries in Johnson County. These forest cover boundaries were used in the further classification of forest type.

47 6. Jackson County Forest Cover Jackson County was used as a test case to determine the value of additional aerial photo interpretation work to map land cover from 2001 natural color aerial photographs. The photography was highly resolved (1 meter pixel resolution), was current to 2001 and covered a major portion of the MARC 8-county study area. This photography represented the best mapping data available for this study at the time. More recent natural color photography has been flown by AerialExpress since this study was initiated. Platte Clay Ray Leavenworth Wyandotte Jackson Johnson Study Area Photo Coverage Area Cass Figure 6. AerialExpress Aerial Photo Coverage area overlaid on the MARC 8-county study area. Portions of Ray and Cass counties in Missouri were not covered. Additional information on this and more recent imagery can be found at Aerial Express.com s web site: Aerial Photo Classification Process of Forest Cover Image classification methods were used to see if improvements could be made on the GAP forest cover data. It was found that the image classification of the aerial photography would significantly improve the GAP classifications both with regard to boundary resolution and temporal accuracy. However, the analysis of natural color imagery is not without issues. The image classification process uses digital natural color imagery. Signatures or groupings within the electromagnetic spectrum are used to help identify or classify ground features such as urban, cropland, water, forests. There are 2 primary types of classification, unsupervised and supervised. An unsupervised classification allows the computer to automatically group

48 reflectance values into classes where supervised classification allows the analyst to select the bands of interest for each class. There are many complications with the two different classification techniques. If you are using several images, each image must be treated separately as reflectance varies with time of day, season of the year, etc. Thus, classification rules vary from image to image. Often classification results are uncertain and/or inaccurate. Refining spectral values is as much an art as it is science-based methodology. Land use classes are rarely discrete with clearly defined boundaries. Often an edge pixel will contain several classes or transition areas between classes. In spite of this, the classification process assigns a single class to every pixel, ignoring uncertainty. The classification of natural color imagery is also problematic as the spectral signatures do not clearly identify specific feature types. Often very different features have similar or identical reflectance signatures. This problem can be overcome, in certain circumstances, by using different types of imagery. For example, color infra-red imagery would be very helpful in extracting vegetation classes. For this project various supervised and unsupervised methods were experiment with all supplying varied results across the study area. In certain locations landscape features were identified cleanly where in other areas the features were not well identified. It was determined that the unsupervised classification method would be implemented on Jackson County in order to determine potential success, failures and costs of this method for the entire study area. Jackson County is covered by 4 full Level C images and 8 partial images. Sixteen unsupervised classes were developed for each of these images. The first 3 to 4 classes which represented the darkest areas within image were used to roughly identify the forested areas. Each image was assessed to determine if any single or any combination of classes would represent a specific landscape feature. A smoothing technique was used prior to classification to remove some of the smaller features. This process incorporated the minor feature into larger adjacent features through a neighborhood pixel averaging process. In most cases this process removed many of the random and difficult to classify pixels that would break up the landscape feature boundaries unnecessarily. In general, forests and wooded areas are darker than most other landscape features. The nature of the forest features was very hetrogenetic, varying according to age, density, slope and time of day. With the exception of water and some dark, freshly tilled soils, most of the darker regions within Jackson County were in forest or woodland. As described above the darker signatures also suggest water and dark soils. We were not able to spectrally tease these different features apart, which required a step of manual inspection and digitizing of edits to the forest layer. This manual visualization and editing process was very time consuming. To begin the process, classified spectral classes that best represented the forest communities were combined to form a single class. This class was then converted to a polygon vector file that represented the outlines of forest boundaries. Transparent shading was assigned to the forested polygons and layed over the 1 meter aerial imagery. On top of this a 5,000 meter grid was laid over the imagery and the forest polygons were edited. Forested polygons in each grid were reviewed checking for omissions and commissions. Heads up digitizing was used to add omitted forest lands and non-forested lands were removed from the forest polygons. Edits were primarily used to clean up edges and to incorporate forested island polygons. Wooded field edges and wooded rural residential areas were captured as well. Many of these problem

49 features were edited and refined while others still need more attention. Many of the polygons required the editor to make an interpretation of forest edge. Because many of the forest edges are not discrete, judgments are required slowing down the editing process. During the initial review of the classification a variety of errors were found. In urban residential areas shadows represented a combination of trees and some shadows from buildings. The urban residential areas with more trees were classified as urban-residential with trees. This is not a perfect classification but it does identify residential areas with relatively more trees. This data layer was ultimately included in the Crosswalk Classification and improved the identification of forested areas in Jackson County significantly. Figure 7. Various textures are found within the Jackson County forest cover boundary representing various forest types and various forest ages. Some recent clear-cut areas as well as some early succession forests are shown here. The overall the accuracy of the Jackson County forest boundary is much greater than any of the low-resolution land cover sources. Data Sets and Classification Methods The Crosswalk Classification throughout the project was a natural progression of give and take between data availability, the technical issues within the GIS and an end product that satisfied resource planners and their ultimate purpose for the data. Keeping this in mind three versions of the crosswalk classification are present within the final GIS land cover data layer named: nri_allcnties (these three classifications are found within the attribute table of the GIS shapefile). The three classifications are simple modifications of what AES and MARC has decided on as a final classification applicable for the purpose of this project called, CROSS_URBF. The CROSS_URBF is used in this project because it is the generalization of the more detailed, but less accurate FINAL_CLAS but also includes urban forests unlike the classification attribute column CROSS_NOUR. In summary, the three attribute columns found in the final natural resource inventory shapefile are as follows: FINAL_CLAS: This classification is the most detailed but also has the greatest chance for error.