Mapping Ecosystems: GeoTechnologies Provide Valuable Tools for Mapping Ecological Units
By Tim Wirth, Paul Maus, Henry Lachowski and Don Fallon

Introduction
Ecosystem management is an integrated approach to managing natural resources where decisions are based on complex interactions of biotic and abiotic factors across the landscape.
      The chief of the Forest Service, Jack Ward Thomas, defines ecosystem management as "... a concept of natural resources management where National Forest activities are considered within the context of economic, ecological, and social interactions within a defined area, or region, over both the short and long term. Its purpose is the management of National Forests to meet human needs while maintaining the health, diversity, and productivity of ecosystems."
      The Forest Service and other federal agencies apply ecosytem management principles to attain their resource management objectives. Defining ecosystem boundaries and mapping ecological units for application in ecosystem management presents many unique challenges. Remote sensing and geographic information systems (GIS) are currently playing an important role in meeting these challenges. Satellite remote sensing provides a unique and comprehensive perspective that compliments aerial photos, field sampling, and other data sources in mapping ecological units. Satellite imagery is in fact the only feasible way of obtaining the synoptic view necessary for large area assessments which may cover many ecological, political, and jurisdictional boundaries.

National Framework for Ecological Unit Mapping
Ecosystems often form a continuum across the landscape in both size and scale. Therefore, in order to map and understand ecosystems, it is often necessary to delineate and describe them at a variety of different scales. To address this need, the USDA Forest Service, along with the National Resources Conservation Service, Bureau of Land Management, Fish and Wildlife Service, U.S. Geological Survey, the Nature Conservancy, and other national and regional agencies, has developed the National Hierarchical Framework of Ecological Units (Table 1) to describe ecological units at various scales. An "ecological unit" is the basic delineation for ecosystems having similar response potential and resource production capabilities at a particular scale. This framework provides a classification and mapping system for stratifying the Earth into progressively smaller areas of increasingly uniform ecological potentials.
      The National Hierarchical Framework of Ecological Units was developed to provide a scientific basis for ecosystem management. The ecological units in the hierarchy were designed to delineate areas of different biological and physical potentials and were developed geographically from both the top-down and bottom-up. Units fit together in a "nested" fashion, i.e., smaller units are subcomponents of larger units at all levels of the hierarchy. Conditions that change at broad scales such as climate and geology are related to conditions that change at finer scales such as biotic distributions and soil characteristics. Units range in size from the Ecoregion scale (Domains, Divisions, and Provinces) - 1,000,000s to 10,000s of square miles, to the Land Unit scale ( Landtype and Landtype Phase units) - 100s to less than 10 acres.
      A number of satellite platforms provide digital imagery at various spatial and spectral resolutions which can be used for planning and analysis at the appropriate scales of the National Hierarchy. For example, the Advanced Very High Resolution Radiometer (AVHRR) located on-board a weather satellite obtains data suitable for displaying broadly defined units such as the Domain, Division, or Province ecological units at the Ecoregion scale (see Table 1). Units at intermediate scales, such as Landscape scales, can be depicted on images from satellite sensors such as the Landsat Multi-Spectral Scanner (MSS) or Thematic Mapper (TM) and the Systeme Probitaire pour l'Observation de la Terre (SPOT). Detailed views at smaller ecosystem scales can be captured from aircraft with on-board sensors such as multi-spectral scanners, video and photographic cameras. In addition, field surveys provide an important link between remote sensing information and the ground. The key to success in using remote sensing for delineating ecological units is to match the appropriate source of information with the appropriate task.
      The units in the hierarchy and the information about each unit can be used by resource managers to plan for and implement management strategies on an ecological basis (i.e. information about vegetation, landform, climate, geomorphic processes, soils, etc. within units and between adjacent units will be used in the decision making process). For example, ecological unit information across levels of the hierarchy might be used for predicting timber harvest effects across stands, forests, and/or regions. In another circumstance, a range manager may use the information on carrying capacity derived from the Landtype level ecological units in his/her decision to remove cattle from a sensitive riparian area. On a broader scale, range managers from adjoining districts might decide to decrease the number of grazing allotments based upon Landtypes or Subsections. Furthermore, several adjacent forests may make management decisions based upon Ecoregion scale information.

Pilot Study
Like many national forests, the Bridger-Teton National Forest in Wyoming is conducting an Ecological Unit Inventory (EUI) where satellite imagery, aerial photographs and GIS are being used to map ecological units at the various levels of the National Hierarchy. The EUI is an ecosystem inventory specifically designed to integrate information on landform, geology, potential natural vegetation, and soils. The Nationwide Forestry Applications Program (NFAP - USDA Forest Service, Salt Lake City, Utah) has been working cooperatively with the Forest EUI Team to develop and test applications in ecological unit mapping using remote sensing and GIS. A pilot project encompassing approximately 900,000 acres is evaluating the use of satellite imagery, digital elevation models (DEMs) and GIS for expediting the challenging ecological inventory process.

Mapping Ecological Units with Remote Sensing
Satellite imagery is being used in ecological unit inventories such as the one on the Bridger-Teton becauses it provides the comprehensive perspective necessary for large area assessments. Various sources of remotely sensed imagery can be applied both directly and indirectly to assist the process of mapping and evaluating ecological units at all levels of the Hierarchy. ECOREGIONS, the broadest units in the Hierarchy, are used in global or continental scale assessments. The ecological units at the Ecoregion scale have been developed for the entire U.S. The pilot study area on the Bridger-Teton National Forest falls within the Dry Domain, Temperate Steppe Regime Mountains Division, and Forest-Alpine Meadow Province.
      A source of imagery appropriate for observing and refining units at these scales is AVHRR. AVHRR satellite imagery has 1.1 km pixels (resolution) and one scene covers a very large area of land (approximately 1.4 billion acres). The multi-spectral imagery provides information about vegetation, geology, and landforms crossing state and country boundaries. The AVHRR image of the Intermountain Western U.S. in Figure 1 shows the Division and Province boundaries as well as state boundaries as an overlay. The Wind River Mountain Range, at the center of the image, is the location of the Bridger-Teton National Forest pilot project.
      Moving down the Hierarchy, SUBREGIONS are characterized by combinations of many factors including climate, geomorphic processes, and topography which in turn affect hydrologic function, soil forming and potential plant communities. Sections and Subsections are the units mapped at this scale. These units are used for statewide, multiforest, or multiagency analysis and assessment.
      Section and Subsection units can be displayed, mapped, and refined on images from satellites such as Landsat. Figure 2 is a Landsat TM image of the entire Wind River Range with Subsection units shown as an overlay. The TM image has a pixel size of 30 meters and covers an area of approximately 31,000 square km. Hence, TM imagery is more appropriately matched to this level of the Hierarchy than AVHRR. It is also more appropriate over these large areas than aerial photography. To get complete stereo coverage of this same area at 1:24,000 scale, over 2,400 photos would be required. In the pilot project, the TM image was used much like a photograph, where the mapping team delineated Subsection boundaries directly on the image. Since the satellite image is registered to a map, the resulting delineations were ready for direct input into a GIS.
      At the LANDSCAPE scale, the next lower level in the Hierarchy (see Table 1), units are defined by general topography, geomorphic process, surficial geology, soils, local climate, and potential natural communities. The Landscape scale is applicable to Forest or watershed level analysis. Landtype Associations are mapped at this level. Repeatable patterns of soil complexes and plant communities are useful for delineating map units at this level.
      In the pilot project, Landsat TM imagery was also used for delineating Landtype Associations. Because the image is digital, one can "zoom" in (with limitations) to various levels of the image and work at the appropriate scale. The same TM image used for a broadscale view at the Subregion scale was used to delineate Landtype Association units at the Landscape scale (Figure 3). Additional information from other sources including state geological maps, historic maps, literature reviews, and input from resource professionals familiar with the study area was used to refine the boundaries.
      The lowest level in the Hierarchy, the LAND UNITS, are based on local topography, rock types, soils, and vegetation. Units at this scale are used for management and planning at the Forest or project level. The ecological units mapped at this level are the Landtype and Landtype Phase units.
      In the pilot project, Landsat TM imagery, Digital Elevation Models (DEMs), aerial photographs, and ground collected data, among other sources, were used to assist in the delineation and analysis of Landtype and Landtype Phase units. Satellite imagery by itself was not adequate for developing units at this scale, but the imagery provided information on vegetation and landform that was useful in the indirect determination of ecological units at the Land Unit scale.
      Delineation of units at the Land Unit scale (Landtypes and Landtype Phases) is time consuming and costly. Deriving units at this level of the Hierarchy is where National Forests are spending the most time and effort. As a result of the pilot project, a process which uses satellite imagery, aerial photos, DEMs, and other data sources together in a GIS, was developed to create a pre-map of the Landtype units. Pre-mapping is an important first step in identifying and delineating final ecological units at the Land Unit scale of the Hierarchy. This pre-map, developed on a computer in the GIS was then used to derive the final Landtype units for the study area.

Developing Landtype Ecological Units On The Bridger-Teton National Forest
To develop Landtype units on the Bridger-Teton National Forest, the study area was first stratified into Landtype Association (LTA) units (Figure 3). The stratification allowed the spectral classes from a classified LANDSAT TM image and DEM classes of slope, elevation, and aspect to be custom fit to each LTA. All further GIS and image analyses were then conducted independently within each of these units.
      Automated processing techniques using the LANDSAT imagery and DEMs were developed and used to generate maps of the entire study area in approximately five weeks. Additional time was required to manually refine the pre-map Landtype units using aerial photos, orthophotoquads, field data, and other ancillary data, and map them onto aerial photographs. The final polygon maps (the pre-maps), were reviewed on a quad-by-quad basis and were taken to the field for further ground data collection and refinement. Summaries of the type and amount of each class for vegetation, slope, and aspect (for unit characterization) were produced by overlaying the pre-map polygons on each of the respective data layers in the GIS. Final edits and updates to the pre-maps using field data, aerial photos, and other ancillary data produced a final digital map of Land Unit scale polygons (Figure 4).

Pilot Projects Results
The use of remote sensing and GIS technologies greatly expedited and enhanced ecological mapping efforts on the Bridger-Teton National Forest. The methods developed for the pre-mapping process allowed for the production of a field ready pre-map in about half the time it would have taken to create a similar type map from aerial photos alone. The satellite image data however could not provide all of the information by itself. When developing units at the Land Unit scale, data from many other sources is needed for unit definition. Satellite imagery does not provide the stereo views one can get using aerial photos. Also, ecological unit lines do not always necessarily coincide with vegetation/slope/ aspect lines derived from the digital data. The satellite imagery was just one of a number of useful tools in a complex, interactive process dependent upon input from many data sources.

Summary
Remote sensing and GIS are proving to be vital tools in the mapping and analysis of ecological units at all levels of the National Hierarchy of Ecological Units. These technologies are equally important in supporting the application of ecosystem management practices in the Forest Service and other agencies. As demonstrated here, these technologies can be applied at many scales in the ecological mapping process; from the project or Forest level, all the way up to large area land assessments which may cover many state and agency jurisdictions. The ability to provide consistent, relatively inexpensive data over large areas encompassing many ecosystem scales and many social, political, and jurisdictional boundaries, is perhaps the most useful and promising application of these technologies for ecosystem management.

About the Authors:
Tim Wirth and Paul Maus are GIS/remote sensing analysts with Pacific Meridian Resources in residence at the Nationwide Forestry Application Program, Salt Lake City, Utah. Henry Lachowski is program leader, USDA Forest Service, Nationwide Forestry Applications Program, Salt Lake City, Utah. Don Fallon is a geologist/GIS specialist, Bridger-Teton National Forest.

Back