Return to List of Research Summaries
TITLE:
A Landscape Approach to Ecological Condition in a Southwestern Watershed (U.S./Mexico)
INVESTIGATORS:
William G. Kepner (EPA), Kurt H. Riitters (TVA), Christopher J. Watts (IMADES), Curtis M. Edmonds (EPA)
BACKGROUND:
There is increasing public concern over the condition of large-scale ecological systems, including landscapes, watersheds, and ecoregions or biogeographical provinces. The public is concerned about cumulative impacts of human activities on natural resources and is particularly focused on changes in biotic diversity and production. Ecosystem fragmentation has been implicated in the decline of biological diversity and ecosystem sustainability at a number of spatial scales, however, traditional methods for measuring and protecting biodiversity have often focused on fine-scale habitat requirements of individual species. Multi-scale monitoring and assessment approaches are necessary to address the wide range of environmental issues, particularly those related to biodiversity. Application of new concepts related to ecosystem management requires the evaluation of landscape pattern and composition at multiple spatial and temporal scales, rather than more traditional approaches of attempting to aggregate fine-scale measurements from site-level assessments. The objective of this project is to evaluate changes in habitat suitability for selected species at multiple landscape scales within the Upper San Pedro Watershed (U.S./Mexico). Landscape composition and pattern metrics will be combined with life history models to determine status, changes, and trend in wildlife habitat suitability over a twenty-year period using Landsat-MSS remote sensing imagery. The research results can be utilized in managing specific species or guilds of species and may provide a working paradigm for watershed and ecosystem management. The research will facilitate ecological risk assessment by providing a process to identify ecosystems or specific biotic components potentially at risk of exposure to human-induced and natural environmental stressors. It is anticipated this project will provide an example of community-based environmental protection by demonstrating how coarse-scale landscape assessment can be used to place finer-scale (communities) needs and resource management options into a watershed context.
It is our hypothesis that ecosystem fragmentation is a leading cause of changes related to habitat suitability which represents one measure of biodiversity. We further hypothesize that landscape composition and pattern metrics derived from remotely sensed imagery, e.g. Landsat MSS or TM, are sensitive to large-scale changes and can be utilized as an effective method for evaluating landscape vulnerability to disturbance associated with human-induced and natural stress.
APPROACH
The determination of habitat suitability using landscape pattern or composition metrics will be investigated in the Upper San Pedro Watershed. The San Pedro River is an important international watershed located in a semi-arid mountainous region that transitions between the Chihuahuan and Sonoran Deserts. The study basin originates near Cananea, Sonora and flows north into the United States near Palominas, Arizona. The upper portion of the watershed encompasses an area of approximately 7,600 km2 (5,800 km2 in Arizona and 1,800 km2 in Sonora, Mexico). Elevations range from 900-2,900 m and biome types include Chihuahuan desertscrub, Sonoran desertscrub, semi-arid grassland, oak savanna woodland, pinyon-juniper woodland, mesquite woodland, riparian forest, coniferous forest, and agriculture.
The study area is internationally renown for its biodiversity, especially in regard to its mammal and bird species richness. It contains the second known richest assemblage of land mammal species in the world and more than 300 bird species. In 1986, the U.S. Department of the Interior acquired nearly 23,490 ha of predominantly riparian habitat between the international border with Mexico and St. David, Arizona. The area was designated a National Conservation Area by Congress in November of 1988 and is currently managed by the U.S. Bureau of Land Management.
The study design includes an approach that begins with the acquisition of Landsat-MSS imagery for the study area and mosaicing the scenes together for further processing into land cover maps for each temporal sequence (early 1970s, mid-1980s, and early 1990s). The land cover classification is based on the Brown, Lowe, and Pase hierarchical vegetation classification system and upon recommendations received from county, state, and federal resource management agencies throughout the watershed. All other relevant GIS spatial coverages, e.g. soil maps, digital elevation maps, digital line graphs, land status, census blocks, hydrologic unit codes, etc. are being acquired from state and federal spatial data clearinghouses, e.g. Arizona State Land Department. The derived land cover maps are further analyzed using ARC/INFO GIS software and spatial statistics to generate landscape composition and pattern metrics. Additionally, life history models for selected wildlife species have been acquired from the literature or derived from empirical data obtained by on-site state and federal resource specialists. This information is used to develop the spatial filter algorithms which combine attributes of composition, minimal area, elevation, etc. and are further utilized to generate habitat suitability maps and descriptive statistics for each of the selected species or species groups (guilds). Comparisons of changes in habitat are evaluated for the three-date period which spans approximately twenty years. The project period will require at least 2 years beginning in October 1996.
STATUS
All remote imagery (Landsat-MSS) has been acquired and pre-processed (i.e. triplicate scenes for three periods 1974, 1987, and 1991 have been coregistered and georeferenced to a 60 x 60 meter Universal Transverse Mercator ground coordinate grid). A derivative product (digital land cover map) has been developed for two of the image sets, i.e. 1974 and 1987, on the U.S. portion of the watershed. Some preliminary landscape metric analysis has occurred and the conceptual approach and results have been presented at professional symposia. A Fact Sheet, poster, and a home page on the EPA/CRD World Wide Web Site (http://www/epa.gov/crdlvweb) have been developed for the project.
Anticipated products include three-date (1970s, 1980s, 1990s) digital land cover maps for the entire Upper San Pedro Watershed (U.S./Mexico) derived from Landsat-MSS remote imagery; journal articles, e.g. Assessing Habitat Suitability at Multiple Scales in a Southwestern Watershed; and a project report, e.g. A Landscape Atlas of the Upper San Pedro Watershed.
Return to List of Research Summaries
homepage