Topographic-Landsat Model of Suitable Spotted Owl Habitat around Los Alamos National Laboratory

Johnson, Terrell

A habitat model is a description of known habitat, and a numerical model is merely a numerical description, which in its simplest form produces a habitat index. It can be no better than the data that went into it, limited by the size and extent of the sample of known habitat that it describes and the parameters available to the description. It is not unique, might not be complete, and can be inaccurate when extended beyond its original sample. The test of a model is how well it performs on an new sample of known habitat. A numerical model that does produce a meaningful habitat index from a widely available data provides a consistent method to identify habitat over large areas of the landscape. All species have habitat requirements that cause them to be found in some situations and excluded from others. Habitat requirements fall into two general categories: physical and biological. Physical factors such as landform, climate, and soil may affect species directly but also determine what flora and fauna may be found there. A topographic model addresses the most important but not all of these factors. Topography describes landfo1ms, provides a rough estimate of climate based on elevation and latitude, and implies a little about soil based on slope and curvature. Biological factors such as seral stage, structure, and patterns of vegetation dete1mine the current suitability of potential habitat. Landsat spectral data may distinguish large scale vegetation cover types and densities, and therefore provide some information on habitat suitability, but they are unable to distinguish small scale vegetation stmcture or composition. The statewide topographic model of potential of Mexican spotted owl (Strix occidentalis lucida) habitat in New Mexico (Johnson 1993) was based on US Geological Survey (USGS) I-degree Digital Elevation Model (DEM) data. These data are less accurate than 7.5-minute DEM data, and fail to resolve some of the topographic structures that spotted owls inhabit. The northern New Mexico topographic habitat model reformulated the statewide model for the higher resolution 7.5-minute DEM data (Johnson 1996) in and around the Santa Fe National Forest. However, the USGS 7.5-minute DEM data have been smoothed and are sometimes inaccurate, particularly in the sharply breaking mesa and canyon topography in and around Los Alamos National Laboratory (LANL). The first objective of the present work was to calculate the topographic habitat model from more accurate DEM data, which were provided for an area surrounding LANL by the LANL Facility for Information Management, Analysis and Display (FIMAD). The second objective was to combine topographic and Landsat spectral color class data, from which LANL developed a land cover map of the area (Koch et. al. 1997), in a model of suitable spotted owl habitat that is specific to this area.