Topoclimatic Habitat Models

Abstract Topoclimatic models use topographic descriptors (elevation, slope, aspect, landscape position...) as primary input to build climatic surfaces that describe spatial and temporal patterns of such physical factors as temperature, incoming solar radiation (insolation), precipitation, soil moisture, and evapotranspiration. Topoclimatic habitat models translate these physical factors into indices of habitat suitability for a particular species or biotic community. Applications range from explaining observed vegetation patterns to predicting changes in biotic distributions under climate change scenarios. Important issues when developing topoclimatic models include data quality and availability, non-independence of input data, and difficulties of model validation. Use of topoclimatic models to predict potential habitat requires knowledge of biology. Potential habitat dynamically changes over time and actual biotic distributions depend upon factors such as dispersal, history, biotic interactions, and time lags. Empirical approaches to topoclimatic modeling establish significant relations between input parameters and predicted climatic factors, but need not require understanding of mechanisms and system complexity. By contrast, mechanistic approaches are based upon physical principles, and can be applied in cases where few empirical measurements are available. We demonstrate how a combination of empirical and mechanistic submodels can be used to predict spatial and temporal patterns of air and soil temperature at a landscape scale. In topoclimatic studies conducted in the vicinity of Rocky Mountain Biological Laboratory (RMBL), Colorado, we utilized a USGS digital elevation model (DEM) and measurements from four weather stations as input. Air and soil temperature regimes were best explained by a modified lapse rate model, which accounted for deviation from a simple lapse rate based on local heating in proportion to insolation. We calculated insolation maps during the growing season using a mechanistic geometric model (TopoView/Solar Analyst). Vegetation patterns predicted using our topoclimatic model (TopoClimate) were more accurate than patterns predicted based merely on slope, aspect, and elevation.