Road-Related Disturbances in an Arctic Watershed: Analyses by a Spatially Explicit Model of Vegetation and Ecosystem Processes

Abstract

Landscape models have proven very useful in assessing historical change in vegetation patterns, for predicting the impacts of human disturbance on ecosystems, and for developing strategies to manage natural resources (e.g., Shugart 1984; Turner 1987; Costanza et al. 1990; Turner and Dale 1991; Wu and Levin 1994). Landscape models may implicitly or explicitly consider the spatial heterogeneity of system properties such as plant biomass, soil nutrient concentration, and topography — defined by either qualitative indices (e.g., patchiness, diversity, contagion) and/or quantitative indices (e.g., autocorrelation, variance, trend) (Li and Reynolds 1995). Accounting for such spatial heterogeneity has been shown to be essential for modeling ecosystem response to disturbance (Turner 1989; Costanza et al. 1990; Turner and Dale 1991; DeAngelis and White 1994). Although the use of simplified, aggregate models to represent vegetation and ecosystem processes (particularly at the scale of a landscape) has inherent dangers (Bonan 1993), the questions posed by resource managers require the development of models that summarize our “state-of-the-art” ecological knowledge and realistically represent the dynamic function of ecosystems in time and space.