Extrapolation Problems in Modeling Fire Effects at Large Spatial Scales: a Review

Abstract

Models of vegetation change in response to global warming need to incorporate the effects of disturbance at broad spatial scales. Process-based predictive models, whether for fire behavior or fire effects on vegetation, assume homogeneity of crucial inputs over the spatial scale to which they are applied. Landscape disturbance models predict final burning patterns, but either do not model mechanistic behavior and explicit spread rates, or require large amounts of data to initialize simulations and predict ecological effects. Empirical data on the ecological effects of fire are not generally available at these scales, and conclusions are often extrapolated upward from stand-level data. Three methods for extrapolating ecological effects of fire across spatial scales and the sources of error associated with each were identified: (1) extrapolating fire behavior models directly to larger spatial scales; (2) integrating fire behavior and fire effects models with successional models at the stand level, then extrapolating upward; and (3) aggregating model inputs to the scale of interest. Extreme fire events present a challenging problem for modelers, regardless of which extrapolation method is employed. No single approach to modeling fire effects is inherently superior; modeling objectives and the characteristics of specific systems will determine the best strategy for each situation.