Coarse scale fire simulation strategies

Development of management-oriented computer models for coarse scale fire simulation is often problematic because of the tradeoff between realism and utility. This study compares three spatial modeling strategies of increasing complexity for simulation coarse scale succession and fire dynamics across the 80 million ha Interior Columbia River Basin (ICRB). In all three approaches, successional dynamics are modeled using a multiple pathway approach where seral community types, called succession classes, are linked along pathways that converge to a stable community type called a potential vegetation type which represents a unique biophysical setting that is static through simulation. Fire dynamics are simulated using three stochastic approaches implemented in three separate computer models. The Columbia River Basic succession Model (CRBSUM) simulates fire using probabilities from a uniform probability distribution that reflects a representative fire return interval CRBSUM was modified to create CRBSUM2 that simulates fire occurrence at the potential vegetation type level and subsequent fire effects at the successional class level. The FIREPAT (FIRE PATtern) model stochastically simulates fire starts from a three parameter Weibull function. Fire is spread across the landscape using a cookie-cutter approach where an exponential probability function determines the size of an ellipse wherein all pixels will be burned. Simulated landscape patterns and disturbance results are compared across the three approaches. Results indicate models of increasing complexity require additional parameterization and computer time, but provide more realistic results.