A fire regime is an emergent property of an ecosystem, generated by the interaction of many factors operating at different spatio-temporal scales. Due to the obvious danger that fire poses to humans and their property, we are typically unable to design and run "experiments" on fire-prone landscapes to understand the mechanisms that control long-term fire regime dynamics. Fire simulation models are possibly the only way we can test the sensitivity of fire regimes to variation in different factors.

We are currently using a new fire simulation model, HFire, which has been parameterized for chaparral-dominated shrublands. This model was created by Marco Morais while he was a graduate student at UCSB working with the Southern California Wildfire Hazard Center. Most fire spread models (e.g., FARSITE) are typically designed to simulate individual fires, instead of running iteratively on a landscape. This is because fire spread models are usually used to evaluate how a single fire would behave in a given set of circumstances.

With the invaluable help of Marco Morais and a research assistant, Lora Summerell, this simulation environment has allowed the testing of fire regime sensitivity to certain input parameters. Our goals have included validation of the model (i.e., generating fire regimes that approximate historical patterns) and identification of thresholds in fire regime dynamics as environmental parameters change.

In collaboration with Jean Carlson (Physics, UCSB) and John Doyle (Control and Dynamical Systems, Cal Tech), we are also investigating what drives changes in fire size distributions and how fire regimes may be organized. In viewing fire regimes as emergent phenomena in complex systems theory, we have applied a new mechanism called "Highly Optimized Tolerance" (HOT) to explain underlying structure in fire size distributions.