Simulating natural fire size distribution in Ontario’s boreal forest biome under assumptions of climate

Background/Question/Methods Recurring wildfires define the spatial mosaic and ecological processes of the boreal forest biome. At broad spatial and temporal scales, characteristics of these wildfires are generalized as fire regimes, and used to infer ecological processes (e.g., carbon cycle, forest succession) and to design applications (e.g., wildlife habitat supply, forest management). Fire size distribution (FSD) is a widely used indicator of forest fire regimes that captures the synoptic relationship between number of fires and their respective sizes. However, most estimates of boreal FSD are based on limited empirical samples of past fires, and do not inform of potential variability under scenarios of climate.  To explore consequences of assumed climate scenarios in a boreal forest landscape, we used the mechanistic fire regime model BFOLDS to simulate the natural fire regime in four large ecoregions (4 -13 million ha) in Ontario, Canada. BFOLDS is a semi-stochastic, spatially explicit, raster-based, event-driven model that incorporates both probabilistic and mechanistic processes. It simulates fire and succession over large forested extents at 1 ha resolution, and long time periods at one year interval overall. We assumed a factorial combination of three climate scenarios and two spatial patterns of fire ignition for simulations of FSD, with 50 replicates of each simulation set.

Results/Conclusions

Our results show that truncated power-law distribution best fits all simulated FSDs, across ecoregions and assumptions. However, the specific parameters of the FSDs varied among ecoregions, with an east-west trend that reflects broad geo-climatic patterns in Ontario.  Overall, the assumption of a drier and warmer climate produced significantly different FSD than that for a wetter and cooler climate, while the magnitude of the difference varied across eco-regions.  With these results we demonstrate the utility of mechanistic simulation of fire regime to explore emergent properties that cannot be discovered using historical fire data.