Static and dynamic controls on fire activity at moderate spatial and temporal scales in the Alaskan boreal forest

Abstract

Abstract Wildfire, a dominant disturbance in boreal forests, is highly variable in occurrence and behavior at multiple spatiotemporal scales. New data sets provide more detailed spatial and temporal observations of active fires and the post‐burn environment in Alaska. In this study, we employ some of these new data to analyze variations in fire activity by developing three explanatory models to examine the occurrence of (1) seasonal periods of elevated fire activity using the number of MODIS active fire detections data set (MCD14DL) within an 11‐day moving window, (2) unburned patches within a burned area using the Monitoring Trends in Burn Severity fire severity product, and (3) short‐to‐moderate interval (<60 yr) fires using areas of burned area overlap in the Alaska Large Fire Database. Explanatory variables for these three models included dynamic variables that can change over the course of the fire season, such as weather and burn date, as well as static variables that remain constant over a fire season, such as topography, drainage, vegetation cover, and fire history. We found that seasonal periods of high fire activity are associated with both seasonal timing and aggregated weather conditions, as well as the landscape composition of areas that are burning. Important static inputs to the model of seasonal fire activity indicate that when fire weather conditions are suitable, areas that typically resist fire (e.g., deciduous stands) may become more vulnerable to burning and therefore less effective as fire breaks. The occurrence of short‐to‐moderate interval fires appears to be primarily driven by weather conditions, as these were the only relevant explanatory variables in the model. The unique importance of weather in explaining short‐to‐moderate interval fires implies that fire return intervals (FRIs) will be sensitive to projected climate changes in the region. Unburned patches occur most often in younger stands, which may be related to a greater deciduous fraction of vegetation as well as lower fuel loads compared with mature stands. The fraction of unburned patches may therefore increase in response to decreasing FRIs and increased deciduousness in the region, or these may decrease if fire weather conditions become more severe.

Keywords

• Alaska
• boreal
• remote sensing
• scale
• wildfire