Can climate change increase fire severity independent of fire intensity?

We tested the idea that climate may affect forest fire severity independent of fire intensity. Pervasive warming can lead to chronic stress on forest trees (McDowell et al. 2008; Raffa et al. 2008), resulting in higher sensitivity to fire-induced damage (van Mantgem et al. 2003). Thus, there may be ongoing increases in fire severity (the number of trees killed), even when there is no change in fire intensity (the amount of heat released during a fire). We examined this question at a subcontinental scale by synthesizing existing information from plot-based prescribed fire monitoring databases across the western United States of America (USA). Prescribed fire data are particularly well suited to exploring the relationship between climate and fire severity because prescribed burns are conducted over a relatively narrow range of fire weather but over a potentially wide range of inter-annual climatic conditions. Specifically, we considered two topics, (i) quantifying the contribution of climate to fire severity (as measured by post-fire tree mortality), and (ii) detecting any secular trends in fire in the climate/fire severity relationship. Statistical models based on data from \u3e330 forest plots showed that across regions and major taxa, probabilities of fire-caused tree mortality were strongly sensitive to pre-fire changes in climatic water deficit, an index of drought. Our downscaled climate data indicated that changes in the climatic water deficit were due to increasing temperatures, without detectable trends in precipitation. These climatic trends were correlated with increasing probabilities of fire-caused mortality over time. Results from this study demonstrate that incorporating measures of pre-fire climatic stress and/or tree health into models of post-fire mortality used by prescribed fire managers may substantially improve their predictive capabilities. The relationships developed here will help managers predict changes in fire severity from large-scale climatic anomalies (e.g., ENSO, PDO) and from secular trends in climate