Evaluation of the Monthly Drought Code as a metric for fire weather in a region of complex terrain and un- certainties in future projections

Evaluation of the Monthly Drought Code as a metric for fire weather in a region of complex terrain and uncertainties in future projections 1 INTRODUCTION Wildfire is an integral part of ecosystems. It is a major source of often positive disturbance in North Ameri-can forests which plays a critical role in determining ecosystem properties such as age structure, species abundance, and landscape patterns. Wildfire activity is, in turn, strongly dependent on climate. Globally, spatial distributions of wildfire occurrence are strongly dependent on spatial patterns in temperature and precipitation, as these variables determine both the fuel moisture and fuel availability of a particular ecosystem (Krawchuk et al. 2009). Previous studies have demonstrated relationships between a variety of climate variables (such as temperature, precipitation and drought severity) and area burned throughout western North America (e.g., Littell et al. 2009). As well, a variety of studies have linked strong anti-cy-clonic conditions (i.e., " blocking highs ") to enhanced fire activity (e.g., Skinner et al. 1999). Given the strong connection between climate and wildfire, it is likely that a changing climate will have a substantial effect on future wildfire severity and frequency. Historically, annual values of total area burned have increased across many regions of North America over the last 50 years (Mouillot and Field, 2005). Some of this increase can be attributed to human activities of the last century, including changes in fire suppression policy, logging practices, and increased population density (Allen et al. 2002). However , in some cases fire suppression has not affected natural fire regimes (Schoennagel et al. 2004). Moreover , Gillett et al. (2004) were able to attribute some of the historical increase in wildfire across Canada to anthropogenic global warming. Numerous studies have attempted to provide future projections of wildfire severity or frequency (see Flan-nigan et al. 2009 for a review and IPCC 2012). In general , these studies use global climate models (GCMs) to provide future climate information from which projected changes in fire weather or actual wildfire activity are inferred. Techniques used in the studies vary in sophistication. Some studies used raw GCM output (e.g. Flannigan et al. 1998), while others have used a " Delta " downscaling approach where monthly GCM anomalies are used to adjust historical daily time-Changes in both temperature and precipitation due to climate change will have a significant impact on future fire weather severity. The availability of a simple but skillful …