Introduction

Wildfire and insect infestations are two major disturbances of forest lands in the United States. Historically, insect infestations and wildfires have had a dominant influence on successional processes in forests of the Western United States (Agee 2003). Fire suppression over the past 100 years has resulted in larger, more severe wildfires and insect outbreaks (Hessburg and others 1994). In 2005, over 0.14 million hectares (.34 million acres) of Federal lands in Oregon and Washington were affected by wildfires (http://www.nifc.gov) and approximately 0.8 million hectares (2.1 million acres) sustained damage from insects such as bark beetles and defoliators (http://www.fs.fed.us/r6/nr/fid). In response to concerns over the size and severity of wildfires and insect outbreaks, Federal land management agencies have adopted forest management strategies that call for restoration activities that include reintroduction of natural and prescribed fire over wide areas of the Western United States. These activities have demonstrated beneficial effects in terms of moderating wildfire, but the potential effects on insect dynamics and insect-caused tree mortality are less clear.

A reciprocal and synergistic association has frequently been described between fire and insects. For example, tree mortality resulting from insect outbreaks has been seen as setting the stage for subsequent wildfires (Geiszler and others 1980, Parker and Stipe 1993). Conversely, wounding and mortality from fire can create focus trees, which act as magnets for bark beetles (McCullough and others 1998). In separate studies, McHugh and others (2003) and Cunningham and others (2005) followed tree mortality and beetle attacks for 3 years after wildfire and found bark beetles more likely to attack trees with fire injury. Similar results were found following prescribed fire (Bradley and Tueller 2001, Wallin and others 2003). However, others have found that beetles did not preferentially attack trees with fire-injured boles, but attack success was higher in injured trees when beetle population levels were low (Elkin and Reid 2004). Additionally, Sanchez-Martinez and Wagner (2002) found low population levels of bark beetles regardless of stand treatment history, including stand replacement by wildfire. These results confirm that the association between insects and fire is a complex one, particularly when evaluated over time and at a large scale.

Numerous studies have begun to examine functional and numerical interactions between insects and fire at the tree and stand level, but few quantitative studies have been carried out that consider the spatiotemporal dynamics of wildfire and insect outbreaks at the landscape scale (Barclay and others 2005, Bebi and others 2003, Fleming and others 2002, Kulakowski and others 2003, Lynch and Moorcroft 2004, Veblen and others 1991). Existing analytical techniques for spatiotemporal analysis of multiple interacting disturbances are not well developed, and large-scale datasets suitable for studying interactions are few (Lynch and Moorcroft 2004).

The present study describes a framework for estimating probabilities, at a landscape level, of various forest disturbances as a tool for quantifying associations between multiple interacting disturbances. The framework builds on related work by the authors and collaborators on estimating wildfire probabilities in relation to weather and fire danger indices and spatial location (Brillinger and others 2003, Preisler and Westerling 2005, Preisler and others 2004).