Evaluating Satellite Imagery for Estimating Mountain Pine Beetle-Caused Lodgepole Pine Mortality: Current Status

Spatial accuracy in the detection and monitoring of mountain pine beetle populations is an important aspect of both forest research and management. Using ground-collected data, classification models to predict mountain pine beetle-caused lodgepole pine mortality were developed for Landsat TM, ETM+, and IKONOS imagery. Our results suggest that low-resolution imagery such as Landsat TM (30 m) is not suitable for detection of endemic level populations of mountain pine beetle. However, good results were obtained for pixels with groups of red beetle-killed lodgepole pine (> 25 trees killed per 30-m pixel), implying that Landsat imagery is most suited to detection of populations at the building or epidemic phase. Preliminary results using high resolution IKONOS imagery (4 m) suggest that detection of individual or small groups of red beetle-killed lodgepole pine can be accomplished with a relatively high accuracy. Introduction The mountain pine beetle (Dendroctonus ponderosae Hopkins Coleoptera: Scolytidae) is one of the most important drivers of vegetation change in lodgepole pine (Pinus contorta) forests. Outbreaks of these insects can be truly impressive events, with annual losses that are often greater than fire or any other natural disturbance. Mountain pine beetle populations can erupt rapidly, resulting in large increases in tree mortality within a few years. Timely forest management is contingent upon population monitoring and detection of beetle-caused tree mortality. Mountain pine beetle populations persist at endemic levels in single attacked trees scattered across a landscape. Population monitoring at this level can be difficult. Given appropriate weather and stand conditions, beetle success increases and groups of trees begin to be attacked. At the outbreak level, thousands of hectares with up to 70% mortality can occur. One promising avenue for detection of tree mortality caused by mountain pine beetles at various population levels is the use of remotely sensed data. Remotely sensed data can be used for detecting visual, and through the near infrared bands, nonvisual physiological changes in vegetation. Numerous studies have investigated the use of satellite-based digital remote sensing for the characterization of forest ecosystems and changes that occur within Mountain Pine Beetle Symposium: Challenges and Solutions. October 30-31, 2003, Kelowna, British Columbia. T.L. Shore, J.E. Brooks, and J.E. Stone (editors). Natural Resources Canada, Canadian Forest Service, Pacific Forestry Centre, Information Report BC-X-399, Victoria, BC. 298 p.