Mountain pine beetle-caused lodgepole pine mortality from the 1980's and subsequent fire occurrence in Colorado

Abstract

A need for understanding the interaction between post-epidemic bark beetle forest stands and subsequent fire occurrence has escalated due to recent (1999-2010) unprecedented epidemic mountain pine beetle in lodgepole pine (Pinus contorta) forests of Colorado. Outbreaks of mountain pine beetle (Dendroctonus ponderosae Hopkins) populations in the early 1980' s resulted in substantial tree mortality across Colorado and provide a means to study past outbreak areas and subsequent fire occurrence. Beetle outbreaks on the Arapaho National Forest (NF) (1980-1987) and the White River National Forest (1981-1987), delineated by the USDA Forest Service Aerial Detection Survey (ADS), indicated approximately 76,900 ha were affected, with approximately 450,000 trees killed per forest. Mountain pine beetle-caused tree mortality is generally thought to increase subsequent fire occurrence and intensity but little scientific research supports this hypothesis. Thus, my objectives were to 1) determine whether there were differences in fire occurrences between lodgepole pine forests in Colorado, impacted or not impacted by previous outbreaks of D. ponderosae and 2) determine if fire occurrences in areas with mountain pine beetle-caused lodgepole pine mortality were related to topographic attributes, ignition type, and meteorological conditions. We used historic USDA Forest Service Aerial Detection Survey maps (1980 - 1990) in conjunction with USDA Forest Service digital fire location records to look for mountain pine beetle and ignition relationships. Sixty eight maps were scanned to spatially identify D. ponderosae-caused mortality in lodgepole pine forests over the Arapaho NF and White River NF. Using a GIS, the spatial relationship between mountain pine beetle caused mortality areas and subsequent fire occurrence was identified. During the summer of 2008, 57 ignition points were field assessed on the Arapaho NF and White River NF to verify the presence of mountain pine beetle-caused mortality prior to the fire as well as confirm the location of the recorded fire. Two of the 57 ignition points had evidence of trees killed by mountain pine beetle prior to the ignition. Tests for independence of fire occurrence and mountain pine beetle-caused mortality were conducted for the Arapaho NF and White River NF. Combined human and lightning-caused fire densities did not differ (a=0.05) between areas with and without mountain pine beetle-caused mortality on the Arapaho NF, however, there were more fires in areas with mortality from the mountain pine beetle than non-impacted areas on the White River NF. Densities of lightning-caused fires alone did not differ between nonbuffered and 50 m buffered mountain pine beetle-caused mortality areas and areas outside the mortality on the Arapaho NF or White River NF. Logistic regression was used to model the probability of an ignition occurring within the area of the 1980's mountain pine beetle-caused lodgepole pine mortality on each forest. Elevation of fire occurrence was the most significant variable explaining the cooccurrence of fire with mountain pine beetle-caused mortality. Spatial autocorrelation was significant within human and lightning-caused fires but not lightning-caused fires alone. Thus, lightning-caused fires were used to eliminate the non-random nature of human-caused fires. Probability density functions were created using elevation ranges of mountain pine beetle-caused lodgepole pine mortality, lightning-caused fires prior to 1980, and elevations of fire occurrences intersecting aerially detected mountain pine beetle-caused mortality post-1980 through 2005 ( during and post-outbreaks) to identify areas of highest probability independent of one another. Maximum probabilities for the occurrence of a lightning fire in a 1980' s mountain pine beetle-caused mortality area within the Arapaho NF was between 2710 m and 2815m while on the White River NF the highest probability was between 2600 m and 2900 m. Probability density functions for a lightning fire occurring in an area with mountain pine beetle-caused mortality were applied across each National Forest using a GIS. Our analysis suggests that 1980's mountain pine beetle-caused lodgepole pine mortality has not contributed to an increase in fire frequency over the subsequent twenty five years. The variability in fire occurrence and the homogeneity of mountain pine beetle outbreak-caused mortality, past and present, between the Arapaho NF and White River NF demonstrates the limited nature of comparability of this study to other locations.