Predictive models of whitebark pine mortality from mountain pine beetle

Abstract

Stand-level and tree-level data collected from whitebark pine ( Pinus albicaulis Engelm.) stands in central Idaho were used to estimate the probability of attack and mortality of whitebark pine caused by mountain pine beetle ( Dendroctonus ponderosae Hopkins) (Coleoptera: Scolytidae). Logistic regression models were calibrated from reconstructed pre-epidemic stand conditions and post-epidemic mortality levels resulting from a widespread mountain pine beetle outbreak that occurred from 1909 to 1940. Basal area (m 2/ha) and stand density index (SDI) were stand-level variables that completely differentiated stands into attacked or non-attacked categories. Whitebark pine stands with basal areas above 10 m 2/ha (44 ft 2/acre) or with an SDI above 80 had a 100% probability of being attacked. Tree diameter, basal area per 0.04 ha, trees per 0.04 ha, and number of stems in a tree cluster were significant predictors of individual tree attack ( p≤0.001) in logistic regression. The tree-level model may be used to estimate anticipated cumulative mortality in currently or potentially infested whitebark pine stands. Stand susceptibility to mountain pine beetle infestation may be identified from density (basal area) or relative density (SDI) thresholds. Predictor variables selected by the models corroborate the susceptible host characteristics identified in other mountain pine beetle–pine systems. This work presents evidence of the generality of host susceptibility characteristics across pine species and over elevation gradients.