White pine blister rust in the central Rocky Mountains: Modeling current status and potential impacts

A survey of limber pine to determine the geographical distribution, incidence, and severity of white pine blister rust (WPBR), caused by the introduced forest pathogen Cronartium ribicola, was performed in central and southeastern Wyoming and northern Colorado in 2002--2004. WPBR was present on 55% of the 504 survey plots. Incidence, the proportion of infected trees, ranged from 0 to 100% and averaged 15.5% over all plots and 28.0% on infested plots. Diameter class and crown class were significantly related to likelihood of infection by WPBR. Incidence varied significantly by elevation and slope position and did not vary by aspect, slope configuration, limber pine density, or degree of canopy closure.; In the summers of 2002--2004, 1258 survey plots were established to determine the distribution of Ribes species, the alternate host for the pathogen, growing within and in the vicinity of native white pine stands in Colorado and Wyoming. Species of Ribes were present in all study areas. The most commonly encountered species were R. cereum, R. inerme, R. lacustre, and R. montigenum. Densities and probabilities of occurrence were related to site variables and varied by Ribes species.; An analysis of canker growth rates was performed on 134 WPBR cankers harvested from limber pine. Mean annual total longitudinal (both up and down the branch or stem) growth rate was 8.4cm/yr. Annual proximal (toward the stem) canker growth rate averaged 4.9cm/yr, and circumferential growth rate averaged 6.5cm/yr. Longitudinal canker growth rate varied by branch diameter, branch height, and condition of the branch distal to the canker, but did not vary by study area.; Data collected in the field surveys was used to develop a series of regression and categorical and regression tree analysis models to predict risk and hazard of WPBR in Colorado. Risk models predicting the presence of WPBR employed meteorological, Ribes, and tree size data and resulted in good agreement between predicted and actual presence. Models developed to predict disease pressure and hazard also employed meteorological, Ribes, and plot-level data. The complexity of modeling disease epidemics across vast landscapes will likely require calibration of models to specific conditions found in Colorado.