A spatial, climate-determined risk rating for Scleroderris disease of pines in Ontario

We used historical distribution data of Scleroderris disease (caused by the fungus Gremmeniella abietina var. abietina (Lagerb.) Morelet) in Ontario t o model its probability of occurrence as a function of climate factors. A l ogistic regression model of the probability of occurrence as a function of the mean temperature of the coldest quarter and the precipitation of the co ldest quarter was a very good fit. The concordance (index of classification accuracy) of the model was 84%. We subsampled the data repeatedly, generat ed new parameter estimates, and tested the predictions against data not inc luded in the model. Classification accuracy was similar for each subsample model; therefore, we concluded that the final model is stable. Gridded esti mates of the climate variables were used to spatially extend the two-variab le logistic regression model and produce a probability of occurrence map fo r Scleroderris disease across Ontario. The predicted map of probability of occurrence fits well with the map of the observed locations of the disease. These results lend credence to previous work that suggests that distributi on of Scleroderris disease is strongly influenced by climate. The classific ation results also suggest that this model is a useful tool for assessing t he risk of Scleroderris disease throughout Ontario.