Validating the Malheur model for predicting ponderosa pine post-fire mortality using 24 fires in the Pacific Northwest, USA
Fires, whether intentionally or accidentally set, commonly occur in western interior forests of the US. Following fire, managers need the ability to predict mortality of individual trees based on easily observed characteristics. Previously, a two-factor model using crown scorch and bole scorch proportions was developed with data from 3415 trees for predicting the probability of ponderosa pine (Pinus ponderosa Dougl. ex Laws.) mortality following prescribed fire. Here, we report validation of that model for broader application using data from 10 109 ponderosa pines in 17 prescribed fires and 7 wildfires, observed for 3 years post-fire, from east of the Cascade Range crest in Washington, Oregon and northern California. The overall rate of correct classification was 87.1% and the rate of correctly predicting mortality was 80.1%. Similar accuracy is reported when testing the model for small trees (<53.3-cm diameter at breast height), wildfire, prescribed fire, and when using a field guide that simplifies application of the model. For large trees (≥53.3-cm diameter at breast height), the overall rate of correct prediction was 93.6% and the rate of correctly predicting mortality was 65.2%. These results suggest the Malheur model is useful for predicting ponderosa pine mortality following fires in this region.