Prediction Model for Postfire Mortality of Pinus yunnanensis in Central Yunnan Province

Prediction model for post-fire mortality based on fire resistance of Pinus yunnanensis and the information at damage trees after forest fire could provide theory basis for estimating damage of forest fire and designing schemes of vegetation restoration in burned area in short time after forest fire. Post-fire mortality and crown-fire characteristics of Pinus yunnanensis trees was investigated in the burned area of ldquo3middot29rdquo Forest Fire in Anning area in Central Yunnan Province in China in 2006. Seven parameters-tree height in short TH, diameter at breast height in short DBH, bark average thickness BAT, rate of blackened trunk in uprightness in short RBTU, the most rate of blackened width at trunk exterior circumference in short MRBC, rate of carbonized average depth at basal trunk in short RCAD and whether flowing resin in short WFR were determined when 208 individual Pinus yunnanensis trees were investigated . The prediction model was established with learning vector quantization in short LVQ in artificial neural network in short ANN, and C# was used to realize the prediction model. The tested results showed that the average accuracy rate of the model was 88.75%, the accuracy rate to dead trees of the model was 93.07% , the accuracy rate to alive trees of the model was 73.21% ,so the model could be applied for discrimination and prediction of post fire mortality of Pinus yunnanensis trees.