基于MAXENT模型的黑龙江大兴安岭森林雷击火火险预测

黑龙江大兴安岭是森林雷击火的高发地区,急需研发精确的火险预测模型对该区森林火灾进行预测.本文基于大兴安岭地区森林雷击火灾数据及环境变量数据,采用MAXENT模型进行森林雷击火的火险预测.首先对各环境变量进行共线性诊断,再利用累积正则化增益法和Jackknife方法评价了环境变量的重要性,最后采用最大Kappa值和AUC值检测了MAXENT模型的预测精度.结果表明: 闪电能量和中和电荷量的方差膨胀因子（VIF）值分别为5.012和6.230,与其他变量之间存在共线性,不能用于模型训练.日降雨量、云地闪电数量及云地闪回击电流强度是影响森林雷击火发生的3个最重要因素,日平均风速和坡向的影响较小.随着建模数据比例的增加,最大Kappa值和AUC值均有增大趋势.最大Kappa值都大于0.75,平均值为0.772; AUC值都大于0.5,平均值为0.859.MAXENT模型的预测精度达到中等精度,可应用于大兴安岭地区的森林雷击火火险预测.

Forest lighting fire forecasting for Daxing’anling Mountains based on MAXENT model.

Daxing’anling Mountains is one of the areas with the highest occurrence of forest lighting fire in Heilongjiang Province, and developing a lightning fire forecast model to accurately predict the forest fires in this area is of importance. Based on the data of forest lightning fires and environment variables, the MAXENT model was used to predict the lightning fire in Daxing’anling region. Firstly, we studied the collinear diagnostic of each environment variable, evaluated the importance of the environmental variables using training gain and the Jackknife method, and then evaluated the prediction accuracy of the MAXENT model using the max Kappa value and the AUC value. The results showed that the variance inflation factor (VIF) values of lightning energy and neutralized charge were 5.012 and 6.230, respectively. They were collinear with the other variables, so the model could not be used for training. Daily rainfall, the number of cloud to ground lightning, and current intensity of cloud to ground lightning were the three most important factors affecting the lightning fires in the forest, while the daily average wind speed and the slope was of less importance. With the increase of the proportion of test data, the max Kappa and AUC values were increased. The max Kappa values were above 0.75 and the average value was 0.772, while all of the AUC values were above 0.5 and the average value was 0.859. With a moderate level of prediction accuracy being achieved, the MAXENT model could be used to predict forest lightning fire in Daxing’anling Mountains.