论文部分内容阅读
森林火灾是山东森林地区严重的环境问题之一。本研究采用山东2001—2010年MOD14A1每日1 km温度异常/火L3级产品与地形、植被、天气、人为和可访问性数据,分析评估了火灾发生原因;收集了林火发生/未发生相关的15个解释变量的空间数据,利用二项Logistic回归模型估计了解释变量的函数与林火存在的概率。结果表明,高火险区域主要集中在黄河三角洲、鲁西北平原,包括德州、菏泽、济宁、枣庄南部、临沂东南部;中火险主要在聊城、滨州、济南北部、淄博北部、潍坊东部、泰安、日照和青岛大部分地区(包括蒙山林区、沂山林区、五莲山林区、徂徕山林区、尼山林区、泰莱林区);低火险主要集中在济南南部、淄博南部、莱芜、青岛南部和胶东半岛(包括济南林区、崂山林区、鲁山林区、昆嵛山林区、牙山林区)。Logistic回归结果表明,影响火灾发生的因素依次是年均温度、CTI、TPI、人口密度、植被类型、年降水量、植被盖度、距道路距离、坡向、距居民地距离、农民纯收入指数、坡度、年平均相对湿度、DEM、年蒸发量。其中前7个EXP(B)都>1,对森林火灾发生的与否贡献较大。这些结果作为战略规划工具来更好预测森林火灾,也可作为一种战术指南帮助森林管理人员设计区域防火措施。
Forest fire is one of the serious environmental problems in Shandong forest area. In this study, the daily 1 km temperature anomalies / fire L3 products and terrain, vegetation, weather, man-made and accessibility data of MOD14A1 from 2001 to 2010 in Shandong were analyzed to evaluate the causes of fire occurrence and the correlation between forest fire occurrence and non-occurrence Spatial data of 15 explanatory variables were used to estimate the probability of explaining the function of variables and the existence of forest fires by using the binary logistic regression model. The results show that the areas with high fire risk are mainly concentrated in the Yellow River Delta and the northwestern plain of Shandong, including Dezhou, Heze, Jining, southern Zaozhuang and southeastern Linyi. The middle fires are mainly in Liaocheng, Binzhou, northern Ji’nan, northern Zibo, eastern Weifang, And most of Qingdao (including Mengshan, Yishan, Wulianshan, Dalaishan, Nishan, Taililain); low-fire insurance is mainly in southern Jinan, southern Zibo, southern Laiwu and southern Qingdao And the Jiaodong Peninsula (including Jinan Forest, Laoshan Forest, Lushan Forest, Kunyu Mountain Forest, Asan Forest). Logistic regression results showed that the factors affecting the fire occurrence were followed by average temperature, CTI, TPI, population density, vegetation type, annual precipitation, vegetation coverage, distance to the road, aspect, distance from residents, net income of farmers , Slope, annual average relative humidity, DEM, annual evaporation. Among them, the first seven EXPs (B) all have a value of> 1, which greatly contributes to the forest fire. These results serve as strategic planning tools to better predict forest fires and as a tactical guide to help forest managers design regional fire precautions.