湖北省杉木栽培区划及气候特征研究

Study on cultivation regionalization and climate features of Cunninghamia lanceolata in Hubei province

通过收集湖北省杉木人工林地理信息和温、湿度相关气候因子数据,利用ArcGIS软件和最大信息熵模型,对湖北省杉木栽培区域进行划分,并找出影响杉木栽培的主要气候因子以及适宜栽培的区域,为杉木人工林的科学布局提供依据。结果表明:1)湖北省杉木气候区划可以分为3个等级:最适宜区、较适宜区和不适宜区,其中最适宜区面积为1.86×10~6 hm^2,较适宜区面积为3.23×10~6 hm^2,适宜区域主要为鄂西山区和鄂东南低山丘陵区;2)模型ROC曲线检测AUC值为0.916,利用二类清查人工林信息验证准确率达到96.22%,说明模型预测效果较好;3)刀切法分析对预测贡献率较大的主要为反应水分条件及水热同步的指标,排名前4位的因子分别为最湿季平均温度、最湿季降水量、年降水量和最湿月降水量。

In order to classify cultivation regions and provide references for layout of Cunninghamia lanceolata in Hubei province,the software ArcGIS and MaxEnt were used to analyze the main climate factors affecting the fir cultivation and the suitable cultivation region with environmental data of temperature and humidity and the known distribution of plantations. The results show that 1) The climate division of Chinese fir in Hubei province can be divided into three grades: the most suitable area, the suitable area and the unsuitable area. The most suitable area was 1.86×10^6 hm^2, and the suitable area was 3.23×10^6 hm^2. The suitable areas were mainly located in the mountainous areas in Western Hubei and hilly areas in southeastern Hubei. 2) The AUC value of model ROC curve detection was 0.916, and the accuracy of verification by using the information of second-class inventory plantation reached 96.22%, which shows that the model has a good prediction effect. 3) The results of knife-cut analysis show that the main factors contributing to the prediction were the reaction water condition and the synchronization index of water and heat, the top four factors were the average temperature in the wettest season, the rainfall in the wettest season, the annual rainfall and the monthly rainfall in the wettest season.