基于MaxEnt模型预测湖南省榧树适宜分布区  

Identification of suitable distribution areas for Torreya grandis in Hunan province based on the MaxEnt model

作　　者：易烜 杨振宇 朱晋梅 谢旭 黄靓 刘文剑 王琢玙 吴磊[2] YI Xuan;YANG Zhenyu;ZHU Jinmei;XIE Xu;HUANG Liang;LIU Wenjian;WANG Zhuoyu;WU Lei(Qingyanghu State Owned Forest Farm,Ningxiang 410627,Hunan,China;College of Forestry,Central South University of Forestry&Technology,Changsha 410004,Hunan,China;The Forestry Department of Hunan Province,Changsha 410004,Hunan,China;Hunan Prospecting Designing&Research Institute for Agriculture Forestry&Industry,Changsha 410007,Hunan,China)

机构地区：[1]湖南省青羊湖国有林场,湖南宁乡410627 [2]中南林业科技大学,湖南长沙410004 [3]湖南省林业局,湖南长沙410004 [4]湖南省农林工业勘察设计研究总院,湖南长沙410007

出　　处：《中南林业科技大学学报》2024年第5期46-55,共10页Journal of Central South University of Forestry & Technology

基　　金：国家自然科学基金项目(31600160);中央财政林业科技推广示范项目([2023]XTO7号);中央财政野生动植物保护项目(65235024);湖南省自然科学基金项目(2021JJ31152)。

摘　　要：【目的】榧树Torreya grandis是我国特有的珍稀经济树种,预测榧树在湖南省的适宜分布区,探究影响其分布的主导环境因子,有助于其保护策略的制定和开发利用适宜地的选择。【方法】以湖南省193个野外调查榧树分布点为基础数据源,选取了22个影响其分布的环境变量,运用MaxEnt最大熵模型和ArcGIS10.7软件,分析影响其分布的主导环境因子,预测其在湖南省的适宜分布区。【结果】1)ROC曲线下训练数据集10次交叉验证拟合的平均AUC值为0.914,测试数据集10次交叉验证拟合的平均AUC值为0.867,标准差为0.074,模型预测精度可靠。2)限制榧树当前潜在适生区分布的4个关键因素依次是海拔、坡度、最干季平均温度以及最热季降水量,其百分比贡献率依次为29.40、16.90、8.60、7.30;置换重要性依次为53.30、14.60、5.70、6.50。43)榧树在湖南省的适生区总面积为9.114×10 km^(2),高适生区、中适生区、低适生区面积分别为0.872×10^(4)、2.659×10^(4)、5.583×10^(4) km^(2);榧树适生区主要集中在湘中、西、东南地区,高、中适生区集中分布在邵阳北部、南部、西南部,益阳西南部,永州中南部,娄底西北部以及怀化中南部。【结论】地形、气候等因子综合影响着湖南省榧树的地理分布格局,邵阳、益阳、永州、娄底、怀化等地区中,海拔较低、地势相对平缓、气候凉爽的溪谷地区可能适合榧树的引种栽培。【Objective】Torreya grandis is a rare economic tree species endemic to China.Predicting the suitable distribution area of T.grandis in Hunan province and exploring the dominant environmental factors affecting the distribution of T.grandis will help to formulate the conservation strategy for the species and select the suitable sites for development and utilization.【Method】193 field survey points of T.grandis in Hunan province were taken,22 environmental variables affecting the distribution of T.grandis were selected,and the MaxEnt model was used in conjunction with the ArcGIS 10.7 software to analyze the dominant environmental factors affecting the distribution of T.grandis and predict the suitable distribution areas of the species in Hunan province.【Result】1)the average AUC value of 10 cross-validated fits under the ROC curve for the training dataset was 0.914,and the average AUC value of 10 cross-validated fits for the test dataset was 0.867,with a standard deviation of 0.074,which made the model prediction results accurate;2)Altitude,slope,mean temperature in the driest quarter,and precipitation of warmest quarter are the four key factors limiting T.grandis’s current potential suitable distribution,with percentage contributions of 29.40,16.90,8.60,and 7.30,respectively;and replacement importance of 53.30,14.60,5.70,and 6.50,respectively.3)The total area of T.grandis suitable area in Hunan province was 49.114×10^(4) km^(2),4 and the areas of high,medium and low suitable distribution areas were 0.872×10^(4) km^(2),4242.659×10^(4) km and 5.583×10^(4) km^(2),respectively;the suitable distribution areas were mainly concentrated in the central,western and southeastern parts of Hunan province,and the high and medium suitable distribution areas were concentrated in the northern,southern and southwestern part of Shaoyang,southwestern Yiyang,central and southern Yongzhou,and the central and southern part of Yongzhou,and the southwestern part of Yiyang,and the southwestern part of Yongzhou.The suitable di

关 键 词：最大熵MaxEnt模型 榧树 适宜分布区 影响因子 

分 类 号：S718.545[农业科学—林学]