摘要
本研究对赤峰市华北落叶松人工林的潜在适宜分布等级区进行估测,确定影响其分布的主要环境因子。基于赤峰市8个旗县区范围内的华北落叶松人工林地理分布点1300个,利用最大熵模型以气候因子、土壤因子和地形因子做为环境变量,通过独立运算和综合运算的方式对赤峰市华北落叶松人工林适宜分布区进行估测。模型运算采取默认参数,设置25%和30%的验证数据比例,以训练AUC和验证AUC作为模型精度评价指标。结果表明:30%验证数据比例的训练AUC及验证AUC与25%的模型运算结果差异不显著,选择30%验证数据比例,综合气候因子、土壤因子和地形因子的模型预测精度最高,训练AUC和验证AUC值分别为0.880和0.877。影响模型预测的环境因子重要性排序及环境变量贡献率排序依次为:地形因子、气候因子和土壤因子。地形因子中的高程变量是影响华北落叶松人工林分布的主要因子。华北落叶松人工林按适生等级划分,高度适生区主要分布在海拔800m~1200m的区域范围,极高度适生区主要集中分布在海拔1200m以上的区域,其预测适宜分布结果与实际分布相符且趋势一致。
Based on the 1300 geographical distribution sites of Larix principis-rupprechtii plantation in the eight districts of Chifeng,combined with climate data,soil data and terrain data as environment variables,the maximum entropy model was used to estimate the potential distribution areas of Larix principis-rupprechtii plantation in Chifeng by independent calculation and comprehensive calculation.The default parameter settings were used to run the model,the validation data were set as 25%and 30%for training AUC and test AUC to evaluate the model precision.The model AUC data results showed that there was no significant difference between validation data 25%and 30%.So validation data 30%was selected as the setting.The results show that the prediction accuracy of the model combined with climate factors,soil factors and terrain factors was the highest,and the AUC values of training and verification were 0.880 and 0.877,respectively.The importance order and contribution rates of environmental variables that affect the model prediction is terrain data,climate data and soil data.Elevation in terrain variables is the main factor influencing the distribution of Larix principis-rupprechtii plantations in Chifeng.MaxEnt prediction results show that the highest suitable areas mainly distributed in the areas with elevation higher than 1200 m,and the higher suitable ecological area mainly distributed in the areas with elevation between 800 m~1200 m.The predict results consistent with the actual distribution and trend.
作者
穆喜云
乌志颜
李显玉
王芳
白晓旭
郭淑文
程瑞春
于胜利
MU Xiyun;WU Zhiyan;LI Xianyu;WANG Fang;BAI Xiaoxu;GUO Shuwen;CHENG Ruichun;YU Shengli(Chifeng Academy of Forestry,Chifeng 024000;Wangyedian Forest Farm of Harqin Banner,Chifeng 024423,China)
出处
《干旱区资源与环境》
CSSCI
CSCD
北大核心
2021年第6期144-152,共9页
Journal of Arid Land Resources and Environment
基金
“十三五”国家重点研发计划,子课题《华北落叶松大径材培育与经营调控技术》(2017YFD060040305)
内蒙古自然科学基金项目(2020BS03019)资助。
关键词
最大熵(MaxEnt)模型
华北落叶松
适生区
现实分布
maximum entropy(MaxEnt)model
Larix principis-rupprechtii Mayr
potential geographical distribution
realized distribution
