首页|基于最大熵模型的云南思茅松潜在分布区

基于最大熵模型的云南思茅松潜在分布区

扫码查看
[目的]预测思茅松在云南省的潜在分布区,探讨影响思茅松潜在分布的主导环境变量,为思茅松适地适树、科学经营提供依据。[方法]基于多次全国森林资源清查云南省固定样地数据确定思茅松现实分布点,选取影响林分生长的19 个生物气候变量、3 个地形变量和11 个土壤变量,采用最大熵模型模拟思茅松在云南的潜在分布区,通过受试者工作特征曲线下面积AUC评价模型精度,使用刀切法分析环境变量对思茅松潜在分布区影响的重要程度,最后通过思茅松存在概率逻辑值采用相等间隔法对思茅松潜在分布区进行分类。[结果]思茅松Maxent模型预测结果准确性极好,10 折交叉验证AUC的均值为 0。921。根据模型各因子的最终贡献率,影响思茅松潜在分布的主要环境变量为温度和降水,其次是土壤、地形因子;刀切法检验结果显示,影响思茅松潜在分布区排名前五的主导环境变量为最干季平均温度(Bio9)、等温性(Bio3)、最湿月降水量(Bio13)、最湿季平均温度(Bio8)和降水量的季节性(Bio15)。思茅松的核心适宜区主要分布在普洱市,其次是临沧市和西双版纳州,玉溪市和红河州有少量分布。[结论]最大熵模型能够很好的预测云南省思茅松的潜在分布区,筛选出影响思茅松潜在分布区的主导环境变量及其阈值合理,形成的思茅松潜在分布适宜性等级图可为思茅松适地适树、科学经营提供依据。
Potential distribution of Pinus kesiya var.langbianensis in Yunnan province based on maximum entropy model
[Objective]To predict the potential distribution area of Simao pine(Pinus kesiya var.langbianensis)in Yunnan province and to explore the dominant environmental variables affecting the potential distribution,so as to provide the basis for the matching the site with trees and scientific management of Simao pine.[Method]The actual distribution of Simao pine was determined using National Forest Inventory data from permanent sample plots in Yunnan province.A total of 19 bioclimatic variables,3 topographic variables and 11 soil variables affecting stand growth were selected to simulate the potential distribution area of Simao pine in Yunnan using the maximum entropy model(Maxent).The accuracy of the model was evaluated by the area under the curve(AUC),and the jackknife method was used to analyse the importance of the influence of environmental variables on the potential distribution area of Simao pine.Finally,the potential distribution area of the Simao pine was classified using the equal interval method based on the logical value of the probability of existence(P).[Result]The Simao pine Maxent model has excellent prediction accuracy,with an average AUC of 0.921 over 10-fold cross-validation.According to the final contribution rate of each variable in the model,the main environmental variables affecting the potential distribution of Simao pine were temperature and precipitation,followed by soil and topographic variables.The jackknife test results showed that the top five dominant environmental variables affecting the potential distribution of Simao pine were mean temperature of driest quarter(Bio9),isothermality(Bio3),precipitation of wettest month(Bio13),mean temperature of wettest quarter(Bio8)and precipitation seasonality(Bio15).The core suitable area of Simao pine was mainly distributed in Puer city,followed by Lincang city and Xishuangbanna Prefecture,and a small amount was distributed in Yuxi city and Honghe Prefecture.[Conclusion]The maximum entropy model could well predict the potential distribution area of Simao pine in Yunnan province,and the selected dominant environmental variables affecting the potential distribution area of Simao pine and their thresholds were reasonable.The formed suitability grade diagram of potential distribution of Simao pine could provide a basis for matching site with trees and scientific management.

environment variablesmaximum entropy modelPinus kesiya var.langbianensispotential distribution

覃阳平、李华、雷相东、罗春林、陈棋

展开 >

国家林业和草原局 西南调查规划院,云南 昆明 650031

中国林业科学研究院 资源信息研究所,北京 100091

环境变量 最大熵模型 思茅松 潜在分布区

2024

中南林业科技大学学报
中南林业科技大学

中南林业科技大学学报

CSTPCD北大核心
影响因子:1.442
ISSN:1673-923X
年,卷(期):2024.44(11)