基于优化的最大熵模型预测中国高山栎组植物的历史、现状与未来分布变化  

Prediction of historical,present and future distribution of Quercus sect.Heterobalanus based on the optimized MaxEnt model in China

作　　者：王晓帆 段雨萱 金露露 王崇云[1,2] 彭明春[1,2] 李云[2] 王旭红 马云飞 WANG Xiaofan;DUAN Yuxuan;JIN Lulu;WANG Chongyun;PENG Mingchun;LI Yun;WANG Xuhong;MA Yunfei(Institute of Ecology and Geobotany,Yunnan University,Kunming 650504,China;College of Ecology and Environment,Yunnan University,Kunming 650504,China)

机构地区：[1]云南大学生态学与地植物学研究所,昆明650504 [2]云南大学生态与环境学院,昆明650504

出　　处：《生态学报》2023年第16期6590-6604,共15页Acta Ecologica Sinica

基　　金：云南省基础研究专项重大项目(202101BC070002);云南省科技厅重大项目《云南植被志》编研项目(KCS101002);云南大学科技创新类项目(202004175)。

摘　　要：气候变化影响着植物物种的地理分布,预测物种潜在适生区对认识其地理分布特征及发挥其生态价值与经济价值具有重要意义。以高山栎组植物为研究对象,运用R语言ENMeval数据包调整调控倍频(RM)和特征组合(FC)优化的最大熵(MaxEnt)模型和ArcGIS软件,基于160条高山栎组植物分布点记录和22个环境因子进行潜在适生区预测,探讨影响其地理分布的主要环境因子,同时预测其在末次间冰期(Last interglacial,LIG)、末次盛冰期(Last glacial maximum,LGM)、全新世中期(Mid Holocene,MH)、当代(Current)、2041—2060年(2050年)和2061—2080年(2070年)3种不同CO_(2)浓度排放路径下潜在适生区的空间分布格局及其质心变化趋势。结果表明:最优模型参数RM=2.5,FC=LQHPT时,MaxEnt模型复杂度和过拟合程度最低,模型预测准确性极高,AUC=0.960±0.15。综合刀切法贡献率、置换贡献率和单因子响应曲线可知影响其地理分布的主要环境因子是气温季节性变动系数(bio4)、高程(dem)、年降水量(bio12)和等温性(bio3),其适宜范围分别为435—625、1792—3978m、670—1050mm和41.5—50.3,累计贡献率高达81.6%,其中温度是影响高山栎组植物最重要的环境因子。当前气候条件下,高山栎组植物总适生区面积62.37×10^(4)km^(2),高适生区面积10.47×10^(4)km^(2),占总适生区面积的16.79%,集中分布于川滇横断山地区、滇中高原北部、藏东南以及黔西部分地区。各个时期间高山栎组植物的适生区面积差异较大,当代潜在适生区面积最小且破碎化最严重,可能由于气候变化及人类影响所致。未来各个时期潜在分布区面积均有增大的趋势,但RCP8.5情景下面积相较于其他情景有所降低,表明全球气候变暖背景下不利于高山栎组植物的长期生长。同时,质心分析表明未来气候条件下其分布有向低海拔和低纬度迁移的趋势。Climate change affects the geographical distribution of plant species,and it is of great significance to predict the potential suitable areas of plant species for understanding their geographical distribution characteristics and exerting their ecological and economic values.In this study,160 recorded occurrences and 22 environmental factors were used to predict the potential suitable areas of Quercus sect.Heterobalanus based on the MaxEnt model optimized by using ENMeval package in R 4.1.1 to adjust regularization multiplier(RM)and feature combination(FC),as well as on ArcGIS software.We probed main environmental factors affecting its geographical distribution.In addition,we predicted the spatial distribution pattern of potential suitable areas and the trend of centroid change in the Last interglacial(LIG),the Last Glacial Maximum(LGM),the Middle Holocene(MH),at present,as well as in future 2041-2060(2050)and 2061-2080(2070)under three different CO_(2)emission paths.The results showed that the parameters of the optimal model were RM=2.5,FC=LQHPT,the model complexity and overfitting were the lowest,but the prediction accuracy of the model was extremely high(AUC=0.960±0.15).Based on the contribution rate of Jackknife method,replacement contribution rate and single factor response curve,the main environmental factors affecting its geographical distribution were seasonal variation coefficient of air temperature(bio4),elevation(dem),annual precipitation(bio12),and isothermality(bio3).And their suitable ranges were 435-625(bio4),1792-3978m(dem),670-1050mm(bio12)and 41.5-50.3(bio13),respectively.The cumulative contribution rate of these four factors was as high as 81.6%.In general,temperature was the most important environmental factor affecting the plants of Quercus sect.Heterobalanus.Under the current climate conditions,the total suitable area of Quercus sect.Heterobalanus was 62.37×10^(4)km^(2) and the high suitable area was 10.47×10^(4)km^(2),accounting for 16.79%of the total suitable area.It was distributed in H

关 键 词：潜在适生区 气候变化 最大熵(MaxEnt)模型 高山栎组 

分 类 号：Q948.1[生物学—植物学]