基于DIVA-GIS的牡丹在中国适生区的预测  

Prediction of suitable habitat of Paeonia suffruticosa Andr.in China by DIVA-GIS

作　　者：圣倩倩 王晨 王立飞[1,2] 刘彦辰 耿兴敏 祝遵凌[1,2,3] SHENG Qianqian;WANG Chen;WANG Lifei;LIU Yanchen;GENG Xingmin;ZHU Zunling(College of Landscape Architecture,Nanjing Forestry University,Nanjing 210037,Jiangsu,China;Co-Innovation Center for Sustainable Forestry in Southern China,Nanjing Forestry University,Nanjing 210037,Jiangsu,China;College of Art&Design,Nanjing Forestry University,Nanjing 210037,Jiangsu,China)

机构地区：[1]南京林业大学风景园林学院,江苏南京210037 [2]南京林业大学南方现代林业协同创新中心,江苏南京210037 [3]南京林业大学艺术设计学院,江苏南京210037

出　　处：《经济林研究》2022年第1期205-213,248,共10页Non-wood Forest Research

基　　金：国家自然科学基金青年科学基金项目(32101582);江苏省博士后科研资助计划项目(2021K377C);中国博士后科学基金面上项目(2021M691602);江苏省高等学校自然科学研究面上项目(21KJB220008);江苏省自然科学青年基金项目(BK20210613);江苏省软科学研究项目(BR2021058);南京市园林绿化行业科技项目(YLKJ202011ZD);南京林业大学大学生实践创新训练计划项目(2020NFUSPITP0052);江苏省高校哲学社会科学研究一般项目(2020SJA0129);南京市绿化园林局科技项目(YLKJ202011ZD)。

摘　　要：【目的】比较分析以DOMAIN模型和BIOCLIM模型预测的牡丹在中国的适生区分布情况,进一步了解牡丹适生区分布现状和未来不同气候情景下牡丹适生区的变化情况,为牡丹的保护、引种与推广提供科学依据。【方法】基于DIVA-GIS软件,利用259个标本分布信息和19个生物气候变量,对牡丹在当前气候条件下的适生区及其在未来气候条件下可能发生的变化情况进行模拟;导出标本点气候数据进行主成分分析,分析出与牡丹分布关联度最高的气候因素。【结果】模型预测结果表明:当前气候条件下,牡丹适生区等级较高的区域分布于我国东部季风区中除东北地区、华南地区南部与云南西南部外的各省区,牡丹高等级适生区在渝、黔、陕、豫、湘、赣、浙、沪、苏、鲁、晋、京、津、冀等省(自治区、直辖市)中的分布尤为广泛,在其余省区中也有良好的分布。主成分分析结果表明:19个生物气候变量中,极端最低温(bio6)、温度季节性变化方差(bio4)、季节降水量变异系数(bio15)、最湿月降水量(bio13)对牡丹适生区的分布起主导作用。使用AUC值和Kappa统计值这两个重要指标,对DOMAIN模型和BIOCLIM模型的预测结果进行评价,结果表明:以DOMAIN模型与BIOCLIM模型对牡丹适生区预测结果的AUC值、Kappa统计值分别达0.887、0.633和0.842、0.594,均可满足预测精度的要求。【结论】当前气候条件下影响牡丹的适生区的最主要因素是温度,牡丹对水分的需求相对温度而言有更高的普适性,牡丹适生区的分布区域需有明显的季节性温度变化特点和适宜的最低气温,而不适宜的温度条件则会导致牡丹生长的停滞。预测认为,未来的牡丹适生区范围会有局部微动趋势,但整体变化不显著;对比分析DOMAIN模型和BIOCLIM模型的ROC曲线和Kappa值可知,DOMAIN模型的两个指标值都高于BIOCLIM模型的,DOMAIN模型对牡丹适生区的模【Objective】By comparing the suitable habitat of Paeonia suffruticosa Andr.in China predicted by DOMAIN model and BIOCLIM model,we can further understand the distribution status of suitable habitat of P.suffruticosa Andr.and the changes of suitable habitat under different climate scenarios in the future to provide scientific basis for the protection,introduction and promotion of P.suffruticosa Andr..【Method】Thisstudy uses distribution of 259 specimens and 19 bioclimatic variablesto predict the suitable habitat of P.suffruticosa Andr.under current climatic condition and possible changes in the future based on DIVA-GIS.Climatic data of sample points are derived for principal component analysis to analysis the climatic factors with the highest correlation with the distribution ofP.suffruticosa Andr..【Result】The results of model prediction showed that areas with higher suitable habitat grades of P.suffruticosa Andr.were mainly distributed in every province,autonomous region and municipality in the monsoon region of eastern China except northeast China,south parts of southern China and southwest parts of Yunnan.In Chongqing,Guizhou,Shaanxi,Henan,Hunan,Jiangxi,Zhejiang,Shandong,Shanxi,Beijing,Tianjin,and Hebei provinces,autonomous regions and municipalities,the areas with higher suitable habitat grades of P.suffruticosa Andr.were more broader,which also occupied a larger area in other provinces,autonomous regions and municipalities.According to the results of principal component analysis(PCA),the min temperature of coldest month(bio 6),temperature seasonality(bio 4),precipitation seasonality(bio15)and precipitation of wettest month(bio13)of 19 bioclimatic variables had the greatest influence on the suitable habitat of P.suffruticosa Andr..The area under receiver operatingcharacteristic curve and Kappa statistic,which wereimportant indicators,were used to evaluate the results of DOMAIN model and BIOCLIM model,and the results were 0.887,0.633 and 0.842,0.594,meeting the requirements of prediction accuracy.【C

关 键 词：牡丹 适生区 生物气候变量 DIVA-GIS 

分 类 号：S688.1[农业科学—观赏园艺]