以广泛分布在中国北方典型草原的建群种长芒草为研究对象,利用Maxent模型对长芒草在中国当前及未来气候变化下的潜在分布区进行预测并对主要影响其分布的环境变量进行分析,结果表明,采用受试者工作特征曲线(receiver operating characteristic curve,ROC)对模型精度进行检验所得到的训练数据与测试数据的受试者工作特征曲线面积(area under ROC curve,AUC)分别为0.962和0.950,表明模型预测结果可靠,当前中国长芒草高适宜性分布区主要有5个,分别为黄土高原分布区、泰山-沂蒙山分布区、横断山分布区、藏南谷地分布区及天山分布区。在RCP2.6(representative concentration pathways 2.6)和RCP8.5(representative concentration pathways 8.5)两种气候情景模式下预测得到的2070年长芒草最适宜的潜在分布区有逐渐缩小的趋势。Jackknife检验对主导环境变量的筛选结果显示,影响长芒草分布的主要环境变量有地形粗糙度指数(terrain roughness index,tri)、9月降水量(precipitation 09,prec09)、气候湿度指数(climatic moisture index,topowi)、2月最高温度(maximum temperature 02,tmax02)、12月降水量(precipitation 12,prec12)和12月平均温度(average temperature 12, tavg12)。结果可为气候变化背景下中国典型草原的可持续管理提供科学依据。
英文摘要:
Stipa bungeanais a dominant species of typical grassland in the north of China.In this study,we project its potential distribution patterns under current and future climate scenarios using the Maximum Entropy Model,and we identify the major factors influencing the distribution patterns.Initial testing of the model indicated results were reliable.In current climate scenarios,there were 5highly suitable areas for S.bungeana: the Loess Plateau region,the Tai-Yimeng mountain region,the Hengduan mountains,the southern Tibet valley and the Tianshan mountain area.Under climate scenarios RCP 2.6and RCP 8.5(RCP denotes'representative concentration pathway')in 2070,the highly suitable areas for S.bungeana would decreased.The results from a jackknife test showed that terrain roughness index,September precipitation,SAGA-GIS topographic wetness,February maximum temperature,December precipitation and December average temperature of December were major environmental variables affecting S.bungeana distribution patterns.These results provide a theoretical basis for the sustainable management of typical grassland in China under climate change.