基于随机森林插值的中亚夏季极端高温变化特征  

Extreme summer high-temperature changes in Central Asia based on interpolated data from random forest

作　　者：孟欣宁 焦瑞莉[1] 刘念 夏江江[2,3] 严中伟 于爽[2] 娄晓 李昊辰 王立志 陈亮 郑子彦[2] 赵娜[6] MENG Xin-ning;JIAO Rui-li;LIU Nian;XIA Jiang-jiang;YAN Zhong-wei;YU Shuang;LOU Xiao;LI Hao-chen;WANG Li-zhi;CHEN Liang;ZHENG Zi-yan;ZHAO Na(Beijing Information Science and Technology University,Beijing 100101,China;The Institute of Atmospheric Physics,Chinese Academy of Sciences,Beijing 100029,China;University of Chinese Academy of Sciences,Beijing 100049,China;School of Science,Beijing University of Posts and Telecommunications,Beijing 100876,China;Peking University,Beijing 100871,China;Institute of Geographic Sciences and Natural Resources Research,Chinese Academy of Sciences,Beijing 100101,China)

机构地区：[1]北京信息科技大学,北京100101 [2]中国科学院大气物理研究所,北京100029 [3]中国科学院大学,北京100049 [4]北京邮电大学理学院,北京100876 [5]北京大学,北京100871 [6]中国科学院地理科学与资源研究所,北京100101

出　　处：《干旱区研究》2020年第4期966-973,共8页Arid Zone Research

基　　金：中国科学院战略性先导科技专项(A类)资助(XDA20020201)。

摘　　要：利用中亚地区65个气象站的逐日最高气温数据,结合ERA-Interim再分析资料以及经纬度、海拔数据,构建了随机森林插值模型,并验证了其可靠性。基于该模型补全了气象站缺失值,获得完整的站点逐日最高气温数据集TStationf,并插值得到中亚1979—2016年空间分辨率为0.75°×0.75°的逐日最高气温格点数据集TRFIMG。基于TRFIMG进一步分析了中亚1979—2016年夏季极端高温指数时空变化特征。结果表明:中亚区域平均极端高温指数增速在0.22~0.30℃·(10a)-1,显著增温的区域主要分布在哈萨克斯坦的西部、土库曼斯坦大部、乌兹别克斯坦东南部等地区。基于TRFIMG得到的夏季极端高温指数增速显著大于基于TStationf得到的结果,这表明用站点观测数据对该地区夏季极端高温趋势的估计明显偏低。本研究得到的数据集可在一定程度上弥补使用站点观测数据片面刻画中亚地区极端高温变化的缺陷,有助于更确切地引导人们在应对极端天气气候事件时采取相应的减缓和适应措施。In this study,a random forest interpolation model was constructed based on daily maximum temperature observations,ERA-Interim reanalysis data,latitude,longitude,and altitude data from 65 meteorological stations in Central Asia.Based on this model,we filled the missing daily maximum temperature data(TStationf)of the meteorological stations.A daily maximum temperature grid dataset(TRFIMG)for Central Asia was created with a spatial resolution of 0.75°×0.75°from 1979 to 2016.Based on TRFIMG,the trends of extremely high temperatures in Central Asia in summer from 1979 to 2016 were then analyzed.The results show that the rate of increase of the extreme high temperature indices range from 0.22°C·(10a)-1to 0.30°C·(10a)-1,and the important warming areas are mainly distributed in western Kazakhstan,most of Turkmenistan,and southeastern Uzbekistan.The increased rates of summer extreme high temperature indices based on TRFIMGare significantly higher than those based on TStationf.This indicates that the estimation of extreme summer high temperatures in this region using the station observations are considerably underestimated.The dataset(TRFIMG)obtained in this study can,to some extent,offset the shortcomings of using station observations to partially describe the extremely high-temperature changes in Central Asia.It can also properly guide people to take corresponding mitigation and adaptation measures in response to extreme weather and climatic events.

关 键 词：随机森林插值 机器学习 夏季极端高温 中亚 

分 类 号：P423[天文地球—大气科学及气象学] P413