阈值选择对高温时空变化特征的影响  

Impact of threshold selection on the spatiotemporal change characteristics of high temperature

作　　者：刘婷婷 朱秀芳[1,2] 孙劭 张世喆[1] 郭锐 徐昆 Liu Tingting;Zhu Xiufang;Sun Shao;Zhang Shizhe;Guo Rui;Xu Kun(Key Laboratory of Environmental Evolution and Natural Disasters,Ministry of education,Beijing Normal University,Beijing 100875,China;Institute of Remote Sensing Science and Engineering,Department of Geographical Sciences,Beijing Normal University,Beijing 100875,China;National Climate Center,China Meteorological Administration,Beijing 100081,China)

机构地区：[1]北京师范大学环境演变与自然灾害教育部重点实验室,北京100875 [2]北京师范大学地理科学学部遥感科学与工程研究院,北京100875 [3]中国气象局国家气候中心,北京100081

出　　处：《地理科学》2023年第4期726-736,共11页Scientia Geographica Sinica

基　　金：国家自然科学基金项目(42077436);国家重点研发计划(2019YFAO606900)资助。

摘　　要：在全球变暖的大环境下,极端高温事件的发生频率正在不断增加。尽管国内外已经有许多关于高温发生频率与时空分布特征的研究,但目前学界对于高温仍没有一个统一的定义。为了分析不同阈值检测出的高温事件的差异,基于中国国家级地面气象站均一化气温日值数据集(V1.0)中的1979―2019年的日最高温和日最低温数据,使用绝对阈值和相对阈值2种方法来定义高温,在此基础上计算绝对阈值高温日数、绝对阈值高温积温和相对阈值高温日数3个高温指标,使用Mann-Kendall检验、Sen's slope、Hurst指数和经验正交函数(EOF),对比分析不同阈值条件下计算的高温指标的变化趋势和时空分布的差异。结果显示:①全国大部分地区的高温呈现增加的趋势,但使用不同阈值的高温指标变化趋势以及持续性存在不同程度的差异。相对阈值法得到的高温日数呈显著增加的站点相比绝对阈值法的更多。绝对阈值法显示高温日数在新疆、华南地区和黄淮海地区的持续性较强,而相对阈值法显示高温日数在东南沿海和西南地区持续性较高。②不同阈值的高温指标的EOF反映的高温指标的时空分布模式同样存在明显不同,绝对阈值高温指标的第二模态反映的是华东地区南部和华南大部分地区与其他地区的反向分布模式,相对阈值高温指标的第二模态反应的是南北地区的反向分布模式。研究者需根据自己的研究目的和研究问题选择合适的高温阈值。Under the background of the global warming,the frequency of extreme high temperature events is increasing.Although there have been many studies on the frequency and spatiotemporal distribution of high temperature around the world,there is still no unified definition of high temperature.In order to analyze the differences in high temperature events detected by different thresholds,and to strengthen the understanding of the multi-dimensional characteristics of high temperature,we detected high temperature by using different threshold methods,including three absolute thresholds(32℃,35℃and 38℃)and two relative thresholds(90%and 95%)based on the daily maximum temperature and daily minimum temperature data from 1979 to 2019 in the daily data set of China's surface climate data.Then,we calculated three high temperature indicators(including the absolute threshold high temperature days,absolute threshold high temperature accumulated temperature and relative threshold high temperature days)and used Sen's slope,Mann-Kendall,Hurst index and empirical orthogonal function(EOF)to analyze their trends and spatiotemporal distribution patterns,and finally compared the difference in trends and spatiotemporal distribution patterns of the high temperature indexes calculated in terms of different thresholds.The results show that:1)The high temperature in most areas of China showed an increasing trend,but there are different in the trend and trend persistence of different high temperature indexes calculated from different thresholds.The number of stations with significant increase in high temperature days obtained by relative threshold method is much more than that by absolute threshold method.The absolute threshold method showed that the trend persistence of high temperature days was stronger in Xinjiang,South China and Huang-Huai-Hai Region,while the relative threshold method showed that the trend persistence of high temperature days was higher in southeast coast and central and western of China.2)There are also significant dif

关 键 词：高温定义 绝对阈值 相对阈值 HURST指数 经验正交函数(EOF) 

分 类 号：P732[天文地球—海洋科学]