| 1960～2018年中国高温热浪的线性趋势分析方法与变化趋势 |
| |
| 引用本文： | 张嘉仪,钱诚.1960～2018年中国高温热浪的线性趋势分析方法与变化趋势[J].气候与环境研究,2020,25(3):225-239. |
| |
| 作者姓名： | 张嘉仪 钱诚 |
| |
| 作者单位： | 成都信息工程大学大气科学学院,成都610225;中国科学院东亚区域气候—环境重点实验室,中国科学院大气物理研究所,北京100029;中国科学院东亚区域气候—环境重点实验室,中国科学院大气物理研究所,北京100029;中国科学院大学,北京100049 |
| |
| 基金项目： | 国家重点研发计划项目2018YFC1507701，中国科学院青年创新促进会2016075 |
| |
| 摘 要： | 高温热浪直接影响人体健康和作物生长。研究全球变暖背景下我国高温热浪发生率的趋势是气候变化研究的基本问题之一，可为人们的生产生活等提供重要的科学信息。目前对于高温热浪趋势的研究大都使用最小二乘（Ordinary Least Squares，OLS）方法估计趋势，结合学生t检验判断趋势的统计显著性。本文审视了以往常用方法在研究我国高温热浪发生率的线性趋势时的适用性。首先，以2018年东北局部地区因当年高温日数异常多而形成离群值的例子展开，说明OLS方法估计趋势时对离群值非常敏感，造成虚假趋势。进一步，通过正态分布检验和自相关计算，发现1960~2018年中国至少有91.14%站点、90.06%格点的高温日数和92.18%站点、87.74%格点的热浪次数的序列不服从正态分布，而且多数存在自相关。采用一种不易受离群值影响并考虑自相关的非参数方法，本文对1960～2018年中国站点和格点、4个典型区域以及全国平均的高温日数和热浪次数的线性趋势做出了更为准确的估计。研究发现，高温日数显著增多的站点主要出现在华南和西北地区，热浪次数呈显著增多趋势的站点目前几乎仅限于华南地区和新疆的个别站点；区域平均而言，仅有华南区域和西北区域的高温日数和热浪次数是显著增多的，华北区域和东北区域趋势并不显著；全国平均的高温日数和热浪次数都是显著增多的。本文对高温热浪的趋势及其显著性估计、统计预测的方法选择上有重要参考价值。
|
| 关 键 词： | 高温热浪 趋势估计 显著性检验 非正态分布 自相关 非参数方法 |
| 收稿时间： | 2019/9/1 0:00:00 |
Linear Trends in Occurrence of High Temperature and Heat Waves in China for the 1960-2018 Period: Method and Analysis Results |
| |
| ZHANG Jiayi,QIAN Cheng.Linear Trends in Occurrence of High Temperature and Heat Waves in China for the 1960-2018 Period: Method and Analysis Results[J].Climatic and Environmental Research,2020,25(3):225-239. |
| |
| Authors: | ZHANG Jiayi QIAN Cheng |
| |
| Institution: | 1.College of Atmospheric Sciences, Chengdu University of Information Technology, Chengdu 6102252.Key Laboratory of Regional Climate?Environment for Temperate East Asia, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 1000293.University of Chinese Academy of Sciences, Beijing 100049 |
| |
| Abstract: | High temperature and heat waves (HT and HW) directly affect human health and crop growth. Investigating trends in the occurrence of HT and HW is one of the fundamental issues of research on climate change and can provide valuable information for life and production. Most of the previous studies on trends in the occurrence of HT and HW used ordinary least squares (OLS) method to calculate the magnitude of the linear trend and then used the Student’s t-test to determine the statistical significance of this trend. This study examined whether traditional methods are suitable for estimating the trend of HT and HW occurrence in China. By showing a case of the annual HT count, with extremely excessive occurrences in 2018 at a station in northeastern China, the authors illustrate that the OLS method is sensitive to outliers and can give a spurious trend. In addition, through normality tests and autocorrelation calculations, we found at least 91.14% of stations and 90.06% of grid boxes for the annual HT count and 92.18% of stations and 87.74% of grid boxes for the annual count of HW in China are non-Gaussian, and majority of them have serial correlation. Applying a nonparametric method that is insensitive to outliers and takes serial correlation into account, we provide a more accurate estimate of linear trends in the annual HT and HW count for each station and grid box, four typical regional averages, and China area-average for the 1960–2018 period. The results show that stations with statistically significant upward trend in HT occurred mainly in South China and northwestern China, and HW stations occurred almost only in South China and in several stations in the Xinjiang Autonomous Region. In terms of the area-averaged time series of the annual HT and HW count, only South China and northwestern China show a statistically significant upward trend, while North China and northeastern China did not exhibit a significant upward trend; those of the average in China are statistically significant. This study provides reference information for choosing the method for estimating trends and their statistical significance and for statistical predicting for the occurrence of HT and HW. |
| |
| Keywords: | High temperature and heatwave Trend analysis Significance testing Non-Gaussian distribution Serial correlation Non-parametric method |
| 本文献已被 万方数据 等数据库收录！ |
| 点击此处可从《气候与环境研究》浏览原始摘要信息 |
| 点击此处可从《气候与环境研究》下载免费的PDF全文 |
