资料非均一性对分析淮河流域平均气温变化趋势的影响

Impact of Data Inhomogeneity on Analyzing Temperature Trends in Huai River Basin

均一性长序列气象资料是研究气候变化的重要基础,评估资料非均一性对分析淮河流域平均气温变化的影响,对于准确理解农业、生态和水资源等领域对气候变暖的响应具有一定指导意义。本研究基于国家气象信息中心提供的淮河流域172个台站1961-2018年逐日平均气温均一化数据和观测数据,采用一元线性回归计算气温变化速率,并通过定义非均一性影响及其贡献率两个术语,定量评估气候资料非均一性对分析淮河流域平均气温变化趋势的影响。结果表明:淮河流域96个气象站点的平均气温序列表现出不均一性,占总站点数的55.8%。均一化前、后,区域年平均气温均显著上升,但非均一性导致增温速率被低估,非均一性影响为-0.015℃·10a^(-1),贡献率达到-6.6%;空间上,20.3%站点受到正影响,升温速率被高估,非均一性贡献率平均为21.3%;35.5%站点为负影响,升温速率被低估,贡献率平均为-43.6%。四季区域平均气温的非均一性影响差异较小,春、夏、秋、冬四季分别为-0.016℃·10a^(-1)、-0.014℃·10a^(-1)、-0.016℃·10a^(-1)和-0.015℃·10a^(-1);但由于夏季增温速度最慢,非均一性贡献率绝对值最大,贡献率为-40.0%,春、秋和冬季的贡献率分别为-4.7%、-8.0%和-4.3%;在春、秋和冬季,非均一性主要影响各站点的气温变化速率,但夏季有11.6%的站点在均一化前后出现了气温升降趋势的转变。

Long homogeneous meteorological data is important to study climate change.Evaluating the impact of data inhomogeneity on analyzing average temperature trends in Huai river basin is value for accurately understanding the response of agriculture,ecology and water resources to climate change.In this study,based on the homogeneous data and observation data of daily average air temperature from the National Meteorological Information Center,the trends of annual and seasonal air temperatures during 1961 to 2018 were calculated using simple linear regression at 172 meteorological stations over Huai river basin.Then the impacts of data inhomogeneity on analyzing mean air temperature trends during 1961 to 2018 were evaluated by using two terms which were inhomogeneity impacts and contribution rates.The results showed that the average air temperature series of 96 meteorological stations in Huai river basin were inhomogeneous,accounting for 55.8%of the 172 total stations.Before and after the homogenization,the regional annual average air temperature both increased significantly.But the increasing rate was underestimated due to the data heterogeneity,and the influence was−0.015℃·10y^(−1)with a contribution of−6.6%.For each station,35 stations(20.3%)was positively affected and the warming rate was overestimated with an average contribution of 21.3%,while 61 stations(35.5%)were negatively affected and the warming rate was underestimated with an average contribution of−43.6%.The influences of inhomogeneity for four seasons showed little difference,which were−0.016℃·10y^(−1),−0.014℃·10y^(−1),−0.016℃·10y^(−1)and−0.015℃·10y^(−1),respectively.However,because of the slowest increasing rate,the absolute value of inhomogeneity contribution was largest in summer with a contribution of−40.0%.The contributions of spring,autumn and winter were−4.7%、−8.0%and−4.3%.Inhomogeneity mostly affected the temperature rising rate at each station in spring,autumn and winter,while led to a turning of temperature trend before and after homogenization at 20 stations(11.6%)in summer.