基于统计降尺度的长江上游流域降水和气温的时空变化趋势研究

Trend Analysis of Precipitation and Temperature over the Upper Changjiang River Basin by Statistical-downscaled CMIP 5 Climate Projections

基于长江流域宜昌站以上集水区的气象站实测数据,采用等距离累积分布函数法(Equidistance Cumulative Distribution Function Method,EDCDFm),对国际耦合模式比较计划第5阶段(Phase Five of the Coupled Model Intercomparison Project 5,CMIP5)的2种排放情景下(RCP4.5,RCP8.5)8个气候模式模拟的降水和气温进行统计降尺度。在此基础上,采用平均绝对误差、相对偏差和相关系数评估EDCDFm模拟结果与历史阶段(1961-2005年)实测站点数据的模拟精度,并对长江上游流域降水和气温的变化进行趋势分析。结果表明EDCDFm降尺度的模拟结果与实测值的拟合程度高,月降水相关系数达到0.89以上,月平均气温的相关系数在0.98以上;2种排放情景下,长江上游未来时期(2010-2099年)的降水、气温均呈增加趋势,且气温显著性增加。

Based on the observed data of weather station above Yichang gauging station of Changjiang River Basin, precipitation and temperature from eight models under two different emissions scenarios （RCP4.5, RCP8.5） of Phase Five of the Coupled Model Intercomparison Project 5（CMIPS） were statistical downscaled using Equidistance Cumulative Distribution Function Method （EDCDFm）. Mean absolute error, relative deviation and coefficient of correlation were used for assessment the accuracy of EDCDFm simulation results during the historical period （1961-2005）. Finally, we analyzed the trend of precipitation and temperature in the upper reaches of Changjiang River. It shows that the EDCDFm downscaling simulation results are very high compared with the observed data, the correlation coefficient of monthly precipitation is about 0.89, the correlation coefficient of average monthly temperature is about 0.98. Under two emissions scenarios, precipitation and temperature over the upper Changjiang river Basin are increased during the period of 2010 to 2099, the temperature significantly increased.