摘要
利用CMIP5中6种气候模式对长江源区降水和气温进行提取,利用最近距离法提取与实测气象站最接近点的降水气温数据。通过4种方法对6种气候模式模拟结果进行集成,对比选出最优集成方法来预测源区未来气候变化。结果表明,BP神经网络法集成效果最好,其次是多元回归法,算术平均法和加权平均法效果相当但效果相对最差。玉树站2020年~2100年降水持续增加,气温虽然也持续升高,但升温速率较历史记录变弱并将在2100年之前达到峰值。
Six kinds of climate models in CMIP5 are used to extract the precipitation and temperature in the source area of Yangtze River,and the nearest distance method is used to extract the precipitation temperature data closest to the measured weather station.The integration of simulation results of six climate models is carried out by four integrated methods,and the optimal integration method is selected to predict the future climate change in the source area.The results show that the BP neural network method has the best integration effect,followed by the multiple regression method,and the arithmetic average method and the weighted average method have the same effect but the worst effect.The precipitation in Yushu Weather Station will continue to increase from 2020 to 2100,and although the temperature also will continue to rise,the heating rate is weaker than the historical record and will reach its peak before 2100.
作者
金浩宇
鞠琴
曲珍
董小涛
郝振纯
JIN Haoyu;JU Qin;QU Zhen;DONG Xiaotao;HAO Zhenchun(State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering,Hohai University,Nanjing 210098,Jiangsu,China;Joint International Research Laboratory of Global Change and Water Cycle,Nanjing 210098,Jiangsu,China;Rikaze Branch of the Tibet Autonomous Region Hydrology Bureau,Rikaze 857000,Tibet,China;Bureau of Comprehensive Development,Ministry of Water Resources,Beijing 100053,China)
出处
《水力发电》
北大核心
2019年第11期9-13,共5页
Water Power
基金
国家重点研发计划项目(2016YFC0402704)
国家自然科学基金资助项目(51539003,41323001,51421006)
关键词
气候变化
CMIP5
集成方法
长江源区
climate change
CMIP5
integration method
the source area of Yangtze River
作者简介
金浩宇(1994—),男,安徽望江人,硕士研究生,主要从事水文预报和水资源管理研究;通讯作者:鞠琴.
