基于环流分型的多站点日降水随机模拟

The increasing global surface temperature has tended to induce a more active hydrological cycle. Assessing the impact of climate change on water cycle has been becoming a research hotspot and demanded urgently by economic and social development. Downscaling technique can be used in transferring climate data generated by General Circulation Models (GCMs) with coarser temporal-spatial resolution to data with higher temporal-spatial resolution required by hydrological models. Downscaling technique is a key technology in the assessment of climate/climate change impact. Stochastic simulation of precipitation is a challenging problem and has attracted more and more attention. Using daily precipitation data collected from 501 meteorological and hydrological stations and runoff data from six hydrological stations distributed over Haihe basin, Wudinghe basin, and Qingjianhe basin located in North China and Loess Plateau and reanalysis data of ECMWF, NCEP and JRA-55, following topics are studied: (1) analyze probability from a wet/dry day to a wet day, and the probability distribution of precipitation; (2) establish spatial correlation within stations by precipitation intensity classification; (3) analyze anisotropy of spatial correlation; (4) establish relationship between circulation patterns and the probability characteristics of precipitation as well as the spatial correlation; (5)compare two simulation frames for building multi-site spatial correlation of daily precipitation. The purpose is to develop a multi-site daily precipitation model conditioned on circulation patterns, which tends to improve the simulation ability for extreme precipitation, low-frequency variability and spatial correlation of stations, etc.. A software is going to be developed based on the research results to convenient the end users for impact assessment of climate/climate change on hydrology process, water resource and soil erosion for large scale basin.

日益升高的全球地表温度可能导致更为活跃的水循环。评估气候变化对水循环的影响逐渐成为研究热点，同时为社会经济发展迫切所需。降尺度方法将气候模式输出的低时空分辨率气候资料转化为水文模式所需的高分辨资料，是气候影响评价的关键技术。降水随机模拟是统计降尺度方法中的难点和热点。基于我国华北和黄土高原地区海河、无定河和清涧河流域共501个气象和水文站日降水及6个水文站径流资料，结合ECMWF，NCEP和JRA-55再分析资料，分析降水发生概率和降水量概率分布；分不同雨强建立站点降水空间相关性；分析降水空间相关性的各向异性；建立环流分型与上述降水概率特征和空间相关性的关系；对比单阶段和双阶段两种多站点日降水随机模拟框架，改进模型在降水极值、低频变率以及空间相关性等方面的模拟能力，发展基于环流分型的多站点日降水随机模型，并形成计算机应用模型，服务于气候/气候变化对大尺度流域的水文水资源和土壤侵蚀评价。

日益升高的全球地表温度可能导致更为活跃的水循环。评估气候变化对水循环的影响逐渐成为研究热点，同时为社会经济发展迫切所需。降尺度方法将气候模式输出的低时空分辨率气候资料转化为水文模式所需的高分辨资料，是气候影响评价的关键技术。降水随机模拟是统计降尺度方法中的难点和热点，而反映热量及其变化的气温也是随机模拟中首先需要考虑的气象要素。研究基于全国2208个气象站点日降水、日气温资料和欧洲中期数值预报中心ECMWF、美国国家环境预报中心NCEP再分析资料，基本阐明了中国大陆地区日降水空间相关统计特征，探究了嫩江流域大气环流和日降水之间的关系，在我国中东部地区4个不同类型区域和流域（嫩江流域，黄土高原区域，湘江流域和龙川江流域）对比评估了研究计划中设计的单阶段和双阶段多站点日降水模拟框架的评估（MulGETS模型和TSWG模型），同时加入了基于频域分析方法的模拟框架的评估（EHS模型），并进一步发展模型，使其能模拟非平稳态日降水和日气温。EHS模型在日降水、日最高气温、日最低气温极值、低频变率和空间相关模拟方面，均优于MulGETS模型和TSWG模型。已发表研究论文11篇，其中SCI论文7篇，核心期刊论文4篇，已接收核心期刊论文2篇；已形成基于EHS模拟框架的多站点日气温随机模拟模型，并形成计算机软件应用模型，获得软件著作权1项；研发的多站点日气温随机模拟模型在2022年初服务于北京冬奥会和冬残奥会高温融雪风险概率预测，通过冬奥河北气象中心为奥组委提供23期雪质风险服务专报，为奥组委部署比赛和雪场管理人员维护雪场提供决策支撑。研究同时可为气候/气候变化对流域水循环和土壤侵蚀的影响评估提供方法和借鉴。