降水资料匮乏地区不同再分析数据降尺度效果的评价  

Evaluation of Downscaling Effect of Different Reanalysis Data in Regions with Insufficient Precipitation Data

作　　者：喻雪晴 穆振侠[1] YU Xue-qing;MU Zhen-xia(College of Water Conservancy and Civil Engineering,Xinjiang Agricultural University,Urumqi 830052,China)

机构地区：[1]新疆农业大学水利与土木工程学院,新疆乌鲁木齐830052

出　　处：《水电能源科学》2020年第9期5-8,23,共5页Water Resources and Power

基　　金：新疆维吾尔自治区自然科学基金项目(2018D01A16);国家自然科学基金项目(51969029,51469034,51569031);新疆农业大学水利工程重点学科研究项目(SLXK2018-02)。

摘　　要：匮乏的降水基础资料不利于分析研究天山西部区域,因此研究外源降水资料的适用性、校验及降尺度,可更好地服务于水资源的高效利用、提高洪水预报的精度及极端水文事件的应对能力等。基于现有的26个实测站点1979~2005年的日降水量数据、同期NCEP/NCAR、ERA-interim、CFSR和CanESM2数据,借助极限学习机法研究了资料匮乏地区降水降尺度及修正方法。结果表明,基于四种不同数据集所建立的统计降尺度模型基本上能够模拟各站点的降水时空分布情况;综合对比四种数据集,ERA数据模拟降水的效果最佳,但不同站点最佳的外源降水存在一定的差异;相较于比例缩放法,分位数映射法修正降水误差更能充分考虑降水的变异性,能有效地提高降水模拟精度。The related research of precipitation laws have been greatly restricted by the lack of basic data in the western region of Tianshan.The implementation of the research on the applicability,calibration and downscaling of external precipitation data can better serve the efficient utilization of water resources,improve the accuracy of flood forecast and the coping capacity of extreme hydrological events.Based on the daily precipitation data between 1979 and 2005 of the current 26 measured stations,NCEP/NCAR,ERA-interim,CFSR and CanESM2 data in the same period,extreme learning machine was used to study the precipitation downscaling and correction method in the data deficient area.The results show that the statistical downscaling model based on four different datasets can basically simulate the spatial and temporal distribution of precipitation at each station.Compared with the four datasets in a comprehensive way,ERA data has the best effect in simulating precipitation,but there are some differences in the best external precipitation of different stations.Compared with the scaling method,the quantile mapping method can enhance the accuracy of precipitation simulation as a result of it can correct the precipitation error and take the variability of precipitation into full consideration.

关 键 词：降水 降尺度 再分析数据 误差修正 

分 类 号：P339[天文地球—水文科学]