遥感降水数据精度检验策略及检验方法综述  

Strategy and Method for Satellite Precipitation Data Evaluation:An Overview

作　　者：郭瑞芳[1] 刘元波[1] Guo Ruifang;Liu Yuanbo(State Key Laboratory of Lake Science and Environment ,Nanjing Institute of Geography and Limnology ,Chinese Academy of Sciences ,Nanjing 210008 ,China)

机构地区：[1]中国科学院南京地理与湖泊研究所湖泊与环境国家重点实验室,江苏南京210008

出　　处：《遥感技术与应用》2018年第6期983-993,共11页Remote Sensing Technology and Application

基　　金：国家自然科学基金重点项目"通江湖泊干旱的多成因机制研究"(41430855);国家高技术研究发展计划项目"星基地综合定量遥感系统与应用示范"(2013AA12A301)资助

摘　　要：卫星遥感反演降水技术为获取全球降水信息提供了途径。遥感降水数据不可避免地存在误差,精度检验是必不可缺的。首先回顾了遥感降水数据全球检验项目;总结检验策略,分为降水事件检验、中小区域尺度检验和大区域尺度检验;然后总结检验方法,包括基于地面观测数据检验方法和基于降水产品检验方法;最后介绍遥感降水数据检验常用流程,包括获取参照数据、检验区域选择和数据对比。遥感降水产品精度检验框架主要存在3个方面的问题,全球遥感降水产品检验策略不完善、参照数据缺乏统一标准和检验指标不统一。未来发展方向可能有完善检验策略、统一参照数据及检验方法。Satellites offer a way of estimating global precipitation.Errors exist inevitably in satellite precipitation products,and have a great deal of difference among products because of the complexity of estimating or retrieving precipitation.This paper firstly outlined global projects of satellite precipitation product evaluationincluding Algorithm Intercomparison Project (AIP),Product Intercomparison Project (PIP),and Program to Evaluate High Resolution Precipitation Products (PEHRPP).These projects concluded that the accuracy of precipitation products is close related with underlying surfaces,latitude,seasons and precipitation patterns.The HRPPs (TMPA,CMORPH,NRLB and GSMaP)underestimate rainfall ranging from 3 to 7 mm/day(10%-67%).PERSIANN overestimates heavy rainfall (200%)while underestimates rainfall (56 %)in the mountains.Then it concluded evaluation strategies including rainfall events evaluation,evaluation at meso-small scale reglon and evaluation at large scale region.Evaluation methods were divided into two categories,evaluation based on ground measurement data and evaluation based on satellite precipitation products.We introduced general evaluation procedure,including obtaining reference data,study area selection and data comparison.The reference datasets mainly include gauge data,Global Precipitation Climatolo- gy Centre (GPCC)datasets and radar data.Finally,we made a prospect on satellite precipitation product evaluation based on the existing problems at present.There currently exist three problems in satellite precipitation products evaluation,imperfect evaluation strategy,non-uniform reference data and evaluation indices.The future research aspects may include optimization of evaluation strategy,normalization of reference data and evaluation methods.

关 键 词：精度 遥感降水数据 检验策略 检验方法 尺度 

分 类 号：P412.27[天文地球—大气科学及气象学] TP75[自动化与计算机技术—检测技术与自动化装置]