EGI OpenIR  > 研究系统  > 空间对地观测与系统模拟研究室
卫星降水产品在天山的适用性评价及降尺度研究
金晓龙
Subtype硕士
Thesis Advisor张弛
2017-05-01
Degree Grantor中国科学院大学
Place of Conferral新疆乌鲁木齐
Degree Discipline理学硕士
Keyword卫星降水数据 精度评价 降尺度 极端降水 时空变化 Satellite precipitation data Accuracy evaluation Downscalling Extreme precipitation Temporal and spatial variation
AbstractHigh accuracy precipitation data is very important for regional climate andhydrological research. Based on 199 meteorological stations, we evaluated 3 sets ofsatellite precipitation datasets (CMORPH, GPM, TRMM) in Tishan mountains usingvarious statistical method. And put forward a new downscaling method based onprecipitation intensity, the precipitation distribution information from GPM and PTPLSGeographic-Precipitation correlation model. Finally, we analyzed the temporal andspatial variation of precipitation in Tianshan mountains based on the downscaledprecipitation data. The main conclusions are as follows:(1) The satellite precipitation data perform better in summer and autumn, amongwhich, GPM has the highest correlation coefficients and minimum relative errors(PB≈10%) with meteorological records. GPM has the minimum error in each elevationzones. In addition, satellite datasets has a better skill to capture weak precipitationevents. In a word, GPM could capture the precipitation changes accurately in Tianshanmountains, and can be applied to the ecology and hydrology research.(2) The annual precipitation of downscaled TRMM data is consistent with theobserved values, with the value of R2 above 0.7, reduce the average error of originalTRMM data effectively, but there is obvious differences between the rainy season (4-10 months) and dry season (11-3 months). The simulated runoff driven by downscaledTRMM data could accurately reflect the change of runoff in Akesu Basin, with thecorrelation coefficient >0.82 and the NS coefficient of 0.48.(3) Precipitation in the northern slope of Tianshan mountains behaves moreregular, with little differences through the whole year, however, The precipitation ofsouthern slope has remarkable seasonality. The results of M-K test show that: there areless precipitation in central and western part of Tianshan, while the precipitation in east,south Tianshan and Yili valley are increase. The PRCPTOT, R95p, R99p and R10mmhave been significantly reduced in the past 20 years, while Rx1d, CWD showed asinificant increase trend. As a whole, the amount and frequency of extreme precipitation in Tianshan are decreasing.
Other Abstract高精度的气象数据对区域气候、水文研究至关重要。本文基于天山山区 2014~2016 年 199 个气象站点数据,应用较为广泛的两套卫星降水产品—— TRMM 与 CMORPH,选用均方根误差(RMSE),相关系数(R),相对误差(PB), 以及分类统计分析指标(FAR, POD, ETS, BIAS)等,评估了新一代卫星降水产品 ——Global Precipitation Measurement (GPM) 在天山山区的适用性。并在此基础 上提出了根据降水强度等级,结合 GPM 降水分布信息和 PTPLS 地理-降水相关 模型,对 TRMM 数据进行统计降尺度的新方法,最后基于降尺度后的降水数据, 分析了山区降水的时空变异特征。得到以下主要结论: (1) 三套产品在降水较多的夏秋季表现较好,相对于 TRMM 与 CMPRPH, GPM 与观测数据的相关系数最高(R≥0.6),相对误差最小(PB≈10%);在整个天山 山区,GPM 相对于其他两套产品表现出较低的误差范围(-55%~50%);GPM 在不 同的高程带内,均表现出同观测站点较低的误差与较高的相关系数;综合分析四 种指数,GPM 表现最佳,能够以较准确的精度和较低的误差估测降水。 (2) 降尺度后 TRMM 数据的年降水量与实测值基本一致,R 2 在 0.70 以上, 且能有效减小原始 TRMM 数据的平均误差,但其误差大小在雨季(4-10 月)和干 季(11-3 月)存在明显差异;由降尺度后 TRMM 数据驱动的水文模型,其模拟的 径流量与实测径流的相关系数高达 0.82,NS 系数为 0.48,能准确的反映阿克苏 流域的径流变化。 (3) 天山北坡的降水较规律,全年差异不大,而南坡的降水具有显著的季节 性,各季的降水差异较大;M-K 检验的结果表明,近 20 年来,天山中、西部降 水减少,而东天山、南天山及伊犁河谷地区的降水增加。大部分山区的年总降水 量、强降水量、强降水日数呈减少趋势,而最大一日降水量和连续湿日指数呈增 加趋势,山区的极端降水总量及频率都在减小。
Subject Area地图学与地理信息系统
Language中文
Document Type学位论文
Identifierhttp://ir.xjlas.org/handle/365004/14879
Collection研究系统_空间对地观测与系统模拟研究室
Affiliation中国科学院新疆生态与地理研究所
First Author Affilication中国科学院新疆生态与地理研究所
Recommended Citation
GB/T 7714
金晓龙. 卫星降水产品在天山的适用性评价及降尺度研究[D]. 新疆乌鲁木齐. 中国科学院大学,2017.
Files in This Item:
There are no files associated with this item.
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[金晓龙]'s Articles
Baidu academic
Similar articles in Baidu academic
[金晓龙]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[金晓龙]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.