EGI OpenIR  > 研究系统  > 荒漠环境研究室
塔里木河流域TRMM降水数据精度评估
沈彬
学位类型硕士
导师李新功 ; 陈亚宁
2015
学位授予单位中国科学院大学
学位授予地点北京
学位专业自然地理学
关键词Trmm降水数据 塔里木河流域 精度评估 时空变化
摘要降水是全球能量和水循环的重要组成部分之一,在不同时空尺度的大气循环中扮演着不可或缺的角色,不同地区降水对全球气候变化的响应是十分明显的,所以正确描述区域降水对人类的生产生活有着极其重要的意义。地面气象站点观测可以精确测量一个地区的降水量,然而,受到经济、地形、海拔、交通等诸多因素的限制,气象站往往架设在平原地区,且密度不大,尤其在山区和沙漠等地形复杂地区布设的站点稀少,且存在着时间和空间不连续等缺陷,所以如何正确描述全球和区域降水特征成为一个难题。TRMM卫星的出现为解决观测站点稀疏地区的降水问题提供了重要支持。TRMM卫星数据作为有效的替代品,其覆盖面积广,在时间和空间上观测较为连续,使估算区域降水量和研究各种尺度天气系统方面得到了很大发展。 本文以塔里木河流域为研究靶区,以地面站点测量的降水数据为真实值,从不同时间尺度,分析了塔里木河流域不同区域下TRMM3B42和TRMM3B43降水数据的精度以及误差分布情况,并通过挑选出均一化降水数据,分析了TRMM数据由于自身空间尺度所带来的影响,对TRMM数据在塔里木河流域降水精度估算进行了评估。得到的主要结论有以下几点: (1)TRMM3B43卫星数据对塔里木河流域所有站点的年均降水量拟合较好(R2=0.8846),流域内24个站点平均年降水量相对偏差为19.02%,其中60%的站点表现为TRMM3B43年降水量高于地面实测年降水量。在塔里木河流域内TRMM卫星数据在相对低海拔的地方(1200 m以下)年降水量表现为高估,在相对高海拔(1200 m以上)的地方大多表现为低估,这可能与TRMM3B43卫星对不同海拔和地形的降水探测能力不同有关。 (2)塔里木河流域24个气象站点位置TRMM3B43月降水数据与测站记录的月降水量14 a平均值对比结果表明,TRMM3B43数据对降水的估计能够体现各个月份的降水情况。大部分站点拟合优度很好或者较好,但也有个别站点较差(于田、且末、乌恰)。 (3)就季节而言,春季拟合效果最好,夏季和秋季的TRMM3B43降水数据存在低估问题,而冬季则偏高估。流域降水主要集中于夏季,冬季降水较少。每年冬季,最大降水区域由流域西部向北偏移,到夏季分布于流域北部,到秋季最大降水区域开始向流域西北部恢复。 (4)塔里木河流域的降水量分布为由东南向西北地区递增,并在流域西北边缘地区增加较显著,形成一个相对丰水带;而向沙漠腹地方向延伸的降水量则呈减少趋势。总得来说,时间尺度越大,一部分误差被抵消,TRMM3B43数据表现得越准确。TRMM3B43数据在流域年降水量的精度大于季度降水量,而月降水量的精度相对最低。 (5)分析TRMM3B42数据在塔里木河流域24个气象站点对应位置的降水事件探测能力。大部分站点TRMM3B42数据对降水事件的发生存在系统性高估,流域北部相对于南部TRMM3B42正确探测降水事件的比例要高,而流域南部TRMM3B42的错误探测降水事件的比例要高。 (6)相比之前的一些研究,塔里木河流域的TRMM3B42日降水数据精度较差,这可能和塔里木流域复杂的下垫面有关:流域内山区、平原、盆地均有分布,地形复杂多变,海拔高度变化明显,地表植被覆盖更是差异巨大。受到不同下垫面的影响,会对卫星数据的反演产生较大的影响。 (7)TRMM3B43和TRMM3B42的检验结果对比可以发现,TRMM3B43月尺度数据无论在相关系数,还是相对偏差方面,都要明显优于TRMM3B42日尺度数据。 (8)各站点均一化降水数据相比均一化前降水数据表现出更好的拟合度,更小的误差,说明TRMM数据由于自身空间尺度问题对其精度评估有着一定的影响。在一定空间范围内的均一化降水事件更能体现出TRMM3B42数据在对应地面降水站点的准确程度。
其他摘要The Tarim River basin, a typical arid area in northwestern China, faces serious water shortage that has adversely restricted local ecological systems and socio-economic development. Since precipitation is the main source of surface water, understanding the spatial-temporal distribution of precipitation is rather important for further water usage. However, precipitation gauge stations are sparse in the basin due to economic imperative, high elevation, complex topography and severe climate, especially in mountains and desert of the basin, making the precise estimation of spatial-temporal precipitation distribution in the region a difficult problem. However, there is great potential to to use satellite estimate of precipitation in the area. As a precipitation radar satellite, TRMM (Tropical Rainfall Measuring Mission) has collected plentiful fine spatial-temporal precipitation data, which makes feasible the use of TRMM precipitation data to study hydrological and climatic characteristics in the Tarim River basin. It is necessary to assess the accuracy of TRMM data before using it. In this study, TRMM Multi-Satellite Precipitation Analysis (TMPA) 3B43 product is evaluated using the precipitation measurement from 24 weather stations (2000-2013) in the Tarim River basin. Results show that: (1) compared with the rain gauges, TRMM3B43 data demonstrated a good accuracy in the whole study area at the annual time scale. The average bias is 19.02% and about 60% of TRMM precipitation is higher than that it from the weather stations. (2) Most sites have a good fit at monthly time scale except at few sites (Yutian, Qiemo and Wuqia). The goodness of fit between TMPA precipitation and observed precipitation is better in spring. In summer and autumn, the TRMM underestimated precipitation, and in winter, it overvalued. (3) Precipitation in the Tarim River basin shows a gradual increasing trend from southeast to northwest, and especially dominant on the edge of the northwest forming a relatively abundant water belt. When moving towards the desert hinterland area, precipitation shows a decreasing trend. (4) Changes of the maximum precipitation in the region have certain regularity in a year. Every winter, precipitation moves from west to north in the basin, and stays in the northern part of the basin in summer. When autumn comes, it moves back to the northwest of the basin. Overall, the TRMM 3B43 data reflects the spatial and temporal distribution of precipitation in the basin. And the coarser the time scale is, the more accurate the TRMM data shows, as errors cancel out with coarser scales. This is why TRMM data in the basin has the highest accuracy at annul scale, while relatively lower accuracy at quarterly and monthly scales. The Tarim River lies between Tianshan mountains and Kunlun mountains, and located in the mid-latitude westerlies control area of northern hemisphere, so changes of the basin rainfall distribution and the maximum precipitation area are influenced by the terrain and seasonal water vapor sources.(5)We analyzed the detection ability of TRMM3B42 data about precipitation event at the corresponding location of 24 meteorological stations in Tarim river basin. Results show that the detections in most stations are systematically overvalued, and the detection accuracy of TRMM3B42 data to precipitation events in northern basin is higher than that in southern basin, while the error rate is higher in southern basin is higher than that in northern basin.(6)Compared to previous researches, the accuracy of TRMM3B42 daily precipitation data is poorer in Tarim river basin. This may be related to the complex underlying surface in Tarim river basin, such as the complex topography including mountains, plains, and basins, the obvious elevation changes, and the enormously different vegetation cover, because the inversion by satellite data will be greatly impacted by the complex underlying surface.(7)The TRMM3B43 month scale data are better than TRMM3B42 daily scale data in both the correlation coefficient and the relative deviation by comparing the inspection results of TRMM3B43 and TRMM3B42.(8)The homogenization precipitation data shows better fitting degree and smaller error than the original data in each site, indicating that the spatial scale limitation of TRMM data has to some extent influenced its accuracy assessment, and within a certain spatial range, the homogenization precipitation events can better reflect the accuracy degree of TRMM3B42 data in corresponding ground rainfall stations.
学科领域自然地理学
语种中文
文献类型学位论文
条目标识符http://ir.xjlas.org/handle/365004/14635
专题研究系统_荒漠环境研究室
作者单位中科院新疆生态与地理研究所
推荐引用方式
GB/T 7714
沈彬. 塔里木河流域TRMM降水数据精度评估[D]. 北京. 中国科学院大学,2015.
条目包含的文件
条目无相关文件。
个性服务
推荐该条目
保存到收藏夹
查看访问统计
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[沈彬]的文章
百度学术
百度学术中相似的文章
[沈彬]的文章
必应学术
必应学术中相似的文章
[沈彬]的文章
相关权益政策
暂无数据
收藏/分享
所有评论 (0)
暂无评论
 

除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。