EGI OpenIR  > 研究系统  > 荒漠环境研究室
天山山区多源降水数据评估及时空格局研究
卞薇
Subtype硕士
Thesis Advisor陈亚宁
2018-06-01
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山区降水 中国天山山区 不同源降水资料 降尺度 降水时空分布 Mountain precipitation China Tianshan mountainous area Different source precipitation data Downscaling Temporal and spatial distribution of precipitation
Abstract近半个多世纪, 全球的增温速率达到 0.175°C/10a, 气候变暖加剧了水循环过程, 降水作为各尺度水循环过程的重要一环, 是重要的气候和气象变量, 在过去半个多世纪里, 降水量级、 频率和降水的时空分布等方面发生了重大变化。 山区降水为干旱内陆地区提供了丰富的水资源, 是重要的补给源区, 山区降水研究对于区域气候变化研究、 生态环境恢复以及水资源配置等领域具有重要意义。 气象观测站点能够提供准确的降水资料, 但山区由于地形复杂, 自然条件十分恶劣,气象站点极少且分布不均, 气象资料缺乏, 山区降水相关研究面临着较大的困难。气象站点插值数据集、卫星遥感降水资料和再分析降水资料是目前降水研究较为常见的替代资料, 因其时间连续性强、 空间覆盖范围广, 在各时空尺度的降水相关研究中得到广泛的应用。中国天山山区是西北干旱区众多河流、 农业生产和生活的重要水资源补给区, 是“中亚水塔” 的重要构成。 天山山区降水受到地形、 海拔、 下垫面等的影响, 表现为强烈的时空异质性。 同时中国天山山区的现有气象站点大多分布在海拔 2000 米以下, 且分布稀疏, 难以满足研究需要, 因此, 利用不同来源降水资料估算山区降水, 进而深入研究山区降水垂直分布的时空变化及其差异性成为广大学者关注的热点。本文则针对中国天山山区气象站点稀少且分布不均, 降水观测资料缺乏等问题, 基于已有气象资料, 使用 MATLAB、 ARCGIS、 SPSS 等软件, 评估不同源降水资料 2000-2015 年间在中国天山山区的适用性(包括 APHRODITE、TRMM3B43、 ERA-Interim 和 GLDAS 数据集) ; 并基于降水与植被之间的密切相关性对不同源降水资料进行降尺度处理和分析; 最后利用精度更高的降水资料分析中国天山山区降水在时间和空间上的分布特征。 研究结论主要如下:1、 APHRODITE、 TRMM3B43、 ERA-Interim 和 GLDAS 降水资料在中国天山山区有一定的适用性。 整体上而言, 基于站点插值获取的 APHRODITE 数据和基于遥感卫星获取的 TRMM3B43 数据的表现优于 ERA-Interim 和 GLDAS 数据, 在天山山区的适用性更高。2、 2000-2015 年间, TRMM3B43 数据呈现最优表现, 并且该数据在降水较多的夏季、 降水较多的海拔高度(500-1290m) 和降水较多的西部与实测数据间的相对偏差较小,对低海拔地区降水存在高估, 对冷季高海拔地区降水存在高估,在暖季则低估高海拔地区降水。 而 ERA-Interim 数据整体上对实测数据存在一定程度的低估, 且在降水较多的月份和区域与实测数据间的相对偏差增大。 GLDAS数据高估实测降水, 且存在“坦化” 现象, 相对来说表现最差。 在对 2000-2007年不同数据集与站点观测降水进行对比时发现, APHRODITE 与站点降水间的一致性最好(年降水达到了 0.95 的显著性相关)。3、 分析空间分辨率为 1km×1km 的 MODIS NDVI 数据的时空分布与降水量间的关系, 发现在月和年时间尺度上, NDVI 数据与降水数据均存在明显相关,在年时间尺度上,与四种降水产品均达到 0.7 以上的显著相关。所以能够利用 1km空间分辨率的 NDVI 数据, 使用求和法对各种来源的降水数据资料进行降尺度处理。4、 利用求和法对四种降水产品的多年月平均降水和年平均降水进行了降尺度研究并进行精度验证, 在时间尺度上, 四种降尺度后的降水数据资料在月时间尺度上与站点实测数据的变化趋势具有一致性, APHRODITE 数据集保持了较高的相关性和较低的相对偏差。在季节时间尺度上,在 2000-2015 年间,TRMM3B43数据整体表现较优, 三种数据整体表现为高估, 冬季表现较差。 在年时间尺度上,APHRODITE 表现最优, 在 2000-2015 年间, TRMM3B43 和 ERA-Interim 数据表现较好。 在空间尺度上, 对 32 个气象站点的实测数据与对应格网数据集的相对偏差做了分析, 发现四种数据对实测降水整体上表现为高估, 对降水较少的吐鲁番站、 达坂城站和七角井站存在明显的高估, 而对昭苏、 巴音布鲁克和吐尔尕特等站的降水存在低估, 5-9 月间格网数据与实测数据间的相对偏差较小。5、 基于经过验证和降尺度的降水资料, 解析中国天山山区降水在时间和空间上分布的特征, 发现在时间尺度上, 中国天山山区三大分区降水主要集中在4-9 月份, APHRODITE、 TRMM3B43、 ERA-Interim 和 GLDAS 的月最大降水量分 别 在 15.7~62mm (2000-2007 年 平 均 ) , 17.8~56.8mm , 3.3~50.7mm 和17.9~51.3mm 之间。 而月最大降水量在不同分区也存在着一定的差异, 伊犁河谷的月最大降水量四种降水产品均显示在 50mm 以上, 东天山月最大降水量最低在20mm 以下, 北天山要高于南天山。 在季节上, 各个分区的夏季降水量最多, 冬季降水最少。 与实测降水最为接近的 APHRODITE 数据显示中国天山山区2000-2007 年均降水量为 450.5mm, TRMM3B43 数据显示出中国天山山区2000-2015 年间的年均降水量为 551.6mm。6、 在空间尺度上, 首先在水平方向上, 受水汽来源的影响, 中国天山山区降水呈现自西向东逐渐降低的趋势, 并且山区降水明显多于平原区, 北天山降水多于南天山多于东天山。 在垂直方向上, 中国天山山区降水的最大降水高度带随着月份不同而存在差异。 在1-3月和12月, 天山山区降水量随着海拔升高呈现增加的趋势, 4-12月间, 在海拔2500-3000m间存在最大降水高度带, 海拔高度小于2500m时, 月平均降水量随高程增加而增加, 当海拔大于3000m时, 降水量随高程增加出现缓慢减少的趋势。中国天山山区年平均降水的最大降水高度带因区域不同而存在差异, 东天山、 北天山和南天山出现在2500-3000m之间, 而典型流域伊犁河谷出现在2000-2500m之间。
Other AbstractOver the past half-century, the global warming rate was 0.175 ° C / 10a, Climatewarming exacerbates the water cycle.Among them, precipitation is an important partof the water cycle process at all scales and it is an important climate andmeteorological variable. Over the past half-century, major changes have taken place inprecipitation magnitude, frequency, and the temporal and spatial distribution ofprecipitation. Precipitation in mountainous areas is an important source of rechargefor arid inland water resources. Research on mountain precipitation is of greatsignificance to the research of regional climate change, ecological restoration andrational allocation of water resources. The meteorological observation stations canprovide accurate precipitation data. However, due to the complex terrain, poor naturalconditions, the sparse and uneven distribution of meteorological stations and the lackof meteorological data in mountainous areas, studies on precipitation in mountainousregions are facing great difficulties. Site interpolation datasets, remote sensingprecipitation products and reanalysis precipitation data are the more commonalternative data for precipitation research at present. They are widely used inprecipitation-related studies on various spatio-temporal scales because of their strongtemporal continuity and wide spatial coverage.Chinese Tianshan Mountain is an important water supply area for many rivers andagricultural production and living in the northwest arid area, and is an important partof "Central Asia Water Tower". Affected by topography, elevation, underlying surface,etc., precipitation in Tianshan Mountains has a strong temporal and spatialheterogeneity. At the same time, most of the existing weather stations in China'sTianshan Mountains are distributed below 2000 meters above sea level and aresparsely distributed to meet the research needs. Therefore, using the precipitation datafrom different sources to estimate the precipitation in the mountainous areas, andfurther studying the spatial and temporal variations of the vertical distribution of precipitation in the mountainous areas and their differences has become the focus ofattention of the majority of scholars.In this paper, aiming at the rare and uneven weather stations in mountainous areasof China's Tianshan Mountains and the lack of precipitation data, based on theexisting meteorological data, using MATLAB, ARCGIS, SPSS and other softwares toassess the applicability of different source precipitation data in China's TianshanMountains in 2000-2015 (including APHRODITE, TRMM3B43, ERA-Interim andGLDAS datasets); Based on the close relationship between precipitation andvegetation, the scaling of different source precipitation data was studied. The temporaland spatial variation characteristics of precipitation in the Tianshan Mountains ofChina were analyzed based on the higher resolution precipitation data. The mainconclusions are as follows:1. APHRODITE, TRMM3B43, ERA-Interim and GLDAS precipitation data havecertain applicability in China's Tianshan Mountains. Overall, the APHRODITE dataobtained from site interpolation and the TRMM3B43 data obtained from remotesensing satellites performed better than ERA-Interim and GLDAS data and are moreapplicable in the Tianshan Mountains.2. The data of TRMM3B43 showed the best performance between 2000 and 2015,and the data showed relatively small deviation from the measured data in the summerwith more precipitation, the altitude with more precipitation (500-1290m) and themore precipitation in the west, there is an overestimation of precipitation in lowaltitude areas, overestimation of precipitation in high altitudes in cold seasons, andunderestimation of precipitation in high altitudes in warm seasons. However, theERA-Interim data as a whole have underestimated the measured data to a certainextent, and the relative deviations in the months and regions with the measured dataincrease. The overall GLDAS data overestimates the measured precipitation, andthere is a phenomenon of "decontamination", which is the poorest relativeperformance. Comparisons of observed precipitation between different datasets andstations during 2000-2007 found that APHRODITE had the best agreement with precipitation at the site (annual precipitation reached a significant correlation of 0.95).3. Analyzing the relationship between the temporal and spatial distribution ofMODIS NDVI data with the spatial resolution of 1km×1km and precipitation, we findthat there is a significant correlation between NDVI data and precipitation data onmonthly and yearly time scales. On the time scale, significantly correlated with allfour precipitation products reaching above 0.7. Based on the NDVI data, it is feasibleto use the summation method to downscaling precipitation products.4. Using the summation method, we study the multi-year average precipitationand annual average precipitation of four kinds of precipitation products and verify theaccuracy. On the time scale, the measured data of four kinds of down-scalingprecipitation products and stations have a more consistent trend on the monthly timescale, the APHRODITE dataset maintains a high correlation and low relative bias. Onthe seasonal time scale, overall TRMM3B43 data performed better over the period2000-2015. All three data were overestimated as a whole and the winter performancewas poor. APHRODITE performed best on a year-scale, and TRMM3B43 andERA-Interim performed better between 2000 and 2015. On the spatial scale, therelative deviation between the measured data of 32 meteorological stations and thecorresponding grid dataset is analyzed. It is found that the four kinds of data aregenerally overestimated for the measured precipitation, the Turpan, Dabancheng andQiaojing stations with less precipitation have obvious overestimation, while there isunderestimation of precipitation in Zhaosu, Bayinbuluke and Tuergat stations, fromMay to September, the relative deviation between the grid data and the measured datais small.5. Based on the validated and downscaling precipitation data, the spatial andtemporal distribution of precipitation in Tianshan Mountains was analyzed. It is foundthat on the time scale, precipitation in the three major parts of the Tianshan Mountainsin China is mainly concentrated in April-September, the monthly maximumprecipitation of APHRODITE, TRMM3B43, ERA-Interim and GLDAS was between15.7 and 62 mm (2000-2007 average), 17.8 and 56.8 mm, 3.3 and 50.7 mm and 17.9 and 51.3 mm, respectively. However, the monthly maximum precipitation in differentsub-regions also has some differences. The monthly precipitation in the Ili Valley isabove 50 mm, the minimum monthly precipitation in the East Tianshan is below 20mm, and the northern Tianshan is higher than that of the southern Tianshan. In theseasons, all sub-districts are dominated by summer precipitation and winterprecipitation is the least. The APHRODITE data, which is closest to the observedprecipitation, shows that the average precipitation in China's Tianshan Mountains was450.5 mm from 2000 to 2007, and the TRMM3B43 data show that the annualprecipitation in China's Tianshan Mountains from 2000 to 2015 was 551.6 mm.6. On the spatial scale, firstly, in the horizontal direction, the precipitation in themountainous areas of China's Tianshan Mountains tended to decrease gradually fromwest to east due to the impact of water vapor sources. The precipitation in themountainous areas was significantly more than that in the plain areas. Theprecipitation in the northern Tianshan Mountains was more than that in the southernTianshan Mountains East Tianshan. In the vertical direction, the maximumprecipitation in China's Tianshan Mountains varies with different months. InJanuary-March and December, the Tianshan Mountains show an increasing trend withthe elevation increasing, between April and December, there is a maximumprecipitation zone between 2500-3000m above sea level, when the altitude is less than2500m, the monthly average precipitation increases with elevation, when theelevation is more than 3000m, the precipitation tends to decrease slowly withelevation increasing. The maximum precipitation of annual mean precipitation in theTianshan Mountains in China varies with different regions, with the eastern Tianshan、the northern Tianshan and the southern Tianshan appearing between 2500-3000m andthe Ili Valley appearing between 2000-2500m.
Subject Area自然地理学
Language中文
Document Type学位论文
Identifierhttp://ir.xjlas.org/handle/365004/14956
Collection研究系统_荒漠环境研究室
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
卞薇. 天山山区多源降水数据评估及时空格局研究[D]. 新疆乌鲁木齐. 中国科学院大学,2018.
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
[卞薇]'s Articles
Baidu academic
Similar articles in Baidu academic
[卞薇]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[卞薇]'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.