IMHE OpenIR  > 数字山地与遥感应用中心
山地典型LAI遥感产品质量验证与不确定性分析
Alternative TitleIntercomparison and validation of leaf area index products and uncertainty analysis in typical mountainous areas
Language中文
杨勇帅
Thesis Advisor李爱农
2016
Degree Grantor中国科学院大学
Place of Conferral北京
Degree Name硕士
Degree Discipline地图学与地理信息系统
Keyword叶面积指数产品 交叉验证 山地 直接验证 时空特征 不确定性
Other Abstract

植被是陆地生态系统中最重要的组成部分,是几乎所有生物生存所需物质和能量的来源。同时,植被对全球变化的反映至关重要,其对气候变化的调节与反馈作用是人类调节气候、减缓大气CO2浓度增加的主要手段,生态系统相关研究对于人类的生存和发展具有重要意义。当前生态系统研究模型如陆面过程模型、生态系统生产力模型以及作物生长模型等都需要关键陆表参数。叶面积指数(Leaf Area Index, LAI)作为表征地表植被生物物理变化和冠层结构特征的重要参数,与植被的蒸腾作用、光合作用以及地表净初级生产力等过程有着非常密切的联系。多种生态系统模型亟需LAI作为其不可或缺的输入数据。卫星遥感技术可以实现频繁而持续的对地观测,基于卫星遥感数据反演成为了获取大区域LAI的有效方法。当前存在多个基于遥感数据反演的大区域、长时序LAI遥感产品。由于遥感数据获取过程和产品生产环节中诸多因素的影响,LAI产品的误差是不可避免的,对其进行质量验证尤为重要。LAI产品的质量验证工作是其广泛应用的前提,验证结果可以为产品的算法改进提供反馈信息,也可以为产品的应用提供理论依据。目前针对LAI产品的质量验证问题已经开展较多的研究工作,但在山地开展的验证工作尤为少见。然而,山地是陆地生态系统重要而复杂的组成部分,是生物多样性最为丰富的陆地单元。山地生态系统为人类提供了不可替代的淡水资源、矿产资源和生态系统服务功能,山地也成为了当前国际研究的热点区域。我国是多山国家,在山地开展LAI产品的质量验证工作对于我国相关生态系统研究至关重要。针对LAI产品在山地的质量验证问题,本文在我国西南地区选取典型研究区,以2001~2013年GEOV1 LAI、GLASS LAI和MODIS LAI产品为研究对象,考虑山地地形特征,选择交叉验证和直接验证两种方法,对LAI产品进行质量验证,并对比LAI产品不确定性的时空变化规律。论文主要包括以下几个研究内容:(1)基于交叉验证方法的LAI产品质量验证,对三种LAI产品进行预处理,使其具有一致的数据投影方式、时间和空间分辨率;以产品缺失值比例描述LAI产品的时空完整性特征,并结合地形因子,分析产品缺失值比例在复杂地形条件下的空间分布规律;对比LAI产品对山地地表植被空间分布特征和时间变化规律的表征能力,验证LAI产品相对准确性;(2)基于直接验证方法的LAI产品质量验证,基于现有LAI反演模型,采用先反演后平均的方法,获取低空间分辨率LAI数据,通过像元LAI的对比完成LAI产品质量验证,定量评估产品精度;(3)以LAI产品提供的不确定性数据集为研究对象,对比LAI产品不确定性的空间分布及其在异质性地表的变化规律,并分析其时间变化规律。本文得到以下主要结论:(1)山地复杂的地形特征容易导致水汽和云的形成,影响山地区域遥感数据质量,造成LAI产品时空完整性变差。研究区内时空完整性从优到劣的顺序为:GLASS LAI产品>MODIS LAI产品>GEOV1 LAI产品;(2)在空间维,GEOV1 LAI和GLASS LAI产品一致性较好,较好的表征了植被的空间分布特征,MODIS LAI产品表征植被空间分布能力较差;由于山地植被结构复杂,地表异质性较高,LAI产品难以准确表征山地植被垂直带谱;在时间维,三种产品难以准确反映冬季常绿针叶林LAI,且难以准确表征农田作物的物候信息;三种LAI产品能较好的表征植被的年内物候变化,GEOV1 LAI产品表征了植被在年际间对气候变化的响应,但GLASS LAI和MODIS LAI产品未能表达该特征;(3)在那曲样区,GEOV1 LAI、GLASS LAI和MODIS LAI产品对草地区域LAI表现出了高估现象,三者RE分别为163.6%、156.5%和133.3%;GLASS LAI和MODIS LAI产品在林芝样区和普定样区植被LAI表现出了低估的现象,在林芝样区,GLASS LAI产品和MODIS LAI产品主算法、备用算法反演结果的RE分别为33.8%、81.1%和90.4%,MODIS LAI产品备用算法反演结果低估水平更严重;在普定样区,GLASS LAI和MODIS LAI产品主算法反演结果RE分别为43.7%和73.2%;(4)LAI产品不确定性的空间分布和时间变化规律和LAI有较好的一致性,在林地区域的不确定性大于草地区域,在时间序列上则表现出夏季大、冬季小的趋势。地表异质性导致草地混合像元LAI的不确定性增大。本文的创新工作在于以典型山地为研究区,从LAI产品对山地植被空间分布特征和时间变化规律的表征能力角度,评估产品的相对准确性,进行LAI产品的质量验证。拟解决的关键问题为评估当前LAI产品对山地植被的时空表征能力,该工作有助于发现LAI产品在山地的问题,并为LAI产品在山地的应用提供理论参考。

;

As the pricipal component of terrestrial ecosystem, vegetation plays an important role in sustaining global and energy cycle, adjusting carbon balance and alleviating the rise of atmosphere CO2 aconcentration and global climate change. In order to advance global and regional land surface models at different scales and improve their predictive capabilities and other applicability to other uses, various space agencies have produced a series of high-level land surface biogeophysical products from different satellite data. Leaf Area Index(LAI), as a key biophysical variable for describing the canopy architecture, determines the transpiration, the interception and absorption rates of solar radiation by vegetation. LAI is an important variable in numerous land surface models. Remote sensing observations acquired with moderate resolution optical sensors allow monitoring the seasonal and interannual variability of LAI fields over regional to global domains. Currently, several global scale and long term LAI data sets has been productd. Intercomparison and validation of LAI products is critical for their proper use in land surface models. However, existing studies were limited over flat terrain, the performance of the LAI data sets in mountainous area is still unknown. Mountain areas, account for approximately 70% of China’s land area, plays an important role in terrestrial ecosystem. Incomparison and validation of LAI products over mountainous area is very significant for carbon cycleof terrestrial ecosystem, global changes and other related studies.The GEOV1, GLASS and MODIS LAI products were indirectly and directly validated over mountainous area, southwestern China. The content of this paper mainly contains: (1) The consistency of the LAI products is evaluated by examining the frequency of data gaps considering the rugged surface. Then the magnitudes of LAI values are compared and the capability to capture phonological characteristics is evaluated. The LAI products are proprecessed to make sure that the LAI product data sets have the same projection and temporal and spatial resolution. (2) The accuracy of the LAI products is evaluated against the fine resolution LAI data. LAI estimation from fine resolution remote sensing data is based on empirical method, which considering the surface reflectance and the characteristics of rugged terrain. The fine resolution LAI values are aggregated to 1-km resolution using averaging method. The RMSE and RE of LAI products and reference LAI value are calculated to identify the accuracy of LAI products. (3) The uncertainties associated with LAI products are intercompared and characterized so that the global LAI products can be better understand. The spatial distribution and time variation of the uncertainties was analyzed over the mountainous area.The main conclusion of this research are summarized as: (1) The rugged surface of mountainous area easily lead to the formation of water vapor and clouds. As a consequence of this circumstances, remote sensing data over mountainous area are easily affected by clouds, resulting the retrieval of LAI from remote sensing data. The best continuity is achieved by GLASS in our study areas and the worst is from GEOV1. (2) GEOV1 and GLASS LAI product shows good spatial distribution for all biome types. The difference between the three products can be neglected over grassland but obvious over forests and cropland. Because of the heterogeneity of mountainous area, LAI products can hardly characterize the altitudinal distribution of biome types. GLASS LAI time series are smooth and continuous. GEOV1 LAI time series have lots of gaps and MODIS LAI time series are very shaky. All three LAI products display unreasonable seasonality for evergreen needleleaf forests and crops. (3) GEOV1, GLASS and MODIS LAI products overestimates the grassland LAI over Naqu area, the RE of three products is 163.6%, 156.5% and 133.3%, respectively. GLASS and MODIS LAI products underestimates evergreen needleleaf forests LAI over Linzhi area, the RE of GLASS, MODIS LAI retrievaled from main algorithm and MODIS LAI retrievaled from empirical algorithm is 33.8%, 81.1% and 90.4%, respectively. GLASS and MODIS LAI products underestimates cropland LAI over Puding area, the RE of GLASS and MODIS LAI retrievaled from main algorithm is 43.7% and 73.2%, respectively. (4) The spatial distribution and time variation of LAI uncertainties associated with LAI products have good consistency with LAI. Uncertainties over forests area are greater than over grassland area. Uncertainties show a seasonal evolution with higher values in summer and lower in winter. The heterogeneous of land surface lead to the mixed pixel of grassland, resulting greater uncertainties over grassland area.The innovation work of this study is that intercomparison and validation of LAI products in the perspective of evaluation the ability of LAI products to capture the spatial distribution and phenology of vegetation in typical mountainous areas. The key issue to be addressed is to evaluate the ability of LAI products to capture the spatial distribution and phenology of vegetation. This study may facilitate the improvement of the quality of LAI product over mountainous areas. 

Document Type学位论文
Identifierhttp://ir.imde.ac.cn/handle/131551/18888
Collection数字山地与遥感应用中心
Affiliation中国科学院成都山地灾害与环境研究所
Recommended Citation
GB/T 7714
杨勇帅. 山地典型LAI遥感产品质量验证与不确定性分析[D]. 北京. 中国科学院大学,2016.
Files in This Item:
File Name/Size DocType Version Access License
山地典型LAI遥感产品质量验证与不确定性(18558KB)学位论文 开放获取CC BY-NC-SAView Application Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[杨勇帅]'s Articles
Baidu academic
Similar articles in Baidu academic
[杨勇帅]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[杨勇帅]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: 山地典型LAI遥感产品质量验证与不确定性分析.pdf
Format: Adobe PDF
All comments (0)
No comment.
 

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