中文版 | English
Arid
土壤水分遥感辅助信息的适用性评价及其遥感应用
其他题名Applicability Research of Auxiliary Spatial Information in the Scaling of Soil Moisture and Its Remote Sensing Application
赵泽斌
出版年2017
学位类型硕士
导师晋锐
学位授予单位中国科学院大学
中文摘要土壤水分是陆表过程的核心变量之一，强烈影响着地表-植被-大气间的能量平衡和水分交换。当前基于星载被动微波传感器的土壤水分遥感产品空间分辨率普遍较粗（一般在25~40km之间），难以满足流域尺度水文气象、生态水文模拟以及水资源管理等研究和应用的需求，而土壤水分的降尺度方法是目前较为可行的解决方案之一；此外，在遥感真实性检验的过程中往往需要开展点尺度观测向遥感像元的尺度上推研究，即对地面观测做升尺度处理。不论在降尺度还是升尺度研究中，辅助信息均起到重要的桥梁作用。但目前对辅助信息的适用性缺乏比较研究。因此，本文通过SiB2（Simple Biosphere Model Ⅱ）模拟和真实遥感数据分别开展对各种辅助指标的比较研究，分析确定每种辅助指标的适用条件，为土壤水分的尺度转换提供基础性的参考信息，最后利用随机森林方法进行土壤水分制图。首先，从模型角度出发，利用2013年5月1日~9月30日黑河中游人工绿洲试验区大满超级站的气象数据驱动SiB2模型，分别模拟了土壤水分、土壤表层温度、植被冠层温度以及地表蒸散发、土壤蒸发等变量；并利用Penman-Mon -teith公式计算了地表潜在蒸散发。利用SiB2模拟结果与Penman-Monteith公式计算结果估算获得六种常用的土壤水分尺度转换的辅助指标：表观热惯量（ATI，Apparent Thermal Inertia）、土壤蒸发（E，Evaporation）、土壤蒸发/实际蒸散发（E/ETa，Evaporation/Actual Evapotranspiration）、土壤蒸发/潜在蒸散发（E/ETp，Evaporation/Potential Evapotranspiration）、蒸发比（EF，Evaporative Fraction）、实际蒸发比（AEF，Actual Evaporation Fraction）。分别在整个植被生长季（5月1日~9月30日）、植被封垄前（5月1日~7月10日）、植被封垄后（7月11日~9月30日）三个阶段分析了以上辅助指标与各层土壤水分之间的关系，结果表明：（1）在整个植被生长季，6种辅助指标与土壤水分之间都具有较好的相关性（0.61E/ETa>EF>E>AEF>ATI；10cm：AEF> E/ETp>EF>E/ETa>E>ATI；80cm：EF>AEF> E/ETp>E/ETa>E>ATI。（2）植被封垄前（株高<1.6m，植被覆盖度<0.8）：2cm：E>E/ETa>E/ETp>EF>AEF> ATI(Ts)>ATI(Tc)；10cm：E>EF>E/ETp>AEF>E/ETa> ATI(Ts)>ATI(Tc)；（3）植被封垄后（株高>1.6m，植被覆盖度>0.8）：2cm：ATI(Ts)>E/ETa>E/ETp>AEF>EF> ATI(Tc)>E；10cm：ATI(Ts)>E/ETa>E/ETp>EF>AEF>ATI(Tc)>E。生长前期，E对土壤水分最为敏感，ATI最不敏感，生长后期则正好相反。综合考虑以上指标与土壤水分的相关性和这些指标的遥感可获取性，对于浅层土壤水分，推荐优先考虑指标EF，对于根区土壤水分，推荐优先考虑指标AEF。此外，本文利用组分温度分解后的土壤表面温度计算的ATI(Ts)与土壤水分在整个植被生长季均具有较好的相关性，相对于之前学者利用地表温度计算的ATI(Tc)仅适应于裸土或低矮植被覆盖区的结论有了进一步改进。在以上辅助指标适用性评价的研究基础上，考虑到前人在应用研究中用到遥感辅助指标：地表温度（LST， Land Surface Temperature）、归一化植被指数（NDVI，Normal Differential Vegetation Index）、温度植被干旱指数（TVDI，Temperature Vegetation Dryness Index）等，因此本文考虑到ATI与土壤表层温度关系密切，EF与地表温度和植被覆盖度等信息有关、AEF与蒸散关系密切，所以选取了与土壤表层温度、植被覆盖度和蒸散有关的指标：LST、NDVI、TVDI、条件植被温度指数（VTCI，Vegetation Temperature Condition Index）以及土壤蒸发效率（SEE，Soil Evaporation Efficiency），并使用ASTER（Advanced Spaceborne Thermal Emission and Reflection Radiometer）遥感影像计算获取以上指标。其中，LST直接获取于ASTER 地表温度产品；TVDI与VTCI计算之前需要先构建NDVI-LST特征空间，在特征空间所确定的干湿边的基础上进行估算；SEE是基于土壤表层温度计算的，因此计算之前需要对LST进行组分温度分解。然后分析了各个遥感辅助指标与土壤水分观测之间的相关性，又分别比较了各遥感辅助指标与PLMR土壤水分产品（700m分辨率）在空间分布模式上的一致性。结果表明：（1）五种指标都与10cm土壤水分的关系较好；定量来看，LST、NDVI、TVDI、VTCI以及SEE与土壤水分的相关性是依次增大的（0.50
英文摘要Soil moisture is one of the core variables of land surface process, and it has a strong influence on the land surface-vegetation-atmosphere’s energy balance and water exchange. However, the spatial resolution of soil moisture products from the passive microwave radiometers are generally coarser (25~40km), which is an obstacle to satisfy the research and application requirements of the hydro-meteorological, eco-hydrological simulation and water resource management at watershed scales. The downscaling of soil moisture is a feasible method now to solve it. In addition, there is a mismatch in the spatial scales between the in-situ observation and the remote sensing retrieve pixel during the validation process of remote sensing products. The in-situ observation should be translated to ground truth at pixel scale through reasonable upscaling method firstly. The auxiliary information is necessary for both downscaling and upscaling of soil moisture, especially in the heterogeneous surface. Unfortunately, the applicability research of auxiliary information is very scarce. This paper aims at analyzing the applicable conditions of each auxiliary indices through SiB2 (Simple Biosphere Model -Ⅱ) simulation and remote sensing images, which can acts as a foundation for the soil moisture scaling. Finally, the random forests method was employed to map soil moisture.From the point of view of model simulation, the SiB2 model was forced by the meteorological observation of Daman superstation from May 1, 2013 to September 30, 2013, which located in the middle reach of Heihe artificial oasis. Then, the soil moisture, soil surface/canopy temperature, evaporation and evapotranspiration are simulated by SiB2, and the potential evapotranspiration is calculated from Penman-Monteith formula. Furthermore, six auxiliary indices are estimated, including ATI (Apparent Thermal Inertia), E (Evaporation), E/ETa (Evaporation/Actual Evapo -transpiration), E/ETp (Evaporation/Potential Evapotranspiration), EF (Evaporative Fraction) and AEF (Actual Evaporative Fraction). Eventually, the correlation between these indices and soil moisture are analyzed during the entire vegetation growing season (from May 1 to September 30), growing season before sealing ridge (from May 1 to July 10) and after sealing ridge (from July 11 to September 30). The results manifest that: (1) During the entire vegetation growing season, all six indices have good correlations with soil moisture (0.61E/ETa>EF>E>AEF>ATI; 10cm: AEF>E/ETp>EF>E/ETa>E>ATI; 80cm: EF>AEF>E/ETp>E/ETa>E>ATI. (2) growing season before sealing ridge: 2cm: E>E/ETa>E/ETp>EF>AEF>ATI(Ts)>ATI(Tc); 10cm: E>EF>E/ETp>AEF>E/ETa>ATI(Ts)>ATI(Tc); (3) growing season after sealing ridge: 2cm: ATI(Ts)>E/ETa>E/ETp>AEF>EF>ATI(Tc)>E; 10cm: ATI(Ts)>E/ETa>E/ETp>EF>AEF>ATI(Tc)>E. E is sensitive to soil moisture all the time, while ATI is insensitive at the growing season before sealing ridge and the situation is reverse after sealing ridge. Considering both the correlation analysis and computability from remote sensing, EF is recommend as auxiliary indices when scaling the surface soil moisture, while AEF is preferential when scaling the soil moisture in root zone. In addition, the ATI calculated based on the soil surface temperature, has a higher relationship with soil moisture during the vegetation growing season, which change the previous knowledge that ATI based on land surface temperature can only be applied in bare soil and low vegetation cover surface. Based on the above applicability assessment and computability by remote sensing, we have selected five auxiliary indices to compare their correlations with soil moisture, including LST (Land Surface Temperature), NDVI (Normal Differential Vegetation Index), TVDI (Temperature Vegetation Dryness Index), VTCI (Vegetation Temperature Condition Index) and SEE (Soil Evaporation Efficiency). These indices are related with land surface temperature, vegetation coverage and evapotranspiration, and can be calculated by ASTER (Advanced Space-Borne Thermal Emission and Reflection Radiometer) images. LST is obtained from ASTER LST products directly. The calculations of TVDI and VTCI are based on the building of a NDVI-LST feature space. SEE is derived from soil surface temperature, which can be provided through component temperature inversion of LST. Then, the relationships between these remote sensing auxiliary indices and soil moisture are compared and analyzed. The consistency of spatial patterns of these indices also are evaluated by using PLMR- retrieved soil moisture products with 700m resolution. The results illustrate that: (1) Five indices have good corrections (0.50
中文关键词土壤水分 ; SiB2模型 ; 遥感 ; 辅助指标 ; 尺度转换 ; 随机森林 ; 制图
英文关键词soil moisture SiB2 remote sensing auxiliary indices scale conversion random forests mapping
语种中文
国家中国
来源学科分类地图学与地理信息系统
来源机构中国科学院西北生态环境资源研究院
资源类型学位论文
条目标识符http://119.78.100.177/qdio/handle/2XILL650/287972
推荐引用方式
GB/T 7714		 赵泽斌. 土壤水分遥感辅助信息的适用性评价及其遥感应用[D]. 中国科学院大学,2017.
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
导出为Endnote文件
谷歌学术
谷歌学术中相似的文章
[赵泽斌]的文章
百度学术
百度学术中相似的文章
[赵泽斌]的文章
必应学术
必应学术中相似的文章
[赵泽斌]的文章
相关权益政策
暂无数据
收藏/分享

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。