Study On Regional Water Deficit Monitoring Method Using Remote Sensing Based On Vegetation Evapotranspiration

Heavy deficit of water is a key problem in semi-arid area of the northwest China, Inner Mongolia. This article mainly studies different degraded degrees grasslands. Aiming at the topic of regional water deficit, this paper has developed methods in MODIS data preprocessing, remote sensing quantitative inversion of land surface parameters, and improvement of regional water deficit monitoring using remote sensing. Main subjects are as follows.According to surface energy balance theory, the semi-arid regional dual-layer evapotranspiration model (SRDEM) is established , using remote sensing monitoring of evapotranspiration at sub-pixel level, combined with the field measurement and meteorological data. Taking the relation between leaf area index and vegetation geometrical roughness into account, a leaf area index is introduced to compute the vegetation geometrical roughness so that the algorithm computing the sensible heat flux can be improved. The surface water deficit index (SWDI) is extracted using actual and potential evapotranspiration to study the regional water deficit condition. The regional soil water content is then assessed.Surface water deficit index with clear definition and specific meanings has higher quantitative precision. But it refers to more parameters and some of them still rely on the meteorological observations of the surface weather station, which leads to limitation to real-time monitoring. Therefore, considering vegetation reflecting, emissive spectrum and the genetic relation of vegetation water, this research has introduced a vegetation water synthetical index (VTWSI) using parameters extracted from remote sensing data, vegetation index, normalized difference water index and vegetation component temperature, which can be used to inverse vegetation water content.The surface water deficit index is extracted based on the dual-layer evapotranspiration model, including soil water evaporation and vegetation transpiration. As vegetation water ismainly provided by soil water in semi-arid area, there is a time lag between the vegetation water provision condition and the vegetation water content. More attention has been paid to the vegetation water content in monitoring regional water deficit and ravages of drought, but it is not enough to only concern vegetation water content. Thus, this article establishes the remote sensing quantitative formula of regional water deficit using genetic plan based on surface water deficit index, surface soil water content and vegetation water content. The result shows that the quantitative formula is ideal in comparison between modeled and true values and through a sensitivity analysis. For vegetation area, the quantitative formula considers not only the soil water but also the vegetation water. It can be used to study surface water deficit, and to improve the method of regional water deficit monitoring using remote sensing.Parameters, such as land surface temperature, vegetation/soil component temperatures, albedo, leaf area index and vegetation coverage etc., are inversed quantitatively using MODIS data as the input of model. To meet the need of dual-layer evapotranspiration model, this study adopts genetic algorithm to inverse component temperatures using two infrared bands of MODIS data, which can provide more accurate parameters for land surface energy balance and evapotranspiration study.Cloud is a large obstacle to processing and analysis of remote sensing image, because it is often cloudy in this area. It is necessary to detect cloud for improving the parameters inversion precision using remote sensing data. Among numbers of cloud detection methods, threshold method is often used to extract cloud information. Considering some subjective factors of threshold methods, this study discusses the cloud detection index and an automatic cloud detection algorithm based on the texture and neural network. The detection result is more accurate and effective.