Soil Moisture Monitoring Model And Method Research Based On RS/GIS

Being an agricultural country, monitoring of soil moisture is very important to flood-control in China. Remote Sensing (RS) is a good technological method for information acquisition of soil surface because of its convenience and speediness. Thus, Remote Sensing introduces important advantages in large-scale monitoring of soil moisture, sometimes inalienable in China. A lot of methods and algorithms have been developed for soil moisture monitoring. In this paper, we have tackled the soil moisture monitoring problem by introducing Remote Sensing strategies and offered a spectrum decomposed model for use, namely, Spectrum Decomposed Model Based Soil Moisture Computation (SDMBSMC). The principium of soil spectrum response was analyzed and the reflected spectral information was decomposed into three parts: dry soil reflectance, water reflectance and mirror reflectance of water membrane. A novel method was brought forward to estimate the soil water contents. The workflow of the method includes: (i) dry soil spectral information collection from existing spectrum database ;( ii) reflectivity extraction from remote sensing images ;( iii) soil moisture computation based on spectrum decomposed model.A case study on Changpin District, Beijing was provided. The implementation steps of soil moisture monitoring based on spectrum decomposed model can be described as following:(1) acquisitions of data, such as climatic data, soil type, spectrum characteristics of soil and water objects, multi-phrases remote sensing images, and statistical charts and data from the local government;(2) Data pre-processing and storage of pre-processed data in GIS database;(3) Processing of original remote sensing images, such as emendation, furbishment, cut, and amplifying. This work is the base for quantitative analysis of soil moisture;(4) Remote Sensing images analysis and interpretation;(5) Partition of soil usage type and vegetation overlay;We use ArcGIS software as a problem solving environment. In ArcGIS software, implementation steps is as following:(1) DN matrix transformation to a luminance matrix based on various sensor response functions for every wave band respectively;(2) Remote Sensing image layer partition based on soil type;(3) Soil moisture computation using SDMBSMC;(4) Storage of computation results into database;The computation results of this case study show that :(1) The approximate precision of soil moisture monitoring on infrequent vegetation land is 89.78% and best;(2) The approximate precision of soil moisture monitoring on a mountainous area is 83.19% and relatively lower due to the impact of the surrounding terrain and vegetation coverage;(3) SDMBSMC is not applicable to water objects and urban areas (areas containing high density architectures) because of the various spectrum characteristics of them;There are many factors that affect the precision of soil moisture monitoring, such as spectral and spatial resolutions of remote sensing images, resolutions of spectral characteristic data, vegetation coverage, the surrounding terrain, land usage types ,etc. Consequently, SDMBSMC has some constraints due to these factors when computing the soil moisture in an area.At last, a detailed conclusion on the whole work and the research results on this paper are summarized, and the improved aspects and the proposals to the future are discussed.