Soil Moisture Inversion In Desert Mining Area Based On SAR Data

Soil moisture is one of the fundamental parameters in studying land water resource cycling and plays an important role in research areas such as climate change,ecological degradation,and land restoration.In the unique ecosystem of desert mining areas,it is difficult to obtain large-scale soil moisture data through direct measurement,and commonly used remote sensing methods for monitoring soil moisture are also limited by factors such as resolution and measured data.Therefore,there is an urgent need for an accurate and reliable remote sensing monitoring method that does not require measured data.This study aims to address the problem of difficult soil moisture monitoring on the surface of desert mining areas without measured data.Taking the desert mining area in the eastern Junggar Basin of Xinjiang as the research area,three soil moisture inversion methods based on SAR data were used to achieve pixel-scale soil moisture monitoring under the condition of no measured data,and the inversion accuracy was compared and analyzed with SMAP products.At the same time,the best method was selected to achieve multi-temporal soil moisture monitoring in the study area,analyze the spatiotemporal variation characteristics of soil moisture in the desert mining area,and explore the impact of coal combustion on the spatiotemporal distribution of soil moisture.In order to improve research efficiency and meet practical production applications,a soil moisture inversion software was developed in this study,which integrates multiple commonly used soil moisture inversion models to meet various polarization data application scenarios.The main research content and conclusions of this article are as follows:(1)This passage describes three methods used to invert soil moisture in the study area,including the empirical model method,transformation detection method,and reverse lookup table method.For the missing surface roughness data required by the reverse lookup table method,the study used the relative roughness during the winter freeze-thaw period in the study area instead of actual measurements.By comparing with optical images,the study found that the spatial distribution of relative roughness corresponds well to the actual distribution of desert Gobi plains and mountains,and can reflect the changes in topography caused by mining and coal combustion in the mining area,making it suitable for the reverse lookup table method to invert soil moisture.By comparing the results of the three soil moisture inversion methods,the study found that the reverse lookup table method based on relative roughness is more suitable for soil moisture inversion in the study area when there is no actual measurement data.The study results show that the multiple inversion results based on the reverse lookup table method using relative roughness are within a low deviation range when compared with the SMAP soil moisture product.In the comparative data,the proportion of points with an absolute error of less than 2% in the Gobi plain area inversion result reaches 90%,and the inversion deviation in other topographies is also small,which can be used for large-scale and temporal monitoring of soil moisture in the study area.(2)In this study,a reverse lookup table was used to invert the multi-temporal soil moisture in the study area,and the inversion results were analyzed.The results showed that the average soil moisture in the working area and the waste rock pile area in the desert mining area was higher than that of the surrounding land surface.This was due to the periodic watering measures taken during excavation work to prevent fires and dust.In addition,for the coal mine landing with severe salinization,the soil moisture remained below 1% for a long time and hardly changed with rainfall,indicating that the land in this area is difficult to naturally restore.For the coal mine with coal fire burning,the spatiotemporal variation of soil moisture of different land cover types in the mining area is different from that in other mining areas due to the combined effects of coal fire and accumulated water.(3)To further investigate the impact of coal fire burning on surrounding soil moisture in the desert mining area,this study selected Jungebi No.1 Mine as the research area and proposed a spatiotemporal abnormal area identification algorithm for soil moisture in the coal mine area based on the spatial distribution results of soil moisture in the 30 th period.By analyzing the abnormal range of the coal mine area,it was found that surface moisture was affected by the combined effects of coal fire and accumulated water.Specifically,this study found that in the water-logged areas without coal fire,there was no obvious abnormal change in soil moisture in the surrounding area.However,in the areas with both coal fire burning and accumulated water coverage,the soil moisture in the surrounding area showed abnormal changes,and the abnormal range gradually decreased as the accumulated water decreased.In addition,the burning areas without accumulated water,and the surface collapse cracks caused by hightemperature burning can also cause abnormal soil moisture on the surface.These findings indicate that coal fire burning has a significant impact on the soil moisture in the surrounding area of the mining area,and the combined effects of coal fire and accumulated water on soil moisture are complex.(4)In the multi-temporal monitoring of soil moisture in desert mining areas,it is necessary to perform vegetation correction,relative roughness inversion,and soil moisture inversion on multi-temporal SAR data.In order to improve research efficiency and meet practical production applications,this article uses Python as the development tool to design and develop a soil moisture inversion software.The software integrates a variety of commonly used soil moisture inversion models,including the water-cloud model,Oh model,AIEM model,etc.,and optimizes algorithms with high computational complexity,using multi-core computing technology to speed up calculations in key steps,thus greatly reducing the inversion calculation time.At the same time,this article has expanded the functions of each functional module that contains inversion models to meet application requirements under different input data.