Experimental Analysis On Retrieval Of Shallow Groundwater Depth In Oasis-desert Ecotone Using Remote Sensing

Water is the source of life, and it’s a precious natural resource for human survival and development. With continuous consumption of water resources, water resources has become an important scientific issues of global sustainable development. Groundwater is defined as the vadose zone located below the ground water in the gap includes not only water between rock cracks and pores, further comprising a cave in the presence of moisture, changes in the environment has a sensitive reaction is not only an important energy environment parameters, but also an important factor in the protection of ecological balance development. It refers to the shallow groundwater with surface water, atmospheric precipitation, which are buried directly interact relatively shallow diving or weakly confined water, buried within 20 m of depth. For arid and semiarid regions, due to strong evaporation, scarce rainfall and uneven distribution, very fragile ecological environment, the oasis is its essence, ecological regulation in arid area it is particularly important. In arid area, many landlocked River has no surface water supplies, thereby forming a meteoric water- plant water- soil water-groundwater hydrology unique ecosystem, where groundwater has become arid area oasis maintain natural vegetation growth remaining water. In the more intense evaporation time, if the water table is high, the evaporation will produce invalid dissipation, resulting in salinization, leading to salt stress phenomenon vegetation; if the water table is low, will produce vegetation water stress, leading to degradation or death of vegetation, resulting in surface desertification.The traditional method of monitoring the groundwater level have taken a longer time and more effort. There are more difficults to achieve a large area dynamic monitoring.Remote sensing technology has a wide range of higher timeliness monitoring features, making it become a new method which can be timely and effective large area groundwater level monitoring.The Macro, dynamic, fast and comprehensive monitoring features of remote sensing technology for fast and efficient large-area groundwater level monitoring provides a new means of detection. This paper attempts by inverting a plurality of surface parameters, combined with the measured groundwater data to explore the method of inverting groundwater depth, and through comparing these methods to find the most suitable and universal algorithm of groundwater depth inversion.The study based on measured data include the oasis soil relative water content(SRWC) and groundwater depth data, as well as multi-spectral and hyperspectral remote sensing data, combined with vegetation and soil conditions, inversed soil moisture with hyperspectral vegetation index and TS-VI feature space( TSNDVI and TS-MSAVI feature space).Based on that, fitted the groundwater depth distribution, found that these two methods to extract groundwater depth is feasible.The main conclusions of this study are as follows:(1) In the first instance, we introduced the use method of HOBO U20 level indicator, and then export the correlation analysis between groundwater depth data and soil SRWC at different soli depths, the results showed that, SRWC at 0-10 cm and 40-60 cm layers with groundwater depth data have more correlation than other soil, which also provide support to through make use of remote sensing technology to estimate SRWC indirectly accept groundwater depth data.(2) Taken pretreatment to Hyperspectral imaging, include atmospheric correction and cut and so on. Be based on 26 spectral vegetation index which related with SRWC, merged the measured field data with hyperspectral imaging HSI hyperspectral data, integrated gray correlation analysis and multiple regression analysis, and established hyperspectral vegetation index SRWC inversion model based on two types of data, the results show that the two models of 0-10 cm layer SRWC inversion effects are better than other soil. To further improve the accuracy of the model, the image hyperspectral vegetation index SRWC inversion model be corrected, results show that the model that be corrected, the coefficient of determination R2 increased from 0.589 to 0.711, and adopted the 0.01 level of significance inspection, RMSE is 0.0014, have better stability. Thus, using this method inverse SRWC is effective and feasible. At the same time, as a basis for groundwater depth inversion, results show that the method is feasible and precision can meet research needs.(3) Taken pretreatment and cut with Landsat 8 image, obtaining MSAVI and NDVI data by inversion, and used split-window algorithm inversed TS data,then TS-NDVI and TS-MSAVI feature space were constructed, and then calculated TVDINDVI and TVDIMSAVI, analyzed their correlation with SRWC, results showed that both with the 40-60 cm layer SRWC have most correlation By TVDINDVI and TVDIMSAVI fitted with dry side, wet side of the equation, inversion 40-60 cm layer SRWC, and accuracy test. On this basis, the study area groundwater depth inversion results show that the method is feasible, and the precision can meet the need for research.(4) By contrasting and analyzing the retrieved precision of these two methods, we figured out that the retroviral of groundwater depth using TVDIMSAVI had a higher precision.