Simulation Of Spatial And Temporal Distribution Of Soil Temperature And Moisture Based On Deep Learning

Surface Water and Heat plays an important influence on the energy exchange between the surface and the atmosphere as well as the moisture cycle.In this thesis,we selected Daqingshan Nature Reserve in Inner Mongolia as the study area,used remote sensing and other data for the 2019 to 2021 growing season,constructed spatial and temporal distribution models of surface temperature and soil moisture based on deep learning algorithms,and based on the constructed model to analyze the complex relationship between soil water and heat conditions and various environmental factors under a single variable and different scenarios.The analysis of the results shows that:(1)The average temperature in the study area over the past 50 years showed an increasing trend with a climatic tendency rate of 0.046℃/a;the total annual precipitation showed an increasing trend with an increasing rate of 0.343 mm/a;the monthly,seasonal and annual scale SPEI all showed a decreasing trend with a tendency rate of change of-0.002/10 a,-0.006/10 a and-0.013/10 a,respectively.(2)NDVI,elevation,mean temperature,slope and surface albedo contribute more to LST(0.486,0.299,0.091,0.089,0.036);the constructed LST model has higher validation accuracy(MAE is 0.0869 and MSE is 0.0580).The LST increases with the increase of mean air temperature and surface albedo(0.106℃/℃ and 0.056℃).The LST decreases with the increase of NDVI,slope(0.477℃,0.976℃/°,and 0.890℃/10m).(3)LST,elevation,NDVI,slope direction and slope contribute more to TVDI(0.360,0.352,0.223,0.033 and 0.032);the constructed TVDI model has higher validation accuracy(MAE is 0.0951 and MSE is 0.0158).TVDI increased with the increase of LST,slope and elevation(0.037/ ℃,0.037/° and 0.030/10m),and TVDI decreased with the increase of NDVI(0.017).TVDI varied less with slope direction The TVDI values were slightly larger on the north,northeast,east and southeast slopes than on the other slopes.This method helps to understand the spatial distribution relationship between environmental factors and soil hydrothermal in mountainous areas,and can provide a new method to study the influence of climate change on hydrothermal processes and vegetation distribution in mountainous areas.