Spatial Downscaling Method Of Soil Moisture Retrieval Based On Multi-source Remote Sensing Data And Its Application

Soil moisture is an important component of both the water cycle and the surface energy cycle;it is also an important indicator for reflecting land degradation and characterizing surface drought information.Soil moisture is related to a number of factors,including vegetation growth,crop growth and food production,as well as important parameters in the fields of hydrology,climate research,agriculture,and ecology.With the development of remote sensing technology,a series of passive microwave soil moisture products have emerged.These passive microwave remote sensing-based soil moisture products provide large areas of soil moisture distributions,which constitute important basic data for global or regional research.However,the spatial resolution of most soil moisture products is 25?40 km,making it difficult for soil moisture data to satisfy the needs of applications such as hydrological modeling.Therefore,passive microwave products must be downscaled to obtain high-resolution soil moisture data.In addition,values of the standardized precipitation evapotranspiration index(SPEI)calculated using station-based meteorological data collected from 1961 to 2013in the middle and lower reaches of the Yangtze River Basin(MLRYRB)are used to monitor droughts.In addition,the SPEI is determined for different time scales(1,3,6,and 12 months)to characterize dry or wet conditions in this study area.Moreover,remote sensing methods can cover large areas,and multispectral and temporal observations are provided by satellite sensors.The temperature vegetation dryness index(TVDI)is selected to permit assessment of drought conditions.In addition,the correlation between the SPEI and TVDI values is calculated,which can provide a scientific basis for early warning and risk management of water resources and agricultural production.The main research work and conclusions of the thesis are as follows:(1)The apparent thermal inertia(ATI)method is simple and convenient,but it is only suitable for bare soil or sparsely vegetated regions,therefore,for high vegetation coverage areas using this method to estimate soil moisture is relatively poor accuracy.The vegetation supply water index(VSWI)method is suitable for areas with high vegetation coverage,when applied to low vegetation coverage areas,it will exaggerate the effects of vegetation.An applicable approach by integrating ATI and VSWI has been proposed to estimate the soil moisture,using NDVI as a threshold.The approach was validated with the field data.The distribution pattern of the soil moisture in the MLRYRB was further analyzed using land use/cover types,elevation and slope.The soil relative moisture distribution shows significant differentiation with the change of land use/cover types,elevation and slope.The results show that the compound ATI and VSWI models can effectively improve retrieval accuracy and remedy the shortage of one-sided method.(2)In the process of soil moisture downscaling,there are two key problems:one is the selection of the downscaling method;the other is the determination of the surface variables,namely the influencing factors of soil moisture.In the downscaling method,this study attempts to introduce machine learning and data mining algorithms into the model,and compares the advantages and disadvantages of random forest model and Cubist algorithm to determine the most suitable soil moisture downscaling method for the MLRYRB.The most commonly employed surface variables are the NDVI,DEM,and surface temperature.In contrast,there are few applications for variables such as albedo and ET;nevertheless,this study considers these factors to have a greater impact on soil moisture.Therefore,the final surface variables utilized in the downscaling process are the longitude,latitude,DEM,slope,aspect,NDVI,LST?D,LST?N,albedo,ET and LC.Based on the moderate resolution imaging spectroradiometer(MODIS)data,this study achieved the 16-day time-scale downscaling process.In the downscaling results of the cubist algorithm,due to the linear relationship among longitude,latitude and soil moisture,blocky features are too obvious.In other words,the relationship for each rule between the soil moisture data and the latitude and longitude displays abrupt changes,which is obviously contrary to the continuity of soil moisture in space.Therefore,in the cubist downscaling model,this study excludes two surface variables,namely,longitude and latitude.A comparison between the random forest model and the cubist algorithm reveals that the R~2 of the random forest-based downscaling method is higher than that of the cubist algorithm-based downscaling method by 0.0161;moreover,the RMSE is reduced by 0.0006,and the MAE is reduced by 0.0014.Furthermore,testing the accuracies of the two downscaling methods reveals that the random forest model is a more suitable method than the cubist algorithm for downscaling AMSR-E soil moisture data from 25 km to 1 km in the MLRYRB,and this finding provides a theoretical basis for obtaining high spatial resolution soil moisture data.(3)Droughts represent the most complex and damaging type of natural disaster,and they have taken place with increased frequency in China in recent years.In this study,the temporal and spatial variations in drought conditions over different time scales(1-,3-,6-,and 12-month)in the MLRYRB are quantitatively detected and analyzed using the SPEI;meteorological data collected at 64 meteorological stations from 1961 to 2013 are used.Moreover,the spatial distribution of drought conditions is further analyzed using the TVDI,and the correlation between the SPEI and TVDI values is assessed to investigate the efficiency of the SPEI.The results showed that the SPEI-1 values clearly reflect subtle changes in drought occurrence and reflect short-term conditions,whereas the SPEI-3 values reflect the occurrence of seasonal droughts.Meanwhile,the SPEI-6 and SPEI-12 values indicate long-term variations.The results show that the SPEI values over different time scales reflect complex variations in drought conditions.Droughts occurred on an annual basis in 1963,1966,1971,1978,1979,1986,2001,2011 and 2013.An analysis of the correlation between the monthly values of the TVDI and the SPEI-3 shows that a negative relationship exists between the SPEI-3 and the TVDI.That is,smaller TVDI values are associated with greater SPEI-3 values and reduced drought conditions,whereas larger TVDI values are associated with smaller SPEI-3 values and enhanced drought conditions.Therefore,this study of the relationship between the SPEI and the TVDI can provide a basis for government to mitigate the effects of drought.