On Spatio-temporal Variability Of Soil Moisture In Babao River Basin As A Typical Cold Alpine Region And Its Downscaling Approach

Soil moisture is an important variable in the global climate system.It controls water cycle,energy exchange and carbon cycle between the land surface and the atmosphere.Soil moisture usually presents high spatial and temporal variabilities due to various factors such as climate,topography and vegetation.In this paper,WATERNET observation datasets in the upper reaches of the Heihe River Basin were employed to analyze the soil moisture’s spatial-temporal variability in Babao River Basin.The relationship between topography,vegetation,soil temperature and soil moisture at different depth were studied.Then we analyzed the scale effect of soil moisture.Finally,Soil Moisture Active Passive(SMAP)data at a 9 km resolution was downscaled to 1 km using the proposed downscaling method,which is based on apparent thermal inertia(ATI)considering sub-grid variability.The main conclusions of this paper are as follows:(1)Soil moisture increases with depth(4 cm,10 cm,20 cm),and decreased from 2013 to 2015.Soil moisture in the three depths shows distinct seasonal variation.(2)Standard deviation,coefficient of variation and semivariogram were used to study the spatial variability of soil moisture at 4 cm.The standard deviation of soil moisture increases and then decreases as the mean soil moisture content increases.The coefficient of variation exhibits moderate variability,ranging from 0.2-0.8.The optimal theoretical semivariogram model in August to October is the Gaussian model,with a range of about 5.2 km and a sill of about 0.018.(3)Daily mean soil moisture at the Babao river basin measured by the WATERNET network was compared against the SMAP data.It was found that the SMAP soil moisture data is underestimated,but its temporal variation is similar to the observations.The correlation coefficient between them is 0.826,and the spatial variability associated with the SMAP data is weak.(4)Correlation analysis between soil moisture and various factors show that soil moisture at 4 cm and 20 cm depths was significantly positively correlated with the altitude.Slope,aspect,topographic wetness index and vegetation had little effects on the soil moisture distribution.There is a significant positive correlation between soil moisture and soil temperature at 4 cm and 20 cm depths.Soil temperature has a lager effect on soil moisture in deeper soils.The correlation between soil moisture and soil temperature is greater at the freeze state than the non-freeze state.(5)The scale effect was observed in soil moisture.The variability of soil moisture decreases along with the increasing spacing.The semivariograms differ on different scales.(6)The linear relationship between MODIS derived ATI and in situ soil moisture measurements was validated and it thus supports the use of ATI as a proxy for the purpose of downscaling the SMAP soil moisture data.The proposed method was evaluated from three aspects.The metrics of correlation coefficient(R),root mean square error(RMSE)and mean absolute error(MAE)between the downscaled results and in situ measurements were 0.636、0.105 cm3/cm3、0.089 cm3/cm3,respectively.