A regional approach was presented in this paper for the spatiotemporally distributed assessment of soil moisture changes in an arid area where ecosystems are highly groundwater dependent. To achieve this, the recent trends (from year 2002 to 2016) of soil moisture in Golmud River Basin were assessed using a combination of remote sensing and CLDAS data.
The Golmud River Basin is located in middle part of southern Qaidam Basin which covers a surface area about 4,566 km2. With average elevation of 2780 m above sea level, the area is a plateau basin and has a typical arid climate. The annual precipitation is only 40 mm and most precipitation is concentrated in summer season from June to August. The annual air temperature is 4.7°C and the winters are cold and windy. The pattern of land cover is as follows: 7% water bodies (mainly salt ponds), 12% shrub and grassland, 1% farmlands, 80% desert and salty areas.
MOD09A1 data of summer (June to September) from 2002 to 2016 formed the core datasets and the Apparent Thermal Inertia (ATI) was employed as the main method for mapping spatial distribution of soil moisture in Golmud River Basin. The Mann-Kendall non-parametric test was used to discuss the time series trends of soil moisture in this research.
The results revealed that the highest ATI is observed in the salt pond areas and it can reach up more than 0.08. Higher ATI areas are mainly occurred in farmlands, followed by grasslands and shrubs. On the other hand, the small ATI areas are mostly clustered in desert areas and ATI can be reduced to 0.02. When we look at the time series change of the annual ATI in the Gulmud River Basin, it generally increases during the period of 2002-2016 and the annual average ATI is varied between 0.02 and 0.03. According to the spatial distribution of the trends, the ATI of most areas has increasing trend and the decreasing trend is only happened in small parts of salty areas in the middle-east river basin. The ATI estimated from remote sensing data are in good agreement with the corresponding CLDAS soil moisture data in all four layers (0-10 cm, 10-40 cm, 40-100 cm, 100-200 cm) and the R-squares between their correlations are 0.80, 0.85, 0.84, 0.76, respectively. The soil moisture is increased with the depth to land surface and the value of layer 0-10 cm is obviously smaller than that of the other three layers. The vegetation cover is an important impact factor for soil moisture and the climate has no big effect on soil moisture due to the small precipitation in the Golmud River Basin.