Evaluating soil moisture status in China using the temperature-vegetation dryness index (TVDI)

In this paper the normalized difference vegetation index (NDVI) and land surface temperature (T-s) derived from National Oceanic and Atmospheric Administration (NOAA) advanced very high resolution radiometer (AVHRR) 10-day composite data were adopted to construct an NDVI-T-s space, from which the temperature-vegetation dryness index (TVDI) is suggested. Soil moisture status in all of China from March to May 2000 was evaluated. To minimize the effect caused by spatial heterogeneity of climatic parameters, TVDI was calculated for each of the three agricultural climatic regions in China. The results show that severe drought was mainly distributed in northwestern China and part of north and south China; the area of severe drought increased from about 67 x 10(4) km(2) in March to 126 x 10(4) km(2) in May. To evaluate the efficiency of TVDI in soil moisture monitoring, the spatial pattern of TVDI was first compared with and verified by the measured soil moisture from more than 90 weather stations in China by the linear regression method. It was found that there is a significant negative linear correlation between TVDI and the measured soil moisture. The paper then evaluates the sensitivity of TVDI to NDVI and T-s. The results show that T-s was more sensitive than NDVI to TVDI and that land surface temperature contains much more drought information than NDVI. Lastly, the paper compares the efficiency of TVDI with the drought index that was deduced from T-s alone, namely the crop water stress index (CWSI). The results show that TVDI has a more significant relation than CWSI to soil moisture measured in situ. All of the aforementioned results indicate that TVDI based on a combination of T-s and NDVI is more accurate than CWSI based solely on T-s in regional drought evaluation.