Land surface temperature derived from brightness temperature dataset and NDVI dataset were used to construct the NDVI/T s space

from which a dryness index-temperature/vegetation dryness index （TVDI） was suggested. The dryness index that combines the land surface temperature with vegetation spectral index is computationally straightforward because it is based on the information derived from satellite data only. Using the TVDI

the surface moisture status in China in March and May in 2000 was studied. The results showed that the severe drought was mainly distributed in North-western China and part of North China and South China; the severe drought area enlarged from about 67×104km2 in March to 126×104km2 in May. TVDI spatial pattern was compared with the measured topsoil moisture from the observatories around China with the linear regression method. A negative linear correlation between TVDI and the measured soil moisture was found

thus TVDI ’s validity in evaluating drought was verified. Afterward

the sensitivity of TVDI to NDVI and T s was evaluated. The results show that the TVDI are more sensitive to the land surface temperature than to NDVI and land surface temperature have more drought information than NDVI which lead to the conclusion that dryness indexes based on land surface temperature are more reasonable than the ones based on NDVI.