Drought Monitoring Models With Remote Sensing And Spatio-Temporal Characteristics Of Drought In China

China, locating in the eastern Asia, has a representative monsoon climatic and drought happened frequently because of the unsteadiness of monsoon. Drought has caused the most severe loss in all of the meteorological disasters in China. As the population growing, reclamation farmland and planting index increasing, the requirement of agro-water and industrial water increasing, drought will be fearful farther in the global change environment.To analyse the drought spatio-temporal pattern in China for a long time with remote sensing, the study focused on Regional Drought Monitoring Models only based on Remote Sensing data. Pure-remote sensing information models are free of the numerous distributional parameters that are difficult to obtain in vast area. The following sections are included in the study.(1) The VCI (Vegetation Condition Index) and SVI (Standard Vegetation Index) based on time-series NDVI data are used to analyse regional drought and the following results were reached: As the vegetation growth status index, the response of VCI and SVI to precipitation was land cover type, temporal and site specified. So VCI and SVI have little information about drought when used to monitor regional drought especially in China that has a mountainous topographty.(2) Based on the interpretation of NDVI-Ts space, NDVI and land surfacetemperature was combined to construct NVDI-Ts space from which the TVDI (Temperature Vegetation Dryness Index) was obtained with line ratio method and normalized slopes method. When compared with the topsoil moisture measured in situ, TVDI are competent in regional drought monitoring. Illumed by the NDVI-Ts space, we combined the NDVI with day/night land surface temperature difference and apparent thermal inertia (ATI) to construct NDVI- A T and NDVI-ATI spaces from which the DVDI and AVDI were estimated respectively with 8-day composite MODIS TS product and 16-day composite MODIS NDVI product. Comparison between TVDI, DVDI and AVDI has showed that TVDI has the more promising potent in drought monitoring than the other two and AVDI is not competent in regional drought monitoring. From the NDVI- A T space, we found that the wet bare soil has the same day/night Ts difference with the well-watered vegetation surface.(3) Considering that land surface temperature was the result of regional energy balance and energy from sun radiance show apparent belt feature in China because of a large latitude range and mountainous topographty, the energy belt feature should have some effect on TVDI when used in drought monitoring. To abate the effect, we will substitute the Ts with the land surface temperature minus air temperature (Ts-Ta) because Ts-Ta is the token of temperature grads between land surface and atmosphere and is the primary drive. To combine (Ts-Ta) with NDVI to construct VIT, estimating air temperature from remote sensing is necessary. In the paper, the relationship between NDVI, Ts and air temperature was analysed, then NDVI-Ts space method and Prihodok & Goward (1997) model were used to estimated air temperature. Comparison results showed that precision of estimated air temperature from the two methods are equivalent and NDVI-Ts method can used the remote sensing information more fully than Prihodok & Go ward model.?(4) With the air temperature estimated from remote sensing data with NDVI-Ts space, the VIT can be constructed by combining NDVI with (Ts-Ta). The Water Deficit Index (WDI) was estimated from VIT and compared with topsoil moisture. The result showed that WDI could be used to evaluate soil water status logically. Compared with TVDI, the WDI has more academic basic. Theoretically, WDI has the similar physical foundation as CWSI. More over, WDI has overcome energy belt feature inherent with TVDI. Though WDI has a lower correlation coefficient with topsoil moisture than TVDI, WDI was still used as the drought indictor in the study on drought spatio-temporal pattern in China.(5) NOAA AVHRR CH4, CH5 and NDVI 10-day composite data with 8km spatialresolution was used to...