Evaluation of Temperature Vegetation Dryness Index on Drought Monitoring Over Eurasia

This study explored the applicability of Temperature Vegetation Dryness Index (TVDI) derived from MODIS (Moderate Resolution Imaging Spectroradiometer) normalized difference vegetation index (NDVI) and land surface temperature (LST) data for drought monitoring in Eurasia from April to September during 2005 - 2014. TVDI was calculated based on the vegetation index/temperature trapezoid eigenspace (VITT) in this study. The Standardized Precipitation Evapotranspiration Index in three time scales (SPEI-01, SPEI-03, SPEI-06) and the Essential Climate Variable surface soil moisture product (ECV-SM) was used to evaluate the performance of TVDI over Eurasia, and to assess the sensitivity of TVDI to different land-cover types. The records of drought events from Emergency Events Database (EM-DAT) were also used to validate TVDI. The results indicated that TVDI negatively responded to SPEIs about 77.43%, 72.01%, and 75.95% of lands respectively, of which 29.84%, 33.93%, and 25.59% were subjected significant relationships (p <; 0.05). A negative correlation between TVDI and SPEIs was detected in most land-cover types, except for evergreen broadleaf forest and deciduous needleaf forest. Similarly, TVDI was also negative correlated with ECV-SM. About 66% of the study areas show negative correlation and 33% of them were statistically significant. There is a highly negative correlation between TVDI and ECV-SM in most types of land cover. Moreover, TVDI could not reflect drought condition correctly when TVDI <; 0.40; it could only reflect long-time excessive drought condition when TVDI > 0.86. A comparison of TVDI with drought records demonstrated that the TVDI could capture drought events in study area.