Characterization and evaluation of MODIS-derived Drought Severity Index (DSI) for monitoring the 2009/2010 drought over southwestern China

This article investigates whether the Moderate Resolution Imaging Spectro-radiometer (MODIS)-derived global terrestrial Drought Severity Index (DSI) had the capability of detecting regional drought over subtropical southwestern China. Monthly, remotely sensed DSI data with 0.05 degrees spatial resolution were used to characterize the extent, duration, and severity of drought from 2000 to 2010. We reported that southwestern China suffered from incipient to extreme droughts from November 2009 to March 2010 (referred to as the "drought period"). The area affected by drought occupied approximately 74 % of the total area of the study region, in which a moderate drought, severe drought, and an extreme drought accounted for 20, 12.7, and 13.2 % of the total area, respectively; particularly in March 2010, droughts of severe and extreme intensity covered the largest areas of drought, which were 16.1 and 18.6 %, respectively. Spatially, eastern Yunnan, western Guizhou, and Guangxi suffered from persistent droughts whose intensities ranged from mild to extreme during the drought period. Pearson's correlation analyses were performed between DSI and the in situ meteorological station-based Standardized Precipitation Index (SPI) for validating the monitoring results of the DSI. The results showed that the DSI corresponded favorably with the time scales of the SPI; meanwhile, the DSI showed its highest correlation (mean: r = 0.58) with a three-month SPI. Furthermore, similar spatial patterns and temporal variations were found between the DSI and the three-month SPI, as well as the agro-meteorological drought observation data, when monitoring drought. Our analysis suggests that the DSI can be used for near-real-time drought monitoring with fine resolution across subtropical southwestern China, or other similar regions, based solely on MODIS-derived evapotranspiration/potential evapotranspiration and Normalized Difference Vegetation Index data.