Remote sensing data are frequently incorporated into drought indices used widely by research and management communities to assess and diagnose current and historic drought events. The integrated drought indices combine multiple indicators and reflect drought conditions from a range of perspectives (i.e., hydrological, agricultural, meteorological). However, the success of most remote sensing based drought indices is constrained by geographic regions since their performance strongly depends on environmental factors such as land cover type, temperature, and soil moisture. To address this limitation, we propose a framework for a new integrated drought index that performs well across diverse climate regions. Our framework uses a geographically weighted regression model and principal component analysis to composite a range of vegetation and meteorological indices derived from multiple remote sensing platforms and in-situ drought indices developed from meteorological station data. Our new index, which we call the station-enabled Geographically Independent Integrated Drought Index (GIIDI_station), compared favorably with other common drought indices such as Microwave Integrated Drought Index (MIDI), Optimized Meteorological Drought Index (OMDI), Precipitation Condition Index (PCI), Temperature Condition Index (TCI), Soil Moisture Condition Index (SMCI), and Vegetation Condition Index (VCI). Using Pearson correlation analyses between remote sensing and in-situ drought indices during the growing season (April to October) from 2002 to 2011, we show that GIIDI_station had the best correlations with in-situ drought indices. Across the entire study region of the continental United States, the performance of GIIDI_station was not affected by common environmental factors such as precipitation, temperature, land cover and soil conditions. Taken together, our results suggest that GIIDI_station has considerable potential to improve our ability of monitoring drought at regional scales, provided local meteorological station data are available.
Read full abstract