Spatio-temporal variability of dryness and wetness based on standardized precipitation evapotranspiration index and standardized wetness index and its relation to the normalized difference vegetation index

Land surface change (LSC) due to human-caused global environmental changes has considerably affected the development of regional droughts. The standardized wetness index (SWI) developed recently by combining the standardized precipitation evapotranspiration index (SPEI) with the evapotranspiration (ET) estimated from the Budyko framework considers the joint effects of climate variability and LSC on the land dryness/wetness conditions. Here, using a 25-year (1984–2008) monthly gridded terrestrial water budget dataset, a comparative global analysis of the spatio-temporal variability of drought characteristics (duration D, severity S, peak K, frequency, and drought area) estimated based on SWI and SPEI is presented. The relationship between dryness/wetness (as indicated by SWI and SPEI) and LSC (as indicated by the Normalized Difference Vegetation Index, NDVI) is explored by correlation analysis and the dynamic time warping algorithm (DTW). The results show that SWI is strongly correlated with SPEI for most global regions (except for the extremely arid and extremely humid regions). Both SWI and SPEI show similar directions and magnitudes in the trends of drought characteristics during 1984–2008. However, the drying trends in SPEI are stronger than that of SWI particularly in arid regions, accompanied with larger global drought areas and higher frequencies of extreme drought. Furthermore, the correlation between SWI and LSC is larger than that between SPEI and LSC particularly in arid regions (e.g., northern Africa, Indus, and southern Australia). As SWI accounts for the effects of LSC, it can be inferred that the underestimation of drying trend indicated by SWI relative to SPEI is partially caused by LSC.