Regional variation of drought parameters and long-term trends over India using standardized precipitation evapotranspiration index.

Analysing historical drought pattern is vital for implementation of efficient drought adaptation and mitigation policies. In this study, we examined the meteorological drought characteristics of India during 1901-2015, using Climate Research Unit (CRU) based Standardized Precipitation Evapotranspiration Index (SPEI) at multiple timescales i.e., 1 month (SPEI01), 3 month (SPEI03), 6 month (SPEI06), 12 month (SPEI12). Here, we applied K-means clustering algorithm on SPEI12 (December) to find out different clusters with distinct drought characteristics. The six different homogeneous regions, i.e., cluster1 (C1), cluster2 (C2), cluster3 (C3), cluster4 (C4), cluster5 (C5), and cluster6 (C6) identified by K mean clustering largely resemble with the clusters mentioned in previous researches. Different drought parameters (duration, frequency, intensity) have been also analysed for each cluster on a monthly, seasonal and interannual basis. The study indicates that northern part of India (C6, C3) experienced frequent droughts at shorter timescales whereas the western (C2) and north eastern (C4) part of the country encountered frequent drought occurrences at longer timescale. It is worthy to mention that the C2 region comprising the semi-arid and arid western part of the country including the great Indian desert, is vulnerable to frequent, prolonged and severe droughts at longer timescale (SPEI12). The study revealed a significant regional variation in drought trends identified by Modified Mann-Kendall (MMK) trend test. The annual trend analysis shows statistically significant (p < 0.05) increasing drought trend over C3 and C4 regions comprising the fertile Gangetic and Brahmaputra plains. The seasonal MMK trend analysis reveals significant increase (p < 0.05) in droughts over C3 (-0.006) and C4 (-0.005) during monsoon. The increasing drought trend over the Gangetic plain (C3) is prominent especially in the months of July (p < 0.05, slope = -0.005) and August (p < 0.001, slope = -0.006). The study provides a region-specific understanding of drought characteristics and long-term trends crucial for preparing adaptive strategies to minimize the cumulative impacts of droughts.