Sub-seasonal prediction of compound hot and dry extremes in India

Abstract

In India’s summer monsoon-dominated climate, compound hot and dry extremes (CHDEs) pose considerable challenges to agriculture and water resources. However, their drivers and predictability at sub-seasonal time scale remain largely unexplored. Using observations, reanalysis datasets, and forecast products, we examine the occurrence, drivers, and prediction skills of soil moisture based CHDEs in India from 1951 to 2020. The observed warming has resulted in a significant (p-value < 0.05) rise in the frequency of CHDEs in four (Himalayan, West Central, Central Northeast, and Northeast) out of six regions from 1951 to 2020. The area affected by CHDEs has also increased significantly during the recent (1986–2020) period compared to the previous (1951–1985). The most severe CHDEs occurred during the major summer monsoon failures in 1972, 1987, 2002, 2009, 2014, and 2015. The summer monsoon breaks are the main drivers of CHDEs, as about 98% of the events occurred during the break periods. We used sub-seasonal to seasonal (S2S) forecasts from the United Kingdom Meteorological Office (UKMO) and Extended Range Forecast Systems (ERFS) to examine the prediction skills of CHDEs at different lead times. While UKMO (from S2S) and ERFS forecasts demonstrate satisfactory skills to predict CHDEs in India at a 15-day lead, UKMO provides better skills than the ERFS, which is currently operational in India. Good prediction skills from both the models for severity and area of CHDEs demonstrate their utility for the early warning of compound extremes in India, which can assist in climate adaptation.