Reassessing the frequency and severity of meteorological drought considering non-stationarity and copula-based bivariate probability

Drought is a multifaceted natural phenomenon with diverse spatial and temporal characteristics and significant environmental, social, and economic effects. The utility of conventional drought indices based on the assumption of stationarity is questionable due to the advent of rapid environmental change. We proposed a non-stationary standardized precipitation index (SPINS) to reassess the nature of meteorological drought. Precipitation observations from eight rain gauge stations for 1961-2017 in South Korea were fitted to a non-stationary gamma distribution using a generalized additive model in location, scale, and shape algorithm. Drought duration and severity were identified by SPINS and joint and conditional return periods were estimated using copula-based bivariate probabilities. The results show that average drought duration, severity, and frequency obtained from SPINS were more severe than those produced by a conventional standardized precipitation index (SPIC). The average inter-arrival time was 41.46 months and 39.86 months for SPIC and SPINS, respectively, at the Seoul station. The joint return periods of observed drought event from March 2008 to June 2008 with a duration of 4 months and a severity of 4.8 were 10.4 and 6.4 years for the SPIC and SPINS, respectively, at the Seoul station. We conclude that return periods calculated using the SPINS are more reliable than those calculated by the SPIC. These findings should prove helpful in drought management and risk assessment under a changing environment as effective management requires prior knowledge of events.

Automated summary

Drought is a multifaceted natural phenomenon with significant environmental, social, and economic effects. The utility of conventional drought indices is questioned due to rapid environmental changes, leading to the proposal of a new non-stationary standardized precipitation index (SPINS). Results show that SPINS calculates more severe drought severity and more reliable return periods compared to conventional indices.