Probabilistic modelling of interarrival time of drought for different operational drought indices used in Pakistan

Abstract

AbstractAdverse climate changes in Pakistan undermine the water resources and give rise to transient and permanent droughts in different parts of the country. The occurrence of drought cannot be prevented, but its detrimental impacts can be brought down by proper pre‐drought planning. Effective planning owes to using scientific and probabilistic approaches. In this paper, the interarrival time of drought for seven operational drought indices in Pakistan is modelled by using the modified distribution of convolution derived from the Bivariate Affine Linear Exponential (BALE) distribution. A new stochastic variable, interarrival time of drought, is generated from the data of drought indices from 1951 to 2016 used in Pakistan by applying the theory of runs. The proposed model identifies Palmer drought severity index (PDSI) and Reconnaissance drought index (RDI) as the most appropriate meteorological indices for the under‐study country. The quantiles of the model are computed that provide information about the interarrival time of drought and help to anticipate this phenomenon. In addition, return periods for all the seven indices are calculated to report the frequency and duration of drought over a certain period. Finally, some recommendations for the National Drought Monitoring Centre (NDMC) and Pakistan Council of Research in Water Resources (PCRWR) are stated for better future planning to avoid the adverse impacts of drought.