Drought variability in Pakistan: Navigating historical patterns in a changing climate with global teleconnections

Abstract

This study investigates the spatiotemporal variability of drought patterns in Pakistan on an annual timescale over 50 years (1971 – 2020) using six distinct drought indices [viz., Standardized Precipitation Index (SPI), Agricultural SPI, Reconnaissance Drought Index (RDI), Effective RDI, Deciles Index (DI), and Percentage Departure (PD)]. Empirical Orthogonal Function (EOF) analyses are employed on the SPI drought index to evaluate interannual variations in drought and their correlation with large-scale ocean–atmosphere circulation patterns. The magnitude of the trends is measured using the non-parametric Sen’s slope estimator, while their statistical significance is evaluated through the Mann–Kendall test. To further explore potential shifts in the correlations between the annual SPI and various climate indices, Rodionov’s regime shift detection test is applied. Our findings revealed six drought years: 1971, 2000, 2001, 2002, 2015, and 2018. The most intense and prolonged episode of drought, reaching an extreme category, occurred from 2000 to 2002, affecting over 60% of Pakistan’s total area. The leading EOF mode of the annual SPI demonstrates a robust relationship with the Pacific Decadal Oscillation (PDO). The second mode characterizes a significant Tropical Southern Atlantic (TSA) pattern, suggesting some level of predictability in drought occurrences across Pakistan. Moreover, regime shift analysis reveals two significant shifts: one in 2006 in the correlation between SPI and PDO, as well as Niño 3.4, and another in 2013 between SPI and TSA. This study can provide valuable insights for policymakers to develop climate-resilient agricultural and water resource management strategies, fostering sustainable development in drought-prone areas of the country.