Droughts’ projections in homogeneous climatic regions using Standardized Precipitation Index in Pakistan

Abstract

This study is conducted to develop homogenous climatic regions (HCR) based on Standardized Precipitation Index (SPI) by combining cluster analysis and L-moment approach. Initially, cluster analysis was performed using site characteristics which resulted in five subjective homogenous regions. To check homogeneity, discordancy and heterogeneity measures were implied based on at-site statistics of the stations using L-moments approach. These statistical measures established validity of five HCR for Pakistan and showed that elevation has a key role in the construction of regions. Best-fit distributions were selected in the form of Generalized Pareto distribution and Pearson type-III distribution using L-Moment Ratio Diagram (LMRD) and goodness-of-fit (GOF) test for drought risk assessment. Projections of drought were obtained for each region in the form of regional quantiles at various return periods using regional frequency analysis (RFA). Finally, at-site frequency analysis (ASFA) was performed to calculate quantiles at different return periods and project drought at each individual site using best-fit distributions. The uncertainty of both types of quantile results was assessed using Monte Carlo Simulation which showed similarity at lower while increasing uncertainty at higher return periods. From our analysis, we found that RFA is statistically better due to many values from sites of the region as compared to ASFA for future drought risk assessment. According to the results, regions one and two are enriched of water resources due to rainfall and glaciers, region five has enough rain in monsoon season while regions three and four include most vulnerable and drought-prone areas of Sindh and Baluchistan provinces and have chances of severe to extreme drought in the near future. The findings of this study are helpful for policymakers, monitoring droughts and its projections, water resource planners, etc. in the study area.