Modelling of Peak Monsoon Rainfall Totals using Regional Frequency Analysis of sites of Punjab, Pakistan

Abstract

Floods are one of the most widely recognized risks on the planet and cause loss of lives, job and property annihilation. The extreme events that affect the environment the most, they play a vital role for investigating climate. Pakistan has been a target to extreme floods and dry spells. It has become a prompt need to investigate such hazards as it will help to reduce the consequences of extreme events. As intense precipitation is associated with floods, so for this investigation extreme precipitation has been statistically analysed to forecast the floods. This work encapsulated regional frequency analysis (RFA) using the Peak Monsoon Rainfall Totals (July and August) variable on 10 stations of Punjab. RFA based L-moments were applied on 10 stations situated in Punjab, Pakistan. According to the discordancy measure demonstrates that the cluster of 10 stations does not contain any discordant sites. The heterogeneity measure (H) provided that the region of 10 stations as a whole is heterogeneous therefore, it is divided into two regions using time series plots. Note that cluster analysis using available site characteristics has not provided desirable homogeneous regions. L-moment ratio diagram and Z distribution statistic results indicated that GLO distribution and GEV distribution are the best fit distributions for Region 1 (consisting of seven stations namely Bahawalnagar, Faisalabad, Jhelum, Lahore, Sialkot, Sargodha and Mianwali) and Region 2 (consisting of three sites namely Bahawalpur, Khanpur and Multan) respectively. Regional and as well as at-site quantiles for various return periods from 1 to 100 years have been calculated.