Extreme Rainfall Frequency Analysis for Meteorological Sub-Division 4 of India Using L-Moments

Extreme rainfall frequency analysis for Meteorological Sub-Division 4 of India was analyzed using Lmoments approach. Serial Correlation and Mann Kendall tests were conducted for checking serially independent and stationarity of the observations. The discordancy measure for the sites was conducted to detect the discordant sites. The regional homogeneity was tested by comparing with 500 generated homogeneous regions using a 4 parameter Kappa distribution. The best fit distribution was selected based on Z statistics and L-moments ratio diagram from the five extreme value distributions GPD, GLO, GEV, P3 and LP3. The LN3 distribution was selected and regional rainfall frequency relationship was established using index-rainfall procedure. A regional mean rainfall relationship was developed using multiple linear regression with latitude and longitude of the sites as variables. Keywords—L-moments, Z statistics, Serial correlation, Mann Kendall test.