Investigating regional distribution for maximum daily rainfall in arid regions: case study in Saudi Arabia

Abstract

The hydrologists need effective procedures to assist them in predicting the implications related to high rainfall risks, both at gauged and ungauged areas. The first step for any assessment is estimating the rainfall values associated with various return periods in years. This information is obtained using rainfall frequency analysis techniques based on observed maximum daily rainfall values. The existence of a regional distribution for the rainfall frequency within a certain region is considered as a precious information for the hydrologists to estimate—with confidence—the expected rainfall at high return periods. The aim of this research is to determine the regional statistical distribution for various regions in Saudi Arabia. Saudi Arabia was selected to represent arid regions as it covers a large area of the Arabian Peninsula in addition to the availability of rainfall data compared with other similar countries. The data of 394 rainfall gauging stations were included in the study, which covers all Saudi Arabia in its 13 main regions. The data was analyzed using multiple frequency analysis methods and was subjected to various statistical tests including Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), Anderson-Darling Criterion (AD), Station-Year Analysis, and Index Flood Method, in order to determine the best regional statistical distribution and define the homogenous regions using regional analysis. The study concluded that the Log-Pearson type III distribution was the best model to describe the distribution of the daily maximum rainfall in the region. Saudi Arabia was sub-divided into seven homogeneous regions. Contour maps for the 2-, 5-, 10-, 25-, 50-, and 100-year rainfall were produced to predict the rainfall at any point within the Kingdom of Saudi Arabia.