Evaluation and Prediction of Precipitation in Qassim Region using Adaptive Neuro Fuzzy Inference System

Abstract

ABSTRACT: Storm wateranalysis is a highly daunting but useful task in arid regions which face seasonal flows from wadies. Flash floods causing threat to life and property require extensive management for their mitigation in such a way that it may help in reducing to some extent the problem of scarcity of water in the region. This vital issue demands in-depth analysis of rainfall and storm water. This research has analyzed rainfall in Qassim Region Saudi Arabia using statistical techniques. Future predictions have been made by the application of Adaptive Neuro Fuzzy Inference System (ANFIS). Three ANFIS models were tested with different architectures and comparatively better performing one was used to simulate the future precipitation. Long historic records of rainfall from 1979 to 2015 collected from Municipality of Buraydah have been scrutinized. Seasonality of rainfall was examined by appraising the seasonality index (SI). The prevailing rainfall intensity was premeditated from Intensity Duration Frequency (IDF) curves. Rational formula was used to find peak of storm-water. It was found that drainage of infrastructure needs improvements especially in downtown of Buraydah city, Saudi Arabia.