Revista de Engenharia Civil e Ambiental

A Comprehensive Statistical Technique for Forecasting of Extreme Hydrological Situations

Abstract

Khurshid Ahmad, Bashir Ahmad and Akhtar Alam

Extreme hydrological situations constantly disturb the earth activities and life, to envisage such extreme activities we need a system that alarms well on time and recognized the expected danger; to prepare such systems one must have knowledge of the significant factors that are actively responsible for such extreme situations and we should have a reliable statistical technique that helps to prepare a useful model for such systems. In this paper we investigate the historical data of peak flood from several gauging stations of river Jhelum in Kashmir, India. A reliable estimation technique (L-moment) is applied for parametric estimation of the probability distributions and a reliable testing techniques are used to check the accuracy of fitting of the distribution, in additional to that L-Moment Ratio Diagram (LMRD) is used to impart information about fitting of distribution. Log Pearsons-III distribution shows better results and satisfies tests of distribution fitting, same probability distribution is globally accepted for flood forecasting.