Three techniques for flow component identification from daily discharge data in Madjez Ressoul catchment, Algeria

Abstract

Empirical models of recession analysis provide information about surface and ground flow processes during periods of drought. The objective of this research is to evaluate the performance of three different techniques (individual segments, filtering method, genetic algorithms) in assessing the contribution of recession flow component in calculating runoff, using daily records of discharge. These techniques will focus on the values of the recession index covering ten flood events in the period of 1973–2003. The outputs of these models were then compared and their results appeared to validate two particular techniques: the filtering method with one parameter and the genetic algorithm method. The coefficient of determination (R2 = 0.7324–0.935) and the cross probability which equals 0.9 confirm the best separation of all events. Absolute similarities between flow types in the filtering method and genetic algorithms present systematic differences in the calibration form and on the consideration of obstacles and limitations.