Predicting Maximum Discharge In Algeria Wadis Using Dimensional Analysis And Statistical Modeling: A Case Study Of Semi-arid Watersheds
Auteurs : Sakhraoui Fouad . Herezi Toufik .
Résumé
Accurate estimation of maximum discharge (Qmax) is essential for hydraulic infrastructure design and flood management in semi-arid regions, particularly where hydrometric data are scarce. This study presents a physically meaningful and regionally calibrated empirical model to estimate Qmax in Algerian wadis, combining approach of dimensional analysis and statistical modeling. Based on data from 60 watersheds provided by the Algerian National Agency for Hydraulic Resources (NAHR), the model integrates three key watershed descriptors: catchment area (A), rainfall intensity (Ptc), and average slope (S). The proposed model demonstrates strong predictive accuracy, with a correlation coefficient of R = 0.97 and a mean normalized error of 15% in the validation phase. Comparative tests against widely used formulas (Giandotti, Mallet-Gauthier, Sokolovsky, Turazza) reveals that the proposed model yields significantly lower errors and better prediction stability. Additionally, 66.6% of the estimates fall within ±10% of observed values, compared to just 3-13% for conventional methods. These results underscore the model’s utility as a reliable, simple, and transferable tool for peak discharge estimation in data-scarce, semi-arid environments such as Algeria.
Mots clés
Maximum discharge ; Statistical modeling ; Dimensional analysis ; Algerian wadis ; Flood management
