Evaluation of additional physiographical variables characterising drainage network systems in regional frequency analysis, a Quebec watersheds case-study

Regional Frequency Analysis (RFA) relies on a wide range of physiographical and meteorological variables to estimate hydrological quantiles at ungauged sites. However, additional catchment characteristics related its drainage network are not yet fully understood integrated in RFA procedures. The aim the present paper is propose integration several characterizing systems RFA, evaluate their added value predicting proposed extended dataset (EXTD) includes characterising characteristics. To new variables, number commonly used approaches applied data representing 151 stations Quebec (Canada) compared standard (STA) that excludes variables. considered include combination two neighborhood methods namely canonical correlation analysis (CCA) region influence (ROI) with regional estimation (RE) models which log-linear regression model (LLRM) generalized additive (GAM). RE also without neighborhood. Results show using lead significantly better flood quantile predictions, especially for large basins. Indeed, variable selection performed EXTD consistently some particular density, stream length ratio, ruggedness number. Two other identified included DHR step: circularity ratio texture ratio. This leads predictions relative errors about 29% EXTD, versus around 42% STA case best approaches. Thus, allow representation physical dynamics within watersheds.