A method for low flow estimation at ungauged sites, case study in Wallonia (Belgium)

Abstract

Abstract. Being able to estimate low flows at any point of a river is really important nowadays for a good integrated management of rivers. Knowing the magnitude as well as the frequency of such extreme events becomes essential. In order to build a model of low flow calculation, usable in ungauged catchments and which takes also into account low flow frequency, we started with a low flow frequency analysis including a comparison of different distributions. Two-parameter Lognormal and Gamma were the most common distributions that fit low flow data in Wallonia. This was followed by a regionalisation of low flows using 25 different climatic and physical catchment variables, and the development of regression models that can be used to estimate the minimum 7-day average flow for different return periods, using catchment characteristics. The variables the most correlated to specific minimum 7-day average flows were the recession coefficient and percolation, regardless of the return period. The determination coefficients of the models ranged from 0.51 to 0.67 for calibration and from 0.61 to 0.80 for validation. Finally, regression coefficients were logarithmically linked to the return period. This enabled us to develop a single model per region and for the whole study area, in function of the return period. In conclusion, the method developed in this study allows us to estimate low flows in gauged and ungauged catchments of a given region for a given return period. The interest of regionalisation and development of regional models is also discussed.