(cest2017_00694) Statistical Comparison of nonlinear rainfall-runoff models for simulation in Africa North-West semi-arid areas

摘要

Determining the relationship between rainfall and runofffor a watershed is one of the most important problems andchallenging task faced by hydrologists and engineers.Conceptual hydrological models represent most suitabletools for this purpose in case of data scarcity. In this workwe set up a comparative study between two conceptualnonlinear models, the GR2M and the ABCD, applied tosemi-arid catchments located in north-west of Algeria.Monthly rainfall, temperature and stream flow data areavailable for the period 1971-2010. Overall, in calibration,the two models perform similarly, whereas the resultsshow that the GR2M model performed better than theABCD in the validation phase. Such circumstance could becaused by different motivations. On one side the differentnumber of model parameters that make the ABCD the lessparsimonious approach, with four parameters to becalibrated. On the other side the inability of the ABCDmodel to capture and describe the groundwater processes,important for the cases study. Moreover the validationphase embeds a large drought period, started in late 1980s,which makes difficult model adaptation to differenthydrological regimes.