This study aims at defining a method for selecting an optimal interpolated climatic dataset in the context of an integrated modelling of the Ebro (85 000 km2 , Spain) hydro-system. Two different sets of temperature and precipitation were chosen according to data availability criteria, and each set was interpolated on an 8 × 8 km grid covering the basin, with and without a monthly or annual altitudinal gradient applied by geographical area (Pyrenees, Cantabrian and Iberian ranges) over 1000 m a.m.s.l. Seven basins (464 to 2975 km2 ) representing different flow regimes of the Ebro catchment were chosen to evaluate the performance of 24 different climatic datasets for modelling water resources. We used a conceptual model (GR4j) at a daily time step to test the sensitivity of hydrological modelling to the different sets of precipitation and temperature. A global score was attributed to each dataset according to the different hydrological modelling criteria in order to discriminate the best-performing dataset.