Estimating the water budget components and their variability in a pre-alpine basin with JGrass-NewAGE

The estimation of water resources at basin scale requires modelling of all components of the hydrological system. Because of the great uncertainties associated with the estimation of each water cycle component and the large error in budget closure that results, water budget is rarely carried out explicitly. This paper fills the gap in providing a methodology for obtaining it routinely at daily and subdaily time scales. In this study, we use various strategies to improve water budget closure in a small basin of Italian Prealps. The specific objectives are: assessing the predictive performances of different Kriging methods to determine the most accurate precipitation estimates; using MODIS imagery data to assist in the separation of snowfall and rainfall; combining the Priestley-Taylor evapotranspiration model with the Budyko hypothesis to estimate at high resolution (in time and space) actual evapotranspiration (ET); using an appropriate calibration-validation strategy to forecast discharge spatially. For this, 18 years of spatial time series of precipitation, snow water equivalent, rainfall-runoff and ET at hourly time steps are simulated for the Posina River basin (Northeast Italy) using the JGrass-NewAGE system. Among the interpolation methods considered, local detrended kriging is seen to give the best performances in forecasting precipitation distribution. However, detrended Kriging gives better results in simulating discharges. The parameters optimized at the basin outlet over a five-year period show acceptable performances during the validation period at the outlet and at interior points of the basin. The use of the Budyko hypothesis to guide the ET estimation shows encouraging results, with less uncertainty than the values reported in literature. Aggregating at a long temporal scale, the mean annual water budget for the Posina River basin is about 1269 ± 372 mm (76.4%) runoff, 503.5 ± 35.5 mm (30%) evapotranspiration, and −50±129 mm (−4.2%) basin storage from basin precipitation of 1730 ± 344 mm. The highest interannual variability is shown for precipitation, followed by discharge. Evapotranspiration shows less interannual variability and is less dependent on precipitation.