Résumé
In a context of water scarcity in Peruvian Pacific catchments added to the paucity of data availability, we propose a methodology which provides new perspectives for freshwater availability estimation as a base reference for unimpaired conditions in the 1970-2010 period. To represent the multi-decadal behaviour of freshwater runoff along the region, a regional runoff analysis is proposed based on hydrological modelling at annual and monthly time step for unimpaired conditions over the whole 1970 ‒ 2010 period. Differential Split-Sample Tests are used to assess the hydrological modelling robustness of the GR1A and GR2M conceptual lumped models, showing a satisfactory transposability from dry to wet years inside the thresholds defined for Nash-Sutcliffe and bias criteria . This allowed relating physical catchment characteristics with calibrated and validated model parameters, thus offering a regional perspective for dryland conditions in the study area (e.g. the anticlockwise hysteresis found for seasonal precipitation-runoff relationship) as well as the impacts of climate variability and catchment characteristics. Under those considerations, a regional discharge of 709 m3/s to the Pacific Ocean from the arid Peruvian Pacific region is estimated with a significant increasing trend of about 43 m3/s/decade over the last four decades (1970‒ 2010). Based on our regional hydrological modelling approach, which allowed describing the unimpaired runoff variability, the Peruvian National Service of Meteorology and Hydrology (SENAMHI) is implementing a downscaled runoff gridded product at sub-basin scale. Among other operational purposes, this product will be applied to characterize hydrological indices for droughts. This offers a great perspective for surveying water deficit in this region which only benefits from 2% of the total national available freshwater while they concentrate almost 50% of the population of the country.