Understanding the past to interpret the future: comparison of simulated groundwater recharge in the upper Colorado River basin (USA) using observed and general-circulation-model historical climate data; [Comprendre le passé pour interpréter le futur: comparaison de la recharge simulée des eaux souterraines dans le bassin supérieur de la rivière Colorado (Etats-Unis d’Amérique) en utilisant des données climatiques historiques observées et issues d’un modèle global de circulation]; [Entendendo o passado para interpretar o futuro: comparação da recarga das águas subterrâneas simulada na bacia superior do Rio Colorado (EUA) usando dados climáticos históricos observados e de modelo de circulação geral]; [El conocimiento del pasado para interpretar el futuro: la comparación de la recarga simulada de agua subterránea en la cuenca superior del río Colorado (EE.UU.), utilizando datos climáticos históricos y modelos generales de circulación]

In evaluating potential impacts of climate change on water resources, water managers seek to understand how future conditions may differ from the recent past. Studies of climate impacts on groundwater recharge often compare simulated recharge from future and historical time periods on an average monthly or overall average annual basis, or compare average recharge from future decades to that from a single recent decade. Baseline historical recharge estimates, which are compared with future conditions, are often from simulations using observed historical climate data. Comparison of average monthly results, average annual results, or even averaging over selected historical decades, may mask the true variability in historical results and lead to misinterpretation of future conditions. Comparison of future recharge results simulated using general circulation model (GCM) climate data to recharge results simulated using actual historical climate data may also result in an incomplete understanding of the likelihood of future changes. In this study, groundwater recharge is estimated in the upper Colorado River basin, USA, using a distributed-parameter soil-water balance groundwater recharge model for the period 1951–2010. Recharge simulations are performed using precipitation, maximum temperature, and minimum temperature data from observed climate data and from 97 CMIP5 (Coupled Model Intercomparison Project, phase 5) projections. Results indicate that average monthly and average annual simulated recharge are similar using observed and GCM climate data. However, 10-year moving-average recharge results show substantial differences between observed and simulated climate data, particularly during period 1970–2000, with much greater variability seen for results using observed climate data. © 2016, Springer-Verlag Berlin Heidelberg (outside the USA).