Snowmelt contributions to green and blue water fluxes: a model-data approach in a snow-dominated mountain catchment

The snow-dominated headwaters of the Colorado River are water towers of the south-western US. Together with an increasingly unsustainable water demand downstream in the Colorado basin, the transition to low- or no-snow conditions upstream in the coming decades has been profoundly altering hydrological regimes and water resources for ecosystems and societies. Looking at the “supply side” of this crucial issue, several studies in the intensively-monitored East River catchment in the Upper Colorado River (spanning shrub-dominated montane valley bottoms, subalpine forests and alpine barren hilltops) have pointed at snowmelt as a main driver of runoff generation and a significant contributor to plant-available green water during the growing season. Yet, a spatially-explicit analysis linking plant water use and runoff generation is lacking. Here we present how observations of stable isotopes of water (²H and ¹⁸O) in the precipitation and stream water, combined with spatio-temporal observations of snow cover and depth with multiple datasets related to the hydrometry and the energy budget, can be used to constrain and evaluate an ecohydrological modelling tool to then track snowmelt contributions to runoff and plant water use. Over the 2014-2020 time period, we deployed a new version of the spatially-distributed, process-based model EcH₂O-iso. The multi-objective calibration yielded a overall good model-data fit across critical zone interfaces and scales (stream, soil, snowpack, groundwater, ET), hinting at the model's ability to capture water fluxes, stores and mixing patterns in the catchment. A set of virtual tracers, tagging snowmelt and lateral saturated flow, further enabled to quantify the large contribution of snowmelt to stream discharge (60-80%) and root uptake (50-70%), much in line with previous independent, spatially-lumped estimates. We further evidence that snowmelt contributions to stream discharge both mobilizes fast surface pathways (runoff in Spring, dominant) and slower lateral groundwater flow downhill and seepage, with corresponding water ages up to several decades. Indeed, this baseflow remains significant in the growing season (~25% of outlet discharge or more), and we find that snowmelt makes up ~60% of groundwater recharge, again in agreement with previous estimates. From this catchment-scale picture, we further explore the spatial patterns of water ages and snowmelt fraction in blue and green water across the ecoclimatic zones of this catchment.