| ABSTRACT | Montane snowpack, with its significant water storing capacity and subsequent modulation of river discharge, is a key control on the available water resources in the Mediterranean region. Consequently, and particularly in the face of future projections, there is much to be gained from diagnosing the extent of these resources at any given time. Unfortunately, and in spite of previous efforts, the accurate estimation of snow water equivalent (SWE) through data assimilation is still a major hydrometeorological challenge. On the one hand, due to strong horizontal heterogeneity across a range of scales, modeling grid-scale SWE is far from trivial. On the other hand, the only widely available observational estimates of SWE are based on passive microwave satellite retrievals and these are not only coarse but known to contain significant bias for wet and deep snowpacks. To constrain the uncertainties inherent in the estimation of SWE we have developed a modular ensemble-based subgrid snow data assimilation framework (ESSDA). Our framework makes use of the deterministic ensemble Kalman filter (DEnKF) with Gaussian anamorphosis to assimilate snow cover fraction (SCF) retrievals into the subgrid snow distribution submodel. Thereby, instead of direct assimilation of the quantity of interest (SWE) based on the aforementioned passive microwave retrievals, ESSDA employs an indirect assimilation of SCF to update the subgrid SWE distribution through an ensemble of non-linear snow depletion curves. We demonstrate the potential of ESSDA through both synthetic (twin) and real experiments. The latter are carried out for a domain in the canton of Valais in south-western Switzerland. In Valais the melt of the alpine snowpack represents the strongest signal in the seasonal discharge of the Rhone river. The domain was chosen as a proof of concept with regards to the use of ESSDA for monitoring SWE resources in an area where these are critical for hydropower and agriculture. Given its modularity and robustness, our framework can readily be applied to diagnose the state of SWE for other montane snowpacks in the Mediterranean region. |