Kristoffer Aalstad, University of Oslo

An ensemble-based subgrid snow data assimilation framework

S. Westermann, L. Bertino


Snow, with high albedo and low thermal conductivity, strongly modulates the thermal state of the underlying ground in addition to the surface energy and water balance. At the same time, the estimation of the seasonal snow cycle at the kilometer scale is a major hydro-meteorological challenge and thus represents a significant source of uncertainty in land surface schemes. To constrain this uncertainty we are developing a modular ensemble-based subgrid snow data assimilation framework (ESSDA).

The framework makes use of the EnKF to assimilate MODIS retrievals of surface albedo and snow cover fraction into the subgrid snow distribution submodel (SSNOWD; Liston 2004). Our problem is particularly challenging from the point of view of the EnKF as both albedo and snow cover fraction are doubly bounded in physical space. In an attempt to alleviate this issue we make use of the technique of Gaussian anamorphosis, as in e.g. Simon and Bertino (2012), to solve this bounded joint state and parameter estimation problem.

The potential of ESSDA is demonstrated both using synthetic (twin) and real experiments. The latter are carried out for the Bayelva catchment near Ny Ã…lelsund (Svalbard, Norway) where independent and accurate ground-based observations of snow cover and snow depth distributions are available. Preliminary validation of the real experiment is promising and indicative of the framework providing robust estimates of the evolution of subgrid snow depth distributions under various conditions. ESSDA is being developed to aid in satellite-based permafrost modeling efforts (see Westermann et al. 2015) and will contribute directly to developing the global permafrost extent product within the recently launched ESA GlobPermafrost project. Being modular, and given the improved snow treatment, the ESSDA approach may also prove valuable for atmospheric reanalysis and forecast initialization efforts.

File available here