The interaction between ice sheets and the underlying bedrock is one of the most important factors driving the dynamics of ice. Quantifying the extent of a subglacial hydrology system can help to obtain more realistic sliding laws, and therefore help to improve the reliability of ice sheet simulations. Recently, several modeling choices have been proposed for subglacial hydrology, covering both a distributed drainage system as well as concentrated channels networks.

In this work, we consider the case of a distributed drainage system, and we explore some of the models already proposed in literature (Schoof et al. 2012, Hewitt 2013, Bueler 2015). Such models are usually formulated in terms of two unknowns, namely the water pressure and the thickness of the water layer, and typically involve on the order of 10 scalar parameters; some are related to the physical processes (e.g., the hydraulic transmissivity), some to an average description of the bedrock topography (e.g., the typical bed bumps dimensions), and some are model fitting parameters (e.g., the power exponent of the non-linear Darcy law in the mass conservation equation). We are particularly interested in studying the sensitivity of the model to some of these parameters, with the goal to identify which parameters most impact the ice sliding law, and therefore need to be accurately estimated from available measurements (via data assimilation techniques).

In our study we focus only on the Greenland ice sheet, and we consider meshes with a medium-high spatial resolution (in the 1km-5km range).