Towards the coupling of a glacier and a hydrological model (OGGM and OS LISFLOOD) for improved loading estimations

Modelling crustal loading deformations is crucial for various geodetic applications, including the realization of precise and stable terrestrial reference frames. Commonly, hydrological, atmospheric and oceanic models are used to predict surface deformations. But also cryospheric deformations are not negligible considering the accelerating melting of glaciers, and glacier models exist to estimate cryospheric mass variations. However, it is not appropriate to just complement the existing hydrological model with the glacier model estimates as part of the glacier induced deformations are already taken into account by the hydrological model via simplified snow routines, which would lead to double-counting of masses.

A consistent way to consider glacier mass variations in deformation studies would be to couple a hydrological model with a glacier model. While on a basin scale this has been done before, large-scale or even global coupling approaches are still rare partly due to the heterogeneous glacier behavior and relatively small extent of glaciers (often smaller than the grid cell size of the global model). The Open Global Glacier Model (OGGM) is designed for global glacier modelling, and thus, a suitable candidate for a global coupling. Here we present first steps towards coupling OGGM with OS LISFLOOD, an open-source global hydrological model running with a global 0.05° spatial resolution previously used for geodetic applications.

As a first case study, we chose the Fraser river basin in North America. We initially conduct model runs separately with OGGM for selected glaciers contained in the study area, and with OS LISFLOOD to obtain mass storage estimates particularly for the snow compartment. Comparison of both model results gives an impression of the potential double-counting of mass if both models were applied separately, and reveals challenges in a possible coupling workflow. For example, OGGM output is stored per glacier, and thus has to be summed per grid cell in order to pipeline it to grid-based OS LISFLOOD. Furthermore, OS LISFLOOD would have to be adjusted to take input from OGGM in glaciated regions with varying extent.