Development of a glacier mass balance model to nest in a RCM

Glaciers play a role in the regional climate system due to their potential ability to reduce the energy balance by reflecting a large part of the downward solar radiation. They are also an important source of freshwater reserves, help to maintain the river flow level and can contribute to the sea-level rise. Changes in surface properties or glacier coverage might also modify surface energy balance. In this work, I developed a methodology to include all individual glaciers of western and Arctic Canada into the Canadian Regional Climate Model 5 (CRCM5) in order to better represent atmospheric feedbacks and glacier hydrological processes.

A glacier mass balance model, initially developed to be driven by observed in-situ meteorological data, has thus been adapted and optimized to be run on pre-processed grids of different spatial resolutions. The resolution has to be fine enough to consider glacier surface heterogeneity (fresh snow/firn/ice, roughness, altitude, slope, aspect, shading…). Techniques to downscale CRCM-meteorological fields to each high-resolution grid cell have been developed. The simulations were performed on a 90 m resolution grid, consistent with that of the Canadian Digital Elevation Data, as well as on aggregated grid cells of 200, 300, and 500 m resolution. Other experiments were conducted using a mosaic approach, in which grid cells were resampled into categories of similar properties.

Validation of the model and methodology is done by simulating glacier mass balance at three study sites in the Rocky Mountains where long series of data exist (Haig, Peyto and Place glaciers) and two in Alaska (Wolverine and Gulkana glaciers). The atmospheric data to drive and test the model are provided by CRCM5 fields. Interactions of glaciers in the climate system through atmospheric feedback mechanisms are not yet considered. A sensitivity test is performed to evaluate the effects of the resolution on the total mass balance and to find a suitable choice of parameters (albedo function, windspeed correction, a.s.o.).

In the figure below, a simulation is shown using the mosaic approach with 16 categories to simulate the total mass balance of the Haig glacier. It appears that the mosaic approach produces similar results compared to a 90 m grid resolution simulation (378 grid cells): the error is of 1%. However, CRCM5 precipitations had to be doubled in order to include orographic effects. This improved drastically the simulated snowpack accumulation.

Figure: Comparison of yearly simulated and observed snowpack accumulation and ice melt over a 12-year period for the Haig glacier. The simulation was performed at a 90 m and 500 m resolution, as well as using the mosaic approach (16 categories based on elevation and aspect).

Comments

It would be nice to see the comparison when driving the glacier model with observed atmospheric fields. That would indicate us where the errors come from.

Cheers

Add new comment