Ice flow modeling experiment

[hoffmaao]

Here I’ve outline an ice flow modeling experiment that I think can naturally build on the tools we’ve already built to test how the resolution of topography near the grounding zone affect ice shelf response to perturbations in basal melt.

first we simulate an ice sheet that has an anisotripically refined mesh 50km inland of a grounding zone with an average element size of 25m in this near grounded zone area and 500m further away. This will define the simulation resolution for two different experiments:

One that uses synthetically generated roughness built from variational autoencoders
and other simpler stochastic methods that capture form drag that represents our best estimate of simulated ice-bed interactions.

The other would use a smoothed topography and surface velocities derived from the control experiment described above to estimate control parameters that parameterize sliding over “unresolved roughness” at the ice bed interface.

The simulations would then be forced with identical climate variability (synthetically generated white noise variability) to understand how assumptions for unresolved bed properties that are prescribed during model initialization affect the response of the ice sheet to climate perturbations.