This is a big project, but one that really needs to be addressed. Currently, in a run with CISM-EVOLVE, the atmosphere never sees the changing topography of the ice sheet. This makes a difference in the precipitation and surface mass balance for growing and shrinking ice sheets. To help the atmosphere see the ice sheet, we currently run a series of scripts known as the "dynamic_atm_topog" to merge data from the ice sheet with the atmospheric topography file. The source for these scripts can be found here: https://svn-ccsm-models.cgd.ucar.edu/tools/dynamic_cam_topography/trunk
These scripts work only off-line, so you either add 30-45 minutes of run time for each iteration of the script (each time the topography is updated) or run it in parallel and the topography update is offset by the number of years since it last ran. These scripts have cause some issues with the JG-BG and ISMIP runs because they failed silently or were not included in the cylc workflow as needed.
The ideal solution would be to send topography from the ice sheet to CAM through the coupler and update the topography within CAM every coupling timestep. This is our eventual goal.
This is a big project, but one that really needs to be addressed. Currently, in a run with CISM-EVOLVE, the atmosphere never sees the changing topography of the ice sheet. This makes a difference in the precipitation and surface mass balance for growing and shrinking ice sheets. To help the atmosphere see the ice sheet, we currently run a series of scripts known as the "dynamic_atm_topog" to merge data from the ice sheet with the atmospheric topography file. The source for these scripts can be found here: https://svn-ccsm-models.cgd.ucar.edu/tools/dynamic_cam_topography/trunk
These scripts work only off-line, so you either add 30-45 minutes of run time for each iteration of the script (each time the topography is updated) or run it in parallel and the topography update is offset by the number of years since it last ran. These scripts have cause some issues with the JG-BG and ISMIP runs because they failed silently or were not included in the cylc workflow as needed.
The ideal solution would be to send topography from the ice sheet to CAM through the coupler and update the topography within CAM every coupling timestep. This is our eventual goal.