Outflow of the Antarctic ice sheet: is it consistente across models for similar geometries?

The alarming acceleration of global sea level rise rate state the importance of resolving the future climate prediction uncertainties. The Antarctic Ice Sheet is one of the main potential contributing factor to future sea level rise, which contribution is still the most uncertain. To account its impact we look a the ice outflow of the AIS in a key region: the Amundsen Sea Sector.

To do so, we test Schoof 2007 [C. Schoof. Ice sheet grounding line dynamics: Steady states, stability, and hysteresis. Journal of Geophysical Research: Earth Surface, 112(F3), Sept. 2007b. doi: 10.1029/2006JF000664. URL https://doi.org/10.1029/2006JF000664.] two dimensional equation defining ice flux at the grounding line (line that separates the part of the ice sheet that is resting on the bedrock and the floating part) as a function of the sliding law, the ice rheology, and the ice thickness. Assuming that those parameters are the same across three dimensional ISMP6 model ensemble, we compare the ice flux at the grounding line.

Therefore, we find similar grounding line geometries, using RMSE, based on defined target: the observed grounding line, a moderately retreated grounding line, and a largely retreated one (both model based). We find that, for three dimensional models, the mean ice flux at the grounding line increase with its retreat. Data assimilation models compute ice flux at the grounding line with a better accuracy than Spin up models for the current grounding line geometry. However, for all the chosen targets the large ice flux spread shows the difficulty with which models compute ice flux at the grounding line, even for models with the same sliding law and sliding coefficient. Leading us to discuss the choice of comparison metric and target, and looking if other regions exhibit the same results.

In this work we define three targets with which we will compare the other models with:

• BedMachine: target based on the observation data, this we allow us to see if the models are able to accuratly reproduce the observed geometry of the grounding line.
• LSCE_GRISLI2 expAE04 2150: target for the limited grounding line retreat.
• UNN_Ua expAE04 2250: target for the large grounding line retreat.

The model-based-target are chosen based on Seroussi 2024 [H. Seroussi, T. Pelle, W. H. Lipscomb, A. Abe-Ouchi, T. Albrecht, J.Alvarez-Solas, X. Asay-Davis, et al. Evolution of the antarctic ice sheet over the next three centuries from an ismip6 model ensemble. Earth’s Future, 12(9):e2024EF004561, Sept. 2024. doi: 10.1029/2024EF004561.], and will allow us to see if the ice flux at the grounding line increases with the retreat of the grouding. As the grounding line retreats, the ice thickness at the GL increases, which leads to an increased flux based (accordingly to Schoof 2007 equation for the ice flux at the grounding line).

Contributors

Marine de Coatpont

Supervisor: Gaël Durand and Cyrille Mosbeux

Material

1. Installation of environment and data In the directory Installation, you will find all the instruction needed to replicate the environment in which I worked. The file install_ISMP6_dataset.md will guide you through the setup installation.

Some data are to be downloaded from and place in the Result directory.

2. Where you want your work Modify config.py first two variables to stipulate where your data are stored with DATA_PATH and where your scripts are (where this directory is) with SAVE_PATH

3. Reproduce all the work of my intership The script in this directory are all the scripts use in my internship, you can either run them individualy in the following order:

• compute_mask.py
• compute_RMSE_global.py
• compute_gl_flux.py
• plot_comp_flux.py
• plot_diag.py
• RMSE_flux.py

Or use compute_all_work.py script. This script will automatize running the other script in your terminal.

4. Use the scripts for your own work You will find in Specifi_script the same script but for a more personal usage, you will be able to chose the target you want using input from the terminal.

Note

It is still important to follow the previous order to avoid mistakes.

For more information on each script please read README_SCRIPT.md.

5. Result and output All the netCDF files, figures and CSV files will be saved in the Result directory folowing this structure:

/all script
	|____Function/
	|	|_____ inti.py
	|		    function.py
	|
	|____ Result/
		|
		|_____ Grounded_mask/
		|	   |_____ {simu}/
		|
		|_____ RMSE/
		|	   |_____ {target_simu}_{comparison_simu}/
		|	   |	      |_____ min_{target}_{exp}.csv
		|	   |          |_____ RMSE_{target}_{year}.png
		|	   |_____ Summary_{target}_{year}.csv
		|
		|_____ Ice_flux/
        |      |_____ {simu}_{exp}_{year_simu}_diag.png
		|	   |_____ {simu}
		|		      |______ ligroundf_{simu}_{exp}_{year}.nc
		|
		|_____ Plot/

Licence

Find in license.txt for more information