I am working on running sea ice perturbation experiments in ACCESS-ESM1.5 using an albedo-based approach in CICE, but I have not been able to see any response in the model output despite modifying the relevant albedo parameters (albicev, albicei, albsnowv, albsnowi) in the cice_in.nml namelist.
I have also tested different albedo_type options available in CICE (e.g. constant, ccsm3, and default) and experimented with modifying the r_ice, r_pnd, r_snw parameters while using the shortwave=dEdd scheme. All experiments complete successfully, however, when compared to the control runs, I am not seeing any differences in the output. I have looked at a range of output variables including SIC, ice thickness, SST but they remain identical to the control run.
As an additional test, I modified the UM namelist (namelists file) by changing l_ssice_albedo from .true. to .false. and this immediately produced differences compared to the control.
I think that the CICE namelist changes are being read in successfully, but the modified albedo values are being overwritten elsewhere in the model, potentially by the UM or coupler setup. I wanted to ask whether there are additional files/namelist settings that also need to be modified so that my albedo values from CICE can be used to implement albedo changes in the model.
Hi @saanvi thanks for getting in touch, the albedo set up in the ACCESS-ESM model is controlled as you have suspected at the end of your message by the UM, which is where the albedos are being controlled in ACCESS-ESM1.5 you cannot control them from the cice_in.nml unfortunately.
If you explain a little more what experiment you are trying to do, I might be able to tell you if it is viable and if you can do it by perturbing the UM albedo parameters.
It’s helpful to know that albedos are being controlled by the UM entirely. Ideally, I would like to constrain/force Antarctic sea ice state towards a targeted extent, using sea ice albedo perturbation as the mechanism to induce those changes. I know that there are other methods like ghost flux and sea ice nudging that can also modify sea ice, but I was holding to explore whether this could be achieved directly through the model albedo configuration as it seemed the simplest.
Also, yes I saw zero response comparing it to the restarts.
Hi @aekiss I have chatted to @saanvi by email and pointed her to the UM settings for the albedo that we used in ACCESS-ESM1.5 which are the same as in the ACCESS 1.3 model, we didn’t update the settings though monitored the control run to see if we needed to.
The reason the CICE parameters are not used is that the albedo calculation is done in the UM along with the UM shortwave radiation code, CICE has included the NCAR shortwave radiation code within its package so it can operate stand alone (or within ACCESS-OM2 say) but for the coupled model you need to use the UM shortwave code.
The settings for ACCESS1-3 are discussed in on this paper by Uotila et al (2013) on page 4