Branching off a run with modified land restart files

[kezhoulumelody]

Hello,

I want to do a set of AGCM abrupt4xCO2+uniform 4K SST warming large ensemble simulation branching off from a control run. The goal of this experiment is to look at the instant atmosphere circulation responses to both SST warming and directive radiative forcing, so changes in every hours (days) after abrupt forcing matter. I want the land and ocean warm all together at the beginning of each branch run, but the land only starts to warm after about 20 days after abrupt4xCO2.

Since land and atmosphere are tightly coupled together, I can't directly modify the land temperature. I have tried to modify all the temperature related variables in the land restart files and apparently this method does not work. Are there any other ways to work around this problem?

Thank you so much!

Melody

 

[erik]

Hi Melody

Interesting problem. I'm thinking you should do some spinup of the land not unlike other spinup exercises that we do for other cases. So do you want the land to be totally spunup for 4XCO2 conditions when you startup these cases? If so you should probably save off something like 20-30 years of a simulation where CAM and CTSM are coupled together, and use that to force the land to be spunup. Depending on if you are using BGC or SP you would then run for 50 to several hundred years.

Here's some examples of spinning up the land model from the CLM5 user's guide...

1.5.5. Spinup of CLM5.0-BGC-Crop — ctsm release-clm5.0 documentation

Since, you want to see the instantaneous surface responses, I think getting the land surface spunup based on longer simulations would make sense. This would for example get the soil Temperatures at depth be spunup for such conditions. It would still respond to the sudden change at the surface, but the deeper soil Temperatures wouldn't just be warming. With the abrupt change the deeper soil Temperatures are just going to be warming so by nature can't be at an equilibrium. Which is why it might make sense to spinup the land surface for 4xCO2 even though your experiments are meant to examine the abrupt changes.

Yeah, I don't think you could just arbitrarily increase the Temperatures uniformly on the restart files and have that work. I would expect it to be out of balance and not be happy. It's possible you could let it be out of balance for a while though and maybe it would reestablish a balance.

@oleson do you have more suggestions about this?

 

[kezhoulumelody]

Hi Melody

Interesting problem. I'm thinking you should do some spinup of the land not unlike other spinup exercises that we do for other cases. So do you want the land to be totally spunup for 4XCO2 conditions when you startup these cases? If so you should probably save off something like 20-30 years of a simulation where CAM and CTSM are coupled together, and use that to force the land to be spunup. Depending on if you are using BGC or SP you would then run for 50 to several hundred years.

Here's some examples of spinning up the land model from the CLM5 user's guide...

1.5.5. Spinup of CLM5.0-BGC-Crop — ctsm release-clm5.0 documentation

Since, you want to see the instantaneous surface responses, I think getting the land surface spunup based on longer simulations would make sense. This would for example get the soil Temperatures at depth be spunup for such conditions. It would still respond to the sudden change at the surface, but the deeper soil Temperatures wouldn't just be warming. With the abrupt change the deeper soil Temperatures are just going to be warming so by nature can't be at an equilibrium. Which is why it might make sense to spinup the land surface for 4xCO2 even though your experiments are meant to examine the abrupt changes.

Yeah, I don't think you could just arbitrarily increase the Temperatures uniformly on the restart files and have that work. I would expect it to be out of balance and not be happy. It's possible you could let it be out of balance for a while though and maybe it would reestablish a balance.

@oleson do you have more suggestions about this?

Hi Erik,

Thank you so much for your reply and it's very helpful!

I want the land to be totally spunup at a warm state with the surface temperature increased to ~4K compared to the parent run. The parent run's compset is FHIST_BGC, so the same one as the CMIP6 AMIP setup. Spinning up the land under 4xCO2 only probably won't let the land TS reaches 4K warming. But I think I can spinup the land using a uniform SST warming scenarios where the land surface reaches ~4K at equilibrium.

I have another concern: if the land starts at a fairly warm condition but the atmosphere does not, is it going to mess up the energy balance of the atmosphere?

Thanks,

Melody