Scheduled Downtime
On Tuesday 24 October 2023 @ 5pm MT the forums will be in read only mode in preparation for the downtime. On Wednesday 25 October 2023 @ 5am MT, this website will be down for maintenance and expected to return online later in the morning.
Normal Operations
The forums are back online with normal operations. If you notice any issues or errors related to the forums, please reach out to help@ucar.edu

How to implement the same climate forcing in different compsets?

S

Shuai Li

Shuai Li
New Member
What version of the code are you using?

We use CESM 2.1.4, running on Derecho.





Describe every step you took leading up to the problem:

We want to explore the effects of land cover change-atmosphere interactions on ecosystem productivity/photosynthesis. We plan to run 4 simulations, the F2000climo component with 1984 land cover data, the F2000climo component with 2014 land cover data, the I2000Clm50Sp component with 1984 land cover data, and the I2000Clm50Sp component with 2014 land cover data. Apart from the land cover change part, we did not make any other changes.



Describe your problem or question:

We found that the initial data/climate forcings of the F2000climo and I2000Clm50Sp compsets are inconsistent. We would like to run the 4 simulations mentioned above with the same initial data/climate forcings. How do we implement this? Do we need a branch run? Which compsets are best to run it with? How long do we need to run it to spin up?
 
Sort by date Sort by votes
slevis

slevis

Moderator
Staff member
- You can specify CLM initial conditions by setting finidat in each case's user_nl_clm. If you have finidat file(s) that you consider consistent, then you can point to them this way. Branch simulations may be the answer for the 2014 cases but, ultimately, the methods that you use are up to you and your collaborators.
- The land-only cases use DATM "climate forcings" while the F cases have an active atmosphere. Please be more explicit regarding the inconsistency that you see.
- Spin up is another methodological choice that's up to you and your collaborators. Sometimes it helps to look at the literature for methods used in past work. Trial and error is also common, where you may start a spin-up thinking that it will take X years, and you end up finding that it takes Y years.
 
Vote Upvote 0 Downvote
O

oleson

Keith Oleson
CSEG and Liaisons
Staff member
You could make the initial data consistent by specifying the same finidat in your user_nl_clm.
Climate forcing will obviously be different because one set is generated by CAM and the other by a type of reanalysis, e.g., GSWP3. You could generate coupler history forcing files from the F-case and use that to force the I-case. Both compsets you've chosen have satellite phenology and should take less than 50 years to spinup from a land perspective. See:

 
Vote Upvote 0 Downvote
S

Shuai Li

Shuai Li
New Member
slevis said:
- You can specify CLM initial conditions by setting finidat in each case's user_nl_clm. If you have finidat file(s) that you consider consistent, then you can point to them this way. Branch simulations may be the answer for the 2014 cases but, ultimately, the methods that you use are up to you and your collaborators.
- The land-only cases use DATM "climate forcings" while the F cases have an active atmosphere. Please be more explicit regarding the inconsistency that you see.
- Spin up is another methodological choice that's up to you and your collaborators. Sometimes it helps to look at the literature for methods used in past work. Trial and error is also common, where you may start a spin-up thinking that it will take X years, and you end up finding that it takes Y years.
Click to expand...
Thank you very much for your helpful response and for clarifying these points. I apologize for not explaining my question clearly earlier. My main goal is to make the simulations from the F compset and the I compset comparable by ensuring that they share consistent external climate boundary conditions/non-land forcing.
 
Vote Upvote 0 Downvote
S

Shuai Li

Shuai Li
New Member
oleson said:
I see that @slevis just posted on this as well...hopefully my contribution will be useful.
Click to expand...
Hi Keith,
Thanks a lot for your detailed explanation, your suggestions and the reference are very helpful.
Based on your advice, I’ve been considering two possible approaches and would greatly appreciate your opinion on which one is more reasonable:

Mode 1:
1. Run the default F2000climo until key state variables (e.g., FSH) reach equilibrium to obtain a reference simulation (F_ref), enabling 3-hourly coupler history outputs.
2. From F_ref, conduct two hybrid simulations using modified fsurdat files for 1984 and 2014 land cover, respectively, running each until equilibrium.
3. Use the 3-hourly coupler history outputs from F_ref (selecting a stabilized model year) as the DATM forcing for two I2000Clm50Sp runs (1984 and 2014 land cover), which are then run to equilibrium for analysis.

Mode 2:
1. Run F2000climo with 1984 land cover until equilibrium and save 3-hourly coupler history outputs.
2. Use those outputs to drive an I2000Clm50Sp run with the same 1984 land cover until equilibrium.
3. Repeat the same procedure for the 2014 land cover case.

In your view, which mode provides a more consistent or scientifically sound setup for ensuring that F and I compsets share consistent external climate boundary conditions/non-land forcing and the simulations are comparable?
 
Vote Upvote 0 Downvote
You must log in or register to reply here.
Top