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

Questions about B_RCP8.5_CAM5_CN run, should I spin-up?

H

happy

happy
New Member
Dear CESM group,
Recently I want to conduct a simulation for 2005-2020 with B_RCP8.5_CAM5_CN. I wonder if I should spin-up. Can I just use "startup" option with B_RCP8.5_CAM5_CN to simulate 2005-2020? Or I should, for example, run B_2000_CAM5 for som years to reach equilibrium and then use the restart file of B_2000_CAM5 run to run the compset B_RCP8.5_CAM5_CN from 2005-2020? Thanks for helping!

Best!
Zhu
 
M

mlevy

Michael Levy
CSEG and Liaisons
Staff member
When you create a case with the B_RCP8.5_CAM5_CN compset, it is already configured with reasonable initial conditions so you don't need to do anything special - just build and run for the 2005-2020 period. (Well, it looks like the compset actually starts at 2006.)

If you are curious about the process, NCAR scientists ran a preindustrial control simulation for a few thousand years to get close to equilibrium, and then ran a B20TRC5CN compset run that had transient forcing through 2005. The RCP compsets all start where the 20th-century transient run ended.

If it is important to do your run from 2005 - 2020 instead of 2006 - 2020, let me know and I will find someone who can provide instructions for running the last year of the 20th-century transient case and then continuing with the RCP forcing afterwards. In that case, it would be helpful to know exactly what version of CESM you are running.
 
M

mlevy

Michael Levy
CSEG and Liaisons
Staff member
Actually, I misspoke... I was confused because

Code: 
$ ./xmlquery RUN_STARTDATE,RUN_REFDATE
env_run.xml: RUN_STARTDATE = 2005-01-01
env_run.xml: RUN_REFDATE = 2006-01-01

So it looks like the case I generated (with CESM 1.2.2.1) is starting from January 1, 2005 but using the 20th century transient run's state from January 1, 2006 as initial conditions. Interesting.
 
H

happy

happy
New Member
Levy, thanks very much, your replies are very helpful. I was also confused, I generated a case with CESM1.2.0, but the RUN_REFDATE is 0001-01-01 by default. Should I change it to 2006-01-01 or 2005-01-01 if I want to run from 2005-2020? And if I want to run from 2006-2020, should I change RUN_STARTDATE and RUN_REFDATE to 2006-01-01?

./xmlquery RUN_STARTDATE,RUN_REFDATE
env_run.xml: RUN_STARTDATE = 2005-01-01
env_run.xml: RUN_REFDATE = 0001-01-01

I also have another question. If I want to run from 1994-2018, But B20TRC5CN can only simulate from 1850-2000, how can I simulate from 2000-2005? Any advice is helpful. Thanks!

Best!
Zhu
 
M

mlevy

Michael Levy
CSEG and Liaisons
Staff member
First, if you really want to run CESM 1.2 then I suggest moving to the latest release (which I believe is 1.2.2.1).

Second, to run from 1994 - 2018 you would just break it up into two runs. Run from 1994 - 2005 in a B20TRC5CN and then use a hybrid run with the RCP compset and REFCASE pointing to your first run for the future scenario portion.

I am going to move this to the Community Projects forum, because the big question remaining is "how do you get 1994 restart files to start your first run" and I don't know the answer to that.
 
H

happy

happy
New Member
Levy, thanks for your reply! It seems like we choose the different -res option. I tried f19_g16 with CESM1.2.0 and CESM1.2.2, both of them showed:

./xmlquery RUN_STARTDATE,RUN_REFDATE
env_run.xml: RUN_STARTDATE = 2005-01-01
env_run.xml: RUN_REFDATE = 0001-01-01

And when I change f19_g16 to f09_g16, both CESM1.2.0 and CESM1.2.2 showed:

./xmlquery RUN_STARTDATE,RUN_REFDATE
env_run.xml: RUN_STARTDATE = 2005-01-01
env_run.xml: RUN_REFDATE = 2006-01-01

So, did this mean that it runs with different initial conditions when we choose different resolution（f19_g16 and f09_g16 for example)? Thanks!
 
M

mlevy

Michael Levy
CSEG and Liaisons
Staff member
Yes, the f19 is a finite-volume 2 degree resolution while f09 is a 1 degree resolution. I don't believe the scientific validity of these RCP runs has been vetted at 2 degrees, so I definitely recommend running at 1 degree if you have the compute resources to allow it.
 
H

happy

happy
New Member
mlevy said:
Actually, I misspoke... I was confused because

Code: 
$ ./xmlquery RUN_STARTDATE,RUN_REFDATE
env_run.xml: RUN_STARTDATE = 2005-01-01
env_run.xml: RUN_REFDATE = 2006-01-01

So it looks like the case I generated (with CESM 1.2.2.1) is starting from January 1, 2005 but using the 20th century transient run's state from January 1, 2006 as initial conditions. Interesting.
Click to expand...
Levy, sorry to bother you again. I'm still confused about initial conditions. You said it uses January 1, 2006 as initial conditions by default when I creat a case with B_RCP8.5_CAM5_CN compset, but how did it get these initial data? I created a case with B_RCP8.5_CAM5_CN and run it successfully, but the input data which listed in $CASEROOT/Buildconf/*.input_data_list does not involve data related to January 1, 2006. So where did it get initial data at January 1, 2006? Looking forward to your reply. Thank you!

Best!
Zhu
 
H

happy

happy
New Member
For example, in my $RUNDIR, part of initial files in atm_in and pop2_in are set as follows:

ncdata =/inputdata/atm/cam/inic/fv/cami-mam3_0000-01-01_0.9x1.25_L30_c100618.nc
init_ts_file = /inputdata/ocn/pop/gx1v6/ic/ts_PHC2_jan_ic_gx1v6_20090205.ieeer8

So I think the initial atmospheric state (ncdata) and initial ocean temperature and salinity (init_ta_file) are from these two files.
 
M

mlevy

Michael Levy
CSEG and Liaisons
Staff member
Your case is a hybrid startup, which means it can use a combination of restart files and initial condition files to initialize the model. If you look at the other init_ts settings in pop2_in, you'll see init_ts_option = 'ccsm_hybrid', which means T & S will be read from the file rpointer.ocn.restart, which should have been written on the 2006-01-01 reference date from the reference case. I'm less familiar with CAM, but believe you can find something similar with ncdata (or maybe CAM is initializing from a file unrelated to the reference case?)
 
H

happy

happy
New Member
Levy, thanks for your reply. I think my case is a startup case. In pop2_in, init_ts_option is set to "ccsm_startup" (for both CESM1.2.0 and CESM1.2.2). If I run hybrid case, I need restart files of 2006-01-01. Do you know where I can download these restarts for CESM_1.2.0.
 
H

happy

happy
New Member
Levy, Thanks for your reply! I had emiled cisl-aws-lens@ucar.edu before, and the CESM Data Manager shared the restarts of CESM1_LENS to me. The restarts are named b.e11.B20TRC5CNBDRD.f09_g16.*. It seems the version of restart fields of CESM1_LENS are closer to CESM1.2.0. So I am confused which one should I use. Is there any sugestions? Thanks!
 
O

ohmpawat

ohmpawat chen
Member
happy said:
Levy, thanks very much, your replies are very helpful. I was also confused, I generated a case with CESM1.2.0, but the RUN_REFDATE is 0001-01-01 by default. Should I change it to 2006-01-01 or 2005-01-01 if I want to run from 2005-2020? And if I want to run from 2006-2020, should I change RUN_STARTDATE and RUN_REFDATE to 2006-01-01?

./xmlquery RUN_STARTDATE,RUN_REFDATE
env_run.xml: RUN_STARTDATE = 2005-01-01
env_run.xml: RUN_REFDATE = 0001-01-01

I also have another question. If I want to run from 1994-2018, But B20TRC5CN can only simulate from 1850-2000, how can I simulate from 2000-2005? Any advice is helpful. Thanks!

Best!
Zhu
Click to expand...
Hello, I am also curious about your first question. If I want to run from 2005-2020, should I set the RUN_STARTDATE=2005-01-01 OR RUN_STARTDATE=2006-01-01？ In other words, if I set the RUN_STARTDATE=2005-01-01, the model's simulation result that I can use is 2005-2020 OR 2006-2020? Thanks for your help!
 
H

happy

happy
New Member
ohmpawat said:
Hello, I am also curious about your first question. If I want to run from 2005-2020, should I set the RUN_STARTDATE=2005-01-01 OR RUN_STARTDATE=2006-01-01？ In other words, if I set the RUN_STARTDATE=2005-01-01, the model's simulation result that I can use is 2005-2020 OR 2006-2020? Thanks for your help!
Click to expand...
Hi， I'm using the hybrid option and running since 2006-01-01 at last. In my opinion, if you want to run from 2005, you need to set RUN_STARTDATE=2005-01-01, and you also need restart files of 2005-01-01. The restart files of 2006-01-01 is available at https://svn-ccsm-inputdata.cgd.ucar..._init/b.e10.B20TRC5CN.f09_g16.001/2006-01-01/, but I don't know where can get restart files of 2005-01-01. Maybe you can look at CESM Large Ensemble Data Sets, this URL provids many datas of CESM_LENS, or you can email cisl-aws-lens@ucar.edu to ask if you can get the restart files of CESM_LENS of 2005-01-01.
 
S

Sreerag

Sreerag
New Member
Hey, I am interested to know about the run type as you discussed. If I am selecting the compset B_RCP8.5_CAM5_CN and giving the GET Reference case as false and Run type as Startup. Where will the model start? Is the model starts from a Cold Start?
 
You must log in or register to reply here.
Top