Main menu

Navigation

ncl scripts to create SST/ICE data files to drive an F compset using B compset history files

12 posts / 0 new
Last post
hannay
ncl scripts to create SST/ICE data files to drive an F compset using B compset history files

Attached are two scripts that allows to create SST/ice files to process B compset history output and create SST/ICE data files to drive an F compset.
There are 2 versions of the script depending if you want to use the icesst tool or not. Both scripts use single variable timeseries of SST and aice (SH and NH)

 

1. create_SST_files_from_B_case_without_bcgen.ncl
Regrid the SST and sea-ice data from the B compset and create files in an appropriate format to run with CAM

2. create_SST_files_from_B_case_with_bcgen.ncl
Regrid the SST and sea-ice data from the B compset and create files in an appropriate format for processing with bcgen (you need to run bcgen afterwards to have a SST/ICE dataset in an appropriate format to run with CAM

 

 

rambhari0123@...

Dear Hannay,


I read you ncl code to create the SST and SIC fields for F compset. at the end of script you used the SST-ICE consistency steps as described in Hurrell et at 2008. I read the hurrell's paper and try to understand each step. I feeel that Hurrell doesn't used the step number 3 (as Written in NCL Script) but you had mentioned here. I am trying to unserstand how step 3 is different from 5. I feel that if i use only step number 5, step 3 will be covered under it.

I tried to avoid step 3 but bcgen throughs error

 latitude =           172  longitude =           110
 does not converge

I request you to please explain the difference between the step number 3 and 5. why is it nacessary to use both ??

Thanking you in anticipation.

Ram

hannay

The reason I am doing these 5 steps is that if I use mean SST and sea-ice without any correction, there will be inconstitency along the ice edge. 

My script doesn't follow the procedure from Hurrell but it uses the relationship in Hurrell. The reason I decided to use the relationship both way if I was concerned there might be instances that step 3 doesn't necesssary capture every inconsistency. 

This is not an official CESM script. This is my own script that I wrote for my own purpose. I share it with others for their convenience. You are welcome to use step 5 instead of step 3 and then look at the fields to see if there is any inconsistency. Or to modify the script in a way that you want. 

 

 

huili7@...

Dear Hannay,

Thanks for sharing the ncl scripts. I found another procedure of creating the SST/ICE data file from the CESM user guide FAQ section. It looks like that method doesn't require regridding to a the 1x1 grid, but instead making changes on the original gx16 grid.  I'm wondering if these two methods would yield the same forcing file for CAM. If not, which one would be the prefered method? 

 

In addition, I found that the SST filed in the B compset output has missing values over land, while the other SST/ICE forcing files from the input repository (e.g. AMIP) do not treat the land as missing. This difference will cause the model to crush. The cesm user guid suggests changing the missing values over land to -1.8, and I'm wondering where does the value of -1.8 come from?

 

Thank you very much in advance. 

Hui

 

yh3019@...

Hi Hui,

May I ask how did you create your SST data in the last? I am troubled by this problem too. When I tried to follow CESM guide, then I found that I need to do many things manually, when I read Hanny's code I found I do not understand why she split the globe into NH and SH and how she defined mid of days. Can you give me some help?

Thank you.

Yu

Yu Huang

yh3019@...
Hi Hanny, I am wondering why you would like to use ice files from North Hemisphere and South Hemisphere instead of just using one ice file for the whole globe? Another thing is that we need to change the SST and ice's time to the mid of each month. What should we do for February with 28 days then? Thank you!Yu

Yu Huang

hannay

The reason we have ice fiels from NH and SH instead of a single global ice files is that we save a lot of space. For most of the the globe, sea-ice is zero. We don't want files with a bunch of zeros that use a lot of space. 

Our  SST files are monthly. So February is monthly. 

yh3019@...

Hi Hanny,

I understand why you split sea ice files now. I modified the code and dealt with sea ice with the same strategy as what you did to SST because I used a global sea ice time series.

Thank you so much for your kind help! 

Yu

Yu Huang

yh3019@...

Hi Hanny, 

Sorry to bother you again. I found that you ncl file generally regrids all data to a 1x1 format. Does CESM automatically change this 1x1 grid of SST to some other grid form I want when I run this F case? For example, if I want to run f19_g16 F_ case, is that ok for me to use your ncl file to creat SST? In this case, my SST/ICE domain should have 1x1 grid but the other domains/grids will still be f19_g16 right?

Thanks a lot!

Yu Huang

hannay

I think that if you want to run a F case with with f19_g16 resolution, you need to have your data on that grid.

Not that we usually run the F case with everybody on the same grid, in your case it would be a f19_f19 resolution.  

yh3019@...

Hi Hanny,

I agree that I should run the F case on the f19_f19 grid but I still have 2 further questions about this.

The first question is I want to use a B f19_g16 case SST/ICE to drive a f19_f19 F case, then should I re-grid the final SST/ICE to f19_f19 or 1x1 grid?

I have a problem with running bcgen package so I checked that NCL script without bcgen you shared with us. The second question is in your NCL script to process the SST/ICE data, how should I modify this /ESMF_regrid_with_weights(SEAICE_tmp,"map_gx1v6_TO_1x1_blin.130423.nc" command to regrid the data to f19xf19 instead of regird to 1x1?

Many thanks to you!

Yu

Yu Huang

yh3019@...

Also I read your create_SST_files_from_B_case_with_bcgen.ncl and create_SST_files_from_B_case_without_bcgen.ncl but I do not see many differences. Is there any reason for the cesm user guide to suggest us to use the bcgen tool?

Yu Huang

Log in or register to post comments

Who's new

  • stans
  • ahadibfar@...
  • jskang@...
  • divanova@...
  • mrostami@...