[Liwg-core] starting the new BG/JG on Monday

Katherine Thayer-Calder katec at ucar.edu
Mon Dec 10 08:25:24 MST 2018


Hi all,

>(1) what version of CESM have you added the revised P1 source mods and
CISM into?
>From the directory is looks like CESM2.1
>We have only test CESM2.0?

Yes, I am using the version of CESM that is being released today. I asked
about this at the LIWG telecon and it seemed ok to everybody. The only
differences between CESM2.0 and CESM2.1 are support for the CMIP
experiments (compsets and tuning parameters mostly). The co-chairs worked
hard to keep actual code changes very minimal between the releases. So, I
think it is ok to use this, even if it is not what we were testing.
However, it is not hard to go back to CESM2.0 either if that is preferred.

Also, we agreed to use the updated Land surface data set with the correct
ice shelves in the antarctic as well, right? I'll need a user_nl mod from
Bill to support that, I think.

Thanks,
Kate

On Mon, Dec 10, 2018 at 7:49 AM Sarah Bradley - CITG <
S.L.Bradley-1 at tudelft.nl> wrote:

> Hi Kate again
>
> I forgot that you also asked about two source mods:
>
>
> /gpfs/u/cesm-scripts/liwg/Coupled_BG_JG_Spinup_CESM2.1.0_Dec2018/SourceMods/src.drv
>
> I have not ran with these included: it was to correct for the blocky
> pattern in the temperature field. As this will not impact on the ice sheet
> evolution, I do not have an opinion on keeping them in or not.
>
> Sarah
>
> ec 2018, at 15:47, sarah bradley <s.l.bradley-1 at tudelft.nl> wrote:
>
> Hi Kate
>
> I have added the revised user_nl_cism in a folder in my home directory
> (the directory you send was read-only)
>
> /gpfs/u/home/sbradley/BG_JG/user_nl_cismDec2018
>
> To reiterate what Laura asked earlier
>
> (1) what version of CESM have you added the revised P1 source mods and
> CISM into?
> From the directory is looks like CESM2.1
> We have only test CESM2.0?
>
> Thanks
>
> Sarah
>
> On 10 Dec 2018, at 12:20, Laura Muntjewerf - CITG <L.Muntjewerf at tudelft.nl>
> wrote:
>
>
> Hi Kate, all,
>
> Kate, thanks for putting it together, especially all the model code and
> changes! I am happy that we can restart the spin-up soon.
> I had a look through the sandbox and there’s some things I notice:
>
> runscript_BG_1.sh:
> * For model version, we are using CESM2.0, right? The variable for the
> model version “CCSMRoot" is pointing to a path that is not there, as Marcus
> also noted. I see the folder that this variable is pointing to has code for
> CESM2.1, that does not seem right to me.
> * line 177, cat to user_nl_cpl. This is not pointing to the user_nl_cpl in
> the case directory, path is missing. "cat >> user_nl_cpl <<EOF" needs to
> be: "cat >> $D/$CaseName/user_nl_cpl <<EOF"
> * lines 165- 168 that involve changing the HIST_N and HIST_OPTION. For
> what I understand, HIST_* sets all the output frequency, not only the
> coupler output. In the user_nls we define for each component the history
> output frequency. I would be worried that specify these HIST_* settings
> would overwrite that, but I am not sure actually. Personally I never set
> these env_run variables.
>
> user_nl_clm
> * to add L2 option, turn off the longwave downscaling in CLM —> glcmec_downscale_longwave
> = .false.
>
> user_nl_cism
> * Sarah will send an mail on the preferred sliding parameters for in the
> CISM namelist that she’s been testing.
>
> runscript_JG_2.sh
> * (This will be an issue when submitting the JG, so not urgent for now).
> from line 150: about concatenating the forcing files. The cpl hist domain
> information is written to the first file: BG_iteration_1_year1_month1. For
> forcing the JG_2, we start with BG_it_1_year5_month1. When pre-processing
> the forcing files, it will be necessary to  also put in the domain
> information. This only needs to be in the forcing file that is used first:
> so to BG_it_1_year5_month1. So somewhere in the pre-processing loop needs
> to be a line like this; that pastes in the domain info: ncks -Ah -v
> doma_lat,doma_lon,doma_area,doma_aream,doma_mask,doma_frac infile.nc
> outfile.nc Let me know if this is clear, or if I can help more.
>
>
> Thanks,
> Laura
>
>
>
> On 9 Dec 2018, at 21:53, Marcus Lofverstrom <lofverstrom at email.arizona.edu>
> wrote:
>
> Hi Kate, liwg-folks,
>
> Thanks a lot for putting this together so quickly! I found a few small
> issues with the current setup:
>
> runscript_BG_1.sh:
> * Variable "Outputroot" is pointing at output directory from September
> * Variable "CCSMRoot" is pointing at model version that doesn't exist
> (perhaps legacy from Sept script?)
>
> runscript_JG_2.sh:
> * Variable "Outputroot" is pointing at output directory from September
>
> user_nls/user_nl_clm
> * Variable "glacier_region_rain_to_snow_behavior": second
> "converted_to_snow" appears to span multiple lines. Not sure if this is a
> problem or not, but might be worth fixing.
>
> user_nls/user_nl_cism
> URL to CISM2 boundary condition appears to have a cut in the middle of the
> word "glissade". This will probably cause a problem at startup.
>
> Best,
> Marcus
>
> On Fri, Dec 7, 2018 at 6:27 PM Katherine Thayer-Calder <katec at ucar.edu>
> wrote:
>
>> Hi everybody,
>>
>> The source code for the simulations is put together and on Cheyenne.
>> Please take a look at the sandbox and let me know what you think by Sunday
>> evening or Monday morning.
>> Source code:
>>
>> /gpfs/u/cesm-scripts/liwg/Coupled_BG_JG_Spinup_CESM2.1.0_Dec2018/Model_Version/cesm2.1.0+cism2_1_66
>> User_nls:
>> /gpfs/u/cesm-scripts/liwg/Coupled_BG_JG_Spinup_CESM2.1.0_Dec2018/user_nls
>> And there are two source mods still floating around here:
>>
>> /gpfs/u/cesm-scripts/liwg/Coupled_BG_JG_Spinup_CESM2.1.0_Dec2018/SourceMods/src.drv
>> One sets the height of the highest glacier elevation class to 1m above
>> the one below it, and the other is a smoothing parameter between real and
>> virtual elevation classes.
>>
>> Please check that all the code you hope to see is in all the right
>> places! Thanks for all the help!
>> Kate
>>
>>
>> On Fri, Dec 7, 2018 at 10:50 AM Katherine Thayer-Calder <katec at ucar.edu>
>> wrote:
>>
>>> Yes, that makes sense. I will include Bill L's block inception changes
>>> with the other commits and make a new cism-wrapper tag pointing to the head
>>> of CISM master + block inception. I just wanted to be clear that there is
>>> additional code going into that tag that was not included in what Bill L
>>> sent out previously. If that's fine with everybody then it's great with me.
>>> This is what I was planning on, too. Just wanted to have clear validation.
>>> :)
>>>
>>> Thanks!
>>> Kate
>>>
>>> On Fri, Dec 7, 2018 at 10:11 AM Sarah Bradley - CITG <
>>> S.L.Bradley-1 at tudelft.nl> wrote:
>>>
>>>> Hi Kate
>>>>
>>>> I noticed that Erik had fixed the wrapper problem: great!
>>>> I think the earlier things that you mention (modified basal melting
>>>> etc) were all for with shelves: we are not going to run with shelves so
>>>> that should be okay to not have tested with the final combined code. But we
>>>> will (for other runs) want to have these in the main master, Can they not
>>>> just all be combined?
>>>> I tested the ‘block inception code’  but  you mention that does not
>>>> include the July/August commits (above) should be fine as these revisions
>>>> are not active in the code that I used to test. Does this make sense?
>>>>
>>>> Sarah
>>>>
>>>>
>>>>
>>>> On 7 Dec 2018, at 17:53, Katherine Thayer-Calder <katec at ucar.edu>
>>>> wrote:
>>>>
>>>> Hi Sarah, so we've added a few small things in the background for CESM
>>>> support. Erik fixed the problem with not being able to restart past year
>>>> 9999 and that is in the cism-wrapper tag cism2_1_64, and then I added in
>>>> the ability to automatically generate the io files for each build (so would
>>>> fix your problem in the previous email chain with trying to generate and
>>>> copy the files over separately). That is tested and added in cism2_1_65.
>>>> Today I am going to make the final cism-wrapper tag to be used in the JG-BG
>>>> simulations, and include the source_cism code to block inception (and
>>>> probably the new CLM branch tag, though I'm going to double check with Bill
>>>> S about that).
>>>>
>>>> Also, just to be clear, Bill L added his block-inception code over the
>>>> July 8 hash that was used in the cism wrapper tag cism2_1_63, and
>>>> previously used (with source mods) in the JG-BG run. So the testing that
>>>> has been done with Bill L's code doesn't include some commits to master
>>>> from last summer, like "Modified the basal melting GLP" and "zero out
>>>> bmlt_float in ice-free ocean for basal melting GLP". The commits that are
>>>> not included in Sarah's testing can be seen on this webpage:
>>>> https://github.com/ESCOMP/cism/commits/master and are the changes from
>>>> July 21 (not included hash 88073c5) through Aug 27 (not included hash
>>>> a8b1538). I was under the impression that we do want to include these
>>>> changes in the JG-BG (ie, work from the head of CISM master this time), but
>>>> that is not specifically what Sarah was testing.
>>>>
>>>> So, let's try to make this clear, do we want to add Bill L's changes to
>>>> the head of Master (and include the July-Aug commits listed above)? Or do
>>>> we want to use Bill L's changes as he has them now, and as they have been
>>>> tested in the last week (not include July-Aug commits)?
>>>>
>>>> Thanks so much,
>>>> Kate
>>>>
>>>>
>>>> On Fri, Dec 7, 2018 at 1:32 AM Sarah Bradley - CITG <
>>>> S.L.Bradley-1 at tudelft.nl> wrote:
>>>>
>>>>> Hi Kate
>>>>>
>>>>> Thats great news.
>>>>> Raymond has ran a quick I  run (repeating  the earlier L2+P1)
>>>>> simulation and nothing glaring different stands out. The T-run with the two
>>>>> new source mods (L2+P1 source mod, and the block inception code) is
>>>>> approximately the same as the earlier version (with source mods version),
>>>>> so that seems all good to progress.
>>>>>
>>>>> I assume for the new tags you meaning the new CLM P1 version and the
>>>>> new block inception code for CISM? Was there an extra revision that I have
>>>>> missed?
>>>>>
>>>>> Thanks
>>>>>
>>>>> Sarah
>>>>>
>>>>>
>>>>> On 7 Dec 2018, at 00:51, Katherine Thayer-Calder <katec at ucar.edu>
>>>>> wrote:
>>>>>
>>>>> Hi all, quick update. In order to set up the sandbox with the most
>>>>> recent CISM, it would be good to do an update to the CESM CISM-Wrapper
>>>>> code. We are working on adding and testing 2 new tags to get this all
>>>>> together. So, I should have the sandbox tomorrow afternoon, not today as I
>>>>> thought.
>>>>>
>>>>> Thanks,
>>>>> Kate
>>>>>
>>>>> On Thu, Dec 6, 2018 at 12:37 PM Marcus Lofverstrom <
>>>>> lofverstrom at email.arizona.edu> wrote:
>>>>>
>>>>>> Hi all,
>>>>>>
>>>>>> Sounds good to me too. Let me know if I can help in any way. Great
>>>>>> job putting all new code changes in place so fast!
>>>>>>
>>>>>> Best,
>>>>>> Marcus
>>>>>>
>>>>>> On Thu, Dec 6, 2018 at 9:20 AM Sarah Bradley - CITG <
>>>>>> S.L.Bradley-1 at tudelft.nl> wrote:
>>>>>>
>>>>>>> Hi Kate
>>>>>>>
>>>>>>> Sounds like a good plan: if you can send out the link to the
>>>>>>> directories for us to look at when they are set up, that would be great.
>>>>>>>
>>>>>>> Sarah
>>>>>>>
>>>>>>> On 6 Dec 2018, at 17:17, Katherine Thayer-Calder <katec at ucar.edu>
>>>>>>> wrote:
>>>>>>>
>>>>>>> Hi Miren, Sounds good to me. I'll set up the sandbox and the scripts
>>>>>>> today. Then everybody will have the weekend to check the code base and make
>>>>>>> sure we are running what they want and expect. I'll aim to start the first
>>>>>>> BG again Monday morning.
>>>>>>>
>>>>>>> Thanks!
>>>>>>> Kate
>>>>>>>
>>>>>>> On Thu, Dec 6, 2018 at 5:51 AM Miren Vizcaino <M.Vizcaino at tudelft.nl>
>>>>>>> wrote:
>>>>>>>
>>>>>>>> Dear Kate, all,
>>>>>>>>
>>>>>>>> Many thanks to Bill L. and Bill S. for the code for
>>>>>>>> no-GIC’s-inception and P1 (cold rain routed into run-off), respectively.
>>>>>>>>
>>>>>>>>
>>>>>>>> Sarah successfully tested the no-GIC’s-inception code (thumbs up).
>>>>>>>> It is only missing to have the code “formally” integrated into CESM-CISM,
>>>>>>>> just a small task, I suppose Kate/NCAR can do this?
>>>>>>>>
>>>>>>>> Raymond and Sarah are setting a 10-year I run & subsequent T run
>>>>>>>> with P1 code from Bill S to test it works fine. They will have the results
>>>>>>>> later this week.
>>>>>>>>
>>>>>>>> @Kate, now that we have all the new code, can you please set the
>>>>>>>> BG/JG to start it on Monday? If you have any questions, Delft is happy to
>>>>>>>> help.
>>>>>>>>
>>>>>>>> Best and thank you everyone,
>>>>>>>>
>>>>>>>> Miren & Delft team
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>> Liwg-core mailing list
>>>>>>> Liwg-core at cgd.ucar.edu
>>>>>>> http://mailman.cgd.ucar.edu/mailman/listinfo/liwg-core
>>>>>>>
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> Liwg-core mailing list
>>>>>>> Liwg-core at cgd.ucar.edu
>>>>>>> http://mailman.cgd.ucar.edu/mailman/listinfo/liwg-core
>>>>>>>
>>>>>>
>>>>>>
>>>>>> --
>>>>>> Dr. Marcus Lofverstrom
>>>>>> Assistant Professor
>>>>>> Department of Geosciences
>>>>>> University of Arizona
>>>>>> 1040 E 4th St
>>>>>> Tucson, AZ, 85719
>>>>>> lofverstrom at email.arizona.edu
>>>>>>
>>>>>
>>>>>
>>>>
>
> --
> Dr. Marcus Lofverstrom
> Assistant Professor
> Department of Geosciences
> University of Arizona
> 1040 E 4th St
> Tucson, AZ, 85719
> lofverstrom at email.arizona.edu
> _______________________________________________
> Liwg-core mailing list
> Liwg-core at cgd.ucar.edu
> http://mailman.cgd.ucar.edu/mailman/listinfo/liwg-core
>
>
> _______________________________________________
> Liwg-core mailing list
> Liwg-core at cgd.ucar.edu
> http://mailman.cgd.ucar.edu/mailman/listinfo/liwg-core
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cgd.ucar.edu/pipermail/liwg-core/attachments/20181210/bc593698/attachment-0001.html>


More information about the Liwg-core mailing list