[Liwg-core] Feasibility of adding some LIWG-relevant diagnostics to Land diagnostics package?

Alice Bertini aliceb at ucar.edu
Mon Mar 13 10:44:23 MDT 2017


Hi Jan,

I think your ideas for adding separate plot sets to the existing LMWG diags
along with corresponding XML settings and obs files is fine. The atm diags
do something similar for chemistry and waccm diags.

Are you planning to do this work on yellowstone or cheyenne? These are the
only 2 machines currently supported
out-of-the-box. If you want to develop on a non-NCAR machine, then you'll
need to do a bit of system work first to
get all the python dependencies installed correctly.

Here are some steps to get you started:

1. If you haven't already done so, please read this page on the github wiki:

https://github.com/NCAR/CESM_postprocessing/wiki/PPDG_Add_Diags

2.  If you're on yellowstone or cheyenne, then start by cloning a copy of
the CESM_postprocessing repo
into your own workspace:

> git clone https://github.com/NCAR/CESM_postprocessing.git

3. Start by perusing the existing code in the lnd_diag and
diagnostics/diagnostics/lnd subdirs to
familiarize yourself with how the current diags are setup in the python
environment.

When you're ready to start coding, then the best thing to do is probably
run a shared screen on-line session
in google hangouts where we can talk through the setup and test / debug
options.

Does that sound reasonable to you?

Alice



On Mon, Mar 13, 2017 at 8:46 AM, Lenaerts, J.T.M. (Jan) <j.lenaerts at uu.nl>
wrote:

> Hi Alice,
>
> Thanks for your advice I have been looking around for a bit, and we would
> be most interested in Set 2 and Set 3. I have a couple questions/findings:
>
> 1.  For Set 2, we would like to copy most settings but opt for a regional
> plot for Greenland and Antarctica, and use different observational
> datasets. This will require to copy the xml variables that are specific for
> Set 2 (such as LNDDIAG_reg_contour), separately for Greenland and
> Antarctica. My questions:
> - should we define a new set for Greenland and Antarctica, and/or one for
> each ice sheet separately? I would prefer two separate new sets because the
> plotting and observational datasets will be different for Greenland and
> Antarctica, but I don’t know if that is acceptable for you.
> - should the observational datasets all be in the same resolution? I see
> there is a an xml variable that specially asks that but is T42 the common
> resolution of all the obs datasets? That is no problem for us but will require
> some regridding prior to using the datasets.
>
> 2. For Set 3, we will define two new regions (Greenland and Antarctic ice
> sheet, those are now defined as land regions but we are only interested in
> the ice covered areas), and we will have some other variables. Also for
> this case, my preference is to create a new set based on Set 3 settings. Is
> that ok for you?
>
> 3. What is the best practice to start adapting the scripts? Where are the
> most recent land diagnostic package scripts located?
>
>
> All the best,
>
> Jan
>
>
>
>
>
>
>
>
>
>
>
>
>
> On 2 Mar 2017, at 22:05, Alice Bertini <aliceb at ucar.edu> wrote:
>
> Sure, here's the python postprocessing set up files for
> case b.e20.B1850.f09_g16.pi_control.all.123_pp
>
> /glade/p/work/aliceb/sandboxes/runs/b.e20.B1850.
> f09_g16.pi_control.all.123_pp
>
> In paticular, you'll want to look at the env_diags_lnd.xml settings.
>
> The corresponding web page with links to all the plot sets:
>
> http://www.cesm.ucar.edu/models/cesm2.0/model_
> diagnostics/test/b.e20.B1850.f09_g16.pi_control.all.123/
> index.20161221-114650.html
>
> HTH,
> Alice
>
> On Thu, Mar 2, 2017 at 1:59 PM, Bill Sacks <sacks at ucar.edu> wrote:
>
>> Thanks, Alice & Keith.
>>
>> Here are my tentative replies; Jan, you can chime in to confirm or
>> correct me:
>>
>> - are there special climo or average files in addition to those already
>>> being created for the lnd diags
>>> that we need to add to the pyAverager?
>>>
>>
>> I'm thinking we can just piggy-back on the current averages – we'll just
>> need to add some more variables that are processed.
>>
>> - what languange are the existing plotting functions written in (NCL,
>>> IDL, matplot, etc...)?
>>>
>>
>> Jan talked about doing this in NCL
>>
>> - what controls do we need to add to the XML settings to control this new
>>> plot set? Is there
>>> an existing script available for generating the Land-ice diags that we
>>> can look at more closely?
>>>
>>
>> Nothing exists currently. Can you point us to an example showing the xml
>> settings you're referring to?
>>
>> Thanks,
>> Bill
>>
>> On Mar 2, 2017, at 1:43 PM, Keith Oleson <oleson at ucar.edu> wrote:
>>
>> Hi,
>>
>> I think Alice covered all the bases nicely, so my answer to question 1 is
>> yes.
>>
>> Keith
>>
>> On Thu, Mar 2, 2017 at 12:11 PM, Alice Bertini <aliceb at ucar.edu> wrote:
>>
>>> Hi Bill,
>>>
>>> It's definitely doable to add a new plot set into the python framework.
>>> In fact, the WACCM group
>>> did just that last summer and wrote up their development experience here:
>>>
>>> https://github.com/NCAR/CESM_postprocessing/wiki/PPDG_Add_Diags
>>>
>>> That being said, I think we'd want to look at a couple of things first
>>> to gauge what needs to be done
>>> before putting it into a time frame:
>>>
>>> - are there special climo or average files in addition to those already
>>> being created for the lnd diags
>>> that we need to add to the pyAverager?
>>> - what languange are the existing plotting functions written in (NCL,
>>> IDL, matplot, etc...)?
>>> - what controls do we need to add to the XML settings to control this
>>> new plot set? Is there
>>> an existing script available for generating the Land-ice diags that we
>>> can look at more closely?
>>>
>>> Just let me know how you'd like to proceed.
>>>
>>> Thanks,
>>> A.
>>>
>>>
>>>
>>>
>>> On Thu, Mar 2, 2017 at 12:00 PM, Bill Sacks <sacks at ucar.edu> wrote:
>>>
>>>> Hi Keith and Alice,
>>>>
>>>> A number of people in the LIWG have expressed interest in having more
>>>> LIWG-relevant diagnostics produced automatically in CESM2 runs. Eventually,
>>>> we'd like to have our own LIWG diagnostics package, but we're not going to
>>>> have time to do that for the CESM2 release.
>>>>
>>>> Most of what we'd want is obtainable from CLM history files. So, in the
>>>> short term – i.e., the next couple of months – we were wondering if it
>>>> would be feasible to add some LIWG-relevant diagnostics fields to the Land
>>>> diagnostics package. At the very least, this would involve adding some new
>>>> fields. Ideally, these would be shown with regional projections centered
>>>> over Antarctica and Greenland; I don't think this is currently done for any
>>>> LMWG diagnostics, though it is for AMWG diagnostics. We might also want to
>>>> include some new observational and reanalysis-derived datasets for
>>>> comparison. Perhaps all of this could be included in a new "set"??
>>>>
>>>> Do you feel this would be feasible over the next two months? Jan
>>>> Lenaerts (cc'ed) is happy to take the lead on this, so this shouldn't take
>>>> much of your time, other than perhaps pointing us in the right direction.
>>>> So the questions are really:
>>>>
>>>> (1) Keith: Does this seem acceptable?
>>>>
>>>> (2) Both: Do you have a sense of how hard this would be? Specifically,
>>>> how hard would it be to (a) add a new set, (b) include regional map plots,
>>>> (c) include new observational datasets?
>>>>
>>>> Thanks,
>>>> Bill
>>>>
>>>>
>>>
>>>
>>> --
>>> Alice Bertini
>>> NCAR - CSEG Software Engineer
>>>
>>
>>
>>
>
>
> --
> Alice Bertini
> NCAR - CSEG Software Engineer
>
>
> - - - - - - - - - - - - - - - - - - - - - - - - - - -
>
> *Jan Lenaerts  *IMAU, Utrecht University || University of Colorado
> @lenaertsjan <https://twitter.com/lenaertsjan> || website
> <http://www.colorado.edu/lab/icesheetclimate/>
>
>
>
>
>
>
>
>
>


-- 
Alice Bertini
NCAR - CSEG Software Engineer
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cgd.ucar.edu/pipermail/liwg-core/attachments/20170313/ba46e627/attachment-0001.html>


More information about the Liwg-core mailing list