Using MPAS-JEDI's own executable to calculate ensemble mean and variance #365
Conversation
| rain_water \ | ||
| snow_water \ | ||
| ) | ||
| set MPASHydroStateVariables = (${MPASHydroIncrementVariables} cldfrac) |
There was a problem hiding this comment.
When I ran test/testinput/eda_OIE120km_WarmStart.yaml the HofX steps all failed with
mpasjedi_vc_model2geovars::changevar: cldfrac must be added to the state variables in order to populate the var_cldfrac geovar with the MPAS diagnostic cloud fraction
I reverted the change to config/mpas/variables.csh and the test case passed.
I also reran test/testinput/letkf_OIE120km_WarmStart.yaml with the reverted variables.csh and it passed.
There was a problem hiding this comment.
I get the same for several tests on the HofX step.
There was a problem hiding this comment.
I used it to also calculate the ensemble mean of cldfrac. Without specified in state variables, the mean of cloud_area_fraction_in_atmosphere_layer could not be calculated. Maybe we need to add both? I am not sure but will try it later.
There was a problem hiding this comment.
I am a bit confused. The model variable name cldfrac has not changed as a generic name yet (i.e., cloud_area_fraction_in_atmosphere_layer). It is also the same for effective_radius_of_xxxx variables. All these variable should go through the "variable change". I think this PR would not work with develop branch of mpas-jedi. @mos3r3n, I wonder if this PR is compatible with some of your feature branch, not with develop.
There was a problem hiding this comment.
It is from the develop branch. I thought those variables were general variables already. I will make a commit to correct them back.
There was a problem hiding this comment.
You can check the list of variables here: https://docs.google.com/spreadsheets/d/1Kh_Ypp8Uwm69yfIhGcAlijgmzU7iY_ZpnBS9SQZoQTg/edit?gid=0#gid=0
There was a problem hiding this comment.
Thanks, BJ!
I modified them to cldfrac, re_cloud, re_ice, and re_snow. Now it works.
There was a problem hiding this comment.
Great, I tested and it HoxF works now.
| re_cloud \ | ||
| re_snow \ | ||
| re_ice \ |
There was a problem hiding this comment.
@mos3r3n , I'd like to confirm that the results are the same (for example, ensMeanVariance) even if we remove these three variables, re_cloud, re_snow, and re_ice.
To clarify the procedure, the list of state variables should come from yaml config. This is not the case in current mpas-jedi's develop branch. See for example https://github.com/JCSDA-internal/mpas-jedi/blob/develop/src/mpasjedi/State/mpas_state_interface_mod.F90#L70-L94 .
This was discussed with @mos3r3n and @jim-p-w, separately before.
The usecase of mpas-jedi is getting more complex, so it would be nice to make this part clear in near future.
There was a problem hiding this comment.
If these variables are not included, their mean will not be calculated. For all-sky HofX of the ensemble mean, these variables should be available.
I haven't tested the DA applications. Maybe we need to tested via 1-week cycling test @jim-p-w .
There was a problem hiding this comment.
I realized that the arithmetic operations only work with vars_ci, which comes from the yaml directly. So this is a necessary change for mpasjedi_ens_mean_variance.x.
| air_potential_temperature \ | ||
| dry_air_density \ | ||
| u \ | ||
| w \ |
There was a problem hiding this comment.
Is this a necessary change ?
There was a problem hiding this comment.
I'm not quite sure whether the w variable is a necessity to initialize the model.
If in the future we assimilate radial velocity data, this variable should be added to the state variables.
config/mpas/variables.csh
Outdated
| air_temperature_at_2m \ | ||
| water_vapor_mixing_ratio_wrt_moist_air_at_2m \ |
There was a problem hiding this comment.
Is this a necessary change ?
There was a problem hiding this comment.
I think for future surface assimilation purpose, these two variables could be added.
|
In my comments above, I am trying to keep the necessary changes that are relevant to this PR in general. Additional features can be done in separate PRs. This will also help us keep track of changes in this repository. |
Thanks @byoung-joo! I agree that T2 and Q2 are not used in DA or prognostic variables of MPAS, so I will not add them in this PR. The w wind is a prognostic variable of MPAS, so I will keep it in this PR. |
Sound good to me. This application seems to be unique. It reads the ensemble using "background" stream and writes out using "background" stream. I think that adding |
|
I confirmed that adding |
Description
Previously, the ensemble mean and variance were calculated using an external executable. This PR uses the mpasjedi_ens_mean_variance.x within MPAS_JEDI to do it.
Issue closed
Closes #(if applicable)
Tests completed
Tier 1:
Tier 2 (optional):