Changes to get snow 2DVar for deterministic and ensemble mean working

[CoryMartin-NOAA]

Changes to JCB templates to allow for 2DVar snow analyses for the deterministic and the ensemble mean for GDAS.

[CoryMartin-NOAA]

@yuanxue2870 @jiaruidong2017 I think this is illustrating how important it is that we keep the offline and GFS workflow YAMLs in sync. Are there any other changes that need to be made? If so, do we need to discuss tomorrow?

[yuanxue2870]

@yuanxue2870 @jiaruidong2017 I think this is illustrating how important it is that we keep the offline and GFS workflow YAMLs in sync. Are there any other changes that need to be made? If so, do we need to discuss tomorrow?

Yes, I would suggest an in-person discussion tomorrow. Since GTS and MADIS yamls will both require another round of changes (e.g., new reject list), all changes we made here are likely not being finalized. But at this stage, I think it is always good to keep uncoupled/coupled model runs (esp. yaml files) in sync (if the observation files and their attributes are identical between these two modes) - I guess this is also one of the reasons Clara wants to introduce JCB into the process to avoid asynchronous changes. Just my two cents.

[jiaruidong2017]

@yuanxue2870 @jiaruidong2017 I think this is illustrating how important it is that we keep the offline and GFS workflow YAMLs in sync. Are there any other changes that need to be made? If so, do we need to discuss tomorrow?

It is good for me now, and I will show some tests for the latest changes in our tomorrow's meeting.

[jiaruidong2017]

@yuanxue2870 @jiaruidong2017 I think this is illustrating how important it is that we keep the offline and GFS workflow YAMLs in sync. Are there any other changes that need to be made? If so, do we need to discuss tomorrow?

Yes, I would suggest an in-person discussion tomorrow. Since GTS and MADIS yamls will both require another round of changes (e.g., new reject list), all changes we made here are likely not being finalized. But at this stage, I think it is always good to keep uncoupled/coupled model runs (esp. yaml files) in sync (if the observation files and their attributes are identical between these two modes) - I guess this is also one of the reasons Clara wants to introduce JCB into the process to avoid asynchronous changes. Just my two cents.

I think the new rejectlist is beyond of this PR, because it requires more efforts. We can update the rejectlist later when it is ready.

[jiaruidong2017]

lgtm. Thanks for your great efforts.

[ClaraDraper-NOAA]

Looks good to me. Later, we can update the configurations.