Add end dates for MetOp-A AMSU-A AVHRR ASCAT

[gmao-msienkie]

Implement changes for #66 - prepare for MetOp-A decommissioning in mksi/sidb

• turn off MetOp-A AMSU-A and AVHRR in active_channels.tbl
• add end date for MetOp-A AMSU-A AVHRR HIRS MHS in available_channels.tbl
• turn off ch 14 iuse=4 in iuseClass-4_channels.tbl
• turn on MetOp-B and NOAA 19 AVHRR

turn off MetOp-A ASCAT (kx=290, satid 4) in gmao_global_convinfo.txt

Using 20210918/12z as transition date (as tested in f5721_metop)

[gmao-msienkie]

non-zero diff even prior to transition date because MetOp-A AVHRR is replaced with MetOp-B AVHRR

[jing-guo]

non-zero diff even prior to transition date because MetOp-A AVHRR is replaced with MetOp-B AVHRR

Shouldn't this be considered as observation-configuration?

[gmao-msienkie]

Oh, I mistyped that - I have to watch it with these late night edits. What I meant to say is that MetOp-A ASCAT is removed and replaced with MetOp-B ASCAT (not AVHRR). Since we do not have a time dependent convinfo configured yet - the switch to MetOpB ASCAT affects all dates, so it is non-zero-diff. The "observation-configuration" was a label I added because I have been doing a lot of edits of the observation database where for dates prior to an observation change the assimilation was zero-diff and after the change the assimilation would not be zero-diff. And, because these have been changes for recent events assimilation tests like the X-experiments would not yet be affected because they are still before the effective dates. Except, of course, for this case as I noted above. Per Matt's note there are specific labels for the label check and 'observation configuration' is not one of the required ones.

[rtodling]

Meta, I just updated the branch, but it seems from the email of last night (and I quote):

So I had a look today at the FP plots, since I wanted to see how the stats and bias correction evolved since the 23 September 12z incident.

What I am thinking is that we can window out the data at 12z Sep 23 – possibly in the obsys-nccs.rc or alternatively in the active_channels.tbl/available_channels.tbl (since we don’t want the bad data affecting channels while they are passive.

that you propose to move the date of turning off. So, can you update this branch accordingly so I can get it to fold onto develop? Thanks

[gmao-msienkie]

Ricardo, the email you referenced is for a separate issue with the AIRS instrument that occurred shortly after the f5271_metop test began. When approved, whichever changes we decide to make will go in as a separate PR. There will also need to be an additional PR for the change to turn on the CrIS-FSR NPP instrument but that issue also needs further investigation.

This PR covers only the changes for the other instruments (turn off MetOp-A, turn on AVHRR for N19 and MetOp-B, turn on MetOp-B ASCAT). While issue #66 mentions other possible dates for the switch I think the ones in this branch are what we are going with.

[rtodling]

Got it!