`d<comp>.streams.xml` inconsistencies

[dougiesquire]

There are inconsistencies between the d<comp>.streams.xml files in some of our configurations and those generated by the scripts in om3-scripts/data_stream_xml_generation. Specifically, in the RYF configurations (at least the ones I looked at):

• drof.streams.xml: all streams have offset=0, but generate_xml_drof.py sets offset=-43200
• datm.streams.xml: the CORE_IAF_JRA55do.PRSN CORE_IAF_JRA55do.PRRN, CORE_IAF_JRA55do.LWDN and CORE_IAF_JRA55do.SWDN streams have offset=0, but generate_xml_datm.py sets offset=-5400

@ezhilsabareesh8, which is right and are there any other inconsistencies?

[ezhilsabareesh8]

@dougiesquire Some variables starts at 1:30 hrs and ends at 22:30 hrs, while the some variables starts at 0:00 hrs and ends at 21:00 hrs. To make the start time uniform for all variable and to eliminate max dt limit error we offset some of the variables, refer here.

[dougiesquire]

So are our RYF configs currently wrong?

[ezhilsabareesh8]

Similarly an offset of -43200 is set only in IAF configs to match the RYF, since IAF starts at noon, we shifted it to midnight similar to RYF

[dougiesquire]

But the scripts don't discriminate on RYF vs IAF - they just always set the offsets to the IAF values

ADDED: So, for example, if I try to generate drof.streams.xml for an RYF config:

$ generate_xml_drof.py 1900 1900

I get a drof.streams.xml file with streams that have offset = -43200.

[ezhilsabareesh8]

There's a bug in the code then, will fix it. The drof.streams.xml files in the configs seems to be fine.

[ezhilsabareesh8]

Thanks @dougiesquire for picking this. error. Fixed denerate_xml_drof.py to set zero offset for RYFs in this PR.

[dougiesquire]

Thanks @ezhilsabareesh8. I think generate_xml_datm.py also needs a similar change.

[ezhilsabareesh8]

I think generate_xml_datm.py also needs a similar change.

DATM is fine, both the RYF and IAF dataset starts at midnight. The offset is required only for few variables which starts at 1:30 hrs, which is already implemented in the code.

[dougiesquire]

Okay, so does that mean what's in our current RYF configs are wrong?

E.g. https://github.com/COSIMA/MOM6-CICE6/blob/070493071c807e26071062ed14be795f150426de/datm.streams.xml#L4-L23

[ezhilsabareesh8]

DATM is fine, both the RYF and IAF dataset starts at midnight. The offset is required only for few variables which starts at 1:30 hrs, which is already implemented in the code.

Apologise, I think I didn't explain clearly. The offset of -5400 for some variables is required only in IAF whereas in RYF everything is uniform and starts at 0:00 hrs. I had a quick look at the dataset to confirm it, datm.streams.xml is consistent with the data. The negative offset of 5400 should be set only for IAF, will modify the code.

[dougiesquire]

Closed by https://github.com/COSIMA/om3-scripts/pull/17