Carra snow depth variable changes level type at midnight june 30th 2022

Hi,
I encountered a strange missing value in the sd ("snow_depth_water_equivalent”) variable for June/July 2022, for
the CarraWest domain (I have not checked the other domain).
I am including the request for the data below.

When use CDO on the data it stops reporting file content after
june 30th, and throws a warning.

cdo sinfo file.grib
…Warning (gribapiScanTimestep): Record 241 (name= id=60.1.0 lev1=0 lev2=0) timestep 241: Parameter not defined at timestep 1!

Using grib_ls on the the file reveals that at the end of june and beginning of july
the “type of level” is changed from “surface” to “heightAboveGround” at exactly
this time:

enmi 20220630 an lambert 0 surface 0 sd
enmi 20220630 an lambert 0 surface 0 sd
enmi 20220701 an lambert 0 heightAboveGround 0 sd
enmi 20220701 an lambert 0 heightAboveGround 0 sd

and this is what throws cdo off. There is a work around to ask for netcdf output
because then the months are split into individual files which can then be combined
afterwards.

This issue arises for both the analysis and forecast stream versions of the variable.

I have not ran into this for other variables that I have been picking up from the CARRA dataset, but my variable-set is not complete.

Not sure how to classify this issue… if this is a feature or a bug in the data set… but it does complicate cdo chains that work with the data.

Sincerely
Halldór


CDS request code:
import cdsapi

dataset = “reanalysis-carra-single-levels”
request = {
“domain”: “west_domain”,
“level_type”: “surface_or_atmosphere”,
“variable”: [“snow_depth_water_equivalent”],
“product_type”: “analysis”,
“time”: [
“00:00”, “03:00”, “06:00”,
“09:00”, “12:00”, “15:00”,
“18:00”, “21:00”
],
“year”: [“2022”],
“month”: [“06”, “07”],
“day”: [
“01”, “02”, “03”,
“04”, “05”, “06”,
“07”, “08”, “09”,
“10”, “11”, “12”,
“13”, “14”, “15”,
“16”, “17”, “18”,
“19”, “20”, “21”,
“22”, “23”, “24”,
“25”, “26”, “27”,
“28”, “29”, “30”,
“31”
],
“data_format”: “grib”
}

client = cdsapi.Client()
client.retrieve(dataset, request).download()

Dear Halldór,

Thank you very much for looking at the CARRA dataset and pointing to a specific problem.

We have recently realised that there are some slight inconsistencies between the data files produced before and after July 2022. The data from July 2022 was produced on our Atos computer, while the older ones on our previous Cray computer. The differences are coming from some differences in some auxiliary software, while the model is unchanged (so scientific consistency is ensured).

We have a similar Forum note at CARRA East - grid change in 2023 and after - #2 by Andras_Horanyi, which points to the same problem.

As a next step our intention is to fully understand the extent and details of the problem and then document it for the benefit of the users.

Thanks again, best regards

                              Andras Horanyi

Thank you for your reply

H

Dear Halldór,

I just posted an answer to another Forum post at CARRA East - grid change in 2023 and after - #10 by Andras_Horanyi

I hope, it also answers to your question.

Indeed for the snow depth water equivalent variable there is a change of the “typeOfLevel” variable from “surface” to “heightAboveGround”.

It was an accidental change what we just realised now thanks to your comment.

I apologise for the problem and thank you for informing us.

Best regards

Andras