rsdt daily

[oligan]

I am surprised rsdt is not requested on a daily basis. This is a very important variable that would be needed at daily resolution for r1 (because it is the same for all members) for at least the historical and one scenario

[taylor13]

rsdt seems to be available on a daily basis for about half (20) of the models that provided daily output for the historical runs. For what purpose is the variable needed on a daily basis?

Given the monthly mean values available in the archive and a knowledge of the earth's orbit, you could calculate it quite accurately for those models not reporting it.

Note also that for surface studies, rsds is available for all but a couple of models.

[oligan]

Hello Karl,

our workflow follows automatically the CDR but rsdt is not in the output for IPSL-CM6A-LR, so my first reaction is to say it is not included in the CDR, and it is unclear to me why some models have provided it. Maybe it is only requested in a recent version of the CDR ?

rsdt is required to compute transmission (rsds/rsdt). rsdt can be reconstructed, but i) it is a pain, ii) it is a function of the calendar used, and iii) we have found substantial (several %) differences in rsdt in polar regions among models on a monthly basis !

Best regards,

Olivier

rsdt seems to be available on a daily basis for about half (20) of the models that provided daily output for the historical runs. For what purpose is the variable needed on a daily basis?

Given the monthly mean values available in the archive and a knowledge of the earth's orbit, you /could/ calculate it quite accurately for those models not reporting it.

Note also that for surface studies, rsds /is/ available for all but a couple of models.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmip6dr/CMIP6_DataRequest_VariableDefinitions/issues/407#issuecomment-656729821, or unsubscribe https://github.com/notifications/unsubscribe-auth/AHUCSU7WY4Q52PPRLCVF7TLR24WJVANCNFSM4OWR2PAA.

--

[taylor13]

fair enough. I agree it would be a pain, and perhaps unreliable.

I will let Martin or others address the question of whether or not we requested rsdt at daily frequency from the historical runs.

Assuming we didn't, we could consider adding it for CMIP7 (assuming that will exist). It would help to know exactly what kind of studies need this field. You mention transmissivity (I presume especially clear-sky), and I suppose clear-sky albedo (at top of atmosphere) might be of some interest to study aerosol direct effects and their variations on shorter time-scales. Are there other use cases?

[oligan]

Assuming we didn't, we could consider adding it for CMIP7 (assuming that will exist). It would help to know exactly what kind of studies need this field. You mention transmissivity (I presume especially clear-sky), and I suppose clear-sky albedo (at top of atmosphere) might be of some interest to study aerosol direct effects and their variations on shorter time-scales. Are there other use cases?

In our case, we have a new dataset of daily (and even hourly) transmissivity dating back to well before WWII (don't ask me how they measured it...), which we would like to compare to model output (or the other way around ;-)

I would argue that CMIP7 or not, the CDR is one element of the CMIP infrastructure (along with new variable definition, forcing datasets, ESGF, and maybe a few other things) that we want to maintain for the benefits of the community. It's my expectation that the WCRP call for a CMIP office should deal with such things.

Olivier

[martinjuckes]

Hello @oligan , @taylor13 ,

sorry for the slow response on this. rsdt as a daily variable is requested as part of the CFMIP (at priority 1) and DCPP (at priority 3) requests. It is likely that the modeling groups submitting it are signed up to CFMIP.

IPSL-CM6A-LR is signed up to CFMIP, so, in principle, daily rsdt should be part of the output.

rsdt is also requested at 3-hour frequency by HighResMIP.