Esmvaltool: Computing ACSIS indicators

[valeriupredoi]

Hi guys, this issue is to let people know that @ed-hawkins and myself are gonna start working on implementing the ACSIS indicators computation implementation in ESMValTool. Since I am by far not a climate scientist and also @ledm asked about these indicators on today's UKESM evaluation call, Ed, could you please add some info here when you got some time? Also, since you are new to the tool, maybe you can tell me what are the exact steps needed to get the computation done and I can start writing code for that? Cheers and welcome to the ESMValTool community, Ed! :beer:

[ed-hawkins]

Great - thanks all!

We are developing a series of observation based climate 'indicators' as part of NERC's ACSIS programme, for example: http://www.acsis.ac.uk/climate-indicators/atlantic-meridional-overturning-circulation

We are starting from available netcdf files and providing a recipe to calculate the indices.

For the AMOC - we start from https://www.rapid.ac.uk/rapidmoc/rapid_data/moc_transports.nc and use the moc_mar_hc10 variable, which is twice-daily values. We will want to calculate the running annual average.

For the NAO - we start from https://www.metoffice.gov.uk/hadobs/hadslp2/data/hadslp2r.pp.gz and calculate the difference between the sea level pressure averaged over a box near the Azores (35W-25W, 30N-40N) and a box near Iceland (25W-15W, 60N-70N). We then calculate a winter (December-March) average of this difference.

[valeriupredoi]

I am starting to look into these now @ed-hawkins ; will probably ask you for more details soon :grin:

[ed-hawkins]

Thanks @valeriupredoi!

The next variables (called Jet Speed and Jet Latitude) are derived from ERA-Interim (and later ERA-5) and are slightly more complicated.

The process is: (1) start with the daily U850 variable and only keep data for 15N to 75N. (2) Average U850 over longitudes 0-60W. (3) Calculate the maximum value for each day = jet speed. (4) The latitude at which the maximum value appears for each day = jet latitude. (5) Average separately for each season.

Hope this makes sense?

Thanks,
Ed.

[valeriupredoi]

cheers @ed-hawkins ! For now, here are the updates:

• started our branch here: https://github.com/ESMValGroup/ESMValTool/tree/version2_development_ACSIS - the utility that I've written is here esmvaltool/utils/acsis; it needs the Greenland/Iceland subsetting and the DJF averaging (will put these in now); __NOTE__ that the moc netCDF file works finr, the hadslp file needs to be netCDF as well, or if it's PP it needs to be correctly formatted for time (no 0000-00-00 dates as the file from the link above has; I can give you the corrected file I created but the correction process is a hack and should not be part of the workflow);
• ERA-Interim data is on Jasmin at /group_workspaces/jasmin4/esmeval/obsdata/Tier3/ERA-Interim ; please apply for esmeval access at https://accounts.jasmin.ac.uk/services/group_workspaces/esmeval/
• I will finish the current utility then will go on and look into the Jet stuff, will ask you if I get stuck :grin:

[ed-hawkins]

Thanks @valeriupredoi - have applied for the esmval access

[valeriupredoi]

yup, just granted you acsis :+1:

[valeriupredoi]

@ed-hawkins we have a PR and a fully functional first executable: https://github.com/ESMValGroup/ESMValTool/pull/731 Data for moc and hadslp in netCDF format is available on Jasmin at /group_workspaces/jasmin4/esmeval/acsis_data, could you pls test the executable when you have some time? No need to install esmvaltool, just source the environment file from /home/users/valeriu/esmvaltool_environment.sh

[valeriupredoi]

also, please check if the variables are correct, see https://github.com/ESMValGroup/ESMValTool/blob/cc8990658be794bb7a998652c4a94a548b1f823c/esmvaltool/utils/acsis/acsismoc.py#L50

[valeriupredoi]

Thanks @valeriupredoi!

The next variables (called Jet Speed and Jet Latitude) are derived from ERA-Interim (and later ERA-5) and are slightly more complicated.

The process is: (1) start with the daily U850 variable and only keep data for 15N to 75N. (2) Average U850 over longitudes 0-60W. (3) Calculate the maximum value for each day = jet speed. (4) The latitude at which the maximum value appears for each day = jet latitude. (5) Average separately for each season.

Hope this makes sense?

Thanks,
Ed.

@ed-hawkins can you pls clarify what exactly is U850 - is it ua (eastward_wind) at 850 hPa? Cheers, sorry, not a scientist here :grin:

[ed-hawkins]

Hi @valeriupredoi - will have a look at the AMOC and pressure soon.

Yes - U850 is the eastward wind at 850Pa - sorry!

[valeriupredoi]

cheers @ed-hawkins - we now have an acsisjet executable too :grin: - needs your review and test tho

[valeriupredoi]

@ed-hawkins this has been stale for a while - do you still want/have time to proceed?

[zklaus]

Just to let you know: AMOC is here, both AMOC and NAO are in CVDP.

[valeriupredoi]

cheers @zklaus :beer: @ed-hawkins do you still want to continue with this work?