Request to add ERA5 re-analysis files on ARCHER-2 for Jan 2019 to Mar 2022 (post-ERA-interim period)

Unified Model
1

Dear NCAS-CMS Helpdesk,

I have run a series of interactive stratospheric aerosol simulations of the Hunga-Tonga volcanic aerosol cloud, with GA4 UM-UKCA CheST+GLOMAP on ARCHER-2 (UM experiments xphy and xphr)

These have already shown some very interesting results, the main one being there is more AOD in the stratosphere right now than can be explained by 0.4Tg of SO2 measured from TROPOMI (and also IR sensors).

I have run a sequence of simulations scaling up the 0.4Tg of SO2 by a factor of 2, of 3 and of 4, and the finding we have suggests a factor of 3 increase in SO2 is consistent with the AOD currently being measured from satellite (e.g. from the OMPS instrument).

These simulations are free-running experiments based on re-starting the model from a state where the QBO was in a similar phase as the current situation, and capture the dispersion of the cloud in quite good agreement with the observations.

But the match of the winds etc. is only approximate.

In discussion with the group at Univ. Colorado, there is a MIP for interactive stratospheric aerosol models, which involves each model running with “specified dynamics” — i.e. to run nudged to meteorological re-analysis files.

The UM-UKCA has this functionality, and we have run nudged simulations of several eruption-clouds, but currently the ARCHER-2 directory has only the ERA-interim re-analysis fields, with data-files only up to June 2019.

The ERA5 re-analysis essentially replaces ERA-interim, and for running the currently meteorological setting, we will need to run with ERA5 re-analysis meteorology.

With the Hunga-Tonga aerosol cloud at this unexpectedly high strat-AOD (given the relative modest SO2 emission of 0.4Tg), and several groups working to understand potential reasons for this – which include already-oxidised-within-volume-plume-sulphate, sea-spray from the detraining vaporised seawater, and/or ultra-fine ash particles.

Each of these is a plausible mechanism for how there can be more strat-AOD than expected from the 0.4Tg emission of SO2.

And then this is quite a priority to get equivalent ERA-5 re-analysis files produced so that the UK community model can run these simulations, both for the MIP, and for UK science.

I think it was Luke Abraham (NCAS-CMS, Univ. Cambridge) and/or Mohit Dalvi (JWCRP Met Office) who originally co-produced this archive of the ERA-interim re-analysis netCDF files currently stored on ARCHER-2 at:

/work/y07/shared/umshared/ukca/analyses/era-in/

Please can someone in the NCAS-CMS team (Luke?) explore how we can plan to get the ERA5 re-analysis files into equivalent netCDF files on ARCHER so that users can run simulations with the nudged UM-UKCA model.

I guess it will require the ERA5 files to be extracted and the scripts used to generate the ERA-interim re-analysis archive be adjusted to produce the ERA5 equivalents.

And I realise it’s not a simple request – and happy to discuss to help with any required communications and planning to achieve this.

Thanks a lot for your help,

Best regards,

Cheers
Graham

Dr. Graham Mann
Lecturer in Atmospheric Science, Univ. Leeds.

2

Hi Graham,

These files are available on Monsoon2 here

/projects/ukca-admin/analyses/era5

However, they haven’t been copied back to ARCHER2 yet. We can copy these across, but there are also necessary code changes that have to be made to be able to read these files correctly.

The code changes are documented in UM ticket #5336. They are for Rose versions of the UM and you will need to back-port these to vn8.4.

https://code.metoffice.gov.uk/trac/um/ticket/5336

Best wishes,
Luke

3

Hi Luke

Thanks for explaining this.

OK, that would be great, yes pls can you (or someone in the NCAS-CMS team) copy the ERA5 directory & files over to ARCHER-2

I will look into the code-change that needs to be back-ported to the v8.4 branch, and if it is a fairly straightforward change I am happy to back-port that to the v8.4 nudging FCM branch.

I will be able to look into this tomorrow or Wed and send an update to this later this week.

Likewise please can you post an update to this helpdesk ticket once the ERA5 files are copied over to ARCHER-2

Thanks again, much appreciated

Best regards

Cheers
Graham

4

The era5 files are on ARCHER here:

/work/y07/shared/umshared/ukca/nudging/era5

Grenville

5

Hi Grenville,

With Sandip working on another project since 1st April, and me having a 3-week CoViD illness in April we didn’t get time in the end before the 16-18 May 2022 international SSiRC workshop.

And it’s only now that we’ve had time to progress to the ERA-5 nudged runs, which we’re keen to include in an ACP letters paper I’ve agreed to lead on the Hunga-Tonga aerosol cloud.

I’ve just checked on ARCHER-2 and I can see the ERA-5 files are there in that directory up to 30th November 2021:

/work/y07/shared/umshared/ukca/nudging/era5/

Please can you or someone in the team copy over the ERA5 re-analysis netCDF files for the period December 2021 to the latest date these are currently available.

One key thing is in relation to the current Antarctic season, and I have a new PDRA working on PSC modelling who we’re keen to run simulations for the timing of the Hunga-Tonga water vapour transported to the Antarctic.

As you may have seen on social media etc., the Southern Hemisphere mid-latitude stratosphere is currently double the concentration of water vapour than usual, and although the Hunga-Tonga water vapour plume did not reach the high latitudes when the Antarctic vortex spun-up in April or so, had the eruption been a few months earlier this might have caused extensive & unusual PSC formation this season.

We’ll run “what if” simulations, repeating the current Hunga-Tonga simulations where we emit the >100Tg of water vapour that was added to the tropical stratosphere, in addition to the SO2 we emitted in the initial simulations – with the UM-UKCA then forming PSCs consistently with any elevation in the water vapour mixing ratio.

It would be a very big effect if the 10 ppmv of water currently at about 10-50 South had reached 55-60 South in March April because then it would be in the Antarctic vortex and PSCs would form at warmer temperatures (5K warmer actually if 10 ppmv had reached that far south at that time of year).

Anyway – whereas those initial runs were OK to use the approximate-QBO, for these timing of PSC runs we really need to run with this year’s ERA5 re-analysis fields.

Please can you or someone in the team copy across those latest ERA5 re-analysis files for the December 2021 to current time, and we’ll then be able to run the simulations nudged on ARCHER-2.

Thanks a lot,

Cheers
Graham

Hunga-Tonga period from

6

Hi Graham, Grenville,

The ERA-5 files have been updated on Monsoon2 here

/projects/ukca-admin/analyses/era5

and are currently available up to the end of April 2022. I’m not sure if more recent files are available.

Note that there is no automated method to produce and distribute these files. These are processed by Mohit Dalvi at the Met Office who then places them on Monsoon, but these then need to be copied by someone else over to ARCHER2.

Best wishes,
Luke

7

Hi Luke,

OK – thanks for this.

That’s great that the ERA-5 files are there up to the end of Apr 2022 on MONSOON2.

Please can someone in the team copy the December 2021 to Apr 2022 files across from MONSOON2 over to ARCHER-2, to the umshared directory here:

/work/y07/shared/umshared/ukca/nudging/era5

Thanks a lot,

Best regards,

Cheers
Graham

8

Dear NCAS-CMS helpdesk team,

Since no-one acted to copy these across, I have this afternoon gone ahead and copied across these ERA5 files myself, but (as far as I am aware) I don’t have permissions to write to the main ERA5 directory on /work/y07/

/work/y07/shared/umshared/ukca/nudging/era5/

The directory I have copied them over to is off my /work area:

/work/n02/n02/gmann/NudgingFiles/ERA5reanalysisFilesFromMONSOON_forDec21toApr22/

Please can someone copy these files across to the central area.

Obviously I can change the path in user-wk-cp of the nudging branch, which I will do for the period I’m running for (I’ve decided best to nudge through the whole of 2021), so I can run the period from this directory.

But please can someone in the NCAS-CMS team (that has the write permissions for that directory) copy the 5 months of the N48L137 ERA5 nc files (1st Dec 2021 to 30th Apr 2022) from where I copied them over to (from MONSOON2), to the central y07 ERA directory.

Thanks a lot,

Best regards,

Cheers
Graham

9

Hi Graham,

We did copy those files over to ARCHER2. However, as I’m sure you can appreciate, we, like everyone else, are busy and unfortunately, occasionally things do fall through the cracks and they didn’t get moved into umshared. I have now copied the files into the umshared era5 directory.

Sorry for the delay.

For the future could you change the permissions on your ARCHER2 directories so that they are readable please:

chmod -R g+rX /home/n02/n02/gmann
chmod -R g+rX /work/n02/n02/gmann

Regards,
Ros.

10

Hi Ros,

OK, no problem – thanks.

For the record, I did check yesterday the files on /work/y07/shared/umshared/ , and that was still only to end of Nov 2021.

But OK that’s great they’re copied over now, to the y07 directory, and I understand it’s August and I imagine people are away etc. the last weeks.

I’ve deleted the local-copies now from my workspace, but have in any case done the

chmod -R g+rX /home/n02/n02/gmann
chmod -R g+rX /work/n02/n02/gmann

That 2nd command hung, but it looks like that open-up of the /work directory has succeeded, with the upper-case S changing to a lower-case s – and now has the same permissions (for example) as Grenville’s

drwxr-s— 17 gmann n02 4096 Aug 24 23:37 gmann/

drwxr-s— 36 grenvill n02 4096 Aug 4 13:48 grenvill/

Thanks again,

Best regards,

Cheers
Graham

11

Hi Sandip,
cc: Luke, Ros,

Just to let you know that I got the ERA5 nudging code-changes included over the weekend,
and those seem to be working OK.

I implemented the code-changes in a branch that “tacks on” the ERA5 directory paths in
addition to the code-changes in Simon Wilson’s v8.4 branch for the nudging_tropopause_fix

And then the v8.4 branch I created has first commit-point as merging in those changes, and
then subsequently the code-changes for the ERA5 nudging.

Rather than implementing the changes in the same way as in the v11.5 branch, I followed a more
expedient approach, essentially adding a 3rd option for L137, and following the approach that had
been followed when the L91 option was added, in addition to the original L60.

But I referred to the v11.5_nudge_use_ERA5 branch when I added the code-changes in.

The branch I created is called vn8.4_nudging_tropopause_fix_OptionToUseERA5 :

UM/branches/dev/gmann/vn8.4_nudging_tropopause_fix_OptionToUseERA5

https://pumatest.nerc.ac.uk/trac/UM/browser/UM/branches/dev/gmann/vn8.4_nudging_tropopause_fix_OptionToUseERA5

I will post the above info also on the FCM helpdesk ticket

The run xphy-m is the initial 2-month ERA5-nudged test run I got going.

Then I managed to do a “work-around” for the problem of getting past 2010, which involves using
GHG trace gas mmr values for 2010, but with climatological SST and sea-ice ancils, and the anthrop
SO2 emissions and single-level user-ancil both set at 2000 values (as in the RJ4.0 TS2000 job).

That work-around will work OK for stratospheric aerosol simulations , because we’re not too fussed
in these simulations for Hunga-Tonga (and other recent eruptions) about getting the tropospheric
oxidising capacity and trop-aerosol regional-hazes exactly as at present – year-2000 for these is OK.

And then the xphy-w job is a job that continues in 2010 with ERA5-nudging.

I am just trying runs in Dec 2019 and Sep 2020, the latter of which will then be the job that I’ll continue
on through to the current mid-2022 (with Hunga-Tonga SO2 emission on 15th Jan 2022).

That will then have run through the specific 2020/21 Arctic polar vortex and 2021 Antarctic polar vortex and will continue then nudging to have the specific 2021/22 Arctic polar vortex and 2022 Antarctic polar vortex.

Thanks again for your help with this – I guess there may be other groups may be interested in picking up this ERA5-nudged GA4 CheST+GLOMAP model for a range of different science.

The inability to nudge to recent years was quite a bad limitation, but with this workaround solution that issue is now resolved, to be able to (in our case) run case studies of recent eruptions such as Raikoke and Hunga-Tonga.

Best regards,

Cheers
Graham