ESMValTool community workshop 11-13 May 2026

[alistairsellar]

Dear ESMValTool community and @ESMValGroup/esmvaltool-developmentteam,

The next ESMValTool community workshop will take place 11-13 May 2026. The workshop will be held at the Met Office in Exeter, UK, with online participation available also.

Agenda

All times are given in BST (UTC+1)

Remote joining instructions have been shared with registered participants by email.

11 May 2026 - Day 1 (Monday)

Morning: Conference room 3 Chair: Birgit Hassler

9:00 am Arrival, registration, coffee
9:30 am Introduction, welcome and workshop agenda (Alistair Sellar)
9:50 am User Engagement Team update and recent requests (Lisa Bock)

10:30 am Coffee

11:00 am Demo: how to run ESMValTool with (fake) CMIP7 data (Bouwe Andela)
11:30 am Pixi introduction and demo (Valeriu Predoi)

12:00 – 1:00 pm Lunch

Afternoon: Collaboration space E1
1:00 – 5:00 pm Coding session
1.00 pm Introduction of coding topics
2:00 pm Coffee

(3:00 pm Parallel session: User Engagement Team Meeting E1-2)

4:00 pm Summary and wrap up of day 1 & discussion (Birgit Hassler)

12 May 2026 - Day 2 (Tuesday)

Morning: Conference room 3 Chair: Ranjini Swaminathan

9:00 am Introduction day 2 and workshop agenda (Birgit Hassler)
9:15 am CMIP7 data

• Changes from CMIP6: introduction, demo, discussion (Matt Mizielinski)
• Status & plans for the CMIP Rapid Evaluation Framework (Birgit Hassler)

10:30 am Coffee

11:00 am ESMValTool consortium updates (Alistair Sellar)

• Exciting news item on consortium membership
• Implementing a Contributor License Agreement – important information for all ESMValTool contributors

11:20 am Styles planned for IPCC AR7 figures public, e.g. icons, color maps. How to include them in ESMValTool outputs? (Katja Weigel)

11:40 am Workshop photo - outside Met Office reception

12:00 – 1:00 pm Lunch

Afternoon: Collaboration space E1
1:00 – 5:00 pm Coding session (all afternoon)
1.00 pm Introduction of coding topics

• CLA reminder

2:00 pm Coffee

(2:30 pm Parallel session: Tech Lead Team Meeting E1-4)

4:00 pm Summary and wrap up of day 2 & discussion (Alistair Sellar)

7:00 pm Workshop dinner. Crave restaurant, Exeter.

13 May 2026 - Day 3 (Wednesday)

Morning: Conference room 3 Chair: Axel Lauer

9:00 am Introduction day 3 and workshop agenda (Alistair Sellar)
9:15 am Configuring data sources and emerging support for data in cloud-native formats (#2690, #2917) (Bouwe Andela)
10:00 am ESMValTool strategy - plans for a journal article (Birgit Hassler)

10:30 am Coffee

11:00 am Technical update (Valeriu Predoi)
Request to discuss - managing recipes for CMIP6 and CMIP7 and revisiting obs data storage and organization on HPC servers (rswamina).
12:00 – 1:00 pm Lunch

Afternoon: Collaboration space E1
1:00 – 5:00 pm Coding session (all afternoon)
1.00 pm Introduction of coding topics
2:00 pm Coffee
4:00 pm Summary and wrap up of day 3 & discussion (Birgit Hassler)

Context and aims

ESMValTool is a community diagnostic and performance metrics tool for evaluation and analysis of Earth system models. ESMValTool workshops are for the whole community, to participate in discussions that shape the future of our community, and to work with ESMValTool to evaluate and analyse climate model data. The workshop schedule will have discussions in the (UK/EU) mornings and coding / drop-in sessions in the afternoons where you can seek help from experienced users and developers.

The focus of this workshop will be getting ready for evaluation and analysis of CMIP7 data. So please join us if you plan to use CMIP7 data and would like to take advantage of ESMValTool’s multi-model support and powerful data processing capabilities.

Visiting the UK
From the UK government website: “Most visitors travelling to the UK need an ETA or a visa... You usually need an ETA rather than a visa if you’re from Europe, the USA, Australia, Canada or certain other countries.”
You can use this site to check whether you need an ETA or a visa: https://www.gov.uk/check-uk-visa.
Please contact Alistair Sellar if you have any questions.

Travelling to Exeter
The Met Office website has useful information about travelling to our Exeter office:
https://www.metoffice.gov.uk/about-us/contact/how-to-find-our-offices#Exeter
The Met Office is about 5km from the city centre. There are 3 hotels within 2km of the office. However, buses in and out of the city centre are more reliable in the day time, so if you plan to socialise in the evening (!?) I would recommend staying at one of the many hotels in the city centre.

Thanks,
Alistair on behalf of the ESMValTool Steering Group

[bouweandela]

Here are some topics I could give a talk about:

• Introduction to using CMIP7 data with ESMValTool Add preliminary CMIP7 support ESMValCore#2935
• Configuring data sources and the emerging support we have for data in cloud-native formats Intake-ESM Integration based on #1218 ESMValCore#2690, Add xcube support ESMValCore#2917

[k-a-webb]

Perhaps this was already a consideration, but if not, it would be great to discuss support for CMIP7 data generated by different versions of the data request.
This is something I've just started working on this week.

[bouweandela]

Our current strategy is to only support the latest version of the data request. Typically, changes are fairly minor, e.g. a new long name, and can be fixed automatically in data generated with an older version of the data request by GenericFix. Do you think this will work for CMIP7 too?

[bettina-gier]

Considering the topic, if you want some ideas for the more practical parts of the workshop: It's been brought up several times that a lot of our recipes are still based on CMIP5 data. We could identify a few that we want to use for CMIP7 and then see if they work out of the box or if some diagnostics need tweaks for the expected CMIP7 data format (or live testing with data if any are available by then)

[bouweandela]

If we have no CMIP7 data available yet, we can fake some CMIP7 data by using the Simple_recmorise_cmip6-cmip7.ipynb notebook to CMORize some CMIP6 data according to the CMIP7 standards.

[bouweandela]

It would be nice to put some focus on the REF recipes too. Especially regarding the timerange for the CMIP7 historical period (until 2021 instead of 2014) and the use/availability of reference observational/reanalysis datasets. Climate-REF/climate-ref#536

[alistairsellar]

This REF suggestion got a few thumbs-up, and I like it to. To help motivate and give context for any coding, it would be good to include an update on REF status and plans as a morning topic. I've pencilled in a topic on this on Monday morning. OK for you @bouweandela?

[bouweandela]

Sure!

[LisaBock]

Requests by the UET:

• One slot in the plenum to discuss UET updates and recent requests, e.g. ESMValTool tutorials
• Dedicated meeting (1hour) for the UET to align on specific improvements of the tutorial and allocate coding tasks for the workshop

[LisaBock]

Please schedule UET-related topics preventively for Monday or Tuesday, as I might not be around on Wednesday.

[katjaweigel]

There are already some styles planned for IPCC AR7 figures public, e.g. icons, color maps: see IPCC-AR7-WG1/figure-design can we link/include them?

[alistairsellar]

Hi all. Registration for the ESMValTool community workshop is now open!
Please register for the workshop now at https://forms.cloud.microsoft/e/FT766GmUmw. Registration closes on Thursday, April 23.

Travelling to Exeter
The Met Office website has useful information about travelling to our Exeter office:
https://www.metoffice.gov.uk/about-us/contact/how-to-find-our-offices#Exeter
The Met Office is about 5km from the city centre. There are 3 hotels within 2km of the office. However, buses in and out of the city centre are more reliable in the day time, so if you plan to socialise in the evening (!?) I would recommend staying at one of the many hotels in the city centre.

Thanks for the suggestions in the discussion above. Keep them coming... (but register first!)

[valeriupredoi]

usual technical update from Old V, please, at a hangover-friendly hour would be well appreciated 😁

[alistairsellar]

Hi all, a final reminder to register for the ESMValTool workshop. The deadline is tomorrow! (Thursday 23 April)
Please register for the workshop now at https://forms.cloud.microsoft/e/FT766GmUmw.

Also, thanks for the agenda suggestions. Please keep them coming...

[alistairsellar]

Hello again. Registration for the ESMValTool workshop has been extended to Sunday (26 April).
Please register for the workshop now at https://forms.cloud.microsoft/e/FT766GmUmw.

[alistairsellar]

Registration for the workshop is now closed. A confirmation email has been sent to registered participants. Please contact me (email address at top of this thread) if you think you registered but didn't get an email just now.

The agenda is under development. Please reply to this discussion ASAP if would like to suggest additional topics for the agenda.

[katjaweigel]

Dear @alistairsellar, I replied to you by email, that I did not get a confirmation mail? I registered for online participation.

[bouweandela]

As discussed at the workshop last year, we plan to switch from mamba/conda to pixi for managing our development environment. This will bring improved reproducibility and is more user-friendly, but there is some learning curve, so I would like to give a short (10 minutes) demo on the first day at the start of the coding session. Assuming we will have a coding session in the afternoon again? Then we can merge the pull requests that will implement this change, ESMValGroup/ESMValCore#3044 and ESMValGroup/ESMValTool#4407, and everyone can immediately start working with it.

[alistairsellar]

Nice suggestion @bouweandela. This feels like important information for all developers, so good to introduce in a morning talk, to reach people who attend the talks but not the coding session. Who could present that? Obviously you could do it but I'm conscious we've already got you down for several items. Or feel free to suggest other people for the other items that I've put your name on.

[bouweandela]

@valeriupredoi volunteered to present pixi

[valeriupredoi]

In collab with the pixifier @bouweandela 🍺

[alistairsellar]

Wonderful - thanks @valeriupredoi and pixifier @bouweandela!

[alistairsellar]

Hi all, really looking forward to seeing you in Exeter or online next week!

The draft agenda is now available at the top of this discussion. Please review and feedback if any issues.

• Named speakers / discussion leads, please check the timings - topic durations are a first guess and plenty of scope to change from what is drafted here. Tagging @bouweandela @LisaBock @katjaweigel @valeriupredoi
• Suggested session chairs, if you are happy to chair please edit to remove the ?. Tagging @hb326 @axel-lauer @rswamina

There are some available slots for discussions, so feel free to suggest additional topics.

[bettina-gier]

Don't see my name in the Agenda anywhere but I'm happy to take up anything as long as I know beforehand ;)

Gonna try and get a project board with current issues and PRs to the mentioned topics set up before the meeting for the coding sessions. The project already exists here if anyone wants to add their own Issues or PRs.

[alistairsellar]

Sorry - rushed use of autocomplete! You're off the hook (for now... your offer is noted).

Great plan on the project board!

[katjaweigel]

Dear @alistairsellar,

thanks for the draf agende! I can talk about "am Styles planned for IPCC AR7 figures public, e.g. icons, color maps. How to include them in ESMValTool outputs?" But I don't think this needs 30 Minutes, maybe rather 15?

[alistairsellar]

Thanks @katjaweigel! Timings updated

[malininae]

Since the schedule is a bit rough for our time zone, I will not join for intro. Hope to hear a bunch of progress in the general sessions though :-)

[valeriupredoi]

Question Answer

RE: @katjaweigel asked about pixi running in an sbatch SLURM job, and, for that I tested on JASMIN, job file:

#!/bin/bash
#SBATCH --job-name=myjob
#SBATCH --time=01:00:00
#SBATCH --cpus-per-task=4
pixi run --frozen --verbose esmvaltool --help

submitted as a small job: sbatch --account=ncas_cms --qos=debug --partition=debug spirun.sh runs fine and dumps the help output to the out file 🍺

[bettina-gier]

I followed this one: https://x-zang.github.io/blog/switch-from-conda-to-pixi/#using-pixi-on-a-job-scheduler-slurm
works on levante
and Katja has another working version ^^.

[katjaweigel]

Or half working, depending on where my native6 extra_facet issue comes from ... (still testing)

[valeriupredoi]

Question Answer

RE: @ehogan asking about overhead time most probably due to network storage and pixi:

$ vim .pixi/config.toml

• create it if it's not there, then see if this added there improves things:

[cache]
netfs-redirect = "never"