iiinova — where your digital self keeps evolving

Reposted by Lukas Brunner

Lukas Brunner @lukasbrunner.bsky.social · 22/09/2026

Nice news article from the CLICCS cluster of excellence on our new study introducting the Variability Atlas 👇 @uni-hamburg.de @esrah.uni-hamburg.de www.cliccs.uni-hamburg.de/about-cliccs...

cliccs.uni-hamburg.de

New online tool helps forecast heatwaves and torrential rains

112

Lukas Brunner @lukasbrunner.bsky.social · 22/09/2026

Here's a short thread on the Variability Atlas with the most important links bsky.app/profile/luka...

001

Lukas Brunner @lukasbrunner.bsky.social · 22/09/2026

Nice news article from the CLICCS cluster of excellence on our new study introducting the Variability Atlas 👇 @uni-hamburg.de @esrah.uni-hamburg.de www.cliccs.uni-hamburg.de/about-cliccs...

cliccs.uni-hamburg.de

New online tool helps forecast heatwaves and torrential rains

112

Lukas Brunner @lukasbrunner.bsky.social · 21/09/2026

Thanks a lot to my co-authors @vidurm.bsky.social Leonard Borchert @janasillmann.bsky.social and Benjamin Poschlod! Thanks also to the World Data Center for Climate (www.wdc-climate.de) for hosting our data, and to Eileen Hertwig for her help in the publication process.

wdc-climate.de

World Data Center for Climate

000

Lukas Brunner @lukasbrunner.bsky.social · 21/09/2026

Main message for the community: internal variability can matter (a lot) for the estimation of extreme indices! We give an overview and discuss some cases, while the Variability Atlas aims to be a user-centered tool for investigating case-specific factors. Paper: iopscience.iop.org/article/10.1...

Effect of internal variability on TX90p for the example of Europe. TX90p from (a) a single member (picked here as the member with median region mean frequency for demonstration purposes) and the 50-member (b) mean, (c) minimum, and (d) maximum as well as from the member with (e) lowest and (f) highest region mean frequency. 50-member (g) standard deviation and (h) coefficient of variation (CV).

100

Lukas Brunner @lukasbrunner.bsky.social · 21/09/2026

For those more proficient in Python and/or in need of more control over metrics, regions, and periods, we also provide: - a GitHub repository with all code used in the paper - an online repository with all ETCCDI indices for the full 50-member MPI ensemble lukasbrunner.github.io/variability_atlas

lukasbrunner.github.io

Variability Atlas for ETCCDI climate extreme indices

The Variability Atlas provides a toolbox to investigate the effects of internal climate variability on different climate extremes. It is introduced and described in the accompanying publication from B...

100

Lukas Brunner @lukasbrunner.bsky.social · 21/09/2026

At the highest level, we provide a fully interactive online Shiny app for an easy overview - no coding experience needed! lukasbrunner.github.io/variability_...

130

Lukas Brunner @lukasbrunner.bsky.social · 21/09/2026

🚨New publication How does internal climate variability affect the calculation of the ETCCDI temperature and precipitation extreme indices? In our paper, we introduce the Variability Atlas, a flexible toolset to address this question for users of all experience levels! doi.org/10.1088/1748...

doi.org

The variability atlas: how internal climate variability affects the estimation of climate extreme indices

The variability atlas: how internal climate variability affects the estimation of climate extreme indices, Brunner, Lukas, Mithal, Vidur, Borchert, Leonard F, Sillmann, Jana, Poschlod, Benjamin

175

Reposted by Lukas Brunner

ESRAH University of Hamburg @esrah.uni-hamburg.de · 10/09/2026

🌇 How different is today's temperature in your hometown from what used to be normal? The new @reuters.com #Climate Monitor lets you check for any place, updated daily. Developed with climate experts, including our #CLICCS researcher @lukasbrunner.bsky.social. Check it out: lnkd.in/efct6nWq

085

Lukas Brunner @lukasbrunner.bsky.social · 03/08/2026

Vorteil des in der App verwendeten gegitterten Datensatzes ist die globale und historische Abdeckung

100

Lukas Brunner @lukasbrunner.bsky.social · 03/08/2026

In der App geht es aber (in meiner Interpretation) auch hauptsächlich um den Unterschied zum historischen Durchschnitt. Bei diesem Unterschied würde ich eine deutliche kleinere Abweichung zwischen Punktmessungen und Flächenmittel erwarten.

110

Lukas Brunner @lukasbrunner.bsky.social · 03/08/2026

Bevor jemand fragt: Der Unterschied in der absoluten Temperatur liegt zb an der größeren Region im historischen Vergleich wie durch die schwarze Box im Screenshot angedeutet

210

Lukas Brunner @lukasbrunner.bsky.social · 03/08/2026

Neuer Hitzerekord in Wien mit über 40°C und mehr als 10°C über den durchschnittlichen Temperaturen für diese Jahreszeit 🥵 Artikel von @derstandard.at Historischer Vergleich von @reuters.com

Screenshot der Startseite von "Der Standard": Ein Foto zeigt eine ausgetrocknete, vertrocknete Wiese mit Bäumen im Hintergrund. Die Schlagzeile lautet "Neuer Hitzerekord: Erstmals 40,1 Grad in Wien", mit Hinweis auf extreme Hitze auch an den Folgetagen und der höchsten Wetterwarnstufe im Osten Österreichs. Darunter folgen Kacheln mit weiteren Nachrichten.

Screenshot eines Reuters-Klimamonitors: Eine orange eingefärbte Karte Österreichs zeigt Wien mit einer heutigen Höchsttemperatur von 36°C, 10,9 Grad über dem historischen Durchschnitt. Ein Diagramm daneben zeigt den Temperaturverlauf 2026 (orange) durchgehend über dem historischen Mittel (schwarz) von Januar bis August. Unten eine Farbskala von -5° bis +5° zur Einordnung der Abweichung.

110

Reposted by Lukas Brunner

Infopoint #Klimawandel #Artensterben @infopointa.bsky.social · 02/08/2026

"Es ist halt Sommer" 🙃 Schönes Tool, schaut mal rein Wie die heutigen Temperaturen im Vergleich zum historischen Durchschnitt ausfallen Reuters Climate Monitor www.reuters.com/graphics/CLI...

Screenshot Reuters Climate Monitor
Temperaturabweichung 02.08.2026 gegenüber 1961-1990
München +6,9°C
Link: https://www.reuters.com/graphics/CLIMATE-AUTOMATED/MONITOR/akpeykqqapr/?lat=48.137&lng=11.576&zoom=6&place=Munich

Screenshot Reuters Climate Monitor
Temperaturabweichung 02.08.2026 gegenüber 1961-1990
Brüssel +8.0C
Link: https://www.reuters.com/graphics/CLIMATE-AUTOMATED/MONITOR/akpeykqqapr/?lat=50.8551&lng=4.3512&zoom=6&place=Brussels

Screenshot Reuters Climate Monitor
Temperaturabweichung 02.08.2026 gegenüber 1961-1990
London +6,8°C
https://www.reuters.com/graphics/CLIMATE-AUTOMATED/MONITOR/akpeykqqapr/?lat=51.5073&lng=-0.1276&zoom=6&place=London

078

Lukas Brunner @lukasbrunner.bsky.social · 02/08/2026

@reuters.com has a new tool to compare today's weather in a climate context. Pretty cool to play around with and pretty sobering at the same time 🥵 www.reuters.com/graphics/CLI...

reuters.com

Reuters Climate Monitor

How today's temperatures compare to the historic average

0167

Reposted by Lukas Brunner

Thomas Rackow 🧊 @trackow.bsky.social · 27/05/2026

Great to see the km-scale models perform well in Lukas' comparison! We bring the same models into an operational framework @ec.europa.eu, with a resolution of 5km. To get started with the latest generation, please look at platform.destine.eu/climate-dt/ @ecmwf.int @esa.int @eumetsat.int

1114

Reposted by Lukas Brunner

Gavin Schmidt @climateofgavin.bsky.social · 22/05/2026

This is a quick summary of the latest CMIP process. It is a complex beast and a consensus project so there are always things that you’ll agree with and those that you might not. 1/

14324

Reposted by Lukas Brunner

Thomas Rackow 🧊 @trackow.bsky.social · 19/05/2026

Happy to see the new #DestinE simulations out in the open! I’d strongly recommend trying the new Jupyter example notebooks, including lazy browsing: github.com/destination-... Looking forward to seeing what people will build and discover with this next generation of #ClimateDT simulations!

github.com

polytope-examples/climate-dt at main · destination-earth-digital-twins/polytope-examples

This repository describes the process for accessing Destination Earth DT data via the Polytope web service hosted on the LUMI Databridge. - destination-earth-digital-twins/polytope-examples

092

Reposted by Lukas Brunner

Lukas Brunner @lukasbrunner.bsky.social · 08/05/2026

If you are like me and prefer to read the typeset version of papers, here you go: the final version is out fresh from the print 😊 www.nature.com/articles/s43...

Screenshot of the title page of the article entitled: Three decades of simulating global temperature patterns with coupled global climate models

1136

Lukas Brunner @lukasbrunner.bsky.social · 08/05/2026

Here's my thread on the main results bsky.app/profile/luka...

010

Lukas Brunner @lukasbrunner.bsky.social · 08/05/2026

If you are like me and prefer to read the typeset version of papers, here you go: the final version is out fresh from the print 😊 www.nature.com/articles/s43...

1136

Reposted by Lukas Brunner

Lukas Brunner @lukasbrunner.bsky.social · 21/04/2026

🚨 New paper: Three decades of simulating global temperature patterns with coupled global climate models 💡 Some of the latest km-scale models perform similarly to obs 💡 Reference choice starts to dominate for the latest models 💡 Resolution matters but is not everything www.nature.com/articles/s43...

nature.com

Three decades of simulating global temperature patterns with coupled global climate models - Communications Earth & Environment

Increasing spatial resolution of climate models has the potential to improve the performance of simulating global surface temperature patterns, based on comparisons between kilometer-scale models, est...

2244

Lukas Brunner @lukasbrunner.bsky.social · 23/04/2026

Thread with main results: bsky.app/profile/luka...

000

Lukas Brunner @lukasbrunner.bsky.social · 23/04/2026

Modern climate models are more accurate than ever : Newsroom : University of Hamburg share.google/SQofonppS6Nl...

share.google

Modern climate models are more accurate than ever

141

Lukas Brunner @lukasbrunner.bsky.social · 22/04/2026

This figure from the supplement of our new paper (see quoted post for details) shows a quite interesting side-result that reverses traditional model evaluation: We investigate which observation-based dataset is closest to how many models and find systematic changes over model generations.

$Fraction of models closest to each reference. Shown is the fraction of models for which each reference dataset (shown on the x-axis) leads to the lowest RMSD.$

0100

Reposted by Lukas Brunner

ESRAH University of Hamburg @esrah.uni-hamburg.de · 21/04/2026

🌡️ How good are #climate models, really? A new study by @lukasbrunner.bsky.social at #CLICCS @uni-hamburg.de examined 176 climate models from 30 years – and shows: the latest models come so close to reality that the measurement data, not the models, are now the limiting factor. 🧪 shorturl.at/j1kbs

0134

Reposted by Lukas Brunner

Universität Hamburg @uni-hamburg.de · 21/04/2026

Wie heiß wird es in der Sahara, wie kalt in der Antarktis? 🌡️ Klimamodelle berechnen die Temperatur-Landkarte der Erde. Aber wie gut sind diese? Eine Studie von Dr. @lukasbrunner.bsky.social zeigt: Im Schnitt sind die Temperatur-Landkarten der Klimamodelle deutlich genauer und zuverlässiger geworden.

uni-hamburg.de

Moderne Klimamodelle sind so genau wie noch nie

052

Lukas Brunner @lukasbrunner.bsky.social · 21/04/2026

Hier geht's zur Pressemitteilung der @uni-hamburg.de über unser gerade erschienenes Papier zum Vergleich von globalen Klimamodellen in den letzten 3 Jahrzehnten. 🧑‍🎓 www.uni-hamburg.de/newsroom/pre...

uni-hamburg.de

Moderne Klimamodelle sind so genau wie noch nie

0105

Lukas Brunner @lukasbrunner.bsky.social · 21/04/2026

Thanks so much to all the people that helped with this, in particular my co-authors Rohit Ghosh, Leopold Haimberger, Cathy Hohenegger, Dian Putrasahan, Thomas Rackow @trackow.bsky.social , Reto Knutti @retoknutti.bsky.social and Aiko Voigt @aikovoigt.bsky.social 🙏🧪🧑‍🎓

150

Lukas Brunner @lukasbrunner.bsky.social · 21/04/2026

We provide an extensive supplement to reveal model differences, also within model generations and spatial resolved the full database is on Zenodo. doi.org/10.5281/zeno...

Root-mean-square distance (RMSD) to the full reference range for all 176
models individually.

130

Lukas Brunner @lukasbrunner.bsky.social · 21/04/2026

We find a clear improvement in model performance with spatial resolution, but note that several other factors (representation of processes, tuning) are also crucial.

Model performance as a function of model resolution. RMSD to the full
reference range as a function of model resolution for all 165 CMIP and 11 km-scale models.

120

Lukas Brunner @lukasbrunner.bsky.social · 21/04/2026

Using a single reference as ground truth for the evaluation of the model temperature we find that ❗For the latest models the reference choice can determine up to 40% of the model bias ❗Models from different generations are systematically closer to certain references Paper: doi.org/10.1038/s432...

Difference between maximum and minimum RMSD divided by the mean RMSD. The RMSD is calculated for each model using each of the 10 different reference datasets in turn. Difference in RMSD when using one of two references: ERA-Interim or its successor ERA5.

140

Lukas Brunner @lukasbrunner.bsky.social · 21/04/2026

We find that km-scale models have the *potential* to match cross-evaluated references (some versions of IFS outperform several reference datasets) but in many cases, they still need considerable work to realize their full potential.

Fraction of Earth’s land surface area for which the models' 20-year mean surface temperature falls within the range of the 10 reference. Two models are highlighted: IFS-FESOM from nextGEMS cycle 4 and from EERIE phase 1.

141

Lukas Brunner @lukasbrunner.bsky.social · 21/04/2026

We investigate the global pattern of surface air temperature from - 176 fully coupled climate models - developed in 5 model generations (CMIP2 to the latest km-scale models) - and 3 decades - compared to 10 observational references

Fraction of Earth’s land surface area for which the models' 20-year mean surface temperature falls within the range of the 10 references and area-weighted root-mean-square distance (RMSD) relative to the reference range.

130

Lukas Brunner @lukasbrunner.bsky.social · 21/04/2026

nature.com

Three decades of simulating global temperature patterns with coupled global climate models - Communications Earth & Environment

2244

Lukas Brunner @lukasbrunner.bsky.social · 28/10/2025

Letzte Woche durfte ich im Rahmen der Ringvorlesung zur Klimakrise der @uni-hamburg.de einen Vortrag zu den naturwissenschaftlichen Grundlagen der Klimakrise geben. Vielen Dank fürs Dabeisein und für die spannenden Fragen! Zum Nachschauen 👇 youtu.be/IuTXbtR1fFw #SystemChangeNotClimateChange

youtu.be

Naturwissenschaftliche Grundlagen der Klimakrise - Lukas Brunner | Fridays for Future

YouTube video by Fridays for Future DE

273

Lukas Brunner @lukasbrunner.bsky.social · 26/10/2025

Vielen Dank für das Interesse und die angeregte Diskussion an alle die (online) dabei waren! Der Vortrag ist zum nachsehen jetzt auch auf YouTube 💻 Die zugehörigen Slides auf meiner Homepage: lukasbrunner.github.io/outreach youtu.be/IuTXbtR1fFw

youtu.be

Naturwissenschaftliche Grundlagen der Klimakrise - Lukas Brunner | Fridays for Future

YouTube video by Fridays for Future DE

050

Lukas Brunner @lukasbrunner.bsky.social · 17/10/2025

Am kommenden Dienstag gebe ich im Rahmen der Ringvorlesung zur Klimakrise an der @uni-hamburg.de einen Vortrag zu den physikalischen Grundlagen der Klimakrise. Der Vortrag ist öffentlich, also gerne vorbeikommen! Freue mich auf den Austausch! #climatecomms www.zfw.uni-hamburg.de/oeffentliche...

zfw.uni-hamburg.de

Physikalischen Grundlagen

2126

Lukas Brunner @lukasbrunner.bsky.social · 02/09/2025

Comparing global km-scale climate models to the resolution of coarser CMIP6 models. The video shows precipitation flux and accumulated precipitation for 2025, ending with a quantification of the difference in resolutions. Animation: Michael Böttinger (DKRZ) Method & paper: bsky.app/profile/luka...

051

Lukas Brunner @lukasbrunner.bsky.social · 21/08/2025

See also the amazing visualization made by DKRZ based on our work 👇👇 bsky.app/profile/luka...

021

Lukas Brunner @lukasbrunner.bsky.social · 19/08/2025

Here's also my thread on the main findings: /profile/lukasbrunner.bsky.social/post/3lrsdhwkev22y

040

Lukas Brunner @lukasbrunner.bsky.social · 18/08/2025

Here is the thread to the original publication: bsky.app/profile/luka...

020

Lukas Brunner @lukasbrunner.bsky.social · 18/08/2025

Amazing visualization of our recent sub-grid variability work by Michael Böttinger (DKRZ) 🤩. It shows the local detail in global precipitation resolved by km-scale models. From 0:50 in direct comparison to CMIP6-type 100km. From 2:07 the sub-grid variability is shown. 🤓🧪 youtu.be/fzA92tSF4ZM?...

youtu.be

The spatial representation of precipitation - a comparison between km-scale and CMIP-type model data

YouTube video by Deutsches Klimarechenzentrum

192

Lukas Brunner @lukasbrunner.bsky.social · 03/08/2025

Local downpour near Oberstdorf, Germany ⛰️

020