Glad you like Weather Lab, Matt!

We also look at all responses we get via the feedback button if you have ideas about how we can make the site more useful.

Research Engineer, AI for Sustainability Mountain View, California, US

We're hiring a Research Engineer based in San Fran or MTV, California, to join our sustainability & weather teams at GDM. Seeking strong engineers at the interface of ML, sustainability, weather, dynamical systems, and/or remote sensing: boards.greenhouse.io/deepmind/job...

📸: Our researcher Kate Musgrave presenting on AI weather predictions

#AMS2025 | @ametsoc.bsky.social

I'll be presenting GenCast, recently published in Nature, tomorrow Tuesday 8h45AM at #AMS2025 in room 339.

GenCast is a diffusion model that outperforms ENS, the top operational ensemble forecast, giving skillful probabilistic forecasts up to 15 days ahead.

ams.confex.com/ams/105ANNUA...

Ferran's GenCast talk on Tuesday morning at #AMS2025 is not one to miss! ams.confex.com/ams/105ANNUA...

I'll be at #AMS2025 next week alongside some other Google DeepMind colleagues behind GenCast/GraphCast. Excited to meet people, discuss ML for weather, and learn!

Welcome, Jeff, and thanks for the links!

Can everyone come over from LinkedIn now 😅 There’s still more of the ML/Earth sciences community active on there than BlueSky I feel

dynamical tests would be in my top 3 too :-)

Elon Musk on X: "Mars will be called the “New World”, just as America was in past centuries. Such an inspiring adventure!" / X Mars will be called the “New World”, just as America was in past centuries. Such an inspiring adventure!

See: twitter.com/elonmusk/sta...

Yeah ofc there are indirect benefits from crewed space exploration and non-Earth-related spacecraft, even if just for the sake of knowledge

I’m speaking more to a shift in myself while at uni, realising there are too many urgent problems on Earth for some Muskian Mars colonisation project lol

I love how William Shatner from Star Trek went to space and realised how horrible it is in contrast to our beautiful home planet.

My life goal used to be to help get humans to Mars. I’m so glad I realised how special Earth is and now work on better living here rather than leaving 🌍

See you at AMS!

It's @ecmwf.bsky.social keeping the tradition this year 😉

bsky.app/profile/rasp...

I ran into this as well - there is a separate button for uploading animations (next to the static image one).

Nice work by the way!

Neural general circulation models optimized to predict satellite-based precipitation observations Climate models struggle to accurately simulate precipitation, particularly extremes and the diurnal cycle. Here, we present a hybrid model that is trained directly on satellite-based precipitation obs...

Can incorporating AI improve precipitation in global weather and climate models?

Yes! In the latest NeuralGCM paper, we show that training on satellite-based precipitation results in significant improvements over traditional atmospheric models:
arxiv.org/abs/2412.11973

My personal opinion (and I think general consensus) is that while purely data-driven modelling already excels from nowcasting to medium-range timescales, it is not the right paradigm for climate forecasting. Very fast ML/physics hybrid GCMs like NeuralGCM will be the way to go.
cc @stephanhoyer.com

Actual rollout schematic animation here:

Looks like Bluesky has a separate button for videos that I missed 🙃 Actual Milton animation here:

Interested in AI weather/climate modeling at #AGU24?

I'll be giving an overview talk on NeuralGCM at 11:30am Wed at the Google booth, and an talk on modeling precipitation with NeuralGCM at 4:25pm Wed in the session A34A.

It's been an honour to work on this study led by Ilan Price with such a talented team ✨: Alvaro Sanchez Gonzalez, Ferran Puig, Andrew El-Kadi, Dominic Masters, Timo Ewalds, Jacklynn Stott, @shakirm.bsky.social, Peter Battaglia, Rémi Lam, & Matthew Willson

Like its predecessor (GraphCast), the weights & code of GenCast have been made publicly available: github.com/google-deepm...

We’re looking forward to seeing how the community builds on this!

A GenCast ensemble member takes 8 minutes on a TPU chip, versus hours on a supercomputer for physics-based models. This opens up the possibility of large ensembles (eg 1000s of members), which could better estimate risks of extreme events. We don't yet know how much this will help.

Cyclone max wind speeds are still underestimated, but this performance on tracks is really promising.

One recent devastating cyclone was Hurricane Milton, which caused >$85 billion in damages. GenCast predicted ~70% probability of landfall in Florida 8.5 days before it struck.

We also extracted cyclone tracks from GenCast and ENS and compared them with ~100 cyclones observed in 2019. GenCast's ensemble mean cyclone track has a 12-hour position error advantage over ENS out to 4 days, and more actionable track probability fields out to 7 days.

For example, we created a dataset of simulated wind power data at wind farm sites across the globe, and found that GenCast outperforms ENS by 10–20% up to 4 days ahead. This is promising, because better weather forecasts can reduce renewable energy uncertainty and accelerate decarbonisation.

It’s vital that we ensure these new ML weather systems are safe and reliable. One thing I'm proud of is our range of evaluation experiments: per-grid-cell skill & calibration, spatial structure, renewable energy, extreme cold/heat/wind, and the paths of tropical cyclones (i.e. hurricanes).

GenCast uses diffusion to generate multiple 15-day forecast trajectories for the atmosphere. It assigns more accurate probabilities to possible weather scenarios than the SoTA physics-based ensemble system from ECMWF, across a 2019 evaluation period.

Probabilistic weather forecasting with machine learning - Nature GenCast, a probabilistic weather model using artificial intelligence for weather forecasting, has greater skill and speed than the top operational medium-range weather forecast in the world and provid...

So excited to share our Google DeepMind team's new Nature paper on GenCast, an ML-based probabilistic weather forecasting model: www.nature.com/articles/s41...

It represents a substantial step forward in how we predict weather and assess the risk of extreme events. 🌪️🧵