Duncan Watson-Parris

Huge congrats to co-leads Ellen H. Davenport & J. Varan Madan!

📦 pip install jcm 📄 Paper: egusphere.copernicus.org/preprints/20... 💻 Code: github.com/climate-anal...

We're just getting started. 🔮

This release sets the stage for a fully coupled, differentiable Earth System Model. We are actively developing this next phase with colleagues at U Copenhagen, Columbia, DLR, and AI2.

Watch this space...

Key Innovations: ⚡️ True Differentiability: Direct gradients for calibration & ML integration. 🧩 Modular: Couples SPEEDY physics to the Dinosaur dynamical core. 🚀 Portable: Seamless on CPUs, GPUs, & TPUs. ✅ Validated: Matches Fortran SPEEDY at T31.

Bridging Physics & AI in Climate Modeling 🌍🤖

Introducing JCM v1.0: A fully differentiable, intermediate-complexity atmospheric model built in Python/JAX.

Training hybrid models is hard because legacy code lacks gradients. JCM solves this.

Ever wondered whether there's a principled way to calibrate all those parameters controlling climate models? In our new paper we show how to calibrate climate model parameterizations using ideas from Bayesian experimental design: doi.org/10.1088/3049... . (1/N)

Just finished @hannahritchie.bsky.social's fantastic book "Not the End of the World" which nicely articulates the things we should, and should not, be worrying about when it comes to sustainability. It's a refreshing and empowering take backed by the latest data. Definitely recommend!

With Philip Stier, Laure Zanna and Tom Beucler
#Climate #ML

AI for Good, All Year - Always Online Our AI powered matchmaking community platform to connect AI innovators with problem owners.

Chris gave one of our opening talks a few years ago and I can't wait to hear about his vision for the future of the field.

The talk will be live-streamed on 23 September at 17:00 CEST! Sign up for free here: neuralnetwork.aiforgood.itu.int/event/ai-for...

Excited to be kicking off another season of "AI and Climate Science" webinars as part of the UN 'AI for Good' Discovery series this quarter with a talk from Chris Bretherton reflecting on 'AI for climate modeling from present to future'.

The Weather & Climate Livestream Join us starting May 28th, as meteorologists and climate scientists from across the US share their research and answer your questions.

One hour into the #Weather and #Climate Livestream! You can join any of the remaining 99 hours right here: wclivestream.com

Climate and weather scientists are joining the anti-Trump resistance in the most ‘scientist-iest’ way | CNN In the face of steep funding cuts for climate and weather research and forecasting, scientists have found a new way of engaging the public.

In the face of steep funding cuts for climate and weather research and forecasting, scientists have found a new way of engaging the public: livestreamed presentations.

Please consider submitting an abstract to the session on 'Understanding reactive gases, aerosols, and land use for air quality and climate change' convened by Fiona O'Connor, @dwatsonparris.bsky.social, Stephanie Fiedler, @flossie-brown.bsky.social and me! Deadline August 13th.

Having made an abortive attempt with Threads I'm now trying BlueSky! Looking forward to interacting with a more positive and engaging community :-)

Industrial Pollution Can Increase Snowfall and Reduce Cloud Cover New research co-authored by Duncan Watson-Parris of UC San Diego’s Scripps Institution of Oceanography and the Halıcıoğlu Data Science Institute finds that industrial pollution in Earth’s colder latit...

Industrial pollution in colder regions increases snowfall and reduces cloud cover, potentially boosting global warming by reducing the sunlight that clouds reflect back into space, according to a new study from Scripps Oceanography researcher Duncan Watson-Parris. (@dwatsonparris.bsky.social)