In case you're working on NSF COA

Really thrilled to share this new publication on the AMORE 2.0 algorithm 💗 This goal of this EPA-funded project was to develop an automated way to reduce the complexity of atmospheric chemical mechanisms using the principles of graph theory academic.oup.com/pnasnexus/ar...

We have a new preprint out! essopenarchive.org/doi/full/10.... In it, Xiaokai Yang, Lin Guo and I explore a method for making atmospheric chemistry surrogate models which are guaranteed to not have exploding errors. This is a step toward generally useful machine-learned models of the atmosphere.

🔔 New Master's internship opportunity 🔔 Come work with me this summer (from the comfort of your own home!) on global urban modeling at Pacific Northwest National Laboratory! Details & apply: careers.pnnl.gov/jobs/10599?l... Due Monday, April 14! More info in 🧵 [1/4]

How to Access the EOT Web Archives Visit web.archive.org Scroll to 🔎Collection Search section Select "End Of Term (US Gov 2024)" from the menu Type in your query term and click search Or download bulk WARC files from eotarchive.org/data/ #EOT2024 #EOTArchive

We have a new paper just published about using machine learning to estimate uncertainty in atmospheric chemistry modeling! dx.doi.org/10.1029/2024...

Research scientist position on AI weather forecasting at UChicago's AI for Climate (AICE) initiative to work with me, Amir Jina, and Michael Kremer. Will work across our new Human-Centered Weather Forecasting Initiative + @dsi-uchicago.bsky.social + Climate Institute

In retrospect, I should have started preparing this material a long time ago...

I did a bluesky thing and made a "list" of atmospheric chemists! Share with your friend and let me know if you want to be added! bsky.app/profile/did:...

We have a new preprint led by student Lin Guo! It is about uncertainty quantification in atmospheric chemistry modeling: arxiv.org/abs/2407.09757

I was quoted in this article! www.npr.org/2023/10/17/1...

We have a new preprint out today! arxiv.org/abs/2309.11035 We find that replacing the advection operator with a machine-learned surrogate may help the model run faster and/or be more accurate in certain situations.

I'm quoted in this article! (Although the quote is me saying something that pretty much everyone already knows)