
Research scientist @allen_ai advancing climate science using machine learning. Coder, storm chaser, and all around weather nerd.
Seattle, WA
Joined February 2019
- Tweets666
- Following889
- Followers275
- Likes9.4K
TroyArcomano retweeted
According to polls, approximately 1% of Americans, many of them with severe incapacities, still think Joe Biden is president. It's disturbing to see that Ari Emanuel is among them and I hope everything is well with him.
What if you could simulate global extreme weather at 3 km resolution for ~1/300th the cost?
Our new HiRO-ACEv2 system does just that, while matching X-SHiELD’s statistics to within 3% even for events as rare as 1 in a million data points. 🧵
huggingface.co/allenai/HiRO-…
TroyArcomano retweeted
Replying to @UNLincoln
@UNLincoln didn’t just cut a program, they cut the future of Nebraska-born students and every student who lives in regions where weather and agriculture define life.
Eliminating the Earth & Atmospheric Sciences program in the middle of America’s ag heartland is reckless. These students live the reality EAS teaches: drought, hail, planting windows, harvest risk, rural safety, climate volatility. This isn’t theory to them, it’s home.
@RodneyDBennett and the UNL Board of Regents chose to dismantle the science that keeps farms operating, communities resilient, and students prepared to serve the industries that raised them. That’s not “prioritizing.” It’s abandoning agriculture, safety, and the students who depend on it.
Nebraska deserves better. Its students deserve better.
#UNL #UNLincoln #NebraskaAg #SaveEAS #WeatherScienceMatters #RuralStudentsMatter #StudentsDeserveBetter #UNLRegents #newx #wxtwitter
TroyArcomano retweeted
I will be at @NeurIPSConf next week to present OmniCast – our latest work on generative modeling for weather forecasting. OmniCast achieves competitive performance across different time scales, while being 10x to 20x faster than leading methods like NeuralGCM or GenCast.
TroyArcomano retweeted
Introducing Swift🕊️ a consistency model with 39× faster inference over diffusion baselines + the first probability flow model capable of autoregressive finetuning on domain-specific objectives.
A step toward extending our generative weather models to seasonal scales...
TroyArcomano retweeted
Excited to share our 2025 ACM Gordon Bell Finalist:
🌎 AERIS, our 1.3–80B parameter pixel-level Swin diffusion transformer, addresses scaling issues in high-resolution weather forecasting using SWiPe parallelism to scale to 121,000 GPUs.
What an absolute embarrassment to have gone to this university
A statement from Texas A&M University President Mark A. Welsh III:
I am reaching out today to follow up on a statement I distributed through social media yesterday regarding an ongoing issue with one of our 16,000 course sections. I directed the Provost to remove the College of Arts and Sciences Dean and English Department Head from their administrative positions.
This afternoon, following full consideration of the facts related to this situation, I directed the Provost to terminate the professor involved, effective immediately. Please understand that these decisions were mine alone and were not made lightly.
Read the full statement to faculty and staff: tx.ag/eAwKu7Y
45
TroyArcomano retweeted
The official NOAA "budget request" to Congress (which is a detailed version of the "President's Budget" would be devastating, and basically end all research and development.
Purportedly, the proposal moves a lot of weather research and development under NWS, but doing so while eliminating the labs themselves and the CIs would get rid of most of the people and infrastructure that does this critical work.
Fortunately, Congress has the final say on this, so I recommend pressuring your Senators and representatives to step in and continue allocating NOAA the resources to pursue life-saving research.
A very sad day for atmospheric science.
Due to the current gap in funding from the U.S. National Science Foundation, the NSF Unidata Program is pausing most operations. Nearly all staff will be furloughed until funds from our existing NSF grant become available. For more information, see buff.ly/d2TC1Oy
2
66
TroyArcomano retweeted
3/16/25 3:15 PM CDT-Our storm survey team has found damage consistent with EF-4 tornado damage near Diaz (Jackson County) with estimated peak winds of 190 mph. Additional information will be provided as surveys continue in the coming days. #arwx
TroyArcomano retweeted
Had to go to NHC this morning to drop off my work computer and CAC to be sent back to EMC. It was a sad moment that hit me with the reality of the injustice that has happened to me, many others in NOAA and elsewhere, and ultimately the American people.
On a happier note I also got to go talk to a middle school science class this morning about weather and hurricanes. That's always a lot of fun and a great way to start building a new generation of leaders in our field :) – at Homestead, FL
TroyArcomano retweeted
Friendly reminder that every operational AI weather forecasting tool that exists today relies on processed data products distributed by @NOAA. This includes fancy MLWP models - can't run them if you don't have the outputs of the ECMWF or NOAA assimilation suites.
TroyArcomano retweeted
Morning all! I've been encouraged by the outpouring of support for all of us in NWS/NOAA.
I wanted to specify a detail that I think is relevant from some conversations I've had. Most of us (at least at EMC) who were terminated were officially "probationary" because we are new federal employees, but have been working for NOAA in some capacity (usually on the contract or university side) for 5, 10, 15 years. So we're the group that has the unique combination of early career motivation/drive but also enough expertise to be quickly effective as we are promoted into new roles. This is not the group you'd get rid of if you wanted to make an organization more efficient - it's what you'd do if you want to destroy an organization. I'd like to assume the best and hope that's unintentional. Hopefully the value of what we do can be communicated effectively. – at Goulds, FL
TroyArcomano retweeted
New #AIFS blog: #MachineLearning (ML) for #weather and climate has continued to evolve at breathtaking pace.
Matthew Chantry takes us on a tour of the year past and gives some previews of the year to come.
➡️ ecmwf.int/en/about/media-cen…



























