The Dept. of Atmospheric Sciences at Texas A&M has a YouTube channel. Check out our first video! Starring @knubez.bsky.social along with many faculty and students! youtu.be/_e57SmX2Ijw?...

Adding my 2 cents to the discussion on Google DeepMind’s performance for Imelda… 🧵 You might see a map like this and be impressed that at no point did it predict landfall in the US. But there is a lot of context lost by simplifying it to just one image.

"In science, to be ‘conservative’ is to understate your findings. In insurance, it means the opposite: erring on the side of overstatement of risks. For a clear assessment of the risks of #climate change, we need these two cultures to meet in the middle" www.cambridge.org/core/books/a...

Most of the "90% of code written by AI" claims come from vendors selling AI tools and lack credibility as a result Armin (creator of Flask, Jinja, Click) is different - when he says 90% of a new significant infrastructure project he's building was AI generated that's worth paying attention to

Quite a satellite image this morning! 🛰️🌀 Clouds extend 1000 miles from #Imelda's center all the way into central Pennsylvania. #Humberto continues to churn in the Atlantic.

My latest effort to be super-popular in @nytimes.com: “Spraying Roundup on Crops Is Fine. Really.” Yes, the world’s most hated herbicide is way less toxic than caffeine. Here’s a gift link where you can leave all your supportive comments: www.nytimes.com/2025/09/28/o...

90 of 111 fatalities from Hurricane Rita happened due to the evacuation. 20 years later, would we do any better? @sarava.net shouting out the value of much improved track forecasts 🎁 link:

I wrote a short piece on the debate about the future of climate modeling: Should we pool resources to do km-scale "masterpiece" runs or focus on using ML to build data-driven models that can still do large ensembles? 1/3

📱 Come make meaningful content with us! This gives a creative, climate-motivated creator the chance to bring science & data to life through digital storytelling. Grow your video & editing skills while communicating climate change impacts. (Probably work in a little fun along the way, too!)

Betteridge's law at work "Do Climate Models Support Claims of Volcanic Global Catastrophes?" agupubs.onlinelibrary.wiley.com/doi/10.1029/...

A quick summary by me of a new instance of an old story. @realclimate.org Time and Tide Gauges wait for no Voortman www.realclimate.org/index.php/ar...

We're hiring! We have an open search for an Assistant Professor in physics of the ocean and/or atmosphere in Columbia's Dept. of Earth and Environmental Sciences. Please circulate this information to good candidates, and apply yourself if you are one! Ad & link to apply: apply.interfolio.com/173819

LLMs introduce a huge range of new capabilities for research, but also make it possible for researchers to "hack" their results in new ways by how they chose to use models for annotation This is a useful pass at quantifying some of the risk, and some mitigation strategies arxiv.org/pdf/2509.08825

I believe in the potential of the AT Protocol, but ideals and reality must be communicated appropriately. If you get on the bad side of Bluesky's moderation team right now, it literally means death. The ideal hasn't been realized yet. (3/3)

I'm happy to share a paper in Science Advances that makes the case for a weak land carbon sink, drawing upon JPL and Chloris biomass time series. To close the budget with a net land sink of 0.8 Pg C/y from 2000-2019, we propose increases in the ocean sink and decreases in fossil fuel emissions.

Another excellent addition to our Focus Issue on Climate Change Informed Cat Modeling at Environmental Research Climate iopscience.iop.org/collections/...

Outrageous! "Just another" corrupt establishment? Theoretical physics is clearly the coolest corrupt establishment we have. (More seriously, it is true that physicists are going to have to do a better job in educating people about what they do and how they operate.)

DOE Climate Working Group RIP Breaking news rogerpielkejr.substack.com/p/doe-climat...

India has a better chance of building a significant domestic solar supply chain than the US by 2035. Though it's a boom right now and will go through a bust.

Today marks the 125th anniversary of the Great 1900 hurricane that devastated Galveston, the deadliest single weather event in American history. But it wasn’t inevitable. The decisions made before and after the 1900 storm and relevance of it today rings loudly if we’re listening.

I had an engagement with blogger, failed California gubernatorial candidate, and self-identified reporter Michael Shellenberger this past week, which started out being about this new paper being heralded by climate skeptics as disproving global sea-level acceleration. (LONG 🧵)

This is AI2ES, my AI institute. No way around it other than to say this is really awful. We are doing so many amazing things! We want to keep doing amazing things but without this level of funding, we will not be able to do so.

cc @bobkopp.net

Our comment on the DOE CWG report is done. It tips the scales at 439 pages, approx. 3x longer than the DOE report. This is related to Brandolini's law: The amount of energy needed to refute bullshit is an order of magnitude bigger than that needed to produce it. Example: refuting one sentence.

An interesting discussion from a scholar of literature the weak points of GPT-5's figurative writing ability, and what this might tell us about the problems of AI-driven evaluation of AI writing (and other uses of LLMs as a judge) when training models. www.christoph-heilig.de/en/post/gpt-...

Ok, so finally, when we talk about collapse of ice sheets nobody credibly believes that collapse is likely by or before 2100. And even in the most extreme collapse scenarios, this is a multi-century process with rates that will depend a lot on climate and the decisions we make. 12/

Notable item from the other site: Google’s AI model killed it on Hurricane Erin 👀 results. x.com/brady_wx/sta...

There seems to be two contrasting goals in modern science: 1. I want to make an important discovery that is valued by the world. 2. I want the world to value me for making an important discovery. (1/2)

Numerical weather prediction models (ECMWF IFS) still outperform AI weather models in forecasting record-breaking hot and cold extremes and unseen wind extremes Preprint arxiv.org/abs/2508.15724 with Zhongwei Zhang @erichfischer.bsky.social @zscheischlerjak.bsky.social and Sebastian Engelke

Meet the Tropical Systems Group @tamuatmo.bsky.social: featuring Dr. Kwesi Quagraine, Stephanie Ortiz Rosario, and Dr. @zmoon92.bsky.social ! We study the weather and climate of the tropics — from ⛈️to TCs 🌀— using km-scale models, obs, & a passion for science comms! sites.google.com/tamu.edu/nun...

The hiatus was a lesson! Can we please stop saying “surprising” when we mean “kind of interesting?” [and completely consistent with the range of internal variability an ensemble of models will produce] www.theguardian.com/environment/... agupubs.onlinelibrary.wiley.com/doi/10.1029/...

This story is insane. Amazing stuff.

I'm writing about insurance and climate change for the first time in ages, and far out those guys have been busy while I wasn't looking

Worth reading. The greatness of a university - or any institution - isn't forever. It can be squandered very quickly.

📢📢 New preprint alert! 📢📢 Radiation is a key piece of climate models, but comes with a trade-off: use an accurate scheme (which is difficult to understand), or a simple scheme (which is not very accurate)... Here I introduce a scheme which bridges this gap! arxiv.org/abs/2508.09353

For the upcoming 20th anniversary of Hurricane #Katrina, I created a track forecast cone comparison for three days before its infamous Louisiana landfall. The purple cone is the actual cone centered on the forecast from 2005 and the green cone shows a hypothetical forecast made in 2025. [1/4]

Another day, another major news outlet publishing First Street criticism of FEMA whole cloth and without mentioning they are in the business of selling an alternative... www.nbcnews.com/science/envi...

No. There is another way, but I don't recommend it. www.nature.com/articles/d41...

What seems to have been happening here is that in order to get the fair to middling improvements in output, it has become necessary to throw much more compute at every query, at a rate which has increased faster than the learning curve for the tokens themselves. That ... ain't good.

How many times does the letter b appear in blueberry ChatGPT: The word blueberry has the letter b three times — positions 1, 5, and 6. How many PhDs does it take to count the letter b in blueberry ChatGPT: Realistically? Zero — any middle-schooler (or a spell-check) can handle it. Checked it myself:

Cat models are *not* forward-looking predictive models. They represent *near-present* risk They are explicitly *not* assessing future wildfire risk. This is routinely misreported, and it means that everything we're seeing in CA now? It only includes the climate change that has happened *so far*

A study predicted economic damages from climate change 3x higher than previous assessments. But it was flawed -- all thanks to the data from just one country. I dug into this strange study and the researchers calling it out: www.washingtonpost.com/climate-envi...

onion getting a bit too real today. 🔭🧪 theonion.com/seek-funding...

Nothing ever changes but the climate Some initial notes on the DOE Critical Review of Impact of Greenhouse Gas Emissions on the US Climate And a 🧵 of other responses I'm aware of (let me know if you know of others) diagrammonkey.wordpress.com/2025/08/01/n... 1/n

Over on a thread launched by @davidho.bsky.social there has been some interesting discussion of oxygen, in part trying to make sense of sound bites like "half the oxygen we breather is made in the ocean." I won't repeat what went on over there, but I thought I'd amplify on a few aspects of ...

"history is quite adept at hiding trends" - too right Nice read on interpreting loss potential (a bit, but don't bet your shirt) from seasonal forecasts.

in case you missed it: Terence Tao calls for pre-registration of methodology for AI mathematical ability tests, and offers an excellent breakdown of why ability even on a single given task is not a binary variable mathstodon.xyz/@tao/1148814...

⚡ A bolt for the record books! Texas A&M Atmospheric Sciences' Dr. Timothy Logan helped confirm a lightning flash that stretched 515 miles—from Texas to Kansas. The rare “megaflash” is now the longest ever recorded. Read more: tx.ag/ATMOLightningRecord

WTF? I was quoted in this story but I have never spoken to the reporter nor would I have said those things. Is everything just generated by LLMs these days? cosmosmagazine.com/earth/oceans...