TODAY IS THE DAY!
Today I anger every member of the weather, water, and climate enterprise by telling them to jump on the AI train or get left behind.
Get your tomatoes and "boos' ready!
Kudos to whoever has started-up the one station in DR Congo. 🙂
Why are there no observations from the Sea of Okhotsk?
It doesn't come with an example for how to fine-tune it using hf transformers, so it will be way more work to use than the three models I started with. My model selection process was heavily biased by the availability of easy-to-modify examples.
But worth considering for the longer term.
Exciting jobs & PhD studentship!
1) postdoc in historical windstorms: jobs.reading.ac.uk/Job/JobDetai...
2) postdoc in extreme event storylines: jobs.reading.ac.uk/Job/JobDetai...
3) PhD in redeveloping the Central England Temperature series (led by Tim Osborn): www.findaphd.com/phds/project...
Three baby robots, dressed up for graduation in hood and gown, leaping in the air in celebration.
AI is indeed the perfect tool for climate data rescue.
The trick is to follow the same process you'd use for citizen science - except, instead of 1000s of human volunteers, you use a trio of fine-tuned baby robots.
brohan.org/Robot_Rainfa...
Two job opportunities on new @copernicusecmwf.bsky.social contract until 30/06/28. Both looking at aspects of data rescue and preparation and provision of global historical land meteorological data. Looking to fill ASAP. First job is for a senior postdoctoral position: my.corehr.com/pls/nuimrecr...
It's not just the UK - global land temperatures can be independently confirmed - agupubs.onlinelibrary.wiley.com/doi/full/10....
A poster demomstrating an extension to the well-known climate stripes graphic.
The future of scientific publishing is YouTube, Github, Medium, Substack, ...
I don't see much of that in the agenda.
MeteoSaver v1.0: a machine-learning based software for the transcription of historical weather data
Preprint in open discussion: egusphere.copernicus.org/preprints/20...
This is a work of art.
It is easy to come up with designs for ground-breaking climate models using ML.
The thing that's very hard is to come up with a ML climate model that looks and operates like an NWP model.
If we can find the courage to move on from NWP, then the future is bright indeed.
They only use ERA5 at all because they are doing NWP imitation. Eventually they will realise that to forecast London precip. (say) an AI model doesn’t have to bother with the whole NWP state vector.
Then we’ll start to see real improvements.
The really interesting result is the reduced observational requirement for DA. How far can we push that?
You could build a model without ERA5 - train it on satellite data, in-situ observations, or what you will. There just isn't much point.
While all the AI models use ERA5, they don't require it.
It really is a supercomputer-free forecasting system.
A screenshot of two application windows open next to each other. On the left is a scanned page of handwritten ocean temperature data from March 1952, and on the right is a Google sheet with most of the data transcribed. There are missing values in places, but the digital data is very accurate.
I'm always looking for ways to make it faster to transcribe historical data, & structured handwritten text has been a real challenge. I just tested Google's Gemini 2.0 Flash model on some Monterey Bay temperature data from 1952. I'm pretty impressed! Not perfect, but better than a blank page. 🌊
Delighted to announce I have taken on a joint position between the Met Office and the University of Leeds:
Research Scientist | Met Office Hadley Centre
Associate Professor | University of Leeds
Looking forward to getting to work on this.
See news article below 👇
Want to do a PhD in climate, with a focus on using machine learning to recover millions of lost weather observations and so better reconstruct climatic changes over the past century?
Open to students from UK and Ireland, funded through @climatecocentre.bsky.social.
www.findaphd.com/phds/project...
🔭 How accurate are historic sea surface temperatures?
NOC's Dr Elizabeth Kent MBE (@lizk40.bsky.social) gives us a rundown and explores a particularly puzzling cold feature on #NOCIntoTheBlue.
Watch, listen and subscribe NOW 🎧 linktr.ee/nocintotheblue
The state vector describing reality has way too many dimensions.
What else can we do?
I won't be following up myself (the data I'm after is almost all in tables). But I do urge you to give it a go - point your browser at aistudio.google.com upload your image and start asking. It could hardle be easier to experiment with.
And as so often it's not perfect, but it's very promising. Definitely worth looking into.
No.
But since you ask.
I've uploaded a page from "A Collection of Voyages Chiefly in the Southern Atlantick Ocean" books.google.co.uk/books?id=sGx...
And asked Gemini a couple of questions:
I was thinking that maybe in a couple of months there would be an upgrade to Gemini that we should experiment with.
I made insufficient allowance for 'AI time'. An update has come out today -
developers.googleblog.com/en/gemini-2-...
I don't know if it makes any difference.
All the code is available at the link, and Gemini's free tier is enough to experiment with.
Go on - give it a try.
" We can say with some confidence that we are better at reading the logs than the original log-keepers were at writing them."
I think we are very close to being able to replace the 'we' with 'AIs'.
This Climate Visualization Belongs in a Damn Museum
In 2018, @blkahn.bsky.social said that the warming stripes graphic 'belongs in a damn museum', even suggesting that it was 'fit for the Museum of Modern Art' (see gizmodo.com/this-climate...).
Well...
Later this month, they will appear in the Museum of Modern Art: press.moma.org/exhibition/p...
