This for me is the Big Hurdle for AI forecasts: predicting extremes in warming world when they're not necessarily trained with them. Promising...
OK so here's some explanation given that chart below blew up for no reason. Here's the same data with 4 strongest El Nino years for Aug-Oct: 1982, 1987, 1997, 2015. It's interesting that we might have a 0.5c head start on these years as the SSTs potentially warm. End of story.
I appear to have started a weird hand-wringing competition purely by flagging up how the ENSO 3.4 region has warmed in the last few days or so.
@norcrosscricket.bsky.social Doubt you'll be reading on comms but nothing much upwind for the time being, assuming more showers don't break out.... FAMOUS LAST WORDS
OK so is it safe to say that "something is happening" ?!
OK so is it safe to say that "something is happening" ?!
Last month FastNet, the AI weather model we're developing with @metoffice.gov.uk, predicted the high temperatures of May's heatwave 84hrs ahead.
It captured the highest midday temperatures more closely than the current operational physics-based global model.
Learn more: bit.ly/47JvdRm
OK so is it safe to say that "something is happening" ?!
The Alan Turing Institute
A powerful El Niño Event is expected to develop during 2026.
How powerful? Well the seasonal forecasting models are currently all over the place.
So, somewhere between a Top 10 event in modern history and super-bonkers extreme crazy land. 🤔
🧪🌊
Open question - should everyone use the same ENSO index to save all the ONI / RONI juggling? If more are using RONI now, is ONI kept on for historic (hyperbolic!) purposes or do the two together serve a purpose?
