My thoughts on the crucial importance of methodology on self-reported AI performance on mathematics competitions, and my policy on commenting on such reports going forward: mathstodon.xyz/@tao/1148814...
My thoughts on the crucial importance of methodology on self-reported AI performance on mathematics competitions, and my policy on commenting on such reports going forward: mathstodon.xyz/@tao/1148814...
Wait, there are math optimizations that are fun and safe?! Sign me up!
--funsafe-math-optimizations
To celebrate the kick-off of #UNOC3 and #OOSC in Nice, we share a striking, mesmerising perspective on Mediterranean Marine Heatwaves βand their clear amplification by climate change.
From @awi.de storyline simulations, visual by @jawijan.bsky.social
πͺοΈ Typhoon spotted in a global 4.4km IFS-FESOM run.
Visualizing cloud structures using only model data β aiming to match the look of satellite imagery. Impressive how close simulations can get.
@awi.de @blender3d.bsky.social @blender.org
For a new visualization project @awi.de, Iβm currently experimenting with satellite imagery, elevation data, and OSM. Exciting!
@blender.org
Exploring ways to visualise wind together with my colleague @sbeyer.net at AWI. Hereβs a glimpse: IFS-FESOM wind speed visualised as coloured particles.
ππ What if we could replay recent extreme weather under different climate conditions? With storyline simulations in the EUβs #DestinE initiative, scientists are reconstructing past, present, and future extreme weather to help us adapt to climate change. Learn more β‘οΈ destine.ecmwf.int/news/replayi...
Experimenting a little bit with #blender particle systems and simulated wind data
Sea ice meets ocean eddies, part II!
Folllowing @oceanographer.bsky.social's idea I used the data from my simulation (still #FESOM2 but 10km this time) to explore the flames of the frozen ocean. π§ π₯
First post on @bsky.app π:
Can #AI-based weather forecasting models (trained on present-day data) provide skillful forecasts also in different colder and warmer states of the climate system?
Our preprint in arxiv explores this question:
doi.org/10.48550/arX...
Here is what we found so far (π§΅1/7)
To celebrate Antarctica Day, here is an image of winds around Antarctica as simulated by a 4 km resolution atmospheric model. π§ͺ
Model: IFS-FESOM | Shown: U component of wind | Project: nextGEMS #SciArt
The maximum 2m-temperature during the peak heatwave period (July 25, 2019) over France for the different storyline experiments (top panels: a, b, c). The bottom chart illustrates the hourly maximum 2m-temperature from the different experiments for the Paris region inside the black dotted box (48.5-49Β°N and 2.0-3.0Β°E). This comparison highlights the varying intensity of the heatwave across different climates.
The storyline approach isnβt new, but for the first time, weβre using it with 10km resolution! π See how extreme events might have looked in the past, without climate change, and how they could unfold in a +2Β°C future. Check out our work led by Amal John: essopenarchive.org/users/852952...