M²LInES is a large international collaborative project with the goal of improving climate projections using scientific & interpretable Machine Learning.
The World
Joined July 2021
- Tweets253
- Following42
- Followers786
- Likes346
M²LInES retweeted
In 2020, we asked researchers: what would it take to make climate models better?
9 teams set out to improve how we see and model every layer of our climate system.
Here are just a handful of major discoveries from 5 years of the Virtual Earth Systems Research Institute (VESRI).
M²LInES retweeted
CDS researchers built Samudra, an AI ocean emulator that runs 150x faster than traditional models.
Unlike past AI attempts, it stays stable for centuries.
Developed at @M2LInES by Laure Zanna, Carlos Fernandez-Granda, & others.
nyudatascience.medium.com/sa…
M²LInES retweeted
Congratulations to Laure Zanna (@laurezanna), who has been appointed as the inaugural recipient of the Joseph B. Keller and Herbert B. Keller Professorship in Applied Mathematics by the University Provost and the Board of Trustees. Read more here:
cims.nyu.edu/dynamic/news/14…
M²LInES retweeted
🚀 Want to learn about the first AI Global Ocean Emulator for Climate that is 150x faster than a numerical model?
Check out our blog post on Samudra!
tinyurl.com/Samudra-AI
M²LInES retweeted
📢Check out this POSTDOC OPPORTUNITY @ LEAP + @LamontEarth!
🌊Work w/@BalwadaDhruv on ?s re: #ocean macroturbulence-flow @ scales of 1-100km, which play an important role in lateral + vertical transport of ocean tracers. (Tent. start: 1/1/2025)
Apply: academic.careers.columbia.ed…
M²LInES retweeted
In honor of @ClimateWeekNYC 2024, we're highlighting a few Courant faculty members who are using to power of mathematics and computer science to combat the climate crisis. Today's feature is Laure Zanna (@laurezanna)!
M²LInES retweeted
The latest @M2LInES newsletter covers a recent publication from Sara Shamekh, Assistant Professor of Mathematics and Atmosphere/Ocean Science. Check it out here:
In this month's newsletter we're highlighting papers by Sara Shamekh and @janniyuval and celebrating Sara and Julia Simpson as awardees 🎉 mailchi.mp/d53f7fd6537d/m2li…
M²LInES retweeted
New Opinion article: @AstroVivi, @atmosbarnes, @DJGagneDos, @OceanCarbon and @BasicScienceSav discuss how a firm ethical basis for #ClimateAI increases explainability, robustness, and trustworthiness
journals.plos.org/climate/ar…
#EthicalAI @Climformatics @LeapStc @M2LInES @ai2enviro
ALT A screenshot of the PLOS Climate article "Ethics in climate AI: From theory to practice" by Acquaviva et al. 2024
M²LInES retweeted
THIS THURS: LEAP Lecture in #Climate #Data Science, w/@YongquanQu2340 (@CUSEAS) + Jiarong Wu (@NYU_Courant)!
📅 5/9/24 @ 12p (EDT)
📍 In-person + via Zoom
📰 Titles + abstracts: bit.ly/3GNAUzh
💻 Register: bit.ly/3U5Tv1h
#LEAPEducation @NSF @columbiaclimate
M²LInES retweeted
Don't miss our Geoscience Director @laurezanna discussing "Accelerating Discovery in Climate Physics with #ML," part of @SimonsFdn Presidential Lecture Series!
Next Wed., 2/7/24 @ 6p (EDT)
Get tix: bit.ly/3HAIGwC
@nyuniversity @NYU_Courant #climate #ocean #physics
M²LInES retweeted
Join @laurezanna to learn about #data #AI #climatemodel at @CUSEAS Colloquium in Climate Science: "Bridging the Gap Between Understanding + Simulating #Ocean Warming"
📅 THIS THURS 1/25/2024 @ 2:45p (EDT)
🗺️ Mudd Hall @Columbia (500 W 120th St, NYC)
bit.ly/3Sq96aQ
M²LInES retweeted
A revolution in the making. Here's my second piece for @AdvSciNews on a study by @willjgregory_ and @laurezanna (among others) about harnessing #AI to improve #climate predictions, with the invaluable help of @vbcorless.
A deep learning neural network is making climate predictions around sea ice loss more accurate than ever before.
@willjgregory_ @Princeton
Find out how 👇
ow.ly/r6EI50QhQOL
M²LInES retweeted
First time seeing ~300 ppl in the MC conference room. Just saw wonderful and insightful #AGU2023 talks by Laure Zanna @laurezanna and Pedram Hassanzadeh @turbulentjet about climate modeling via ML and the physical consistency of recent weather forecast emulations!
M²LInES retweeted
Thank you #NeurIPS for elevating our benchmark ClimSim with a Best Paper Award! A passion project spanning my roles at @UCI @LeapStc, @exascaleproject & @NVIDIAAI with @berkeley_ai @M2LInES & more. @SungdukYu speaks on Wed am or find me at our poster #AGU2023 the next morning.
🏆 At #NeurIPS, #NVIDIAResearch with partner universities, were awarded a Best Paper Award for ClimSim, a massive dataset for physics and ML-based climate research to help scientists produce better predictions of storms. nvda.ws/41mRyz5
📚 arxiv.org/abs/2306.08754
🧵3/3
M²LInES retweeted
Today we dive into exploration of the oceans. 🌊
Our first speaker is @laurezanna with a wonderful talk on Ocean and Atmosphere Dynamics, and its future developments in the age of data and AI.
M²LInES retweeted
Please consider our #AGU23 Fall Meeting Session on Machine Learning Subgrid Scale Parameterizations for Earth System Modeling, with invited speakers Chris Bretherton (Allen Institute) and Sara Shamekh (Columbia University) agu.confex.com/agu/fm23/prel…
M²LInES retweeted
David Kamm kicked off the quick fire talks with a presentation about reducing uncertainty in climate models through improved parametrizations of small scale processes @M2LInES
M²LInES retweeted
The authors of this article propose a new statistical framework to study changes in shapes of probability distributions and application to the sea level rise problem.
ALT An image of the article impact statement
Exploring the nonstationarity of coastal sea level probability distributions
@FabriFalasca10 @AndrewBrettin @LaureZanna S M Griffies, Jianjun Yin et al
→ doi.org/10.1017/eds.2023.10
@NYU_Courant @NOAA_GFDL @UAzGeos
#ClimateModels #ProbabilityDistributions #SeaLevel #SeaLevelRise
ALT Figure G1. (a) Time series showing the largest changes in kurtosis in the Mediterranean sea for the GFDL-CM4 model 1pctCO2 run. (b) Linear quantile slopes for the time series shown in panel (a). The slopes are computed for every and indicated as blue dots. The black dashed line indicates the projection onto polynomials as defined in Eq. (7). Green “check” marks indicate statistically significant (95% confidence level) changes in moments; red “crosses” indicate nonsignificant changes. (c,d) Histograms for the first and last 40 years of data of the time series shown in panel (a).
M²LInES retweeted
📢 POSTDOCS! 📢
Interested in a project using #MachineLearning + #ArtificialIntelligence to emulate ice flow physics in ice sheet models? Come aboard @LamontEarth @Columbia @columbiaclimate to work w/@JKingslake, @laurezanna, @robinbell + David Porter: academic.careers.columbia.ed…