27 papers in the archive.
CNN classifier with explainable AI identifies which climate models and resolutions produce winter surface temperature fields, showing ocean eddy resolution drives model similarity.
Climate analysis & attributionMet OfficeMax Planck Institute for MeteorologyUniversity of Oxford
Review arguing machine learning and data assimilation are converging, with ML accelerating DA and DA supplying uncertainty quantification and physical constraints.
Data assimilationBureau of Meteorology (Australia)Colorado State University (incl. CIRA)ECMWFMax Planck Institute for MeteorologyNASANOAA (incl. GFDL, PSL, GSL, NCEP/EMC)University of Oklahoma (AI2ES)University of Reading
A deep learning approach using stage-consistent ocean dynamics claims to extend ENSO prediction beyond 18 months; the abstract is missing, so details are unverified.
Sub-seasonal to seasonalChina Meteorological Administration (CMA)Max Planck Institute for MeteorologyNanjing University of Information Science & Technology (NUIST)
Intercomparison of statistical and machine learning methods for estimating forced climate response from single realizations; skilful in models but divergent when applied to observations.
Climate analysis & attributionColorado State University (incl. CIRA)ETH ZurichMax Planck Institute for MeteorologyNSF NCAR / UCARNOAA (incl. GFDL, PSL, GSL, NCEP/EMC)New York University (M2LInES)Stanford UniversityUniversity of California, Los Angeles (UCLA)University of Oxford
ML boundary-layer height scheme trained on coarse-grained high-resolution ICON runs; stable 36-year AMIP coupling reduces biases in coarse-resolution climate simulations.
Hybrid physics-ML & parameterizationMax Planck Institute for Meteorology
Neural-network replacements of JSBACH4 stomatal conductance and photosynthesis parameterizations, trained on FLUXNET, reduce transpiration and GPP residuals in ICON-ESM land.
Hybrid physics-ML & parameterizationMax Planck Institute for BiogeochemistryMax Planck Institute for Meteorology
Feed-forward networks trained on FLUXNET replace JSBACH4 stomatal and photosynthesis parameterizations inside ICON-ESM, reducing flux residuals against observations.
Hybrid physics-ML & parameterizationMax Planck Institute for BiogeochemistryMax Planck Institute for MeteorologyTechnical University of Munich (TUM)
A VAE plus spatiotemporal-transformer latent diffusion model generates large climate ensembles on the fly from the MPI Grand Ensemble, avoiding storage of pre-computed simulations.
Bidirectional LSTM emulates ICON radiative heating rates accurately, and SHAP analysis shows the learned relationships match radiative physics.
Hybrid physics-ML & parameterizationColumbia University (LEAP, Lamont-Doherty)Max Planck Institute for MeteorologyNew York University (M2LInES)
mloz is a machine-learning ozone parameterization that runs about 31 times faster than UKESM chemistry, enables interactive ozone in CMIP-style runs and transfers to ICON.
Hybrid physics-ML & parameterizationDeutscher Wetterdienst (DWD)Karlsruhe Institute of Technology (KIT)Max Planck Institute for MeteorologyUniversity of Cambridge (Aardvark)
ML emulators with history matching tune ICON-A parameters, reaching a top-of-atmosphere net radiation budget within 0–1 W/m² with fewer simulations.
An ML radiation emulator in ICON runs four times faster than RRTMGP and stays stable over 10-year AMIP-like runs and +4 K SST warming.
Hybrid physics-ML & parameterizationColumbia University (LEAP, Lamont-Doherty)Max Planck Institute for Meteorology
Machine learning trained on simulations detects past volcanic eruptions from post-eruption temperature patterns, with explainable AI identifying regional fingerprints including the unidentified 1809 event.
Climate analysis & attributionMax Planck Institute for Meteorology
Perspective on how machine-learning emulators, reanalyses and compute are reshaping operational global weather prediction, and the strategic scenarios national services should prepare for.
SuperdropNet, a neural emulator of warm-rain microphysics, is coupled into the ICON model via three strategies, with stable, physically consistent warm-bubble results.
Hybrid physics-ML & parameterizationMax Planck Institute for Meteorology
A deep-learning burnt-area model coupled into the JSBACH4 land surface model of ICON-ESM improves global fire simulation (monthly r 0.79 vs −0.07 for the standard scheme), with LRP attribution.
Hybrid physics-ML & parameterizationMax Planck Institute for BiogeochemistryMax Planck Institute for Meteorology
Trains U-Net and gradient-boosted tree convection parameterizations for ICON from filtered high-resolution data; ablated U-Net runs stably for 180 days online with better precipitation extremes.
Hybrid physics-ML & parameterizationColumbia University (LEAP, Lamont-Doherty)Max Planck Institute for Meteorology
Nature Reviews review of deep learning's impact on weather and climate prediction; abstract absent, judged from title and venue.
Global medium-range NWPECMWFETH ZurichMax Planck Institute for MeteorologyUniversity of Oklahoma (AI2ES)
Position paper on Earth Virtualization Engines that pair high-resolution physics-based climate models with machine learning and federated access to exabyte-scale climate data.
Software, scaling & compressionCSCS - Swiss National Supercomputing CentreETH ZurichMeteoSwissMassachusetts Institute of Technology (MIT)Max Planck Institute for MeteorologyNSF NCAR / UCARNVIDIAUniversity of Reading
Explainable neural network on 100 MPI Grand Ensemble members identifies spring tripolar North Atlantic SST patterns as precursors of European summer heatwaves.
Climate analysis & attributionColorado State University (incl. CIRA)Max Planck Institute for Meteorology
WMO prize challenge winners used AI to post-process ECMWF S2S forecasts, beating bias-corrected references for weeks 3-6 precipitation and temperature.
Sub-seasonal to seasonalColumbia University (LEAP, Lamont-Doherty)Environment and Climate Change Canada (ECCC)ECMWFMax Planck Institute for MeteorologySwiss Data Science Center (SDSC)
A statistical GAM emulator predicts local surface temperature responses to tree cover change from ESM output, offering a cheap complement to greenhouse-gas emulators.
Climate emulationETH ZurichKNMI (Royal Netherlands Meteorological Institute)Max Planck Institute for BiogeochemistryMax Planck Institute for Meteorology
Supervised classifier trained on 18 observation and reanalysis datasets labels Central and Eastern Pacific El Niño types, then shows the observed CP increase lies within internal variability in large ensembles.
Climate analysis & attributionMax Planck Institute for MeteorologyNSF NCAR / UCAR
Self-organizing maps link North Atlantic spring SST to summer jet-stream teleconnections, identifying windows of opportunity for seasonal European summer predictions with MPI-ESM hindcasts.
Sub-seasonal to seasonalMax Planck Institute for Meteorology
Neural network cloud cover parameterizations trained on ICON storm-resolving simulations generalise across regional and global data, with SHAP used to explain their feature use.
Hybrid physics-ML & parameterizationColumbia University (LEAP, Lamont-Doherty)Max Planck Institute for Meteorology
Crowd-labelled about 50,000 shallow cloud clusters in satellite images to train deep learning that maps four organization patterns and builds global climatologies.
Observations & remote sensingMax Planck Institute for MeteorologyTechnical University of Munich (TUM)
Nature review on deep learning for data-driven Earth system science, covering weather, climate and process understanding; abstract absent, judged from title and venue.
Climate analysis & attributionLawrence Berkeley National Laboratory / NERSCMax Planck Institute for BiogeochemistryMax Planck Institute for Meteorology
