ESA-ECMWF WORKSHOP 

7 - 10 May 2024| ESA-ESRIN 



Agenda

Please find below the agenda of the workshop. Time-slots are indicated in Central European Time (CET)

Tuesday, 07 May 
 (no live streaming)

StartEndSite
Min
15:00
16:00
Arrival of poster presenters - Poster Session 1
16:00

19:00Poster Session 1 and Icebreaker
150
Wednesday, 08 May   
(live streaming)
09:00

09:10Opening
Chair: Rochelle Schneider (ESA), Massimo Bonavita (ECMWF)
10
09:10

09:20on-site             ESA Welcome
Rochelle Schneider (ESA)
10
09:20

09:30onlineECMWF Welcome
Andy Brown (ECMWF)
10
09:30

10:15on-siteKeynote speaker
Prof. Xiaoxiang Zhu (Chair professor for Data Science in Earth Observation TUM)
45
10:15

11:00on-siteKeynote speaker
Prof. Martin Schultz (Professor in Computational Earth System Science at the University Cologne)
45
11:00

11:20Coffee Break20
Session 1.1 (TA5): New Generation Computing for AI (HPC, edge and quantum computing)
Chairs: Alessandro Sebastianelli (ESA) and Marcin Chrust (ECMWF)
11:20

11:40onlineMARS-S2L: machine learning for operational CH4 super emitter detection at UNEP's IMEO 
Anna Vaughan (UNEP)
20
11:40

12:00on-siteNVIDIA HPC Trends in AI for Earth System Science
David Hall (NVIDIA)
20
12:00

12:20on-siteMachine Learning for Very Low-Latency Storm Nowcasting with Onboard Satellite Processing: EO-ALERT Final Results
Álvaro Morón Elorza (Deimos Space)
20
12:20

12:40on-siteThe Earth Observation Training Data Lab
Juan Pedro B. (EarthPulse)
20
12:40

13:00on-siteQuantum Advances in Earth Observation: Unlocking the Potential of Quantum Machine Learning Algorithms
Francesco Mauro (University of Sannio)
20
13:00 13:30Working Group Discussion30
13:30

14:30Lunch Break60
Session 1.2 (TA2): Multidomain ML4ESOP
Chair: Maryam Pourshamsi (ESA) and Anna Jungbluth (ESA) 
14:30

15:00on-siteGlobal AI-Powered School Connectivity Prediction with Earth Observation & Modelling Priority Areas for Improving Global School Access: A Geostatistical Machine Learning Approach
Kelsey Doerksen and Abi RIley (UNICEF & ESA)
30
15:00

15:20on-siteImproved situational awareness along power lines using satellite imagery and machine learning
Michele Gazzea (Stormgeo AS)
20
15:20

15:40on-siteObserving cloud vertical structure in geostationary satellites: advances using a one-dimensional temporal convolution in deep learning
Mathilde Ritman (University of Oxford)
20
15:40

16:00on-siteAdvancing the Study of Extreme Weather Events with Data, Deep Learning Methods and Climate Analysis
Sol Kim (University of Chicago)
20
16:00

16:20on-siteTowards the inversion of plumes from power plants and industrial sites in satellite CO2 images using deep neural networks
Joffrey Dumont Le Brazidec (ECMWF)
20
16:20

16:30Coffee Break10
16:30

16:50on-siteEnhancing Earth Observation Feature Identification with Transfer Learning and Community-Driven Tools
Roope Tervo (EUMETSAT)
20
16:5017:10on-siteIntegrated machine learning approaches for volcanic cloud tracking in EUMETSAT SEVIRI images
 Federica Torrisi (INGV, Catania)
20
17:10

17:30on-siteAntarctic Sea ice detection using concurrent multispectral and synthetic aperture radar imagery.
Martin Rogers (British Antarctic Survey)
20
17:30

17:50on-siteDetection and source-attribution of methane super-emitter plumes using Deep Learning
Berend Schuit (SRON Netherlands Inst4Space)
20
17:50

18:10on-siteBiomass estimation and modeling using Machine Learning and Remote Sensing 
Hanan Abdul (Western Norway)
20
18:10

18:30onlineForecasting Permafrost Carbon Dynamics in Alaska with GeoCryoAI
Gay Bradley (NASA JPL)
20
18:30

18:50Working Group Discussion
Thursday, 09 May
(live streaming)
Session 2.1 (TA3):  End-to-End ML4Weather and Climate Prediction
Chair: Matthew Chantry (ECMWF) and Mihai Alexe (ECMWF)
09:00

09:30on-siteAIFS – ECMWF’s Data-Driven Probabilistic Forecasting System
Mihai Alexe (ECMWF)
30
09:30

09:50on-siteWeatherBench 2 - Updates on the state-of-the-art in AI-weather models
Stephen Rasp (Google Research)
20
09:50

10:10on-siteA Graph-Based Latent Variable Model for Probabilistic Weather Forecasting
Joel Oskarsson (Linköping University)
20
10:10

10:30on-siteGTPCast: Explainable generative AI for precipitation nowcasting 
Gabriele Franch (Fondazione Bruno Kessler)
20
10:30

10:50on-siteData-driven sea-ice modelling with generative deep learning
Tobias Finn (CEREA, ENPC)
20
10:50

11:10on-siteImproving cross-site generalizability of vision-based solar forecasting models with physics-informed transfer learning
Quentin Paletta (ESA)
20
11:10

11:20Coffee Break
Recording
10
11:20

11:40onlineWhy are machine learning forecast systems so fast?
Christopher Subich (Environment and Climate Change Canada)
20
11:40

12:00on-siteEmulating 3D-Var data assimilation using autoencoder-like methods
Boštjan Melinc (University of Ljubljana)
20
12:00

12:20on-siteAtmoRep: a probabilistic foundation model for atmospheric dynamics
Ilaria Luise (CERN)
20
12:20

12:40on-siteHybrid physical-data driven and surrogatesea-ice modelling
Alberto Carassi  (University of Bologna)
20
12:40

13:00on-siteAI-based weather and climate prediction with NeuralGCM and beyond
Stephan Hoyer (Google Research)
20
13:00

13:30Working Group Discussion30
13:30

14:30Lunch Break60
Session 2.2 (TA1): ML4DestinE
Chair: Rochelle schneider (ESA) and Mariana Clare (ECMWF)
14:30

14:50on-siteAI4DESP: Unblocking Artificial Intelligence power to explore potential and challenges in Destination Earth Service Platform.
Tona Calogera talk by Gaia Cipolletta (ESA & SERCO)
20
14:50

15:10on-siteMachine Learning and the DestinE digital twins
Mariana Clare (ECMWF)
20
15:10

15:30on-siteDestination Earth Data Lake unlocking big earth data processing with AI/ML capabilities
Michael Schick (EUMETSAT)
20
15:30

15:40Coffee Break
Recording
10
15:40

16:00hybridGuiding Principles and Concepts for the use of AI in DTEs
 Louise Mercy talk by Jonny Langstone (Telespazio UK)
20
16:00

16:20on-siteTowards Forest Digital Twin: Deep Learning-based Tree Height Estimation using Monocular Multimodal Satellite Data
Meruyert Kenzhebay (Western Norway University Of Applied Sciences)
20
16:20

16:40onlineGradient boosting-based soil wetness for forestry climate adaptation in HarvesterSeasons service -training a model to forecast soil water index SWI from a comprehensive set of IFS model predictors in Destination Earth
Mikko Strahlendorff (Finnish Meteorological Institute)
20
16:40

17:10Working Group Discussion30
17:10

19:10Poster Session 2 and Icebreaker
120
Friday, 10 May
(live streaming)
Session 3.1 (TA4): Hybrid ML-NWP/Climate 
Chair: Massimo Bonavita (ECMWF) and Patrick Ebel (ESA)
09:40

10:00on-siteOnline model error correction with neural networks - from theory to the ECMWF forecasting system
Alban Farchi (ENPC) 
20
10:00

10:20on-siteOn the potential and limitations of data-drivenforecast models
Massimo Bonavita (ECMWF) 
20
10:20

10:40on-siteSummer drought prediction in Europe through hybrid climate simulations and remote sensing
Laia Romero (Lobelia Earth) 
20
10:40

11:00on-siteRole of Large-  and Small-Scale Dynamics on Tropical Convective Rainfall in DYAMOND Models 
Lucile Richard (EPFL)
20
11:00

11:10Coffee Break
Recording
10
11:10

11:40on-siteML applications for forecasting - some orientation in a rapidly changing landscape
Olivia Martius (University of Bern)
30
11:40

12:00onlineImproved subseasonal prediction of South Asian monsoon rainfall using data-driven forecasts of oscillatory modes
Eviatar Bach (CalTech)
20
12:00

12:20on-siteScore-based generative modeling meets data assimilation
François Rozet (University of Liège)  
20
12:20

12:40on-siteComplex network and machine learning to improve marine ecosystem modelling and data assimilation
Ieuan Higgs (University of Reading)
20
12:40

13:00on-siteMachine Learning-Based Observation Operators to Assimilate Microwave and SIF SatelliteObservations into the ECMWF Integrated Forecast System
Sebastien Garrigues (ECMWF)
20
13:00

13:30Working Group Discussion
Recording
30
13:30

14:30Lunch Break60
14:30

15:30Working Groups plenary discussion and close
60