BEGIN:VCALENDAR VERSION:2.0 PRODID:-//CERN//INDICO//EN BEGIN:VEVENT SUMMARY:2026 Workshop on Machine Learning for Land and Hydrology DTSTART:20261103T090000Z DTEND:20261105T180000Z DTSTAMP:20261006T022800Z UID:indico-event-530@ecmwfindicostage.vs.mythic-beasts.com CONTACT:events@ecmwf.int DESCRIPTION:\nBackground\nUnderstanding and accurately simulating the terr estrial water cycle is essential for predicting hydrological extremes\, ma naging water resources\, and assessing the impacts of climate change. Land -surface and hydrological processes also help constrain the atmosphere and \, through their longer memory\, can enhance the reliability of forecasts at longer lead times. Physically based hydrological and land-surface model s underpin many operational forecasting systems\, yet they often struggle to represent nonlinear interactions\, parameter uncertainty\, and heteroge neous and scale-dependent land-surface processes—particularly at the spa tial and temporal scales required for real-time decision-making.\nRecent a dvances in machine learning (ML) provide complementary pathways to improve hydrological and land-surface prediction. These include fully data-driven ML models trained directly from observations\, hybrid approaches that com bine ML with process-based models\, and physics-informed methods that embe d physical constraints into learning algorithms. Together\, these approach es offer new opportunities for scalable\, computationally efficient\, and observation-driven prediction systems suitable for operational and pre-ope rational use. In addition to improving predictive components\, ML also ena bles advances in data assimilation\, which can help integrate emerging sat ellite and in situ datasets into forecasting systems more effectively and at reduced computational cost.\nWorkshop overview\nThe workshop will bring together experts in machine learning\, hydrology\, and land-surface model ling to discuss the development and use of ML-based and ML-enhanced foreca sting systems. It will focus on datasets\, methods\, applications\, and ev aluation strategies for machine-learning approaches\, covering fully data- driven\, hybrid\, and physics-informed models. Contributions will span bot h global and regional perspectives\, addressing ML-based prediction of str eamflow\, floods\, soil moisture\, and key land-surface processes\, with a n emphasis on robustness\, scalability\, and operational relevance.\nDiscu ssions will address scientific and operational challenges such as data ava ilability and latency\, robustness under extremes and non-stationarity\, u ncertainty quantification\, evaluation and benchmarking\, and the transfer ability of ML models across regions\, climates\, and time scales. By expli citly linking observation-driven ML approaches with operational requiremen ts\, the workshop aims to inform the development of the next generation of hydrological and land-surface forecasting systems.\nExpected outcomes\n\n \nA synthesis of emerging ML methods and use cases for hydrological and la nd-surface modelling.\n\n\nCommunity recommendations for data standards\, benchmarking\, and evaluation practices.\n\n\nIdentification of key resear ch priorities and future directions.\n\n\nStrengthened collaboration betwe en the hydrology\, land-surface\, and ML communities\, including ECMWF Mem ber States.\n\n\nFormat\nThis workshop is designed exclusively for in-pers on attendance. There will be no virtual or online attendance option availa ble. We look forward to welcoming attendees on site and engaging face-to-f ace throughout the event.\nThe workshop runs over 2.5 days and will to be structured around the full pipeline of ML for land surface modelling. Day 1 will focus on novel ML methods and applications for land surface process es\, with sessions on new methods and ML applications. Day 2 will tackle " Closing the Reality Gap" — covering datasets and observations\, and eval uation and benchmarking\, with discussions on data assimilation challenges and whether we need ML-specific evaluation approaches. Day 3 will be a ha lf-day focused on moving to operations\, with a discussion on what's still needed from both modellers and users.\nEach session will combine keynotes \, submitted short talks\, themed discussions and debates (e.g. "What is t he role of physical models in land surface ML?")\, and working group break outs to scope concrete outputs. We're also planning a poster session with reception\, an ECR panel on the future of scientists in the field\, and a final synthesis session to capture research priorities and next steps.\nAt tendance\nThe workshop will take place at ECMWF's headquarters in Reading\ , UK over 2.5 days from 3 to 5 of November. Participation is open to resea rchers\, operational scientists\, and practitioners working in hydrology\, land-surface modelling\, Earth system science\, and machine learning. Pri ority will be given to registrants from organisations in ECMWF's Member St ates.\nWhilst there are no registration fees\, please note that we are una ble to provide funding to support attendance. Participants are expected to cover their own travel\, meal\, and accommodation costs.\nRegistration is now closed and the workshop is fully subscribed.\nCall for abstracts\nAbs tract submission is now closed.\nTimeline\n5 February 2026: Registration a nd abstract submission open 26 June 2026: Abstract submission deadline 3 1 July 2026: Notification of abstract acceptance 4 September 2026: Regist ration closes \n\nhttps://ecmwfindicostage.vs.mythic-beasts.com/event/530 / IMAGE;VALUE=URI:https://ecmwfindicostage.vs.mythic-beasts.com/event/530/lo go-1882613348.png LOCATION:ECMWF URL:https://ecmwfindicostage.vs.mythic-beasts.com/event/530/ END:VEVENT END:VCALENDAR