WP4: Difference between revisions
Created page with "= WP4 – Data-driven models for forecasting, downscaling and data fusion = == Overview == This Work Package develops data-driven Artificial Intelligence (AI) and Machine Learning (ML) methods to improve the quality, resolution, and computational efficiency of reanalysis and forecast products for offshore wind farm (OWF) applications. The objectives are: (i) To design AI methods for improved local forecasts and working window planning. (ii) To develop AI tools for..." |
No edit summary |
||
| (One intermediate revision by the same user not shown) | |||
| Line 17: | Line 17: | ||
* '''Co-lead:''' HEREON | * '''Co-lead:''' HEREON | ||
* '''Contributing partners:''' FMI, SINDBAD, NIVA, SMHI, all WP4 partners | * '''Contributing partners:''' FMI, SINDBAD, NIVA, SMHI, all WP4 partners | ||
---- | |||
== WP4 common Drive Folder == | |||
[https://drive.google.com/drive/folders/1JwRU15WlN4Y71mp0O6CUj52ETFuUmF-z WP4 Folder in Google Drive ] | |||
---- | ---- | ||
== WP4 Meeting Notes == | == WP4 Meeting Notes == | ||
[[WP4_MeetingNotes|WP4 Meeting Notes]] | [[Meetings:WP4_MeetingNotes|WP4 Meeting Notes]] | ||
---- | ---- | ||
Latest revision as of 11:29, 26 January 2026
WP4 – Data-driven models for forecasting, downscaling and data fusion
Overview
This Work Package develops data-driven Artificial Intelligence (AI) and Machine Learning (ML) methods to improve the quality, resolution, and computational efficiency of reanalysis and forecast products for offshore wind farm (OWF) applications. The objectives are: (i) To design AI methods for improved local forecasts and working window planning. (ii) To develop AI tools for downscaling coarse-resolution data. (iii) To create AI-based data-simulation fusion methods for reanalysis and bias correction. (iv) To implement AI-driven emulators to assess OWF impacts on sea ice.
WP4 bridges physical modelling and data science, providing fast and efficient surrogate solutions to support digital twin-based ocean forecasting and OWF impact assessments.
Lead and Participants
- Lead partner: DMI
- Co-lead: HEREON
- Contributing partners: FMI, SINDBAD, NIVA, SMHI, all WP4 partners
WP4 common Drive Folder
WP4 Meeting Notes
Tasks
The WP4 is divided into four main tasks:
- WP4.T1 – AI/ML-based methods for local forecasting (Lead: DMI / HEREON)
Description: Development of optimized wave forecast systems for OWF sites that integrate ML-based forecast error correction and ensemble generation. These models combine physical model outputs and observed data to produce improved site-specific forecasts and weather windows for OWF operations.
- WP4.T2 – AI/ML-based methods for downscaling (Lead: DMI / HEREON / SINDBAD)
Description: Training of deep neural networks (CNNs, GANs) to downscale low-resolution model outputs to fine-scale OWF-relevant domains. The resulting tools will provide high-resolution, low-cost forecasts for decision-making at local OWF sites.
- WP4.T3 – AI/ML-based methods for data-simulation fusion (Leads: DMI / NIVA / SMHI)
Description: Fusion of observations and model data using advanced AI algorithms such as 4DVarNet for improving reanalysis quality and bias correction of climate simulations. Includes: – WP4.T3.1 – Data fusion for improving reanalysis – WP4.T3.2 – Bias correction for climate simulations
- WP4.T4 – AI for assessing the impact of OWFs on sea ice (Lead: FMI / TalTech)
Description: Development of deep learning emulators (CNNs, GNNs) to reproduce sub-kilometer-scale sea ice states influenced by OWFs. These models will help assess physical impacts of turbine foundations on ice dynamics.
Page Links
- WP4.T1 – AI/ML-based methods for local forecasting
- WP4.T2 – AI/ML-based methods for downscaling
- WP4.T3 – AI/ML-based methods for data-simulation fusion
- WP4.T4 – AI for assessing the impact of OWFs on sea ice
Source: DTO4OWE Full Proposal – Research Programme (Work Packages and Tasks).