Accelerating AI Wave Forecast Development with a Shared Platform

TurboSWAN is an innovative machine learning model designed to accelerate the wave forecasting process. The iML-WKW development team is collaborating with Rijkswaterstaat to make the model’s development more reproducible, scalable, and manageable.

To that end, the team will introduce shared tools on the platform to track experiments, automate workflows, and standardize data storage. The team will also link the model to powerful computing environments (high-performance computing), such as Snellius, the Dutch national supercomputer. The goal is not only to make TurboSWAN suitable for operational use but also to develop reusable building blocks for the platform that future AI and machine learning projects in the water and climate domain can utilize.