New Deep Learning method accelerates marine pollution predictions in coastal environments
A new paper published in Marine Pollution Bulletin, co-authored by eScience Center RSE Sonja Georgievska and former colleague Yang Liu, introduces an efficient deep learning model to predict the transport of pollutants in coastal environments. The model significantly reduces the time and cost of traditional forecasting methods. The paper results from a project, a collaboration between the Netherlands eScience Center and researchers from international institutions, that focuses on applying deep learning to predict the movement and spread of pollutants like oil and microplastics in coastal areas. This innovative approach forecasts pollutant advection and dispersion, achieving predictions in seconds—orders of magnitude faster than conventional models. The project resulted from the eScience Center’s “Small Scale Initiative” call that supported collaborations between domain experts and machine learning specialists.
The publication highlights the growing importance of computational efficiency in marine pollution monitoring, especially in the aftermath of marine accidents. By dramatically speeding up predictions, the method holds the potential for improving real-time environmental monitoring and decision-making. As the eScience Center continues to engage in projects that merge machine learning, or AI, with environmental science, this collaboration exemplifies the Center’s role in bridging the gap between advanced computational techniques and real-world scientific applications. Follow-up publications and events are expected that further explore the application of machine learning in marine science, underscoring the long-term impact of these research efforts.
Authors and link to full article
Fajardo-Urbina, J.M., Liu, Y., Georgievska, S., Gräwe, U., Clercx, H.J.H., Gerkema, T., Duran-Matute, M., 2024. Efficient deep learning surrogate method for predicting the transport of particle patches in coastal environments. Mar. Pollut. Bull. 209, 117251. https://doi.org/10.1016/j.marpolbul.2024.117251.
Who are we?
The eScience Center is a research organization dedicated to applying research software to answer research questions in any scientific domain through project collaborations. It has the largest concentration of dedicated, high-level research software expertise in the Netherlands. The eScience Center also has a Fellowship Programme and makes all of its software and training materials openly available online. For more information about what we offer, visit esciencecenter.nl.