A better way to study ocean currents
Source: MIT News
Published on the MIT News website, this article looks at the creation of a machine learning-based model capable of improving the prediction of ocean currents, with the advantage of being able to trace more easily various types of pollution (oil, plastic...)
Another advantage is that it helps rescuers to better locate shipwreck victims. This model is also capable of enabling oceanographers to improve biomass monitoring and better understand climate change.
Figure 1 - First column: ground truth predictions (upper) and divergence (lower). Second column: current predictions.
Third column: divergence estimates. Fourth column: posterior divergence z-values.
To achieve this, MIT computer scientists and oceanographers have collaborated to develop this predictive model. The research team includes Renato Berlinghieri, Brian L. Trippe, David R. Burt, Ryan Giordano, Kaushik Srinivasan, Tamay Özgökmen, Junfei Xia, Tamara Broderick.
MIT computer scientists have proposed to model buoy velocity (longitude and latitude) using Gaussian processes.
More technical details on the design of this learning model are available at: Gaussian processes at the Helm(holtz): A more fluid model for ocean currents, on ArXiv.org.