Maya Avida ’26



Project Title

Deep Learning for Prediction of Ocean Turbulence

Certificate(s): Statistics and Machine Learning, Sustainable Energy

Sea surface height (SSH) is a critical metric for understanding ocean eddies and currents. However, satellites are only able to measure SSH in one dimension, along the track in which they pass over the ocean. The standard method of estimating ocean currents uses optimal interpolation, which is an imperfect deterministic statistical method. This year, my research mentor Scott Martin published a paper demonstrating a new method for gridding SSH that significantly outperformed optimal interpolation by using machine learning to synthesize observations of SSH and sea surface temperature. My project used this new model to predict sea surface height up to 30 days into the future from raw satellite data. Being able to predict SSH forward in time would be very useful to oceanographers for a broad range of benefits, for example aiding oceanographers during field research by predicting the location of future eddies.

Internship Year


Project Category

Oceans and Atmosphere


School of Oceanography, University of Washington - Seattle, Washington


Georgy Manucharyan, Assistant Professor, School of Oceanography, University of Washington; Scott Martin, Ph.D. candidate, School of Oceanography, University of Washington