Modeling Oceanic Variables with Dynamic Graph Neural Networks

TL;DR

This paper introduces a data-driven approach combining sequence models and graph neural networks to predict oceanic variables, outperforming traditional numerical methods in complex, time-sensitive environments.

Contribution

The paper presents a novel end-to-end framework integrating LSTM, Transformers, and GNNs for ocean variable prediction, reducing domain knowledge dependency.

Findings

Our model outperforms the Santos Operational Forecasting System (SOFS).

The approach effectively captures temporal and spatial dependencies.

Results demonstrate improved accuracy and flexibility.

Abstract

Researchers typically resort to numerical methods to understand and predict ocean dynamics, a key task in mastering environmental phenomena. Such methods may not be suitable in scenarios where the topographic map is complex, knowledge about the underlying processes is incomplete, or the application is time critical. On the other hand, if ocean dynamics are observed, they can be exploited by recent machine learning methods. In this paper we describe a data-driven method to predict environmental variables such as current velocity and sea surface height in the region of Santos-Sao Vicente-Bertioga Estuarine System in the southeastern coast of Brazil. Our model exploits both temporal and spatial inductive biases by joining state-of-the-art sequence models (LSTM and Transformers) and relational models (Graph Neural Networks) in an end-to-end framework that learns both the temporal features and the spatial relationship shared among observation sites. We compare our results with the Santos Operational Forecasting System (SOFS). Experiments show that better results are attained by our model, while maintaining flexibility and little domain knowledge dependency.