A Sentinel-3 foundation model for ocean colour

2025-09-25

AI Summary

Addressing the scarcity and high acquisition cost of labeled data for remote sensing retrieval of chlorophyll concentration and primary productivity in marine science, this paper introduces Ocean-Prithvi—the first Earth observation foundation model tailored for marine optics. Built upon the Prithvi-EO vision transformer architecture, it employs self-supervised masked autoencoding pretraining using unlabeled Sentinel-3 OLCI imagery, effectively integrating large-scale unlabeled data with sparse high-quality annotations. This enables precise modeling of spatial fine-grained structures and spectral response characteristics of ocean color. After task-specific fine-tuning, Ocean-Prithvi achieves significant performance gains over state-of-the-art baselines in both chlorophyll-a concentration estimation and net primary productivity prediction. Results demonstrate its strong generalization capability under few-shot settings and practical utility for ecological monitoring.

Technology Category

• Computer Vision: Remote Sensing / Geospatial AI
• Machine Learning: Deep Neural Architectures and Foundation Models
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Security and Privacy: Data transparency and provenance
• Search and Retrieval-Augmented AI: Large language models for search

Abstract

Artificial Intelligence (AI) Foundation models (FMs), pre-trained on massive unlabelled datasets, have the potential to drastically change AI applications in ocean science, where labelled data are often sparse and expensive to collect. In this work, we describe a new foundation model using the Prithvi-EO Vision Transformer architecture which has been pre-trained to reconstruct data from the Sentinel-3 Ocean and Land Colour Instrument (OLCI). We evaluate the model by fine-tuning on two downstream marine earth observation tasks. We first assess model performance compared to current baseline models used to quantify chlorophyll concentration. We then evaluate the FMs ability to refine remote sensing-based estimates of ocean primary production. Our results demonstrate the utility of self-trained FMs for marine monitoring, in particular for making use of small amounts of high quality labelled data and in capturing detailed spatial patterns of ocean colour whilst matching point observations. We conclude that this new generation of geospatial AI models has the potential to provide more robust, data-driven insights into ocean ecosystems and their role in global climate processes.

Problem

• Developing a foundation model for ocean color using Sentinel-3 satellite data
• Addressing sparse labeled data challenges in ocean science applications
• Improving chlorophyll concentration and ocean primary production estimates

Innovation

• Pre-trained Prithvi-EO Vision Transformer architecture
• Reconstructs data from Sentinel-3 OLCI instrument
• Fine-tuned for chlorophyll and primary production tasks