The Challenge: A Recurring and Costly Threat

HABs are one of the most severe environmental risks for Chilean salmon farming. In 2016, one of the worst blooms wiped out around 12% of national production, with losses above US$800 million, and other events followed in 2021, 2023 and 2024.

Yet farmers still rely on fragmented monitoring based on local sensors and periodic sampling, which offers limited anticipation and little time to act. What they need is:

Early and reliable warnings of HAB events
High spatial resolution, at farm and ACS level
Robust, validated risk indicators
Intuitive dashboards designed for operational decision-making

The Solution: An Early-Warning Service to Anticipate HAB Risks

PREDIFAN combines data sources in a single, user-friendly platform:

Satellite data for offshore phytoplankton activity
In-situ phytoplankton measurements for local accuracy
Hydrodynamic models & machine learning

Together, they make it possible to simulate how blooms grow and drift. The results feed site-specific risk indicators, with daily maps, 72-hour forecasts and automatic alerts when critical thresholds are exceeded.

The key innovation is twofold. Technically, PREDIFAN is a hybrid solution merging space data, field data and modelling. Functionally, it shifts farmers from reactive monitoring to risk anticipation, giving them time to adjust feeding, activate aeration or relocate cages before impacts occur.

A European Consortium Strongly Anchored in Chile

PREDIFAN is led by BiOceanOr, alongside Inria, the French national research institute for digital science and technology, and ACRI-ST, a specialist in Earth observation, with ESA funding.

On the ground, the service is developed hand in hand with Chilean salmon producers such as Blumar and Invermar, involved as pilot partners. Their feedback keeps the tool anchored in real operational needs, from design to testing.

The project draws on BiOceanOr’s expertise in:

AI and marine data science
Phytoplankton dynamics and HABs
Oceanography and hydrodynamics
Fish health biology

Where It Happens

PREDIFAN focuses on southern Chile, in the Los Lagos and Aysén regions, which host the highest concentration of salmon farms and face very frequent phytoplankton issues. This makes them the areas where early warning is most needed. The platform, however, is intended to serve the whole industry, and active collaboration with salmon farmers is key to the project’s success, in Chile and beyond.

Next Steps

The HAB risk models are being finalized as well as the first product interface. The coming months will focus on bringing the service to users, with key milestones being:

December 2026: first beta testers start using the platform
Early 2027 to June 2027: operational testing with beta testers in real conditions, through the high-risk season

 

Project Highlights at a Glance