The system uses a combination of regional weather models, physical and biogeochemical oceanographic models and advanced AI-powered algorithms that ingest real-time data from Innovasea's environmental sensors and open data.
This allows for in-pen predictions of dissolved oxygen (DO) levels and can warn farmers of low DO events up to 48 hours before they occur. Users can easily access the forecasts and set customisable thresholds in an interactive chart located in the environmental monitoring dashboard in Realfish Pro – a cloud-based platform that enables fish farmers to monitor, manage and control key aspects of their farms and is active on more than 700 sites, in 25 countries.
Low DO levels can stress fish populations, leading to reduced appetites, slower growth and a negative impact on overall health and welfare, ultimately leading to increased mortality levels.
"Being able to predict low dissolved oxygen events empowers farmers to determine when to feed with confidence and helps them mitigate against environmental threats before they happen," said Tim Stone, Innovasea vice-president.
"Through this partnership with BiOceanOr, our users will be able to plan feedings better, address changing water conditions, and take a more proactive approach to improve fish health and growth."
The algorithm, trained on over 200 million data points to date, was developed by BiOceanOr in collaboration with several international research agencies. This vast data set allows for more accurate predictions and will continue to improve over time as additional information is collected worldwide.
"Our predictions model is a result of our unique combined expertise in AI-powered analytics, marine biology, fish health and oceanography," explained Gaëtan Fabritius, chief business officer at BiOceanOr. "By integrating predictive services on top of Innovasea's real-time sensors and operation-driven Realfish Pro platform, we can provide farmers with a more complete and actionable service to improve operations and drive sustainable practices."
