One of the biggest fears of shellfish farmers are tiny algae. Their abundance can now be predicted four weeks in advance.
Areas where shellfish is either harvested or farmed for human consumption can often be affected by harmful algae blooms. In Norway, one of the culprits is called Dinophysis acuminata, a type of algae that can produce dangerous toxins which become a problem in large numbers, such as during algae blooms. When shellfish take them up, and these are being consumed by people, disease outbreaks can occur. Monitoring locally in the water around a farm is possible to avoid public health risks, but the farmers still suffer economically. By being able to forecast these harmful algae blooms, farmers and monitoring authorities have the option to plan strategies to avoid public health risks and economic impacts. But forecasting these has not been easy.
Edson Silva, together with colleagues at the Nansen Center and the Institute for Marine Research, assessed the environmental conditions leading to blooms and developed a forecasting model using machine learning to predict the abundance of the algae affecting northern Norway. They used satellite observations and local monitoring information for toxic species and tested their method for a shellfish farm in the Lyngen fjord. Their model can produce reliable forecasts of the harmful algae up to four weeks in advance.
Their model uses publicly available satellite observations and local monitoring information and can be adapted to basically any region in the world where both are available. This is great news for shellfish farmers and monitoring authorities across the globe!