A biological early-warning system for marine pollution that uses fish larval swimming behaviour, tracked automatically via AI, to detect contamination before it becomes lethal.
A biological early-warning system for marine pollution using larval fish swimming behaviour as a sensitive indicator, combining automated video tracking with AI-based behavioural classification, aimed at real-time deployment in ports and coastal environments. Laboratory exposure experiments have been completed, PAH and microplastic datasets generated, and the automated tracking pipeline is operational. The key finding is that behavioural changes — reduced activity and mobility, increased immobility, abnormal turning — occur before mortality, supporting use as an early-warning tool. Next steps are to complete the AI classifier, validate in mesocosm and harbour conditions, and build a real-time monitoring prototype.
Dr. Igal Berenshtein leads the work at the University of Haifa’s Marine Ecology & Ocean Health Lab, in collaboration with CAMERI, the Coastal and Marine Engineering Research Institute.
