World Aquaculture Singapore 2026

June 2-5, 2026

Singapore

FISH WEIGHT ESTIMATION USING MANUAL SAMPLING AND BIOENERGETICS MODELING: A DATA ASSIMILATION APPROACH

Alaa Jamal*,  Mohamad Saadi, Ofer Hadar, Jaap van Rijn

Institute of Agricultural and Biosystems Engineering, Agricultural Research Organization

alaa@agri.gov.il



Accurate fish weight estimation is vital for sustainable aquaculture, yet manual sampling—the most common practice—is labor-intensive, disruptive, and often inaccurate when based on small subsamples. Computer vision methods offer alternatives but face limitations in commercial settings, while bioenergetics growth models remain constrained by static, non-site-specific parameters. We introduce a Bioenergetics–Extended Kalman Filter (BE-EKF) framework that integrates bioenergetics modeling with data assimilation. By coupling the model with the Extended Kalman Filter, the system dynamically updates both fish weight and key parameters whenever new measurements are available, allowing real-time adaptation under variable farming conditions. A 197-day trial with Nile tilapia in a recirculating aquaculture system showed that BE-EKF consistently outperformed static bioenergetics models and manual sampling. Current weight estimates improved by 20–60%, forecast errors were reduced by over 60%, and in cases of small subsamples (<40% of the population), BE-EKF even surpassed manual measurements by correcting bias and variability. Requiring only limited sampling and no additional hardware, BE-EKF provides an affordable, transferable tool for aquaculture. The method enhances monitoring, forecasting, and decision-making, and can be extended to other species, offering a practical pathway toward precision aquaculture.