Introduction
In the current context of climate change, aquaculture is facing new challenges. Extreme events are increasing, and variations in temperature and dissolved oxygen, the main drivers in feeding decisions, become more difficult to predict. This impacts the processes and routines of fish farms, forcing farmers to react to environmental changes rather than to anticipate them . Based on AI and machine learning coupled with biology, Bioceanor has developed a new bio-guided approach to adapt the feeding time window of fish within the next 24 hours, based on environmental condition predictions
Results
Based on temperature and oxygen prediction, we can identify the ideal and less favorable periods for feeding for the next working day . Figure 1A is illustrating a pre-deployment analysis where 57.9% of feed is given outside the optimal periods, resulting in inefficient productivity and economic loss. Figure 1B is illustrating the recommendations given to the farmer for the next days. By adjusting DFI according to oxygen levels, it is possible to optimize feed conversion and reduce the percentage of feed given outside of the optimal period. Proactive and adaptive feed management based on oxygen level forecasts not only ensures that feed is provided at the right time, but also improves farm performance and minimizes environmental impact by reducing waste.
Discussion and conclusion
Our value proposition is unique to anticipate future oxygen concentration to assist farmers not to feed fish when oxygen is below a given threshold, and rather use optimal oxygen windows.