ESTIMATION OF RAINBOW TROUT (ONCORHYNCHUS MYKISS) RESPIRATION RATE WITHIN A COMMERCIAL RACEWAY USING A DATA ASSIMILATION APPROACH

Innovations and decrease in costs of sensors are making it possible to apply to aquaculture the concept of “Precision Livestock Farming”, introduced in the agrifood sector in early 2000. The implementation of the “Precision Fish Farming” (PFF) framework (Fore et. al., 2018) is likely to revolutionize the aquaculture industry, leading to a new generation of softwares and decision support tools, based on dynamic data driven models. In a previous work (Royer et al., 2021), we proposed a PFF based model of oxygen mass-balance for rainbow trout within a commercial raceway, showing that it is possible to dynamically estimate hourly fish respiration rate in commercial farming condition and, on this basis, to improve current control systems. In this study, the estimation method was improved by introducing a data assimilation procedure (Kalman Filter) that allows one to correct the respiration rate as data acquisition goes by and, on this basis, to obtain more accurate short-term predictions of DO concentration.