46 Priyantha I. Hathurusingha and Kenneth R. Davey
Quantitative mathematical models, widely used in the chemical
engineering and process foods industries, offer the potential for insight,
management, and ultimately control of taste- taint in RAS farmed fish.
Modelling of taste-taint chemicals in fish with time is a fairly new concept in
aquaculture management but is gaining wider acceptance due to the associated
lower cost than with direct experimentation.
However, it is important to develop an adequate model to take account of
the dynamic and complex nature of RAS. Here we report a comprehensive,
critical review of taste-taint modelling in RAS fish that includes: an overview
of predictive models for chemical congeners in fish-flesh; models for
predicting accumulation of GSM and MIB in fish-flesh; development and
limitations of predictive models; factors to be considered in developing an
adequate model for RAS fish; benefits of models to the supply chain, a critical
review of an extensive recent experimental validation of one predictive model,
and; risk analyses of predictive models.
AN OVERVIEW OF PREDICTIVE MODELS FOR CHEMICAL
CONGENERS IN FISH-FLESH
Predictive models, widely used in chemical engineering, provide a basis
for evaluating environmental toxicology and risk assessments in farmed fish.
They are increasingly being used by regulators, scientists, and
toxicologists for decision making (Arnot and Gobas, 2004). This include
evaluating the possible environmental hazards from new and existing
chemicals through trophic levels (Arnot and Gobas, 2003), quantifying
maximum daily loading of congeners of interest (Gobas et al., 1998), assessing
potential impacts on biota due to various pollution sources, determining the
effectiveness of control measures (Gobas and Z’Graggen, 1995; Morrison et
al., 1998), providing site-specific information about chemical concentrations
(Arnot and Gobas, 2004), and; quantifying the toxicological fate of chemicals
in the environment (Wen et al., 1999).
However, difficulties are to determine sufficient and adequate data,
particularly because of the complexity of the natural environment, and the
intensive labour and costs needed (Peters, 1991).
In model development an understanding is required of the biological,
physio-chemical and structural factors of the biota that can influence the
kinetics of chemical accumulation. The rate of chemical accumulation is