2.4.1. Phenomenological Model

The first approach was to apply the phenomenological model developed by Cucci [42] to directly predict the variation of a* in ribeye steaks during storage. This model was based on data acquired in laboratory conditions following the previous experimental design. In a similar way to approaches employed in predictive microbiology, there are two steps in the modeling. First, a Gompertz model describes the shape of the change in parameter a* as a function of two parameters, i.e., lag time and maximum rate of change, as well as initial and final values of a* (Equation (1)).

Then, a linear regression model provides the change in these two parameters as a function of the four experimental factors: oxygen content, storage temperature, aging time, and cutting direction (Equation (2) and (3)).

Sensitivity analysis was performed on the parameters, and confidence intervals on the time to color shift were evaluated through Monte Carlo analysis, the detailed methodology and figures for which are detailed in Cucci [42]. All parameters were used in the analysis. However, Angle and Mat were found to have much less influence than the other factors [42]. Therefore, for the present study figures, cutting direction and meat aging were set at 0° and 14 days, respectively.