Application of logistic statistical modelling in the evaluation of suitable conditions for the supply of fresh tomatoes in selected South African supply chains

In the present work, a novel statistical modelling approach was employed to develop tomato quality models based on their physicochemical and subjective quality changes during transportation and storage, and to predict the chances of tomato fruit marketability. Seven disinfection treatments, two storage environments and three transportation conditions were subjected to tomatoes of three maturities harvested in summer and winter seasons. A binary variable, based on fruit marketability, was used to predict the probability of tomato fruit marketability under various disinfection treatments, storage and transportation conditions. This approach is ideal for analysis due to the large number of experimental factors involved. The probability of fruit marketability was comparatively lower for tomatoes transported on rough roads, as compared to those transported on smoother roads. However, tomatoes transported on moderately rough roads that were furthest from the market had the lowest probability of marketability. Tomatoes harvested at the green maturity stage, transported on the shortest, smoothest road (designated by low International roughness index (IRI) values), stored under refrigerated environment and treated with anolyte water, combined with biocontrol, resulted in tomatoes with the highest probability of marketability. The hue angle (h), firmness, pH and mass loss were good predictors of the probability of tomatoes marketability. The firmness and h, however, contributed heavily to the model’s predictive ability. Humidifying ambient storage rooms during winter was also shown to be a critical operation that can potentially increase the probability of tomatoes marketability harvested during winter. The models that were developed can be used by tomato industry players to aid in the selection of appropriate fresh tomato supply conditions.