Modelling of inactivation kinetics of Escherichia coli and Listeria monocytogenes on grass carp treated by combining ultrasound with plasma functionalized buffer

•DBD plasma system was a significant source of reactive chemistries.•US and PFB caused 0.26–2.88 log CFU/g reductions of bacterial population in fish.•UPFB presented synergy with improved reductions of 1.98–3.92 log CFU/g.•Non-linear models were more accurate for describing bacterial inactivation.•Biphasic model was the best fit for predicting bacterial inactivation kinetics. Linear (first-order) and non-linear (Weibull, biphasic, and log-logistic) models were evaluated for predicting the inactivation kinetics of Escherichia coli and Listeria monocytogenes on grass carp treated by a novel technique (UPFB) combining ultrasound (US) with plasma functionalized buffer (PFB). Results showed that UPFB was more effective for inactivating bacteria when compared with individual applications of US or PFB with reductions of 3.92 and 3.70 log CFU/g for Escherichia coli and Listeria monocytogenes, respectively. Compared with the linear model, the three non-linear models presented comparable performances and were more suitable for describing the inactivation kinetics with superior adj-R2 (0.962–0.999), accuracies (0.970–1.006) and bias factors (0.995–1.031), and by assessing the strengths of evidence, weights of evidence and evidence ratios for the models, the biphasic model was identified as the best fit model. The current study provided new insights into the effective evaluation of decontamination methods.