Analysis of the fluctuating patterns of microbial counts in frozen industrial food products

Sequences of industrial microbial counts of frozen apple concentrate, ground beef, carrots and ice cream (two flavors), which included standard plate count (SPC), yeast/molds, coliforms and in beef potential pathogens all had very irregular fluctuating patterns. In almost all the cases the fluctuations were independent, i.e. they had no significant autocorrelation for any lag. All the counts were considered as having a lognormal distribution as a first order approximation and its parameters were used to estimate the frequencies of future events where the counts exceed predetermined levels. Comparison of the estimates with the actually observed frequencies in fresh data sets showed that they were in reasonable agreement. That the same general probabilistic model was applicable to very different microbial populations types in four very different kinds of frozen foods suggests that the irregular fluctuating pattern of the counts is a manifestation of the interplay of many factors, some partly or fully unknown, whose effect varies randomly. Usually they roughly balance one another and the fluctuations remain within a characteristic range. But there is a probability, which can be estimated, that many of these factors will act in unison creating an event of an unusually high (or low) count. Therefore, the irregular fluctuating pattern of the counts should not be considered as a noise to be smoothed or discarded, but as a useful source of information, and the basis for quantitative predictions. (C) 2000 Elsevier Science Ltd.