The microbial quality of water that comes into the edible portion of produce is believed to directly relate to the safety of produce, and metrics describing indicator organisms are commonly used to ensure safety. The US FDA Produce Safety Rule (PSR) sets very specific microbiological water quality metrics for agricultural water that contacts the harvestable portion of produce. Validation of these metrics for agricultural water is essential for produce safety. Water samples (500 mL) from six agricultural ponds were collected during the 2012/2013 and 2013/2014 growing seasons (46 and 44 samples respectively, 540 from all ponds). Microbial indicator populations (total coliforms, generic Escherichia coli, and enterococci) were enumerated, environmental variables (temperature, pH, conductivity, redox potential, and turbidity) measured, and pathogen presence evaluated by PCR. Salmonella isolates were serotyped and analyzed by pulsed-field gel electrophoresis. Following rain events, coliforms increased up to 4.2 log MPN/100 mL. Populations of coliforms and enterococci ranged from 2 to 8 and 1 to 5 log MPN/100 mL, respectively. Microbial indicators did not correlate with environmental variables, except pH (P<0.0001). The invA gene (Salmonella) was detected in 26/540 (4.8%) samples, in all ponds and growing seasons, and 14 serotypes detected. Six STEC genes were detected in samples: hly (83.3%), fliC (51.8%), eaeA (17.4%), rfbE (17.4%), stx-I (32.6%), stx-II (9.4%). While all ponds met the PSR requirements, at least one virulence gene from Salmonella (invA-4.8%) or STEC (stx-I-32.6%, stx-II-9.4%) was detected in each pond. Water quality for tested agricultural ponds, below recommended standards, did not guarantee the absence of pathogens. Investigating the relationships among physicochemical attributes, environmental factors, indicator microorganisms, and pathogen presence allows researchers to have a greater understanding of contamination risks from agricultural surface waters in the field.

The purpose of this study was to evaluate the performance of artificial intelligence tools for the prediction of Salmonella presence and absence in agricultural surface waters based on the population of microbiological indicators (total coliform, generic Escherichia coli, and enterococci) and physicochemical attributes of water (air and water temperature, conductivity, ORP, pH, and turbidity). Previously collected data set from six agricultural ponds monitored for two growing seasons were used for analysis. Classification algorithms including artificial neural networks (ANNs), the nearest neighborhood algorithm (kNN), and support vector machines (SVM) were trained and tested with a 539-point data set for optimum prediction accuracy. Classification accuracy performances were validated with data set (400 samples) collected from different agricultural surface water sources. All tested algorithms yielded the highest accuracy around 75 ± 1% for generic E. coli followed by enterococci (65 ± 5%) and total coliform (60 ± 10%). Classifiers calculated 6-15% higher accuracy ranging from 62 to 66% for turbidity than all other tested physicochemical attributes.Based on E. coli populations measured in other water sources, trained algorithms predicted the presence and absence of Salmonella with an accuracy between 58.15 and 59.23%. The classification performance of ANN, kNN, and SVM algorithms are encouraging for the prediction of Salmonella in agricultural surface waters.

The U.S. Food and Drug Administration (FDA) has defined standards for the microbial quality of agricultural surface water used for irrigation. According to the FDA produce safety rule (PSR), a microbial water quality profile requires analysis of a minimum of 20 samples for Escherichia coli over 2 to 4 years. The geometric mean (GM) level of E. coli should not exceed 126 CFU/100 mL, and the statistical threshold value (STV) should not exceed 410 CFU/100 mL. The water quality profile should be updated by analysis of a minimum of five samples per year. We used an extensive set of data on levels of E. coli and other fecal indicator organisms, the presence or absence of Salmonella, and physicochemical parameters in six agricultural irrigation ponds in West Central Florida to evaluate the empirical and theoretical basis of this PSR. We found highly variable log-transformed E. coli levels, with standard deviations exceeding those assumed in the PSR by up to threefold. Lognormal distributions provided an acceptable fit to the data in most cases but may underestimate extreme levels. Replacing censored data with the detection limit of the microbial tests underestimated the true variability, leading to biased estimates of GM and STV. Maximum likelihood estimation using truncated lognormal distributions is recommended. Twenty samples are not sufficient to characterize the bacteriological quality of irrigation ponds, and a rolling data set of five samples per year used to update GM and STV values results in highly uncertain results and delays in detecting a shift in water quality. In these ponds, E. coli was an adequate predictor of the presence of Salmonella in 150-mL samples, and turbidity was a second significant variable. The variability in levels of E. coli in agricultural water was higher than that anticipated when the PSR was finalized, and more detailed information based on mechanistic modeling is necessary to develop targeted risk management strategies.