Validation of geospatial algorithms to predict the prevalence and persistence of pathogens in produce fields to improve GAPs.

Foodborne pathogen contamination of produce in the production environment continues to present a considerable concern and can lead to recalls or even outbreaks. There is a need for further development of science based approaches to assist growers in minimizing the risk of produce preharvest contamination. The purpose of this project is to validate a GIS-based modeling tool that identifies specific locations and times on a produce farm where the prevalence of foodborne pathogens is elevated, and as a result, the risk of produce contamination is higher. This GIS tool can be applied to any location because it utilizes a farms unique combination of landscape characteristics (e.g., proximity to domestic animal operations), soil properties (e.g., soil moisture), and climate (e.g., precipitation) in its prediction process. The implementation of GIS by the produce industry will increase the understanding of factors that promote foodborne pathogen prevalence and persistence on fields, and will assist growers in focusing their food safety efforts using risk-based strategies. Growers will be able to target areas within their farms that are at high risk for contamination and implement more informed field management decisions and science-based strategies (e.g., alteration of cropping schemes) to limit potential produce contamination.

The risk of produce contamination can be reduced if contamination is minimized in the production environment, specifically during the growing and harvest stages. The implementation of science based strategies to limit produce contamination on the farm aligns with the Food Safety Modernization Act’s foundation of prevention-based food safety practices. On-farm produce safety is complicated by the fact that each farm has a distinct combination of topography, land-use interactions and climate. Combinations of these factors influence the ecology and transmission of pathogens, and subsequently impact the risk of produce contamination. Our research group has conducted several extensive field studies and identified key factors predicted to increase the likelihood of pathogen presence in the produce production environment. Based on these identified factors and our group’s expertise in using Geographic Information Systems (GIS) tools, we have developed geospatial algorithms (GA), which can be used to generate predictive maps that identify areas that are more or less likely to be reservoirs of specific pathogens (e.g., Listeria monocytogenes, Salmonella). Therefore, the overall goal of the proposed research is to increase our understanding of pathogen transmission in the produce production environment. This will be accomplished by (i) validating predictive maps developed for individual produce farms by implementation of our GAs using remotely-sensed data in a GIS platform and (ii) examining the risk of produce contamination during and after precipitation events by quantifying the time after precipitation and amount of precipitation where the frequency of pathogen detection is elevated. We will focus on the foodborne pathogen L. monocytogenes in this project, as we have found it at considerably higher prevalence (15%) than Salmonella (4.6%) or shiga toxin-producing E. coli (STEC, 2.7%) in the produce production environment. A statistically robust (e.g., increased power) field study can, thus be conducted at a substantially lower cost, compared to a pathogen of low prevalence. If the GA can be validated (i.e., statistical evidence that L. monocytogenes prevalence is different within areas on a farm), then the GIS framework can be adapted to evaluate potential contamination of Salmonella or STEC at the farm level. We hypothesize that areas identified to have a significantly higher prevalence of L. monocytogenes will likely pose a higher risk for produce contamination than areas not identified to be L. monocytogenes reservoirs. The validation of GIS-enabled modeling will be invaluable to the produce industry because it allows for specific and science-based food safety plans to be developed for individual farms based on identifying specific or likely hazards (e.g., proximity to domesticated animal operations). Knowledge of such hazards will assist growers in managing contamination risks on their farms by evaluating their current prevention-based programs (e.g., Good Agricultural Practices) and implementing new preventive measures. For example, the proposed research may define a level of soil moisture or proximity from domesticated animal operations where the prevalence of L. monocytogenes is significantly lower. This field scale project directly addresses CPS priorities: 1.2 buffer zones from domestic animals to fruit and vegetable production and 1.5 climate, environment, production practices.

The proposed research will use GIS modeling and statistical tools in combination with extensive field sampling to accomplish the following three objectives: 

1. Implement geospatial algorithms to develop predictive maps identifying environmental reservoirs (i.e., well defined spatial areas) for Listeria monocytogenes on produce farms. 

2. Independently validate each geospatial algorithm’s power to predict areas that have a significantly higher prevalence of L. monocytogenes (i.e., high risk areas), as compared to areas identified as having significantly lower prevalence of L. monocytogenes by the algorithm (i.e., low risk areas). 

3. Quantify the effect of precipitation on the frequency of L. monocytogenes detection in high and low risk areas identified by the geospatial algorithms and the risk of L. monocytogenes transfer to produce during or after precipitation events.

Aim I Key findings 

• Proximity to surface water and pasture were significantly associated with L. monocytogenes isolation from produce production environments. 

• By validating two of the risk factors identified in previous models (proximity to water and proximity to pastures) that can be used to predict field areas with increased risk of L. monocytogenes detection, our study demonstrates that geospatial models can be used to accurately and prospectively predict fields and areas in produce fields with an increased risk of pathogen detection. Recommendations 

• Growers for whom L. monocytogenes is a pathogen may want to carefully manage growing areas in proximity to surface water and pastures. Similarly, processors or growers that conduct trace-back investigations (e.g., based on a finished product positive) may want to more heavily focus sampling on field sites in proximity to surface water and pastures when trying to identify pre-harvest L. monocytogenes sources. 

• To facilitate identification of additional risk factors and additional control strategies, future models should account for temporal (e.g., changes in management practices or meteorological factors over time) and farm size. Aim II Key findings 

• Proximity to surface water and roads (i.e., two rules that were the basis for predicted risk in Aim II) were associated with an increased likelihood of isolating L. monocytogenes from soil samples collected in produce fields.

• The likelihood of isolating Listeria spp. and L. monocytogenes was greatest during the 24 h immediately following rain or irrigation events. 

• The diversity of Listeria spp. and L. monocytogenes subtypes (ATs) was lower after irrigation events than after rain events. Recommendations 

• Interventions to reduce the risk of pathogen contamination in fields may need to take into account the water source (i.e., surface water versus rain). 

• Waiting 24 h after irrigation and rain events to harvest crops may significantly reduce the risk of L. monocytogenes contamination. 

• The use of land-use factors to predict risk and to tailor cropping schemes to reduce risk (e.g., planting high risk crops in low risk areas) may be useful for developing targeted on-farm food safety risk management plans.