The role of riparian zones in bacteria dispersal to produce farms.

Riparian buffer zones have been implicated in the transmission of foodborne pathogens to produce fields and fresh fruits and vegetables, but no one currently understands how the growth of produce in close proximity to riparian zones influences the risk of contamination. The proposed work will measure the movement of fecal bacteria through riparian zones and onto produce fields by detecting the movement of genes from those bacteria. The measured movement of genes from field sampling will be compared to models that represent competing ideas about how the bacteria move. These models produce maps for farms and surrounding lands. They tell us how bacteria move across the land in a manner similar to what roadmaps tell us about the movement of cars. The maps that agree best with the movement patterns of fecal bacteria will be used to advise growers about when, where and how riparian zones increase risk of foodborne pathogen dispersal onto produce. Ultimately, a web-based tool can be developed to apply the best model to new lands and help the produce industry evaluate crop planting decisions, pre-harvest surveillance practices and harvest practices to prevent product contamination.

Riparian buffer zones in agricultural landscapes provide essential ecosystem services while increasing species diversity and habitat connectivity for wildlife and plants. Unfortunately, improved wildlife habitat adjacent to crop production areas may increase risk for contamination of fruits and vegetables. Therefore, buffer zones could serve as transport pathways across landscapes for foodborne pathogens. Accurate information about the role of buffer zones in dispersing foodborne pathogens is essential for the produce industry to strike a balance between the use of buffer zones to enhance soil and water quality while controlling the risks associated with infiltration of produce by wildlife. In this proposal, we will apply an innovative approach, using techniques that have been pioneered in the emerging field of landscape genetics, to systematically answer questions about pathogen movement, sources, and sinks within agricultural landscapes. We propose to conduct causal modeling of the effect of riparian zone and produce field qualities on the movement of fecal bacteria across agricultural landscapes. Competing models will be formulated using remotely-sensed and field-collected data, e.g. vegetation density and type, buffer width, soil habitat quality, produce type, and infiltration rate of host animal feces. These models will provide data to predict produce contamination risk in the landscape context. The data will produce maps of the most efficient paths for fecal bacteria across the landscape. The predictions from these maps will then be compared to genetic data from fecal bacteria isolates collected in the field. The set of models producing good fits to the genetic data will define landscape attributes that promote pathogen dispersal. A key advantage of analyzing fecal bacteria directly, as opposed to tracking host animals, is that this analysis can account for unmeasurable events such as transmission of bacteria among host individuals or even among host animal species during dispersal. These models can also account for persistence across extrahost environments. The product of this research will be: i) a set of landscape attributes that effect dispersal, and ii) a set of characteristic dispersal patterns of fecal bacteria within buffer zones and from buffer zones to crop production areas. These results will inform spatial modeling of food contamination risk associated with buffer zones. This quantitative assessment of the role of riparian buffer zones in the movement of pathogens can be used to modify pre-harvest product sampling strategies and produce harvest methods to account for the spatial structure in contamination risk in a produce field. For example, our project may define a certain proximity to a buffer that carries an increased risk of pathogen presence and thus may not be suitable to grow crops that are susceptible to contamination from the soil. This project directly addresses three specific questions under priority 1.3-Co management of Food Safety and the Environment and questions about movement efficiency of foodborne pathogens between domestic animal and produce operations. This project will specifically support the produce industry by providing scientific data and tools to predict the risk of produce contamination on lands impacted by riparian zones.

Main Objective: Improve the prediction of produce contamination risk by modeling the rules of pathogen dispersal through riparian zones to produce fields. 

Objective 1: Define candidate models for pathogen dispersal based on remotely-sensed and fieldcollected data to predict the dispersal efficiency of fecal bacteria across an agriculture landscape. 

Objective 2: Determine the genetic distribution of fecal bacteria from fecal and soil samples at participating farms in the landscape using multilocus sequence typing. 

Objective 3: Conduct goodness-of-fit tests comparing the fecal bacteria genetic distribution on the landscape with predictions of high efficiency dispersal pathways calculated from the models.

Key findings 

• The dispersal dynamics of E. coli varied by landscape (with the landscapes studied differing by the proportion of forest land) and forested areas may act as a reservoir of extrahost E. coli. 

• Different dispersal behaviors were exhibited by different E. coli subgroups. Our ability to distinguish these groups was crucial to finding the strongest possible signals of overland dispersal. 

• E. coli groups D and F were frequently isolated from environmental samples and their populations were structured by extra-host environmental factors. When these groups were included in dispersal analysis of all E. coli, the signal of environmental selection on these groups distorted the pattern of dispersal associated with other E. coli. 

• E. coli group B1A was one of the most frequent isolates, and this group exhibited a relatively strong signal of dispersal with some preference for riparian corridors. 

• E. coli group B1A may be a good target for examining the potential for dispersal of E. coli pathogens with wildlife for numerous reasons. 

• Riparian forests played a role in the overland movement of E. coli. Our dispersal observations indicate that in less forested landscapes, detectable dispersal is largely limited to riparian forests. However, a different set of models may be needed to test the role of riparian zones in less-forested, more agriculturally intensive landscapes. Recommendations based on these findings 

• Our results indicated that future attempts to understand E. coli movements on farms might do better if they focus on E. coli groups B1, B1A and E which are closely related to the E. coli pathogens of greatest concern and exhibit the best correlation to wildlifebased dispersal models. 

• Groups D and F are less relevant from a risk perspective (though they do contain some pathogens) and exhibit strong signals of environmental selection, suggesting that dispersal of these groups is harder to track. If understanding dispersal and/or risk of STEC contamination is the goal, it may be better to not include E. coli isolates representing groups D and F in the data analyses. 

• While the dispersal and persistence of E. coli is complex, our analyses indicate that terrestrial wildlife likely played a role in the movements of E. coli among produce fields. This suggests that future projects should further evaluate the effects, in E. coli dispersal, of barriers that reduce terrestrial wildlife movement. 

• The observation that dispersal dynamics of E. coli varied by landscape suggests that generalization from studies in a specific landscape to other landscapes may provide potentially misleading information on E. coli and pathogen dispersal.