Evaluation of multiple disinfection methods to mitigate the risk of produce contamination by irrigation water.

Water used for frost protection and irrigation is one of the most likely points of pathogen contamination during fruit and vegetable production. Previous studies have focused on chemical rather than microbiological water-quality parameters. Consequently, a knowledge gap exists regarding surface water sanitary quality and the risks associated with the timing and method of application. In response to FDA’s proposed standards for surface water quality, we propose to evaluate the adequacy of three in-line methods for disinfecting frost protection and irrigation water. An ultraviolet light module, a chlorine dioxide injection system, and a peroxyacetic acid injection system will be evaluated based on the reduction of indicator microorganisms (E. coli and fecal coliforms) and the presence or absence of pathogens (Shiga Toxigenic E. coli [STEC]) in a double- cropping system with strawberry and cantaloupe. These disinfection techniques will be compared to non-disinfected, pond water with cattle access and populations of all organisms of interest. In addition to evaluating populations of microorganisms pre- and post-treatment, we will also sample plant tissue during flower, early fruit, peak fruit, and late harvest to determine transfer rates of foodborne pathogens. Yield and quality characteristics for each crop among disinfection treatments will also be compared.

There is currently a need to evaluate disinfection systems in real-world scenarios that can be readily implemented to mitigate food safety risks with irrigation water. This project will conduct in-field evaluations of three irrigation-water disinfection methods: ultraviolet light (UV), chlorine dioxide (ClO2) and peroxyacetic acid (PAA). While these systems have been utilized in wastewater treatment, food processing, and in agriculture water systems a lack of knowledge exists regarding how they will perform when utilized for in-line treatment of surface water to mitigate foodborne pathogens, in this case Shiga Toxigenic E. coli (STEC). The project team will establish a series of plasticulture and bare-ground research plots containing double-cropped strawberries and cantaloupes, and irrigate these crops from a pond that also serves to water beef cattle, with populations of fecal coliforms, generic Escherichia coli as well as Shiga Toxigenic E. coli (STEC). This same pond water and experimental configuration will also be used to frost protect the strawberries. During irrigation and frost protection, water will pass through a sand filter and then be divided across four water treatment systems (UV light, ClO2, PAA and a non-disinfected control). An in-line UV light module will be mounted into the UV irrigation system. Likewise, in-line injectors will be used for the ClO2 and PAA systems. Each research plot will have the ability to be drip and overhead irrigated. During frost protection and irrigation, water samples will be taken from the pond and from the irrigation discharge to evaluate disinfection effectiveness. The team’s research will determine the inactivation of STEC, E. coli, and fecal coliforms in treatment systems as compared to the control. In addition, the transfer and survival of these organisms onto strawberries and cantaloupe in a drip and overhead irrigation setup will be determined. Plant tissue samples will be taken during flower, early fruit, peak fruit, and late harvest. Strawberries will serve as the model crop for frost protection; cantaloupes will provide a late season model crop and will demonstrate differences in potential pathogen contamination between drip and overhead irrigation, as well as bare-ground and plasticulture systems. Yield and quality for each crop among disinfection treatments will also be compared to evaluate treatment impacts on total and marketable yield. The project deliverables will include inactivation rates of STEC, E. coli, and fecal coliforms for each irrigation water disinfection system as well as information regarding transfer of these organisms to produce and impact on produce yield and quality when utilizing indirect and direct irrigation methods and plasticulture and bare-ground cultivation techniques.

Overall, our objective is to determine the applicability of incorporating disinfection technologies into produce irrigation and frost protection systems in order to mitigate surface-source water that has been potentially contaminated with pathogenic organisms. 

Objective 1. Determine inactivation of indicator organisms (E. coli and fecal coliforms) and STEC from a surface-water irrigation source after treatment by sand filtration followed by: 1) UV dosage of 10,000 µW·s/cm2, 2) ClO2 dosage of 20 ppm with 2 minutes of contact time, 3) PAA dosage of 20 ppm with 2 minutes of contact time, or 4) no further treatment (control). 

Objective 2. Determine transfer of pathogen (STEC) and indicator organisms (E. coli) from irrigation water to the fruit of model crops (strawberry and cantaloupe) with the three mitigation strategies as compared to no treatment, utilizing both overhead and drip irrigation delivery.

Objective 1. Determine inactivation of indicator organisms (E. coli and fecal coliforms) and STEC from a surface-water irrigation source after treatment by sand filtration followed by: 1) UV dosage of 10,000 µW·s/cm2, 2) ClO2 dosage of 20 ppm with 2 minutes of contact time, 3) PAA dosage of 20 ppm with 2 minutes of contact time, or 4) no further treatment (control). Calcium Hypochlorite - Ca(ClO)2 Calcium hypochlorite was the chlorine source used during 2014. Stock concentrates were produced using 454 g (1 lb) packs of 68% calcium hypochlorite (i.e., swimming pool shock) to get a stock solution of 12% available chlorine. At this concentration, not all of the inert ingredients within the packs are soluble and thus form a significant precipitant. The precipitant was removed to prevent clogging of the metering pumps. For this project, a 12% concentration was chosen because the PAA solution was also 12%; this allowed both metering pumps to be operated at the same setting. Contact time was provided by adding sufficient pipe volume such that two minutes elapsed before the water was applied to the crop. Overall, calcium hypochlorite performed very well. This product significantly inactivated generic E. coli and STEC organisms as compared with the non-treated control (Table 1 and 2). Generic E. coli and STEC were not detected in 2014 strawberry and tomato crops irrigated with water treated with calcium hypochlorite. The target available chlorine concentration was 20 ppm to ensure the satisfaction of the chlorine demand created by the organic matter in pond-1 and the short contact time; this dosage was higher than needed. It is recommended that producers have an injection system that can provide 10 to 20 ppm of available chlorine, and then the dosage can be lowered until 3 to 5 ppm chlorine residual remains in the water that is applied to plant surfaces. Chlorine Dioxide - ClO2 Chlorine dioxide was the chlorine source during 2015 and was injected at a rate to produce a 10 ppm concentration in the irrigation water. As shown in Tables 3 and 4, this product performed similarly to calcium hypochlorite, and inactivated STEC below detection limits. Generic E. coli was detected twice, but populations were at or below 11 MPN/100 ml. Some plant damage on the chlorine dioxide plots was attributed to the sodium content of the disinfectant solution. This product must be manufactured on site; however, there are now vendors that will provide the chlorine dioxide precursors in smaller packets (as opposed to a shipping container) that produce final product volumes that are reasonable for producers to use for disinfection. It is recommended that producers have the capacity to inject chlorine dioxide at a rate that can produce a 5 to 10 ppm chlorine concentration in the irrigation water. Peroxyacetic Acid - PAA PAA performed very well as a disinfectant of raw surface water, even with a short contact time. PAA seems to have a slightly greater affinity for oxidizing microbes than for the dissolved organic matter, which reduces the potential for dissolved organic matter to interfere with disinfection. PAA also readily decomposes to carbon dioxide and water in the environment. This product is commercially available in several concentrations. It is somewhat difficult to compare the various PAA formulations; this product is a mixture of peracetic acid, hydrogen peroxide, acetic acid and water. Peracetic acid is the primary active ingredient; however, hydrogen peroxide and acetic acid also have disinfectant properties. The solution used for this project was 12% peracetic acid, 18.5% hydrogen peroxide, and 20% acetic acid. The disinfectant performance was very good, but care must be taken when using this product. Initially, the 12% concentrate was used as the stock solution. However, the concentration produced sufficient volatilization that the metering pumps would frequently vapor-lock, allowing large water volumes to pass without treatment. As shown in Table 1, when injection was properly controlled, the compound performed moderately well. Tables 2, 3, and 4 show the improvement with PAA performance achieved with changes in management. This problem was alleviated by diluting this concentrate by 50% (thus doubling the injection rate), and by replacing the diaphragm metering pump with a peristaltic metering pump. A second potential issue with using PAA is the change in water pH; after treatment, the irrigation water dropped from approximately 6.8 to approximately 4.5. The pH change has the potential to acidify the soil and change the availability of nutrients. The issue of pH change needs further research to determine whether this concern is warranted. Ultraviolet Light - UV The results of using UV light are shown in Tables 1, 2, 3, and 4. A particular advantage of using UV light is that a module can be installed on the irrigation pipeline to treat all the water. However, this is also a disadvantage because the pathogen kill-zone is limited to the volume within the module – there is no downstream residual treatment. UV systems are designated by water flow rate and UV intensity. For point-of-use drinking water treatment, the U.S. EPA recommends a UV dosage of 40,000 µW·s/cm2. This value has a two-fold safety factor. Further, a NSF-certified UV system must be able to provide this exposure when about 50% of the transmitted radiation is blocked by a dirty quartz sleeve or by turbid water. Because the UV module has a fixed volume, as the flow rate changes, so does the UV dosage. As such, the user must size the UV device based on the greatest flow rate expected to be treated. For this project, the greater flow rate was during overhead irrigation, for which the water received a 35,000 µW·s/cm2 dosage. During drip irrigation, the dosage was 47,000 µW·s/cm2. As mentioned, UV is very sensitive to turbidity. This project was able to successfully remove pathogens from surface water with turbidities as high as 35 NTUs. There is a notable difference in UV disinfection success between Table 1 and Tables 3, 5, and 7. While UV provided successful treatment in all cases, when the source water was switch to pond-2 (less turbidity, see Table 5), the pathogen reduction was more complete. Based on these findings, it is recommended that a UV device should be selected that can provide a minimum dosage 40,000 µW·s/cm2 at the required flow and that the maximum turbidity should be limited to approximately 30 NTUs. The UV module should be placed for easy maintenance and include an intensity monitor let the operator know when the UV transmission is cannot provide the required dosage. Overall, all disinfection treatments performed better than the untreated positive control and were found to be similar to municipal water (Table 6). This demonstrates that these disinfection methods are promising mitigation strategies that can be applied by growers to reduce risk. They should especially be considered when water will contact the edible portion of the crop. Operator Observations It is important to consider irrigation system start-up. At start-up, the pipelines are not under pressure and the pump will transfer water at a greater rate than during steady-state conditions. As discussed in this report, injection rates and UV dosages have been based on steady-state conditions. Either the start-up water can be diverted until the water flow rate reaches steadystate, or additional disinfection capacity can be added to account for the increased flow rate. If fertigation is used, then the fertigation system needs to be disinfected. Water used to dissolve the nutrients must be from a sanitary source and the injection equipment must be sanitized before use. An alternative is to place the fertigation system before the disinfection system. Objective 2. Determine transfer of pathogen (STEC) and indicator organisms (E. coli) from irrigation water to the fruit of model crops (strawberry and cantaloupe) with the three mitigation strategies as compared to no treatment, utilizing both overhead and drip irrigation delivery. The original intent of this objective was to evaluate the movement of pathogens from the irrigation water onto the model crops, with the anticipation that the plots irrigated with treated water would demonstrate less contamination as compared to the non-treated control plots. To minimize cross-contamination, the plots were separated by curtains and there were blank rows between treatments to increase the separation. As seen in Table 7, the STEC contamination was fairly well distributed across all treatments in the 2014 strawberry trial (24–40%). The untreated control has slightly more contaminated samples (40%) than the treated irrigation blocks, but it is certainly not significant (p > 0.05). These plots were located 100 m (300 ft) downwind of a pasture containing beef cattle (stocking density approximately 1 cow per acre). It was assumed that other environmental pressures (insects, small mammals, birds, and bioaerosols) played a significant role in crop contamination such that any irrigation treatment effects were eliminated. No conclusions could be drawn from this trial. The research plots were reestablished 850 m (2,800 ft) away from this pasture, but still downwind for subsequent trials. Tomatoes were transplanted into the new plots. The tomato trial resulted in only one positive sample, isolated from the control plot, but no significant difference was found between treatment blocks (Table 8; p > 0.05). This crop was only drip irrigated, and the results indicate that drip irrigating tomatoes is a good agricultural practice. Strawberries were transplanted the fall of 2014 and grown out in 2015. As shown in Table 9, the 2015 strawberries were as contaminated as the 2014 crop. This crop did not receive frost protection, thus the drip irrigated plots were not overhead frost protected (a confounding factor from 2014), and still there was equal contamination across all treatments (43–32%). The project team is attempting to find more information about the contaminant sources beyond the irrigation water. Isolates collected from plant samples are being submitted for molecular genetic analysis to hopefully provide more information as to the pathogen reservoir. The final model crop was cabbage. As shown in Table 10, there was minimum contamination (4–7%), but it was found among all treatments, with the exception of municipal water. These plots only received drip irrigation, but were significantly closer to the soil than the tomato crop. On the basis of the above results, we cannot draw specific conclusions about pathogen transfer from irrigation water. Other environmental factors are apparently a greater cause of contamination. While the specific vectors are unknown at this time, it appears that their effect overwhelmed any potential treatment effect provided by disinfecting the irrigation water.