In this study, six open-field plot experiments were conducted during two (summer and fall) growing seasons (2018 and 2019) at the University of Maryland Eastern Shore (UMES) Agricultural Experiment Station, within the 2.02 ha NOP-certified organic field in Princess Anne, MD. A randomized complete block design was used on the two sites within the NOP-certified areas for both growing seasons. Each treatment used in this study consisted of four replications with three poultry products: two crops in summer, two crops in fall, and two cover crops. Cucumber and cantaloupe plots were amended with either aged poultry litter (APL) or heat-treated poultry pellets (PP) for crops planted in May as summer crops and previously cropped with hairy vetch (11 kg/ha) or forage radish (5 kg/ha) in October 2017 and 2018. Poultry litter products were applied to plots at 2300 kg/ha, consistent with nutrient management practices and regulations for cantaloupe and cucumber. Aged poultry litter (APL), poultry litter compost (PLC), or heat-treated poultry pellets (PP) were applied to the spinach and radish plots. Spinach and radishes were planted in August (2018 and 2019) as cool-season crops.

The PL was obtained from the UMES poultry open-sided litter storage shed, and PP (NPK, 4–3-4) was obtained from the distributor (Mighty Grow, Inc. P.O Box 526, 870 Edward Loper Rd. Fruitdale, Alabama, USA) and collected for weight measurement and field distribution. The aged poultry litter, APL, was aged in a stack for 6 months, and the composted poultry litter, PLC, was aged approximately 10 years after thermophilic composting was conducted by a validated process. These poultry litter products were well decomposed and did not contain elevated levels of ammonia that would ‘burn’ crops after planting. Treatments were analyzed for dry matter, N, and P content (Moebius-Clune et al., 2016). The Mighty Grow pellets (PP) were applied sub-surface (ss), in 5 cm deep trench bands (2/plot), flanking seed rows, and covered with plot soil. The PL was surface applied likewise in bands flanking seeded rows and then manually tilled into the soil before inoculation with a sterile PL extract with a 3-strain gEc^rif-R cocktail. Cantaloupe and cucumber plots were left bare for 10 days, thereafter black plastic mulch and drip lines were installed prior to planting.

Seedlings of hybrid cantaloupe (Cucumis melo cv. Divergent F1 OG) and slicing cucumbers (Cucumis sativus cv. Marketmore 76 OG) seeds (Johnny’s Selected Seeds, Winslow, Maine) were produced in the UMES greenhouse. Seedlings were transplanted to plots on the NOP-certified organic research/teaching farm in June 2018 and 2019. Plants were irrigated/checked daily and replaced as needed.

Of 16 cantaloupe plots (each 26.79 m²), eight received 8.1 kg of PL, and eight others received 7.5 kg each PP. Amendments were applied either as flanking bands or subsurface trenches adjacent to each seeded row. Prior to planting, plots had been cover-cropped either with Hairy Vetch (HV) or Forage Radish (FR) as part of the soil quality/health maintenance program. Cover crops also preceded cucumber crops. Of 16 cucumber plots (each 7.44 m²), half were amended with PL product by flanking bands and the other half were amended by subsurface trenching with either 2.8 kg/plot of APL, or 2.7 kg/ plot of PP.

Spinach and radish were grown in subsections of 20 plots, 2 m × 1 m/plot, with 1 m x 1 m subplots, in 2018. Four full plots received 0.94 kg of TPL (tilled in by hoe). Four other plots also received 1 kg of PLss (subsurfaced)/plot, four plots received 1.0 kg PP (subsurfaced, ss)/plot, four plots received 1.2 kg of PLC (subsurfaced, ss)/plot and the remaining four plots received 1.24 kg PLC/plot. Poultry product amounts for different treatments were calculated based on the recommended nitrogen requirement for each crop type, the amount supplied by the type of treatment to be applied, and plot sizes.

Three rifampicin-resistant (Rif ^R) strains of non-pathogenic E. coli (TVS, 353 and 354 and 355) on tryptic soy agar with rifampicin cultured plates (TSAR) (Neogen, Lansing, MI; Fisher Scientific, Fair Lawn, NJ) were received from BARC (from Dr. Trevor Suslow, University of California, Davis). These E. coli strains were isolated from different sources: TVS 353 (E. coli W778) was isolated from irrigation water, TVS 354 (E. coli P149) was isolated from lettuce leaves, and TVS 355 (E. coli S19) was isolated from sandy loam soils (Tomás-Callejas et al., 2011). Two colonies of each E. coli strain were sub-cultured onto fresh MacConkey + rifampicin (MAC-R) agar plates (Neogen, Lansing, MI; Fisher Scientific, Fair Lawn, NJ) and incubated at 37°C. Cultured plates were retrieved after 24 h. Isolated colonies were then selected and transferred into 50 mL of half (½) strength tryptic soy broth (TSB + rifampicin)-(Becton, Dickinson, and Company, Sparks, Maryland/Gold Biotechnology, St. Louis, MO) and incubated at 37°C, 24 h. Retrieved turbid cultures (50 mL) were added to their own separate 750 mL sterile PLE (1PLE: 1dH₂O). The solutions were replaced in the incubator for 48 h. at 37°C. Equivalent cell densities for low inoculum (6 log CFU/mL) by combining PLE cultures that were used in field experiments were selected. Bacteria cultures were enumerated by dilution plating of each strain in sterile half-strength TSB on MAC-R agar plates (2 plates /dilution) and incubated at 37°C for 24 h. Bacterial cell counts were performed to determine the dilution factor that was used in inoculation for plots.

Poultry litter extract (PLE) and inoculant preparation as well as application to plots, followed the procedure described for a three-strain cocktail of E. coli rifampicin-resistant (gEc^rif-R) (Sharma et al., 2019). Calculated amounts of each E. coli at 6 log MPN/mL were added to sterile distilled water in a sterile carboy (20 L). This mixture was used to make up 18 L for cucumber and cantaloupe plot inoculant (sterile water and PLE was used as diluent). The contents were mixed thoroughly and poured into a backpack sprayer, which had been time calibrated to spray 1 L of the inoculum (Black and Decker Automatic Sprayer, battery operated 4 gals., 15.14 L).

A 2.54 cm diameter probe was used to collect five soil cores (composited) per plot per sampling event in an “X” pattern across each plot. Black plastic mulch was applied to cucumber and cantaloupe plots and slotted to provide access for soil sampling; access sites were closed after each soil core sampling event. Soil samples were collected from amended, inoculated areas on plots with gloved hands and a sterilized stainless-steel corer. Samples were stored in sealed, labeled sterile 680 g sterile sample bags and transported to the laboratory in an insulated chilled cooler. Soil samples collected on day zero before inoculation (D0B) served to evaluate soil health before amendment application and planting of crops. Samples were collected on D0B to determine the presence or absence of any E. coli or other bacterial pathogens in the soils/plots before inoculation or poultry litter product amendment. Soil samples were also collected immediately after manure and inoculum application (D0A). Subsequent soil sampling was conducted over 30, 60, 90, 120-, and 150-day post-amendment periods. The soil textural class of 2018 plots was sandy loam, and in 2019, plots were repeated on silty loam soil. This difference resulted from in-field rotations as required for organic cultivation by the National Organic Program (NOP).

Soil temperatures were measured periodically by placing a WatchDog weather station (Spectrum WatchDog 1,000 Series Micro Station, (Spectrum Technologies, Inc. East Plainfield, IL 60585) in the open fields for cantaloupe and cucumber. The weather station was pre-calibrated with unique external plug-in sensor ports. The instrument logs data from these sensors at a selected measurement interval (1–60 min). Soil moisture content reported in this study are based on wet and dry weight data from 25 g of each ‘as received’ sample were re-weighed after 24 h drying at 70°C. Dry weights were calculated by subtracting the recorded grams dry weight (gdw) from the original recorded grams wet weight. The moisture content of the soil was calculated as g/dw. This procedure was used to measure the soil’s gravimetric moisture percentage to normalize the bacterial content within the soil. Gravimetric soil moisture (percentage) was calculated as:

Bacterial concentrations from assayed soil samples (1, 30, 60, 90, and 120 dpi) were expressed as Log₁₀ from MPN values (US EPA, 2024). Data were exported, and Analysis of variance (ANOVA) was conducted using the Statistix 9 (Analytical Software, Tallahassee, FL) program to separate the observed data into different components and determine the relationship between the dependent and independent variables. Mean comparisons by controlling the error rate to the specified level were performed using the Tukey HSD All-Pairwise Comparisons Test. Means differences were considered significant at a level of (P < 0.05).

Results for soils cropped with cantaloupe, cucumber, spinach, and radishes showed that survival and persistence of E. coli populations involve several environmental factors as well as on-farm management practices. Sharma et al. (2019) reported that seasonality also contributes to gEc ^rif-R survival duration in manure-amended soils. Populations of E. coli in these soils are affected by temperature which may account for population decline patterns (Semenov et al., 2007). Except for the 2018 season (Figures 1A, 4A), where there was a significant difference in E. coli population for HVPL compared to the other treatments on day 60, no other difference was observed for treatments to population survival for all amended plots (cantaloupes, cucumbers, spinach, and radishes) in the 2018 and 2019 seasons. Pre-season cover cropping with hairy vetch or forage radish and placement methods used as a combined treatment type did not show any major impact on the decline of E. coli populations in cantaloupe or cucumber-cropped soils. In another study, incorporation of three cover crops, mustard greens (Brassica juncea), Sunn hemp (Crotalaria juncea), and buckwheat (Fagopyrum esculentum) resulted in significantly reduced populations of the inoculated generic E. coli strain compared to the control within 10 to 30 days indicating a role for those cover crops in mitigating E. coli survival in soil (Zhao et al., 2023). When compared, aged poultry litter did not show any significant difference in population declines or survival patterns relative to thermophilic poultry litter compost and heat-treated poultry litter pellets (Figures 1A,B, 4A,B, 7B). There were no other forms of animal manure included in this study for comparison. However, studies have reported longer survival for E. coli in poultry litter-amended soils when compared to other types of animal manure such as dairy or horse (Sharma et al., 2016). Other studies by Sharma et al. (2016, 2019) have shown that even though the differences in E. coli survival were significantly noticeable in poultry litter compared to other untreated animal manure, the die-off timelines were within NOP-stipulated harvest wait times. Results from other studies also have confirmed that the NOP stipulations of 90 and 120-day harvest wait times are appropriate for crop production on soils amended with treated and untreated poultry manure.

In general, the decline in gEc^rif-R populations in both seasons (2018 and 2019) for manure placements may be due in part to the similarity in application types. All three methods involved poultry products being applied at a particular depth (approximately 6.35 cm) and eventually being covered by soil. All plots also had been cover cropped in this study. Incorporation of product into the soil at this depth could facilitate litter product decomposition and thereby cause stimulation of a competitive/inhibitory microbial population or a depletion of available, suitable nutrients to sustain E. coli populations (Mahler et al., 1994; Mahler and McDole, 2001). Likewise, the production of decomposition products inhibitory to E. coli may occur to a greater extent than the cumulative conditions affecting E. coli populations closer to the soil surface, such as with the broadcast of BSAAO products (Mahler et al., 1994; Mahler and McDole, 2001). According to studies carried out by Sharma et al. (2019), differences were observed in the survival of E. coli when applied on the surface in comparison to those involving subsurface or minimal tillage. These results indicate that incorporating raw manure into the soil is a technique that may help to limit the survival and persistence of E. coli bacteria. Hutchison et al. (2004) concluded that the amount of time raw manure remains on the soil surface influences the rate at which pathogen populations decline. Conversely, incorporating manure into the soil would prolong the time these pathogens remain viable. However, leaving manure on the soil surface may lead to increasing the risk of spreading pathogens in the environment and potentially to the fresh produce grown on the soil.

For both experimental seasons (2018 and 2019), data in this project showed a relationship between the gEc^rif-R survival and temperature. Higher temperatures supported populations of E. coli, but as temperatures began to decline, E. coli populations declined in cantaloupe and cucumber plots (Figures 2A,B, 4A,B). Similar temporal and moisture patterns were observed in spinach and radish plots (Figures 8A,B) for both experimental seasons (2018 and 2019). Temperature and temporal data showed that E. coli populations survived in warmer months (Figures 2A,B, 5A,B, 8A,B). However, the contrasting soil moisture data from 2018 and 2019 relative to E. coli population changes, suggests the potential influence of confounding factors not measured in these experiments. One such factor is the different soil types present in the experimental plots for 2018 (sandy loam) and 2019 (silty loam). These soils have different structural components which may impact these organisms differently. Clearly, E. coli populations survived well in warm months when the temperature was at or above the maximum daily growing degree temperature (30°C); and populations declined as temperature declined. When temperatures fell below the minimum daily growing temperature (10°C), populations started to die off and eventually achieved undetectable levels. Fluctuations were most noticeable in cantaloupe and cucumber plots, but consistent patterns were present for the spinach and radish plots. The E. coli populations continued to decline after initial temperature declines. However, data for the 2019 experimental season in this project showed that E. coli population decreased along with a decrease in soil moisture. Environmental conditions at a given location and time impact pathogen survival and persistence. Temperature is a major factor, the pattern of which often coincides with the survival pattern of these microbes. Even though this study examined the interrelationship between temperature and soil moisture with the pattern of survivability, these findings also show that no single factor accounts for E. coli survival when levels hover in a moderate range. The data also show that one factor may have a different effect on the survival pattern in contrast to another (Williams et al., 2015). This means that these factors can be re-evaluated, and a more in-depth comparison made with other investigations carried out in other locations that have different environmental trends and patterns such as topography and the type of macro- and micro-flora and fauna present. These environmental factors, especially temperature patterns are complex and differ by region. Moreover, based on the findings in this study, the stipulations will satisfy the regulations of the NOP for the Delmarva region, but the experiment could be repeated in a different region to confirm the strength of these relationships and effectiveness in the agriculture industry.

Commonly, fresh produce production has occurred in an open field where they are continuously exposed to pre-harvested microbial contamination via irrigation water, agriculture soil, and raw or improperly composted manure. The agricultural soil used in production may naturally contain certain pathogens or have been populated by the application of animal manure, such as poultry litter (as used in this research). Animal manure can directly contaminate the food crops grown, especially if the fruit is damaged by bruises or cuts (Iwu and Okoh, 2019). Physical damage can wound the integrity of the produce surface and create a path for microbial contamination. The abrasions not only create areas that harbor contaminants but may also decrease the efficacy of washing this fresh produce for consumption. Therefore, it is important that farmers take into consideration the growth and harvesting of their produce in such a way that minimizes the damage and limits pathogen contamination. In a previous study conducted by Jensen et al. (2006), the transfer and survival of E. coli to lettuce in direct contact with soil amended with raw manure and grown under natural on-farm management practices was determined to be contaminated by E. coli from the surrounding environment and wildlife. In the same research, it was also determined that the fields with the shortest wait time intervals between application of manure and harvest supported the highest percentage of E. coli contamination. Several studies have reported on risk and protective factors associated with farm management procedures, environment, and weather that can collaboratively contribute to the contamination of fresh produce grown on manure amended soils (Park et al., 2014; Pang et al., 2020; Phan-Thien et al., 2020, Pires et al., 2023, Strawn et al., 2013; Xing et al., 2019). The cucumbers and cantaloupes used in the research were placed on plastic mulch to prevent direct contact with the soil, but contamination was detected at 90 dpi. This data explains the contamination of the produce from observed splash of soil particles during heavy rainfall onto the plants and fruits (Allende et al., 2017). In contrast, spinach and radish plots had no plastic mulch and thus were susceptible to contamination from direct contact with amended soil. Radish bulbs directly contacted the soil, so they were more susceptible to microbial contamination. However, as the E. coli population declined, so did transfer of E. coli. By 120 dpi, contamination potential, and its occurrence declined significantly from those immediately post-inoculation, resulting in no contaminated cucumbers, cantaloupes, and radish, with only two spinach plants (from 20 sampled) detected positive for E. coli. As the E. coli population declined, so did the potential of E. coli contamination. Clearly, contamination potential and occurrence declined by 120 dpi, resulting in undetectable contamination in cantaloupes and cucumbers. The probability of irrigation water as a source of microbial contamination was excluded since none of the E. coli inoculum strains (TVS-353, TVS-354, and TVS-355) were detected in the well water used for irrigation, which was tested at the initial stage of this project. Water quality was also tested every 30 days throughout the experiment in 2018 and 2019 to ensure contaminant-free irrigation water. Although contamination occurs at the pre-harvest stages in production, contamination can also happen at post-harvest stages through improper hygiene and handling, equipment, and processing. The pathways of contamination occur mainly through cuts and bruises on the edible portion of the produce (Iwu and Okoh, 2019).