Summary
Although it is known that flooding can introduce microbiological, chemical, and physical hazards onto croplands, little data is available on the presence and persistence of contamination post-flooding over time. This rapid response project aims to quantify select microbiological indicators and pathogens in flooded fields over time. While pathogens in the soil will usually die-off rapidly over time due to drying conditions or fluctuations in temperature, floodwaters have the potential to contain large amounts of human sewage and runoff from animal production areas that could greatly impact die-off over time. Currently, LGMA recommends a waiting period of 60 days before replanting to minimize the risk of pathogens persisting in the soil into the growing season. This waiting period can be shortened to 30 days with the inclusion of soil testing. In general, comprehensive testing for pathogens is not recommended for all flooding situations, but if there is a reason to believe that the soil is heavily contaminated with human pathogens, food crop producers may want to consider microbial testing following a flood. Depending on the flooding circumstances, pathogens of interest may include the following: -Bacterial pathogens, such as Salmonella, E. coli O157:H7, other Shiga toxin-producing E. coli, and Clostridium perfringens -Viral pathogens, such as hepatitis -Parasites, such as Cryptosporidiumand Giardia NOTE: Resources: Symposium PPT is provided as "other" item.
Technical Abstract
Flooding poses significant challenges for crop production due to the introduction of microbiological, chemical, and physical hazards onto croplands. However, there is a lack of data on the presence and persistence of contamination post-flooding. This study aimed to address this knowledge gap by quantifying select microbiological indicators and pathogens in flooded fields over time. The University of Arizona conducted a three-month (90 days) longitudinal study to evaluate the presence and persistence of microbiological indicators in soil samples from four different ranches. These ranches represented flood impacts from the Salinas River (Ranch S, T) or from creeks/tributaries (Ranch F, H). The study focused on three microbiological indicators: Total Coliform bacteria, Fecal Coliform bacteria, and generic Escherichia coli. Additionally, soil and water samples were analyzed for the presence of two pathogens: Shiga toxin–producing E. coli (STEC) and Salmonella. Variable levels of Fecal Coliform bacteria were detected in soil samples across space and time, ranging from <1 to 1986.3 MPN/gram. The study observed log reductions ranging from -0.28 to 0.34 for Total Coliform bacteria, log reductions ranging from 0.04 to 0.80 for Fecal Coliform bacteria, and log reductions ranging from 0.00 to 0.95 for E. coli bacteria across all fields assessed over the 13-week study. Comparing the presence of pathogens to indicator organisms, approximately 8% of soil samples were presumptive positive for STEC linked targets (stx + eae) using droplet digital PCR. For Salmonella, roughly 5% of soil samples were positive after the first flood, with 3% positive after the second flood, suggesting minimal impact of flood waters on the presence of Salmonella in soils. In contrast, the study recorded very few positive samples for STEC after the first flood. However, after the second flood, up to 47% of samples from all ranches combined showed presumptive-positive results for STEC in soil samples, as determined by culture. These findings raise important questions about the prolonged flushing effect of flood waters, the force of flood waters on bacterial adhesion to soil particles, and the impact of nutrient loading on the survival of pathogens over time. Understanding these factors is crucial for developing effective strategies to mitigate microbiological contamination in flooded fields as well as appropriate metrics to ensure the safety of agricultural products.
Research Objectives
1. To work alongside industry partners to conduct a longitudinal study on the impact of flood water(s) on the presence and persistence of microbiological indicators and pathogens on lands used to grow fresh produce.
Findings & Recommendations
Grower Key Findings:
• Fecal Coliform bacteria may not be the best indictor of pathogen risk.
– Highly variable across space and time
– Not correlated to STEC or Salmonella
• Generic E.coli are much more consistent across space and time.
• Not all flood waters are equal risk. 1.3
• Flood waters from adjacent creeks/tributaries indicated a greater likelihood of detecting pathogens (STEC) in soils.
• Bacterial indicator numbers declined or “recovered” before 30-day interval in all ranches.
• LGMA guidance on re-planting post flood at 60 days is conservative.
See this project's final report PDF to view the tables referenced below. Over the course of the study, our team was able to calculate log reductions of Total Coliform bacteria, Fecal Coliform bacteria, and generic E. coli in soil samples collected in each of the four ranches evaluated as part of the longitudinal study. Log reductions across all fields assayed over the 13-week study ranged from -0.28 to 0.34 for Total Coliform bacteria, 0.04 to 0.80 for Fecal Coliform bacteria, and 0.00 to 0.95 for E. coli bacteria. While these reductions may seem minimal, it is important to recognize that initial bacterial concentrations were not orders of magnitude above those anticipated, thus indicating that at the start of the study concentrations had already declined to relatively “low” values, or that that the impact of flooding on indicators was not as detrimental as originally thought.
The following figures (Figures 2–5) show the average concentrations of each of the three indicator organisms evaluated on each sampling date across the four fields sampled. Fields “T” and “S” are indicative of Salinas River flooding while fields “F” and “H” are representative of tributary/creek flooding from adjacent lands. The date of January 15th, 2023 on each of the graphs below identifies the date of initial flood water receeding from each ranch evaluated.
Figure 2. Bacterial die-off over time at Ranch S impacted by the Salinas River.
Figure 3. Bacterial die-off over time at Ranch T impacted by the Salinas River.*(*Samples collected on 03/08/23 were lost in the mail upon shipping from California to Arizona.)
Figure 4. Bacterial die-off over time at Ranch T impacted by a tributary or creek.
Figure 5. Bacterial die-off over time at Ranch F impacted by a tributary or creek*.(*Samples were not collected on 02/09/23 due to the ranch still being saturated.)
In each of the four figures it can be easily seen that as time passes, with each week of sampling all bacterial indicator organisms decline. With the additional flooding event that occurred after the week of 03/08/23, one can visually note the increase in all parameters measured, followed by a period of decreasing concentration. One important observation is that the variability in the Fecal Coliform bacteria across any individual field is quite high. This was observed across all four ranches evaluated.
Figure 6 below represents a heat-map of Fecal Coliform bacteria measured in individual or composite soil samples collected at field transects of 100, 200, 400, 800, or 1600 feet on each individual sampling date. Blocks indicated as red signifiy higher concentrations of Fecal Coliform bacteria, while blocks highlighted in green indicate a lower concentration of Fecal Coliform bacteria. As a note to the reader, the current LGMA guideline for acceptable criateria to re-plant a previously flooded field is <100 MPN fecal coliform bacteria per gram of soil. Variable levels of Fecal Coliform bacteria were detected in soil samples across space and time, ranging from <1 to 1986.3 MPN/gram. Alternatively, looking at the distribution of generic E. coli from the same ranch (F) over the same sampling period (Figure 7), it is quickly seen that MPN values are much more consistant across space and time.
When evaluating generic E. coli, the threshold value of 10 MPN/gram of soil has been suggested as an alternative criteria for acceptance to re-plant previously flooded fields. We do not see significant differences in grab sample versus composite sample strategies. Figure 6. Fecal Coliform heat-map for Ranch F Figure 7. Generic E. coli heat-map for Ranch F When evaluating the presence of pathogens in soil samples when compared to indicator organisms, 8% of soil samples in total were considered presumptive positive for STEC linked targets by ddPCR. Of the total samples collected, roughly 5% of soil samples were positive for Salmonella after the first flood, with 3% positive after the second flood, indicatiing little impact of flood waters on the presence of Salmonella in soils. Alternatively, after the first flood the research team recorded very few STEC positive samples (Table 2). However, after the second flood, up to 47% of samples from collective ranches had soil samples presumptive positive for STEC by culture. This raises questions regarding the impact of sustained flushing over an extended period of time, the potential force of flood waters on bacterial adhesion to soil particles, and the impact of nutrient loading of pathogen survival over time.
Results indicate that fields adjacent to creeks/tributaries with overland flow had increased likelihood of detecting pathogens than those adjacent to the Salinas River. In total, the research team evaluated five different water sources adjacent to flooded fields, including the Salinas River, Pajaro River, Miller Creek, Alisal Creek, and an unnamed drain. In the evaluation of water sources, Salmonella was found in all sources, whereas STEC was detected exclusively in the creeks/tributaries and rivers. The prevalence of pathogens was significantly higher in the two creeks/tributaries, with positive samples comprising 80% of the total. In contrast, the two rivers showed a lower incidence, accounting for only 20% of positive samples. Notably, no STEC was detected in the collected drain samples, offering further insights into the pathogen distribution across the different water sources.
Additionally, the research team was able to confirm STEC SerO groups more often in samples collected from fields adjacent to flooded creeks/tributaries. While some enteric bacterial strains cause acute outbreaks linked to specific sources, other strains—referred to as reoccurring, emerging, or persisting (REP) strains can reoccur and periodically cause acute outbreaks. They can also emerge and increase in frequency or persist and cause illnesses over periods of months or years, despite investigation and prevention efforts. It is important to note that O157:H7 was not confirmed from any sample collected in the study and that none of the previously reported REP strains, REPEXH01 and REPEXH02, were identified (https://www.cdc.gov/ncezid/dfwed/outbreakresponse/rep-strains.html). When confirming the presumptive pathogens, the following SerO groups were identified as predominant from soil samples (see Table 3).
As mentioned above, physical and chemical parameters were also evaluated for soil samples collected in the study. One parameter that was identified as particularly useful for industry is the gravimetric water content (GWC). Gravimetric water content (%) is the mass of water per mass of dry soil. It is measured by weighing a soil sample, drying the sample to remove the water, then weighing the dried soil. When analyzing the data of GWC collected across all samples and comparing values with those of indicator organisms, the Pearson correlation coefficient (r) = 0.54 for E.coli and gravimetric water content % was calculated. This indicated that as soil moisture increased, E. coli MPN also increased (positive correlation). While notable for this study, this finding has been reported previously in the literature, as soil pH and moisture content are primary drivers of E. coli O157 survival (5). It is important to consider that gravimetric water content is an inexpensive and relatively straightforward parameter to monitor and could be a possible additional monitoring parameter used by LGMA and industry to inform field re-entry post flood.
Table 4 outlines ranges in various soil property measurements collected across all four unique ranches. Little to no sodicity-salinity problems were detected at the 0 6” profile. While high levels of sodium (Na) were detected, no ‘sodicity’ problem was detected; higher levels of Calcium and Magnesium possibly helped to ‘neutralize’ the sodium. It was identified that the pH at two ranches was very high, which can indicate that those ranches may be more prone to Na-problems in the future. However, it should be noted that flooding seemed to help with salinity, indicating that water pushed the salts (Na) down and away from the root zone.
Table 4. Soil physical and chemical analysis.