Escherichia coli is a facultative anaerobic bacterium. It is a mesophilic species with optimum growth from 37°C to 42°C, with growth greatly impeded at lower temperatures and declining rapidly at higher temperatures. E. coli has been intensively studied as it serves as a model organism for understanding fundamental biological processes, including gene expression, replication, and metabolism. It is also widely used in medicine for producing proteins and enzymes, as well as containing several pathogenic strains.

E. coli in soils has mostly been investigated in survival studies under laboratory conditions, with soil samples from the field inoculated with specific strains of the bacterium. In these studies, E. coli abundance, commonly measured in CFUs, tends to decrease rapidly after inoculation before levelling off until the bacterium is found to be below detection limits. Reported survival periods range from weeks to months to 336 days, with longer survival times found for soils that were wetter and for generic rather than pathogenic strains (Murphy et al., 2024).

Traditionally, E. coli is considered an indicator of recent fecal contamination. Survival rates of E. coli in soils have been found to decline with high temperatures and aerobic conditions and low soil clay and dissolved organic carbon (DOC) contents (van Elsas et al., 2011). Microbial community structure and soil management practices were also found to have complex influences on E. coli survival rates. Nonetheless, evidence has suggested that certain strains of E. coli can persist in soil environments, particularly in relatively wet and cold pasture soils of Ireland (Brennan et al., 2010a) and the Alps (Texier et al., 2008). Studies have shown that E. coli can survive for extended periods of time in soils amended with manure, suggesting potential for establishment of naturalized populations. The timeframe of survival is an important consideration, given E. coli’s pathogenic potential, and that restrictions on manure and sewage sludge application to soil are time restricted when growing fresh produce (i.e., soft fruits and vegetables; Department for Environment, Food & Rural Affairs, 2018). Factors such as soil moisture, clay content, vegetation density, and rooting can contribute to enhanced persistence of E. coli in soils. Some hydrochars have been shown to adsorb E. coli in column experiments (Chung et al., 2017). However, the impact of biochar on E. coli survival in soil is variable, ranging from no effect with six different biochar types (Alegbeleye and Sant’Ana, 2023) to having biocidal properties against the pathogenic strain E. coli O157:H7, with the efficacy of the biocidal properties being dependent on biochar production temperature and biochar age (Gurtler et al., 2020). In rainfed pastures, the presence of grazing livestock can contribute further to the introduction and maintenance of E coli populations in soil (Texier et al., 2008; Muirhead, 2023). Naturalization of such communities may affect farm soil productivity, as well as have broader off-farm implications for environmental quality and public health.

In their mini-review, van Elsas et al. (2011) questioned the assumption that E. coli shows natural declines in open environments and stated that any environmental reservoir of the bacterium can cause risks to human health: “The factors that determine the rate of survival of particular E. coli strains, and thus the risks (to human health), are not (yet) easily predictable, and our capability to understand the effects of the secondary habitat (especially soil and water resources) on E. coli behaviour will be paramount to our abilities to manage the organism from both environmental and public health perspectives.”

Therefore, to help reduce this knowledge gap, in this study, we aimed to quantify the relative abundance of E. coli in topsoil samples of a rainfed biodiverse pasture, collected 24 months after exclusion of grazing animals, and to investigate correlations between E. coli survival and soil physicochemical variables, including biochar amendment, in a Mediterranean pasture soil.

E. coli was present in composite topsoil samples from all 20 plots (Supplementary Figure S1). The estimated marginal means were 17.0% and 20.2% for control and biochar-amended treatments, respectively, without any significant difference between treatments (Figure 1a). No significant correlations were found between E. coli relative abundance (RA) and soil pH or soil organic matter contents before or after biochar application (Figures 1b–d) or metrics of soil temperature during the growing season (Supplementary Figure S1). For biochar-amended plots, E. coli RA significantly increased with increasing soil moisture content, whereas there was no significant trend for the control plots. Potential indicators of E. coli inputs from animal feces did not show any significant correlations with E. coli RA (Figures 1e,f).

One of the samples was a statistical outlier (2.7 standard deviations removed from the mean in the biochar-amended treatment). Within-treatment variation was low for the control soil (coefficient of variation (CV) of 0.15) and approximately double that for the biochar-amended soil (0.28), after omission of the outlier.

The prevalence of E. coli (100%) in our pasture topsoil samples is higher than in previous work, which did not use a “without grazing” exclusion, that found forest and pasture land use to be the most prevalent for the presence of E. coli with 73% prevalence, compared to 10% prevalence in cropland and 15% in shrubland (Dusek et al., 2018). The analysis used in this study can only report on relative abundance rather than population growth per se. Nevertheless, the results demonstrate that the E. coli population was still maintained at significant levels, as a proportion of the total microbial community investigated in this study over timeframes previously thought to be untenable without further inoculation from inputs such as manures or sewage sludges, especially in such a warm environment (Tran et al., 2020). The higher prevalence in our study may be because the previous study focused on the landscape scale, whereas our study was in a relatively homogeneous, small area, that is, 0.25 ha. In addition, direct observations of runoff during an intense rainfall event, as well as observed soil surface evidence of erosion/deposition at the <1 m² scale throughout the field site following rainfall, suggest that bacteria may have been distributed from cow pats to other areas of the field site before grazing exclusion, thereby contributing to the observed E. coli prevalence. This is also supported by the lack of observed significant correlations between indicators of animal feces inputs (available N and P) and E. coli RA. Texier et al. (2008) found >97% E. coli prevalence in Alpine grassland soils without grazing exclusion using the 16S rRNA method. Another grazing-exclusion field study, in Ireland, using lysimeters, found the prevalence of E. coli in leachate samples to vary from 3% to 48%, with the higher value being for a poor-draining Stagnic Luvisol (Brennan et al., 2010b). One possible reason may be that their use of the CFU method underestimated E. coli presence when it was in a viable but non-culturable (VBNC) state, to which our DNA sequencing method is not sensitive.

One of the biochar-amended treatment plots was a clear statistical extreme outlier for which we could not find any explanation in our auxiliary data. It is also unlikely to be caused by a point source, for example, wild animal excrement, because of the sampling design of six subsamples across each 9-m² plot. Potentially, it was caused by an auxiliary variable that we did not measure, or by a DNA analysis error. Apart from the extreme outlier, the remaining 19 plots showed relatively consistent E. coli RA, that is, 9.3%–24.1%, with no significant difference between the control and biochar-amended treatments. Other biochar amendment studies have found increases, decreases, and no effect in E. coli survival depending on combinations of biochar and soil properties (Ye et al., 2018; Ahmadimoghadam et al., 2023; Alegbeleye and Sant’Ana, 2023). Although the biochar treatment in our study had similar E. coli RA after 24 months of grazing/manure exclusion as the control treatment, it did have double the range and almost double the within-treatment variation (after omitting the outlier). However, increases in range and within-treatment variation are not uncommon for studies where the treatment consists of incorporating a foreign material into the soil (Olmo et al., 2016), and a CV of 0.28 for a field study is still quite low.

Our findings do not conform with the reported environmental causes linked to E. coli decline, that is, high temperatures, aerobic conditions, low soil clay contents, and low DOC contents (van Elsas et al., 2011). If anything, two factors indicate reduced survivability. Specifically, clay content, which was 18% in the soil studied here, can be considered relatively low in this context, because soils with <18% clay are classified as “coarse” (Feeney et al., 2024). Aerobic conditions can be considered relatively high because the soil is not compacted (exclusion of grazing animals), and the good drainage conditions cause a quick decrease in SMC during dry periods. Nevertheless, we still found evidence indicating that in this relatively aerobic soil, there was a trend for E. coli to increase with increasing soil moisture content during the growing season. We only found this trend for biochar-amended plots, which may be in part because of the twice-as-large range in E. coli RA compared to the control treatment. More field data, covering a wider range of soil moisture contents, are required to explore whether this trend is specific to biochar-amended soil or also occurs in control soil. Although we do not have data on DOC, the relatively low SOM content (before biochar amendment) would likely mean that DOC was also likely low, with this factor also likely to decrease E. coli abundance in the soil over time.

Due to the restricted inputs through exclusion of grazing by livestock and rabbits, the observed 17%–20% relative abundance of E. coli after 24 months is suggestive of not only survival, but naturalization of E. coli. The field site has been under the same management since its conversion from a vineyard around 40 years before this study. Antibiotic use in the grazing animals is so sparse that the potential of an antibiotic-resistant strain of E. coli entering the soil that way is extremely unlikely. Various bacterial adaptation mechanisms have been suggested recently to potentially favor E. coli, for example, stress resistance in nutrient-poor environments (Somorin et al., 2016), such as the soil environment included in this study, and other microbiological DNA or transcriptional mechanisms (Sazykin et al., 2024; Nakamoto et al., 2025), which are beyond the scope of this brief research report.

While our results demonstrate that E. coli can persist and even thrive relative to other microorganisms in rainfed pasture soils long after the removal of grazing animals, we acknowledge that our study did not include strain-level identification or functional gene profiling. As such, we cannot determine whether the E. coli populations observed harbor virulence factors or antibiotic resistance genes, which are critical for assessing their potential risks to public and environmental health. As such, any discussion of environmental and public health implications is, therefore, speculative at this point. Nevertheless, the long-term persistence of E. coli in soils raises important questions about its ecological adaptation and the possibility of these populations becoming naturalized components of the soil microbiome.

While the specific impacts are not fully understood, E coli naturalization in these pastoral ecosystems can have both on-farm and off-farm implications (Jones, 1999; Muirhead, 2023), which should be taken into consideration under a One Health approach. At the farm level, this could reflect in altered microbial and nutrient dynamics, thus impacting soil productivity and other soil-based ecosystem services (Cheong et al., 2024). Environmental health implications include potential soil/water contamination (Jones, 1999), splashing during rainfall events, as well as surface runoff into downstream freshwater systems and leaching into groundwater reservoirs (Muirhead, 2023). In addition to directly affecting human communities that rely on these resources, other direct human exposure routes are linked to occupational health and safety for individuals working in or near contaminated soil or water via ingestion, inhalation, or dermal contact, thus potentially leading to gastrointestinal illness if pathogenic strains are present. Concerns in relation to food safety may also arise upon E. coli entering the food chain, particularly in integrated crop-livestock systems (Jones, 1999; Muirhead, 2023; Cheong et al., 2024).

Specific recommendations for research and management involve filling current knowledge gaps in elucidating the potential long-term consequences to soil ecosystems, livestock, and public health, as well as the underlying mechanisms for E. coli survival and human/animal exposure. It is also necessary to further understand the ecological role of naturalized E. coli populations in soil, their interactions with native microbes, and the factors influencing their persistence and pathogenicity. Specifically, we recommend that future studies incorporate metagenomic, whole-genome sequencing, or functional gene assays to explore the phylogenetic diversity, pathogenic potential, and resistance profiles of soil-resident E. coli strains. Such approaches would significantly strengthen risk assessments and guide appropriate monitoring and management strategies.

The findings of this study support the tentative conclusions from two previous field studies that E. coli may become naturalized in the soil (Brennan et al. , 2010a; Texier et al., 2008). This is contrary to a wealth of laboratory and greenhouse studies that report weeks to months of E. coli survival in soil (van Elsas et al., 2011; Murphy et al., 2024). Considering the implications for OneHealth, it would seem prudent to further investigate E. coli presence and survival in field soils. Specifically, the hypothesis that E. coli survives for longer, or has an increased chance of establishing itself in the soil microbial community, in soils that are year-round predominantly wet, is questioned by the results from this study, in which the soils are dry for extended periods of time. Therefore, we recommend that E. coli, including generic and pathogenic strains, be monitored in soils of other Mediterranean field sites, under any land use, that receive organic fertilizers where E. coli may be present (manures, slurries, and low-grade treated sewage), and pastures with and without the presence of grazers up to 2 years after the final amendment with those fertilizers. In addition, monitoring E. coli in the same samples with multiple measurement techniques, for example, culturing and various types of differential abundance methods (Nearing et al., 2022), may provide further insights into how these quantifications relate to each other.

We conclude that E. coli not only survived, but apparently thrived, either growing or maintaining its prevalence relative to the abundance of other microorganisms characterized in this study, in both the control and biochar-amended pasture soil, 24 months after grazing/manure input exclusion. This has implications for monitoring and management requirements as well as policy implications, whereby currently the exclusion period for growing fresh produce globally varies between 90 days and 365 days following manure application. Future research incorporating genomic or functional profiling is key for determining whether soil-resident E. coli represent a benign, naturalized community or a latent risk to human and animal health.