the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 15 Oct 2025
Modeling E. coli fate and transport in and around a cattle pond
Abstract. Contamination of surface water is a concern for public health. Lands used for animal production are sources of fecal microorganisms that can reach water bodies, impact their quality, and adversely affect their potential uses. Understanding the mechanisms of microbial transport through surface/subsurface flow is imperative to predict surface water contamination and to assign management strategies for enhanced water quality. The aim of this work to develop and test a mechanistic numerical model to simulate watershed-scale surface/subsurface water flow, bacteria release from cow manure, and their fate, as well as transport to a cattle pond. The integrated surface-subsurface hydrological platform HydroGeoSphere (HGS) was the basis for the site-specific model. The pond and its environs were monitored for 15 months for E. coli concentrations, which remained relatively high throughout the study The model was applied to simulate Escherichia coli (E. coli) bacteria transport in a grassed drainage basin grazed by a permanent herd of approximately 50 cattle. Most model parameter values were adopted from the literature. The model explicitly accounted for cow excretion to the pond as a source of microbial contamination. The latter was estimated from the time spent by cows in the pond, which in turn was estimated from imagery obtained with eight trail cameras installed to cover the pond surface. Images were obtained every 15 min. Simulations for two years showed that the non-calibrated model replicated spatiotemporal patterns and peak E. coli concentration reasonably well. The E. coli cumulative flux loaded by cattle excretion directly to the pond was around two orders of magnitude greater than that with the surface flow. The results of this work indicate the opportunity and show the approach to obtaining a moderately accurate forecast of microbes in cattle ponds using only readily available data.
- Preprint
(1954 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-4138', Anonymous Referee #1, 27 Feb 2026
-
AC1: 'Reply on RC1', Yakov Pachepsky, 15 Mar 2026
Reviewer 1’s Comments
Comment
The paper entitled “Modeling E. coli fate and transport in and around a cattle pond” deals with pond water contamination by cattle. The authors developed and tested a mechanistic numerical model to simulate watershed-scale surface/subsurface water flow, bacteria release from cow manure, and their fate, as well as transport to a cattle pond. The paper is generally well written but need some clarifications as listed below.
Response
Thank you for your detailed and constructive feedback on our manuscript. We greatly value the reviewer's suggestions, as they highlight important opportunities to strengthen the presentation and interpretation of our results.
Comment
My main comment is about the consideration of suspended sediment in the model. As bacteria are mostly transported with eroded soil and manure particles, and as it can survive longer in the sediment, I would expect a more thorough discussion on bacteria and sediment. It is not clear to me (1) how the authors considered the effect of sediment and bacteria resuspension by, e.g., cattle trampling, on bacteria concentration in water, and how they simulated soil and manure erosion along with suspended sediment settling and resuspension, and (2) if the authors considered different die-off rates for water column, soil, and pond sediment.
Response
(1) We thank the reviewer for raising this important point.
We agree that sediment-associated bacteria and mechanical resuspension (e.g., cattle trampling) can significantly influence in-stream concentrations. However, the present study focuses on dissolved-phase bacterial transport and does not explicitly simulate sediment erosion, settling, or resuspension processes. In the current version of our model, we have not explicitly accounted for sediment erosion and transport processes, nor for the resuspension of bacteria from bed sediments (including potential effects from cattle trampling or other disturbances). The microbial migration component focuses primarily on direct overland transport from manure application sites (e.g., via runoff and attachment/detachment to soil surfaces) and in-stream decay/inactivation processes, without coupling to a dynamic sediment transport. This simplification was made to maintain computational tractability and to align with the primary objectives of integrating the microbial model into the existing general watershed framework, which itself does not include detailed sediment dynamics.
To address the reviewer's concern, we have added:
(a) a new paragraph to the Mathematical model section (lines 156–159 in the revised manuscript):
"The present model does not explicitly simulate soil and manure erosion, suspended sediment transport, settling, or resuspension. As a result, bacterial transport is represented primarily through direct runoff and overland flow pathways, without accounting for attachment to or release from suspended or bed sediments. However, the model accounts for the release of bacteria from cowpats uploaded onto the soil surface.”
(b) A new paragraph to the Discussion section (lines 439–445 in the revised manuscript) explicitly acknowledging this limitation:
"The present model does not incorporate explicit simulation of soil and manure erosion, suspended sediment transport, settling, or resuspension processes. This omission may underestimate bacterial concentrations in receiving waters during high-flow events, where sediment resuspension, potentially exacerbated by cattle trampling or other disturbances, can act as an important secondary source of fecal indicator bacteria. Numerous studies have highlighted the significance of sediment-associated bacterial transport and resuspension in agricultural watersheds (e.g., Jamieson et al., 2005; Pandey et al., 2012; Bradshaw et al., 2021). Future model extensions could benefit from coupling a sediment transport sub-module to more comprehensively capture these interactions."
We believe this addition provides transparent acknowledgment of the limitation, places our modelling choices in context, and highlights avenues for future improvement without overclaiming the current model's capabilities.
(2) We considered different die-off rates for the water column and soil, as shown in Tables 1 and 2. The pond sediment die-off rate was assumed to be the same as that for the soil.
Comment
- L52 “postharvest processing” not clear
Response
We changed it to “habitats for wildlife”.
Comment
- L70 “so the scope of the problem is not well known” not clear
Response
We deleted this from the text.
Comment
- L248-249 This sounds more like method
Response
We prefer to keep it here because “grassland area of around 60000 m2” is used in the following calculations.
Comment
- L256 “each year” as fare as I understand, only 1 winter period was monitored, not clear
Response
We deleted “each year” from the text.
Comment
- L261 a verb is missing
Response
The statement was rephrased: “The influx of source terms representing E. coli loading from direct cattle excretion into the pond was calculated as described in Section 2.4.3.”
Comment
- L274 “keep the pond from emptying” not clear
Response
We have revised the relevant section of the manuscript to improve precision and readability. Specifically, we replaced the unclear phrasing and added the following explanatory paragraph (lines 287–292 in the revised manuscript):
"The simulated water level in the pond is controlled by a balance of inflows (precipitation and overland runoff) and outflows (evaporation, infiltration through the pond bottom and dam, and any overflow during high-precipitation events). To ensure realistic pond persistence during the multi-year simulation period—preventing unrealistic complete drying while avoiding excessive overflow, we fitted the hydraulic conductivity of the clay liner at the pond bottom to a value of 0.0002 m/day. Increasing this parameter above 0.0002 m/day resulted in excessive seepage losses, leading to a significant and unrealistic decline in the simulated pond water level over the simulation period, which was inconsistent with observed or expected pond behaviour in the study area."
This revision more clearly explains the role of the hydraulic conductivity parameter in maintaining a stable pond water balance without implying an artificial constraint unrelated to physical processes. The value of 0.0002 m/day (equivalent to approximately 2 × 10⁻⁹ m/s) falls within a realistic range for compacted or amended clay liners used in agricultural ponds or waste storage facilities, where low seepage is desired to retain water while allowing minimal infiltration.
Comment
- L278 “Simulations show that it usually occurs during and after rain events” What does the authors mean? Surface runoff occurring during and after rainfall events is rather obvious. The number of days with rainfall was small? Some rainfall events do not generate surface runoff in the catchment, depending on rainfall typology and soil antecedent conditions?
Response
We removed this statement from the text. Also, see our response to the following comment 8.
We also explained a small but persistent surface water inflow to the pond during dry periods (lines 305-309): “This small but persistent surface water inflow to the pond during dry periods (no precipitation), attributable to slow drainage of shallow subsurface lateral flow (interflow or return flow) from upslope areas that daylights and reaches the pond via surface pathways. This is distinct from precipitation-driven overland flow/surface runoff, which occurs only during or immediately following rainfall events (Tarboton, 2003).”
Comment
- L285-296 does it mean that, in the model, surface runoff also occurs when there is no rain? Do the authors call “runoff” the surface runoff or the overall water yield to the pond? Besides, is the dam completely impervious?
Response
No, the model does not generate surface runoff during periods without precipitation.
In the manuscript (lines 297–300), we explain what we refer to as runoff: “The HGS model calculates fluid fluxes across the pond boundary. Water fluxes are computed between active nodes (located on the pond boundary—dark blue dots in Figure 3a) and contributing nodes (located just outside the pond boundary). The calculated surface water flux represents the simulated runoff to the pond (Figure 7).”
In the simulations, the dam is build of clay with the hydraulic conductivity of 0.0002 m/day (lines 226-227).
Comment
- L318 “Concentration in the pond source locations” is it what the authors call “interior locations” on L365?
Response
Changed to “Concentration in the source locations, simulating cattle excretion in the pond (Figure 3a),”
Comment
- L324-325 What was the sampling protocol regarding the presence of cows in the pond or on the shores at sampling time? Were some samples taken in the vicinity of a cow that would resuspend sediment by trampling and mixing water?
Response
We added the following statement (lines 131-132 of the revised manuscript: “All samples were
taken between 10:00 and 12:00 in the absence of cows.”
Comment
- L337, 338, 339 “shoulder” not clear. On L339, the point in “After the "shoulder" period. The” must be revised.
Response
We changed the jargon term “shoulder” to the commonly used “lag period.”
Comment
- L355 “entering the pond by manure excretion” do the authors mean “by direct excretion of feces into the pond”, as suggested in L359? I would suggest a similar revision of the sentence on L420-421.
Response
Corrected following the reviewer’s comment.
Comment
- L356-357 If I understand well the meaning, I would rephrase as such: “Simulations show that water flowed mainly from the pond into the subsoil, so that E. coli concentrations in the subsoil did not affect the water quality in the pond.”
Response
Rephrased as suggested by the reviewer.
Comment
- L359, 363 remove capital C to “Coli”
Response
Removed.
Comment
- L368-373 what about the role of temperature?
Response
While the effect of temperature was not explicitly studied, it was accounted for in the simulations by introducing a temperature-dependent E. coli die-off rate (lines 237-240).
Citation: https://doi.org/10.5194/egusphere-2025-4138-AC1
-
AC1: 'Reply on RC1', Yakov Pachepsky, 15 Mar 2026
-
RC2: 'Comment on egusphere-2025-4138', Anonymous Referee #2, 01 Mar 2026
The manuscript coded and entitled “egusphere-2025-4138. Modeling E. coli fate and transport in and around a cattle pond” presents interesting results about E. coli inputs and concentrations inside a cattle pond during a 15-month period. The overall presentation is well structured and clear. Methodologies for characterizing the inputs and concentrations are sufficiently described. The number of data and measurements of inputs are adequate. However, the part regarding mathematical modelling could be significantly improved. The authors could conduct a more in-depth literature review to obtain die-off rate values from other references and better fit their value, taking into account a broader bibliographic range. Then they could calibrate their model (presented currently as non-calibrated). Performing a sensitivity analysis of the model to the different parameters included in the equations (tables 1 and 2) is recommended. This sensitivity analysis provides a better understanding of the system behaviour. In addition, authors may wish to consider including more removal processes and comparing the goodness of fit considering a single process (current version) or several (improved version to better understand the functioning of the ecosystem). Considering more processes helps to improve understanding of how the system works and even to be able to make recommendations to promote greater removal of bacteria in the pond.
Comments on formal aspects:
Line 14: the aim of this work is (verb is missing).
Line 17-18: The complete name Escherichia coli should be placed the first time the abbreviation appears in line 17.
Line 338-339. The sentence is cut by a dot.
References are not alphabetically ordered and some of them are cut, making their search difficult.
Citation: https://doi.org/10.5194/egusphere-2025-4138-RC2 -
AC2: 'Reply on RC2', Yakov Pachepsky, 15 Mar 2026
Response to Reviewer 2’s Comments
Comment
The manuscript coded and entitled “egusphere-2025-4138. Modeling E. coli fate and transport in and around a cattle pond” presents interesting results about E. coli inputs and concentrations inside a cattle pond during a 15-month period. The overall presentation is well structured and clear. Methodologies for characterizing the inputs and concentrations are sufficiently described. The number of data and measurements of inputs are adequate.
However, the part regarding mathematical modelling could be significantly improved. The authors could conduct a more in-depth literature review to obtain die-off rate values from other references and better fit their value, taking into account a broader bibliographic range. Then they could calibrate their model (presented currently as non-calibrated). Performing a sensitivity analysis of the model to the different parameters included in the equations (tables 1 and 2) is recommended. This sensitivity analysis provides a better understanding of the system behaviour. In addition, authors may wish to consider including more removal processes and comparing the goodness of fit considering a single process (current version) or several (improved version to better understand the functioning of the ecosystem). Considering more processes helps to improve understanding of how the system works and even to be able to make recommendations to promote greater removal of bacteria in the pond.
Response
Thank you for your detailed and constructive feedback on the mathematical modeling aspects of our manuscript. We greatly value the reviewer's suggestions, as they highlight important opportunities to strengthen the presentation and interpretation of our results.
We fully agree that an expanded literature review on bacterial die-off rates, model calibration, sensitivity analysis, and the inclusion of additional removal processes (e.g., sedimentation, filtration, predation, or UV inactivation in pond systems) would provide deeper insights into fecal indicator bacteria dynamics in manure-impacted systems. Similarly, performing calibration against site-specific observations, conducting formal sensitivity analyses (e.g., using methods such as Latin-Hypercube One-at-a-Time or global approaches), and comparing single- versus multi-process formulations would offer valuable understanding of key controlling factors and potential management implications for improving bacterial removal in ponds or similar ecosystems.
However, the primary objectives of the present study were focused on the conceptual integration of the microbial migration model into the existing general watershed modeling framework and accounting for the microbial water quality in the cattle pond, with an emphasis on demonstrating the feasibility of this coupling, basic process representation, and qualitative evaluation of simulated microbial fate under representative scenarios. As such, formal model calibration against field measurements, detailed sensitivity analyses of the parameters in Tables 1 and 2, and systematic comparison of alternative removal process formulations were not among the aims of this research phase. The die-off rate employed was selected based on representative values from the literature applicable to similar agricultural settings (with appropriate citation), while keeping the model parsimonious to facilitate initial testing and integration.
To address the reviewer's recommendations transparently, we have revised the manuscript as follows:
In the Methods section (around lines 238–245), we have expanded the justification for the chosen die-off rate by referencing published datasets:
“E. coli die-off/inactivation rates exhibit considerable variability depending on environmental conditions, e.g., temperature, moisture, pH, organic matter content, attachment to particles, solar radiation, predation, and matrix type such as manure, soil, runoff, or pond water (Lim and Flint, 1989; Soupir, 2007; Ravva and Korn, 2007; Muirhead, and Littlejohn, 2009; Oliver et al., 2010; Tran et al., 2020). The E. coli die-off parameters in Table 1 and 2 were taken as average over the parameters for datasets presented in published databases. The Q10 model parameters of E. coli survival in water were averages over 16 survivals in wastewater datasets presented in the work of Blaustein et al (2013), and the Q10 model parameters of E. coli survival in manure were averages over seven experimental datasets with bovine manure published by Martinez et al. (2013).”
In the Discussion sections (new paragraph, lines 446–454), we have explicitly acknowledged the absence of calibration and sensitivity analysis:
"The current implementation does not include site-specific calibration or formal sensitivity analysis of key parameters (e.g., die-off rates, sorption coefficients as listed in Tables 1 and 2). While these steps would provide valuable insights into parameter influence and model behavior—and are recommended for future applications, the present work prioritizes proof-of-concept integration and qualitative assessment over quantitative optimization. Similarly, the model currently incorporates a single dominant removal process (first-order die-off/inactivation) to maintain simplicity during initial coupling; additional mechanisms (e.g., settling, resuspension, or biotic interactions) were not included but represent promising extensions. Such enhancements could better elucidate ecosystem functioning and inform management strategies to enhance bacterial removal in pond systems, as supported by prior modeling efforts in watershed-scale microbial fate and transport (e.g., Ferguson et al., 2010; Bradford et al., 2013; Cho et al., 2016)."
These additions aim to provide a clearer context for our modeling choices, recognize the limitations constructively, and position the work as a foundation for more advanced analyses in subsequent research.
Comments on formal aspects:
- Line 14: the aim of this work is (verb is missing).
Response
Corrected.
Comment
- Line 17-18: The complete name Escherichia coli should be placed the first time the abbreviation appears in line 17.
Response
Corrected.
- Line 338-339. The sentence is cut by a dot.
Corrected.
Comment
- References are not alphabetically ordered and some of them are cut, making their search difficult.
Response
Following your comment, we reviewed the references and arranged them alphabetically.
Citation: https://doi.org/10.5194/egusphere-2025-4138-AC2
-
AC2: 'Reply on RC2', Yakov Pachepsky, 15 Mar 2026
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 273 | 175 | 37 | 485 | 28 | 27 |
- HTML: 273
- PDF: 175
- XML: 37
- Total: 485
- BibTeX: 28
- EndNote: 27
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The paper entitled “Modeling E. coli fate and transport in and around a cattle pond” deals with pond water contamination by cattle. The authors developed and tested a mechanistic numerical model to simulate watershed-scale surface/subsurface water flow, bacteria release from cow manure, and their fate, as well as transport to a cattle pond. The paper is generally well written but need some clarifications as listed below.
My main comment is about the consideration of suspended sediment in the model. As bacteria are mostly transported with eroded soil and manure particles, and as it can survive longer in the sediment, I would expect a more thorough discussion on bacteria and sediment. It is not clear to me (1) how the authors considered the effect of sediment and bacteria resuspension by, e.g., cattle trampling, on bacteria concentration in water, and how they simulated soil and manure erosion along with suspended sediment settling and resuspension, and (2) if the authors considered different die-off rates for water column, soil, and pond sediment.
L52 “postharvest processing” not clear
L70 “so the scope of the problem is not well known” not clear
L248-249 This sounds more like method
L256 “each year” as fare as I understand, only 1 winter period was monitored, not clear
L261 a verb is missing
L274 “keep the pond from emptying” not clear
L278 “Simulations show that it usually occurs during and after rain events” What does the authors mean? Surface runoff occurring during and after rainfall events is rather obvious. The number of days with rainfall was small? Some rainfall events do not generate surface runoff in the catchment, depending on rainfall typology and soil antecedent conditions?
L285-296 does it mean that, in the model, surface runoff also occurs when there is no rain? Do the authors call “runoff” the surface runoff or the overall water yield to the pond? Besides, is the dam completely impervious?
L318 “Concentration in the pond source locations” is it what the authors call “interior locations” on L365?
L324-325 What was the sampling protocol regarding the presence of cows in the pond or on the shores at sampling time? Were some samples taken in the vicinity of a cow that would resuspend sediment by trampling and mixing water?
L337, 338, 339 “shoulder” not clear. On L339, the point in “After the "shoulder" period. The” must be revised.
L355 “entering the pond by manure excretion” do the authors mean “by direct excretion of feces into the pond”, as suggested in L359? I would suggest a similar revision of the sentence on L420-421
L356-357 If I understand well the meaning, I would rephrase as such: “Simulations show that water flowed mainly from the pond into the subsoil, so that E. coli concentrations in the subsoil did not affect the water quality in the pond.”
L359, 363 remove capital C to “Coli”
L368-373 what about the role of temperature?