the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 11 Mar 2025
Technical Note: Extending the SWAT2012 and SWAT+ models to simulate pesticide plant uptake processes
Abstract. The SWAT model is widely used for simulating pesticide fate and transport in agricultural watersheds but currently lacks the ability to represent chemical uptake by plants, which is a significant pathway particularly relevant for stable compounds that can persist in the root zone. To address this limitation, the publicly available SWAT code was modified to incorporate pesticide plant uptake processes, building upon recent improvements in chemical subsurface transport pathways. The implementation calculates chemical plant uptake based on plant water uptake, substance-specific uptake factors, and concentrations of the chemical in soil pore water. The enhanced model was tested in two agricultural catchments using a stable pesticide soil metabolite with known plant uptake characteristics. Results demonstrate that including plant uptake processes reduced metabolite concentrations in streamflow by 5–17 %. The implementation reveals the importance of plant uptake as a sink, particularly for persistent compounds, and provides new capabilities for assessing agricultural pesticide management practices or mitigation strategies and their effects on environmental fate. The functionality has been implemented in both SWAT2012 and SWAT+, with code provided as an electronic supplement to this technical note.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(1417 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1417 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-877', Anonymous Referee #1, 07 Apr 2025
- AC2: 'Reply on RC1', Jens Kiesel, 29 May 2025
-
RC2: 'Comment on egusphere-2025-877', Anonymous Referee #2, 05 May 2025
This is a well written manuscript with the objective clearly stated, which is to describe the work extending SWAT to simulate pesticide plant uptake. This addition represents a major update to the SWAT model which has been widely used for water quality research and assessments throughout the world. Descriptions of the model implementation, calibration and evaluation are clear. It is publishable with minor technical corrections as suggested below.
- P7, ln205. The model without plant uptake looked under-estimating Catchment 2 data in 2013 onward (Fig 3). How would you justify the need for the plant uptake in this case? Over pre 2013, the model without plant uptake seemed performing generally well (albeit at some level of over-/under- predictions). Improvement wasn't obvious with the case of “with plant uptake” (i.e., comparing Fig 3 Plot b and c). Some more explanation should be useful than simply stating “not all sources of the metabolite were considered”.
- P7, ln210. Adding plant uptake to the model essentially includes another “sink” term in the mass balance system. There shouldn’t be surprises to see reduced concentration predictions. For c2, such reduction seems not needed as comment above. Some data interpretation including potential impact of different factors between c1 and c2 should strengthen the statement in this paragraph. An open discussion otherwise is needed if this result does suggest plant uptake is a catchment- (or crop-) dependent process.
- P15-16, Figs.2 and 3 captions. There appeared to be duplicated “b” denoting Plots b and c.
Citation: https://doi.org/10.5194/egusphere-2025-877-RC2 - AC1: 'Reply on RC2', Jens Kiesel, 29 May 2025
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-877', Anonymous Referee #1, 07 Apr 2025
This manuscript presents the integration of a function representing the absorption of pesticides by plants in SWAT (version 2012) and SWAT. It then presents its application for a metabolite on 2 catchments. It concludes that it is important to take this process into account, since the simulations carried out lead to decreases in concentration of 5 to 17% compared to simulations that do not take it into account.
The subject is an interesting one, given the growing interest in the role of metabolites in contamination pressure on aquatic environments or on the human supply of drinking water. However, beyond the interest of providing a code integrating this new functionality, it seemed to me that the manuscript would benefit from being more detailed.
- More details would be appreciated for the description of the soil profile, flow components in it, the formation of metabolites in the profile (for SWAT+) and the way in which the abstraction of water (and pesticides) is modulated according to depth. Are the parameters given for the equations describing the abstraction of water by plants fixed (lines 95 to 105), or can they be varied? Because if it is not possible to modulate the fact that 50% of the water is taken from the top 6% of the soil, pesticides will also be taken as soon as they enter the soil, which may underestimate the risk of infiltration into less biologically active horizons.
- It is surprising that the parameter describing the uptake of pesticides by plants is linked to the basins.bsn file, when it is explained before that it is a parameter that varies with both the plant and the molecule in question. This choice, which may be constrained by the structure of the model, could have been discussed? In the same way, the fact that it is not possible to represent the return of pesticides to the soil at the time of harvest, when the crop residues are left in place, could have been justified.
- The fact of always referring to the PEARL model is also surprising. Whenever PEARL is a model widely used for the approval of substances, the choice of process representation should rather refer to knowledge about the processes rather than to another model (unless it has been validated for the aspects represented, but there is no reference cited on this subject?)
- The part dealing with application and validation in the basins is quite frustrating, because it deals with anonymous basins (only one of the two basins is described in the article by Rathjens et al 2023 to which the reader is referred), for an anonymous metabolite emerging from an anonymous substance. Furthermore, it is said that the model has been recalibrated because new data are available, without further details: it would have been useful to clarify this, especially since, depending on the nature of the data, there is no need to recalibrate the hydrological part, but only the part relating to pesticides. Very little is said about the sampling method used (punctual, time-controlled, flow-controlled? ) which nevertheless influences the interpretation that can be made of the data and their use for calibration.
- The authors could have justified the fact that they only worked on concentrations, and not also on pesticide flows. The representation of observations and simulations in double cumulative form (cumulative curves of the quantities of metabolite exported as a function of the cumulative curve of the volumes discharged) would be interesting to better judge the adequacy between data and model. The analysis of a few targeted periods, in addition to or instead of the entire period, would also be more convincing of the fact that it is the integration of uptake by plants that allows for the improvement of the representation of the observed concentrations.
- In fact, the spatial and temporal scales used to present the application of the developed model are not the most suitable for convincing of the merits of representing plant uptake: at the scale of a 10 or 38 km² catchment, PUF can be seen as an additional calibration parameter... Performing an analysis at a more restricted scale would have been appreciated.
- In Figure 1, it seems that the new routine also modifies the representation of the flow in the drains: yet plants do not withdraw water in this drain section?
Citation: https://doi.org/10.5194/egusphere-2025-877-RC1 - AC2: 'Reply on RC1', Jens Kiesel, 29 May 2025
-
RC2: 'Comment on egusphere-2025-877', Anonymous Referee #2, 05 May 2025
This is a well written manuscript with the objective clearly stated, which is to describe the work extending SWAT to simulate pesticide plant uptake. This addition represents a major update to the SWAT model which has been widely used for water quality research and assessments throughout the world. Descriptions of the model implementation, calibration and evaluation are clear. It is publishable with minor technical corrections as suggested below.
- P7, ln205. The model without plant uptake looked under-estimating Catchment 2 data in 2013 onward (Fig 3). How would you justify the need for the plant uptake in this case? Over pre 2013, the model without plant uptake seemed performing generally well (albeit at some level of over-/under- predictions). Improvement wasn't obvious with the case of “with plant uptake” (i.e., comparing Fig 3 Plot b and c). Some more explanation should be useful than simply stating “not all sources of the metabolite were considered”.
- P7, ln210. Adding plant uptake to the model essentially includes another “sink” term in the mass balance system. There shouldn’t be surprises to see reduced concentration predictions. For c2, such reduction seems not needed as comment above. Some data interpretation including potential impact of different factors between c1 and c2 should strengthen the statement in this paragraph. An open discussion otherwise is needed if this result does suggest plant uptake is a catchment- (or crop-) dependent process.
- P15-16, Figs.2 and 3 captions. There appeared to be duplicated “b” denoting Plots b and c.
Citation: https://doi.org/10.5194/egusphere-2025-877-RC2 - AC1: 'Reply on RC2', Jens Kiesel, 29 May 2025
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 2,757 | 99 | 27 | 2,883 | 42 | 60 |
- HTML: 2,757
- PDF: 99
- XML: 27
- Total: 2,883
- BibTeX: 42
- EndNote: 60
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Hendrik Rathjens
Jeffrey Arnold
Gerald Reinken
Robin Sur
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1417 KB) - Metadata XML
This manuscript presents the integration of a function representing the absorption of pesticides by plants in SWAT (version 2012) and SWAT. It then presents its application for a metabolite on 2 catchments. It concludes that it is important to take this process into account, since the simulations carried out lead to decreases in concentration of 5 to 17% compared to simulations that do not take it into account.
The subject is an interesting one, given the growing interest in the role of metabolites in contamination pressure on aquatic environments or on the human supply of drinking water. However, beyond the interest of providing a code integrating this new functionality, it seemed to me that the manuscript would benefit from being more detailed.
- More details would be appreciated for the description of the soil profile, flow components in it, the formation of metabolites in the profile (for SWAT+) and the way in which the abstraction of water (and pesticides) is modulated according to depth. Are the parameters given for the equations describing the abstraction of water by plants fixed (lines 95 to 105), or can they be varied? Because if it is not possible to modulate the fact that 50% of the water is taken from the top 6% of the soil, pesticides will also be taken as soon as they enter the soil, which may underestimate the risk of infiltration into less biologically active horizons.
- It is surprising that the parameter describing the uptake of pesticides by plants is linked to the basins.bsn file, when it is explained before that it is a parameter that varies with both the plant and the molecule in question. This choice, which may be constrained by the structure of the model, could have been discussed? In the same way, the fact that it is not possible to represent the return of pesticides to the soil at the time of harvest, when the crop residues are left in place, could have been justified.
- The fact of always referring to the PEARL model is also surprising. Whenever PEARL is a model widely used for the approval of substances, the choice of process representation should rather refer to knowledge about the processes rather than to another model (unless it has been validated for the aspects represented, but there is no reference cited on this subject?)
- The part dealing with application and validation in the basins is quite frustrating, because it deals with anonymous basins (only one of the two basins is described in the article by Rathjens et al 2023 to which the reader is referred), for an anonymous metabolite emerging from an anonymous substance. Furthermore, it is said that the model has been recalibrated because new data are available, without further details: it would have been useful to clarify this, especially since, depending on the nature of the data, there is no need to recalibrate the hydrological part, but only the part relating to pesticides. Very little is said about the sampling method used (punctual, time-controlled, flow-controlled? ) which nevertheless influences the interpretation that can be made of the data and their use for calibration.
- The authors could have justified the fact that they only worked on concentrations, and not also on pesticide flows. The representation of observations and simulations in double cumulative form (cumulative curves of the quantities of metabolite exported as a function of the cumulative curve of the volumes discharged) would be interesting to better judge the adequacy between data and model. The analysis of a few targeted periods, in addition to or instead of the entire period, would also be more convincing of the fact that it is the integration of uptake by plants that allows for the improvement of the representation of the observed concentrations.
- In fact, the spatial and temporal scales used to present the application of the developed model are not the most suitable for convincing of the merits of representing plant uptake: at the scale of a 10 or 38 km² catchment, PUF can be seen as an additional calibration parameter... Performing an analysis at a more restricted scale would have been appreciated.
- In Figure 1, it seems that the new routine also modifies the representation of the flow in the drains: yet plants do not withdraw water in this drain section?