the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 19 Jun 2025
Identification of erosion hotspots and scale-dependent runoff controls on sediment transport in an agricultural catchment
Abstract. Understanding how agricultural land management influences sediment transport is crucial for identifying critical source areas (CSAs) and improving erosion mitigation strategies. While numerous studies focus on in-stream sediment concentrations, fewer investigate overland flow on the hillslopes. We monitored streamflow and sediment fluxes at an overland flow station (E2) and an in-stream station (MW) across 55 runoff events (2011–2022) in the Hydrological Open Air Laboratory (HOAL), Austria. The catchment was segmented into four distinct areas (A, B, GW9, C) based on topography, hydrological connectivity, and proximity to the stream, allowing a spatially explicit assessment of erosion hotspots. Temporal patterns of sediment transport were analysed to infer spatial variability, and differences in sediment transport dynamics among areas were quantified using Kruskal-Wallis tests and effect size analysis. Results suggest that at E2 (hillslope scale), non-erosive cultivation significantly reduced peak turbidity (~9.5 times) and sediment load (~3.8 times) in flat agricultural areas (7.2 % slope, <500 m from the stream) but had no measurable effect in steep (10–12 % slope) or distant (>1000 m) agricultural areas. Across all field types, conversion to non-erosive cultivation did not affect peak flow. At MW (catchment scale), compared to E2, peak turbidity at MW decreased (~5.4–7.7 times) due to dilution from subsurface flow contributions, while peak flow increased (~2.8–11 times) due to additional inputs from wetlands, springs, and subsurface flows. Sediment load at MW was ~2.4–5.4 times higher than at E2, likely due to unmonitored diffuse overland flow and sediment inputs from tile drainages. Our findings indicate that non-erosive cultivation alone in steep terrains or distant agricultural areas is insufficient to effectively mitigate sediment transport. Effective sediment management in agricultural catchments requires spatially targeted erosion control strategies that account for topography, hydrological connectivity, and field proximity to streams.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Hydrology and Earth System Sciences. The authors also have no other competing interests to declare.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(3016 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 04 Jan 2026)
-
RC1: 'Comment on egusphere-2025-2541', Anonymous Referee #1, 24 Jul 2025
reply
-
AC1: 'Reply on RC1', Christopher Thoma, 31 Jul 2025
reply
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-2541/egusphere-2025-2541-AC1-supplement.pdf
-
CC1: 'Reply on AC1', José Carlos de Araújo, 08 Oct 2025
reply
The text is sound and can be accepted for publication at HESS. It tackles an interesting and relevant topic (erosion hotspots and their relation to runoff and sediment transport) and uses primary data for the analysis. the text is well written and the figures are meaningful. Despite its merits, the manuscript needs some improvement before publication.
Major review.
Lines 117-119. You mention "sediment load change across spatial scales" as one of your aims. However, the experimental catchment and the stream reach are small (66 ha and 620 m, respectively). Do you consider that this experimental setup allows you to investigate different "spatial scales"?
3. Methods. A table with the data availability (e.g., number of samples, monitoring period, method...) would be desirable. Besides, neither a table, nor a figure with the events' characteristics is presented, which hampers the readership possibility of analyzing the results.
Line 234. How did you measure the total kinetic energy?
Lines 271-272. Please justify the selection of the trigger values (5 Ls-1, 2 Ls-1, 100 mgL-1) and comment how these values may intefere on the results.
Figures 8 and 9. Why should erosive land cover (especially at this spatial scale) correlate with EI30? Please consider revising the figures.
6. Conclusion. We expect to read some conclusive statements about hydrological and sedimentological processes (e.g., dillution, sediment sources, and macropores, among others).
Minor review.
Lines 56-57. I understand that the numbers (20%, 80%) refer to a specific example, they cannot be used in generalized terms. Please consider rephrasing the sentence.
Figure 2. Please identify the height of the isolines.
Lines 287-290. Please check for the correct reference to Figures 4a, 4b, 4c, and 4d.
Figures 4b and 4d. Flow volumes vary two orders of magnitude for similar peak flow. How valid is the use of peak flo as a proxy in these cases?
Table 2. Please provide the size of each area. This is particularly relevant to compare the results at E2 and MW.
The term EI30 is (correctly) defined as 'erosivity', but throughout the text and figure captions, we read that EI30 is "intensity". Please revise it.
Lines 657-658. Where were the data from Frau 1 and Frau 2 presented? What do you mean by "previously assumed"?
Citation: https://doi.org/10.5194/egusphere-2025-2541-CC1 -
AC2: 'Reply on CC1', Christopher Thoma, 17 Oct 2025
reply
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2025/egusphere-2025-2541/egusphere-2025-2541-AC2-supplement.pdf
-
CC2: 'Reply on CC1', José Carlos de Araújo, 24 Oct 2025
reply
The authors have provided a complete review of our requests and suggestions. We have no further comments and agree with the publication of the manuscript.
Citation: https://doi.org/10.5194/egusphere-2025-2541-CC2
-
AC2: 'Reply on CC1', Christopher Thoma, 17 Oct 2025
reply
-
CC1: 'Reply on AC1', José Carlos de Araújo, 08 Oct 2025
reply
-
AC1: 'Reply on RC1', Christopher Thoma, 31 Jul 2025
reply
-
RC2: 'Comment on egusphere-2025-2541', John Quinton, 01 Dec 2025
reply
This is an interesting paper that provides a lot of detail on the transfer of sediment through the HOAL catchment. However, because of the volume of information the paper struggles to move from a detailed study of the HOAL catchment to a paper which highlights findings that have wider interest to the hydrological community beyond those working on HOAL. This is a major weakness and will require a major revision.
The analysis is largely based around means and deviations around the means for different characteristics of the hydro/sedigraphs. Not surprisingly there is a lot of variability which makes it hard to see if there are any differences. I wonder if there are better ways of analysing these time series which pair the data in some way. For example paired ratios or erosive to non-erosive land use for individual events. This requires a significant effort.
Currently the paper reads like a report or chapter on the HOAL catchment rather than a paper suitable for publication in HESS.
I have made a large number of comments below.
L142 Figure 2.
It is hard to see these points on the map. Perhaps include an inset map with the stream
Is the non erosve cultivation always in the same place?
Needs to separate out the pathways from the monitoring points. For example: 'this pathway is important for these reasons and is monitored at this point using this kit'
L149 Figs 3a and b are flumes not gullies. E1 and E2 are presumably flumes. Separate the monitoring from the features
L154 Same point. these are flumes not tile drainage systems.
L156 Is this from the surface or subsurface?
Why aren't Frau 1 and Frau 2 monitored?
L180 Again separate monitoring from the form/pathway. e.g. a) Flume used to monitor an overland flow pathway
L188 Is this a combination/rotation of crops or is only one crop grown per year ? Not clear
L189 Is it permanent grassland?
L195 Precipitation was measured
L207 How good was this relationship? Was it affected by particle size? You need to tell us.
L221 It would be good to have the planting and harvesting schedule as supplementary information. Could be a data base file
L249 Table 1. Can you give us a spatial feel for where the erosive cropping takes place? I realise that it might be impractical to have ten maps in the main paper, but perhaps in the suplementary information/
L254 define direct flow. Is it the same as overland flow? make sure you are consistent with your terms
L317 Replace This with Thus, the ...
L320 can you provide a distance in m?Long and shorter could mean 1 cm or 1 km
Are these slope are so different? The range of slopes appears b to be quite narrow (9.7 to 11.5%) - I certainly wouldn't describe 11.5% as very steep. I suggest you just refer to the slope steepnesses in %
L351 Avoid these subjective steepness terms
L354 You have this information in the text. Either have it in a table or text, but not both
L392. Avoid subjective descriptors of steepness and distance
L401 quote to p<0.01
L404 How did you get g/l as your turbidity unit. Did you calibrate and if so how good was the calibration and did you take account of that uncertainty when testing for differences. This needs to be described
L407 Figure 6. I wonder if there is a better way of looking at this data since you clearly have a lot of variability caused by different event sizes which makes it hard to see differences. Could you for example look at the ratios of Erosive:Non Erosive and see how that relates to event size?
Otherwise you have lot of graphs which aren't very interesting and could probably be dropped from the manuscript
L416 But this is data from across all events where the variability will prevent you finding significant differences. The data needs to be paired in some way - see my previous comment.
L444. I am not convice by the graphs in Figure 7. They seem to repeat the information in Table 4.
L471 Figure 8 looks like a Table to me. Which location does the analysis relate to? Some of the correlations reported are non sensical. Why for example would you expect a correlation between EI30 and erosive area? Do not present correlations for things which you know are not correlated. Comment also applies to ‘Figure 9’
L550 The areas you refer to are not flat! They have a slope of 7.2%!
L560 Here and in other places in the discussion (Figure 10, Figure 11) you introduce new results in the form of observations. These need to be in the results.
L580 This is an important finding. Put it at the front of the paragraph then discuss it.
L627 This seems like a an important finding which has more generic value than some of the very site specific findings that have been discussed above. I would recommend making it more prominent.
L637-646 Reads like results rather than discussion
L671 A key point that is worthy of discussion. Place it at the top of the paragraph then discuss.
Citation: https://doi.org/10.5194/egusphere-2025-2541-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 962 | 75 | 34 | 1,071 | 27 | 41 |
- HTML: 962
- PDF: 75
- XML: 34
- Total: 1,071
- BibTeX: 27
- EndNote: 41
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
General comments
The manuscript presents an experimental analysis of the factor contributing to runoff, erosion and sediment load, at the small (66 ha) headwater catchment scale. The results suggest that considering both catchment structural connectivity and crop type (erosive vs non-erosive) is needed to assess the effect of management pratices on sediment load and peak flow.
The assessment of sediment source and field-to-stream connectivity at the catchment scale is a current research question. The additional effects of agricultural conservation practices on water and sediment dynamics at the catchment scale is an additional interesting and relevant scientific question. The studied catchment presents a high-quality database of traditional hydrological gauging stations, including high-frequency rainfall, runoff, streamflow and tile drainage monitoring of water and sediment load.
However, the manuscript presents major issues that preclude publication.
First, calculations are hard or not possible to understand. Particularly, the assessment of sediment load values, a central point in this study, is unclear. Was the turbidity-sediment concentration rating curve of good quality? How was noise on turbidimeter values processed? The authors alternatively used turbidity and sediment load values in the analysis, but what is the point in analysing turbidity if sediment load values were available? Evaluating the robustness of the results is therefore not possible.
The methodology used for analysis is unclear. From my understanding, the authors chose to focus on peak values for flow and sediment/turbidity, which is surprising. To analyse the catchment dynamics, why not study the event-scale water volume and sediment load? How did the authors account for hysteresis effects? How was the noise on turbidity values processed? Both may have significant implications for the robustness of the results, particularly considering the significant scattering presented in the log-log plots (Figure 4).
The land use classification, which serves as a basis for analysis, is questionable. Defining winter wheat and winter barley as ‘non-erosive’ crops would require a strong justification, particularly in a study addressing the runoff event scale. What about the intra-annual variations of crops growing and agricultural practices, e.g. storm event occurring on ploughed fields vs crusted fields? It is questionable to propose general results such as those proposed in this manuscript without combining the analysis of both soil surface and rainfall dynamics.
Moreover, it is unclear how the authors labelled the different areas (A, B, C, GW9) as ‘erosive’ or ‘non-erosive’ (e.g. in Table 2 and 4), considering that these areas included a mix of erosive and non-erosive crops (e.g. Area A in Figure 2). It is therefore not possible to assess if the main results are supported by the data.
Last, the main message of the manuscript, as indicated in the abstract and conclusion, i.e. the need to consider both cultivation practices and catchment connectivity, lacks novelty, particularly considering that only part of the catchment structural connectivity was considered in the analysis.
As a conclusion, given the lack of novelty, the unclear analysis procedure, and the inability to assess whether the results support the claims presented, I would not recommend publication in HESS.
Specific comments
The title is misleading: ‘identifying erosion hotspots’ would require dedicated studies using e.g. sediment tracing and/or distributed modelling.
Figure 2 is hardly readable, please consider bringing monitoring stations to the foreground and / or to increase dots size. It may help readers to use intuitive names for the monitoring stations over the manuscript, i.e. what is the difference between ‘Sys’ and ‘Frau’? Would it be relevant to change these names for TD (Tile Drain) and other monitoring stations for e.g. S (Streamflow), R (Runoff)… ?
It is misleading to provides Pearson’s r on a scatterplot including regression lines. It is suggested to add correlation coefficients to the correlation matrices, and to indicate determination coefficients in the figures.
Table 3: It is surprising that the peak flow is not significantly affected by tile drainage. Literature underlines the importance of tile drainage in modulating peak flow.
l.493-497: It is not clear how a correlation coefficient can be used to deduce that ‘rainfall erosivity exerts a dominant control over hydrological and sediment transport mechanisms’. It is also surprising that the correlation between EI30 and the sediment dynamics is better at the catchment scale than at the plot scale, while increasing scale should results in a higher complexity.