Sediment Fingerprinting Enables the Determination of Soil Erosion Sources and Sediment Transport Processes in a Topographically Complex Nile Headwater Basin

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The topic of soil erosion is highly relevant and is addressed with an appropriate approach within a specific context, contributing to the identification of intervention areas and providing tools for soil management in these regions. Although the study is limited to a specific region and there is no novelty in the methods, diagnostic approach, or erosion modeling, the reporting of regional progress in addressing soil loss issues is valued.
However, several concerns arise:

• Some formatting aspects are not properly followed, including the organization of sections, references, and figures. A more careful preparation of the manuscript is recommended before submission.
• The main concern is that a similarity check revealed more than 40% overlap with the technical report "Rwanda Integrated Water Security Program (RIWSP) Sediment Fingerprinting for the Nyabarongo Upper Catchment in Rwanda" (under USAID/Rwanda Cooperative Agreement No. AID-696-LA-11-00001).
• Upon reviewing the referenced technical report, it is disappointing to find that the same data and figures are presented without proper citation of the original work.
• It is essential to properly cite the original project and report, not only acknowledge the institutional support. Although the authors mention the support received from the Ministry of Natural Resources (MINIREMA), the Government of Rwanda, USAID, and the Global Water for Sustainability Program in the acknowledgments section, the study itself does not formally cite the original technical report from which the data, figures, and results are derived. Proper citation of the original report is necessary to ensure transparency. Moreover, the article should provide additional value or a new contribution to the scientific community beyond what is already available in the technical report. Otherwise, it remains a summarized version of an existing publicly accessible report.

To be considered for publication, the manuscript requires substantial improvement, including:

• Proper citation of the original technical report.
• Additional analysis or approaches, such as a historical analysis of land use change, soil erosion and sediment load modeling using USLE (Universal Soil Loss Equation), or a more detailed development of conservation practices and proposals for management instruments.

In its current form, it is not possible to recommend the publication of the manuscript.

Author Response

Dear Reviewer,

Firstly, we thank you for your valuable comments and suggestions. We have taken them into consideration and revised the manuscript accordingly. We really appreciate the time you have taken for the review. Specific responses are italicized below.

 

Comments and Suggestions for Authors

The topic of soil erosion is highly relevant and is addressed with an appropriate approach within a specific context, contributing to the identification of intervention areas and providing tools for soil management in these regions. Although the study is limited to a specific region and there is no novelty in the methods, diagnostic approach, or erosion modeling, the reporting of regional progress in addressing soil loss issues is valued.
However, several concerns arise:

• Some formatting aspects are not properly followed, including the organization of sections, references, and figures. A more careful preparation of the manuscript is recommended before submission.

>>>Thank you, we have now ensured that the formatting of the paper is in accordance with journal requirements.

• The main concern is that a similarity check revealed more than 40% overlap with the technical report "Rwanda Integrated Water Security Program (RIWSP) Sediment Fingerprinting for the Nyabarongo Upper Catchment in Rwanda" (under USAID/Rwanda Cooperative Agreement No. AID-696-LA-11-00001). Upon reviewing the referenced technical report, it is disappointing to find that the same data and figures are presented without proper citation of the original work.

>>>Yes, you are totally right in that this study has been described in the technical report which is also authored by us. However, a technical report is not peer-reviewed outside the circle of authors, unlike a scientific publication. The reason for us to want to publish this as a scientific paper is to increase visibility of the work so that this can contribute to the literature to inform other studies, especially in the under-represented world where soil erosion is the greatest. A technical report, while available to all to download, does not often appear under technical searches using indexing tools or the internet. The summarization of a technical report as a scientific paper is a widespread practice, because technical reports are required by funding/governing institutions as a repository of all work done, which is often too voluminous for a scientific paper. For instance, in this case the technical report (105 pages long) has results for each suspended sediment sample at each subcatchment, which is useful information for the Government of Rwanda, but may not be that useful to audiences elsewhere – hence we include results just from one subcatchment at the individual sample level, to illustrate the process.

 

It is essential to properly cite the original project and report, not only acknowledge the institutional support. Although the authors mention the support received from the Ministry of Natural Resources (MINIREMA), the Government of Rwanda, USAID, and the Global Water for Sustainability Program in the acknowledgments section, the study itself does not formally cite the original technical report from which the data, figures, and results are derived. Proper citation of the original report is necessary to ensure transparency. Moreover, the article should provide additional value or a new contribution to the scientific community beyond what is already available in the technical report. Otherwise, it remains a summarized version of an existing publicly accessible report.

>>>Yes, thank you for this suggestion. We have now cited the original technical report, which was also authored by us. We agree that this is good, not only for transparency and continuity, but also as a further resource for interested readers. We have also included two additional figures, and improved the maps ( Figs 2 and 3).

To be considered for publication, the manuscript requires substantial improvement, including:

• Proper citation of the original technical report.
• Additional analysis or approaches, such as a historical analysis of land use change, soil erosion and sediment load modeling using USLE (Universal Soil Loss Equation), or a more detailed development of conservation practices and proposals for management instruments.

>>>We have included a discussion of the effects of deforestation of the original Afro-montane forests upon continued soil erosion, despite the existence of terraces and other soil conservation measures. This is borne by results from several catchments that have all constituent geological types represented in similar quantities in suspended sediment. Whereas those from other catchments show the preponderance of a few dominant types, suggesting localized disturbances such as mining.

>>>We agree that sediment load modelling is needed to add to the catchment level understanding of water quality and reservoir lifespan; however sediment load modeling was out of the scope of this study, because of the lack of river discharge data, as mentioned in the report. As the purpose of this study was to identify potential sources for remedial action, we did not have the time nor the resources to obtain discharge data at each suspended sediment sampling point. It is a good suggestion for future work.

Reviewer 2 Report

Comments and Suggestions for Authors

General comments:

The manuscript quantified sediment sources from ten geological types in the Nile Nyabarongo Upper Catchment in Rwanda using sediment fingerprinting methods. Overall, the manuscript is well written and structured. This reviewer has the following comments for the authors to consider.

1. The introduction section should be expanded to include a more detailed review of the fingerprinting methods or approaches widely used in the literature, including single composite fingerprints vs. multiple composite fingerprints as well as different optimization schemes/methods for the various mixing models.
2. Methodology needs to be expanded. For example, the size of source samples and number of sediment samples should be added to the tables for each geological source type and to each subcatchment. The sample size is crucial for producing reliable estimates and reducing uncertainty of the estimated proportions. My main concern for the work is the small number of soil and sediment samples and the uneven distributions among subcatchments. Some subcatchments only have a few soil samples. Five sediment sampling events were taken during January to May, but how many samples were taken for each event at each sampling location was not stated. For the Monte Carlo simulation for Bayesian inference using the MIXSIAR Mixing Model, the sufficient sample number is required for adequate probability distribution fitting. A few samples cannot allow a good fit of a probability distribution, leading to unreliable estimation.  Along the line of uncertainty, please explain how the 95% credibility interval in the figures was calculated. The impact of the small size on the estimated proportions should be discussed in the manuscript.
3. Please present the selected composite fingerprints using the discrimination function analysis for each subcatchment, which were used in the mixing model. How many elements were used in each subcatchment? What are their discriminating powers?

 

Author Response

Dear Reviewer,

Firstly, we thank you for your valuable comments and suggestions. We have taken them into consideration and revised the manuscript accordingly. We really appreciate the time you have taken for the review. Specific responses are italicized below.

 

Comments and Suggestions for Authors

General comments:

The manuscript quantified sediment sources from ten geological types in the Nile Nyabarongo Upper Catchment in Rwanda using sediment fingerprinting methods. Overall, the manuscript is well written and structured. This reviewer has the following comments for the authors to consider.

1. The introduction section should be expanded to include a more detailed review of the fingerprinting methods or approaches widely used in the literature, including single composite fingerprints vs. multiple composite fingerprints as well as different optimization schemes/methods for the various mixing models.

>>>We have presently expanded this portion to include a brief review of sediment fingerprinting approaches and different kinds of tracers. In the interests of space and length, we could not include a full review of different schemes for mixing models, as the main purpose of the paper is to describe the application in a remote mountainous catchment undergoing severe erosion, and it is hoped that this can inform the realm of possible ways to tackle the issue. We did include some text describing the statistical methodology in greater detail, and the reasons for the selection of the statistical methods.

2. Methodology needs to be expanded. For example, the size of source samples and number of sediment samples should be added to the tables for each geological source type and to each subcatchment.

>>>We have improved the explanation - that it is 5 samples per geological type across NNYU catchment, and each sample is further composed of 5 soil samples taken within a 50 m radius, so as to exclude possible nugget effects. We have also included some results of cumulative error % in the selection of tracer elements from these samples. As there were 13 geological types the number of samples was 5*13 = 65, plus some additional samples were included for a total of 73 soil samples. The geological type Ho was not sampled, as its alluvial in nature, which could have come from any point upstream. Re: suspended sediment samples, in each campaign, based on the river system, 14 locations were selected, typically just before a tributary meets the main river. And there were 5 such campaigns across dry and wet seasons, which makes it a total of 14*5 = 70 suspended sediment samples.

>>>The methods section has been expanded with more detail, a map of the catchment location and river flow direction has been created, the list of elements that were used as tracers has been included, statistical analysis section described in more detail, and the two types of analysis at the subcatchment level ( single suspended sediment sample, pooled samples for al campaigns) has been better described as analysis A and B respectively.

3. The sample size is crucial for producing reliable estimates and reducing uncertainty of the estimated proportions. My main concern for the work is the small number of soil and sediment samples and the uneven distributions among subcatchments. Some subcatchments only have a few soil samples.

>>>We agree that more the samples, the better is the confidence about representation, especially as geological type classification is not binary - there can always be variation of the element composition. However, analysis of the elemental composition of the first set of soil samples showed that the elemental composition within a geological type did not have a high degree of variation, allowing the selection of a set of elements to be reliably used as tracers or fingerprint of that geological type. Also logistical issues influenced the location of sampling within a geological type. For instance, in Fig 3, in the Sk geological type - yellow color , three sampling location dots appear bunched together close, while a large part of the type is unsampled. Well, the entire area is very mountainous, roadless most of it, and the three dots took two whole days of driving and hiking up steep pathless mountainsides to sample. This is not only a large area, it is very rugged. You can look at this area on Google Earth to get a sense of the topographic ruggedness that is a serious challenge. Also some subcatchments were dominated by few geological types (based on the official geological survey map we used).

 

4. Five sediment sampling events were taken during January to May, but how many samples were taken for each event at each sampling location was not stated. For the Monte Carlo simulation for Bayesian inference using the MIXSIAR Mixing Model, the sufficient sample number is required for adequate probability distribution fitting. A few samples cannot allow a good fit of a probability distribution, leading to unreliable estimation.  Along the line of uncertainty, please explain how the 95% credibility interval in the figures was calculated. The impact of the small size on the estimated proportions should be discussed in the manuscript.

>>>Thank you for the suggestion for better clarity, we have now stated this explicitly in the manuscript. Sorry, instead of credibility interval, this should read confidence interval - we have corrected that in the text. We have also added a brief discussion on the benefits of increased sampling in improving the accuracy of the results, while also being subject to logistic realities of sampling in a rugged and inaccessible terrain, plus the financial costs in sending soil samples overseas for analysis.

5. Please present the selected composite fingerprints using the discrimination function analysis for each subcatchment, which were used in the mixing model. How many elements were used in each subcatchment? What are their discriminating powers?

>>>We have expanded and added the following: The following elements passed the TEAF laboratory analytical QCquality control checks and were evaluated for their use as a tracer: Li, Mg, Al, P, K, Ti, V, Cr, Mn, Fe, Cu, Zn, Ga, As, Rb, Mo, Sn, Ba, La, Ce, Nd, Sm, Eu, Dy, Th, and U. Li and Ga failed the range test and were excluded from consideration as a tracer. All remaining tracers were approximately normally distributed and appropriate for input into the MixSIAR mixing model.The Kruskal-Wallis H-test, step-wise discriminant function analysis and the jackknifed discriminant function analysis were performed individually on each subcatchment (Tables 1-4 in SMs). The optimum set of tracers for each subcatchment varied between three elements (Nd, Sm, and La) for the Secoko with a cumulative error rate of 17% and 21 elements (Al, Ba, Zn, V, As, Sn, Ce, Mn, Mg, Rb, K, Dy, Eu, Sm, Fe, La, Th, Cr, P, U, Nd) for the Lower Nyabarongo with a cumulative error rate of 13% (Table 1, SM). The highest cumulative error rate, 42%, was for Upper Mwogo using six elements (Sn, Al, V, Ba, Cr, and Fe). Prediction accuracy using the leave-one-out jackknifed discriminant function analysis quantified the uncertainty in predicting each of the potential sources within each of the individual subcatchments as between 0% (unable to accurate predict the Nw geologic type in the several of the subcatchments) to 83% (high prediction ability for the Gi geologic type in the Satinsyi subcatchment) (Table 2 Series).

We are also referring to the complete set of results that were included in the technical report we published for this study. That is a lot of data and figures to report in a scientific paper. The report can be accessed at https://www.researchgate.net/publication/322056731_Sediment_Fingerprinting_in_the_Nile_Nyabarongo_Upper_Catchment_Rwanda

 

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Authors

 Authors need to:

In the discussion there is a lot of information that will need to be put in the introduction.

The methodology needs to be revised and the data and results reorganised.

The figures are inadequate and need to be revised.

The discussion is repetitive and ambiguous. There are no conclusions.

I attach the pdf of the article with comments.

Parts of the text are not understandable.

The supplementary data is incomplete

Best regards

Comments for author File: Comments.pdf

Author Response

Dear Reviewer,

Firstly, we thank you for your valuable comments and detailed suggestions. We have taken them into consideration and revised the manuscript accordingly. We really appreciate the time you have taken for the review. Specific responses are italicized below, and are included as comments in your pdf version.

 

Comments and Suggestions for Authors

Dear Authors

 Authors need to:

In the discussion there is a lot of information that will need to be put in the introduction.

• Some of it we have put in the Introduction, while some repetition is present in the Discussion to better explain the context of the results.

The methodology needs to be revised and the data and results reorganised.

• Yes, we have significantly added detail and revised the methodology text

The figures are inadequate and need to be revised.

• We have revised Figures 2 and 3, added Figure 1 and 4 as per your suggestion.

The discussion is repetitive and ambiguous. There are no conclusions.

• We felt that some repetition could better explain the results in context of the discussion. 

I attach the pdf of the article with comments.

We have replied to your comments in your attached pdf, as that would be easier for you to refer to the context of your comments and our responses. 

Parts of the text are not understandable.

We hope it is much more lucid now, as we have majorly revised it.

The supplementary data is incomplete.

We have also referred to the 105 page technical report that includes a full set of graphs and data.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I acknowledge the detailed explanation provided by the authors. The revised version can be recommended for publication, as it incorporates most of the recommendations and improves the overall presentation. Minor issues related to text size and figure quality, especially Figures 1, 7, and 8 could be revised.

Author Response

Dear Reviewer, once again we thank you for your valuable comments that have made the manuscript clearer and more focussed. In accordance with your suggestions, we have improved Figs 1, 7 and 8.

Fig 1 - we have redone this figure using a Public Domain map from the Rwanda Water Control Board showing all catchments designated Level 1 - the NNYU catchment is one of them. One can recognize the shape of this catchment as shown in Fig 2, 3 and 9 in Fig 1 (the pink catchment labelled NNYU). We hope this is clearer than the previous map.

Figs 7 and 8 have had their contrast increased, which has made the labels and concentric rings sharper. 

In addition, for the benefit of readers, we have uploaded some additional data that shows the sediment concentrations at each sampling event - Fig 7 shows the average over all 5 sampling campaigns, but the data shows more detail in sediment concentration variation in each subcatchment.

Best regards,

the authors

 

Reviewer 2 Report

Comments and Suggestions for Authors

appreciate your efforts in fully addressing my comments.  i have no further comments.  thanks.

Author Response

Dear Reviewer,

we thank you sincerely for your time, your guidance and very valuable comments, that have made the manuscript much clearer and focused. 

best regards,

the authors

Reviewer 3 Report

Comments and Suggestions for Authors

Dear authors

The authors tried to improve the text, but the conclusions remain very brief. They could include some of the results obtained and referred to in the discussion.

Most of the figures are from the “technical report [36] “and could and should be improved. The paper could have had much more impact and the conclusions could have been clearer if the authors had created new images using geographic information system (GIS) tools.

Most of the images used do not help either the authors or the reader to understand the purpose and conclusions of the articles.

The resolution of figures 5 to 8 is very low.

Figure 9 should be comparable with Figures 2 and 3 and overlap at least the information relating to “Map of NNYU catchment. Subcatchments are delineated from suspended sediment sampling stations (orange circles) in 220 the Nyabarongo river and its tributaries,”

 

Best regards

Author Response

Dear Reviewer,

we once again thank you for your exhaustive comments and valuable time, that has helped focus and clarify the manuscript. 

In accordance with your most recent suggestions, we have expanded the Conclusions with a point thats also included in the Discussion, and that suggests how the existence of a few or all geological types in a suspended sediment sample can indicate whether erosion is a result of localized disturbances, or diffuse human activities spread widely over the subcatchment. It is helped that for someone just reading a quick conclusion, this can be a useful item. 

We have created a new Fig 1 to clearly show the NNYU catchment location, whose shape can be recognized from Figs 2, 3 and 9. This map is a Catchment level 1 map obtained from Rwanda Water Control Board (map is public license-free). This map also shows the Nyabarongo and Akagera river flowing to Lake Victoria, to illustrate these headwaters of the Nile in a more recognizeable manner.

We have improved the picture quality (contrast and sharpness) of Figures 5-8.

For Figure 9, we tried overlaying the main river layer along with soil and sediment sampling points using GIS, but it got very dense with details, making it illegible for the given size - if it were a 2 meter by 1 meter poster, certainly your suggestion to add rivers, sampling locations would add value. Instead, Figure 2 can be placed adjacent to Figure 9 by a reader who wishes to see the cells in relation with the subcatchments and geological types. Tech Report [36] which we authored for the Governmen of Rwanda also has land use maps at subcatchment resolutions for those readers wishing more spatial detail, as well as a topography map. 

In addition, for the benefit of readers, we have uploaded some additional data that shows the sediment concentrations at each sampling event - Fig 7 shows the average over all 5 sampling campaigns, but the data shows more detail in sediment concentration variation in each subcatchment.

Thank you again.

Kind regards.

the authors