Comparative Study on Sediment Delivery from Two Small Catchments within the Lena River, Siberia

Round 1

Reviewer 1 Report

Comments for author File: Comments.docx

Author Response

Many thanks for the careful editing of our article and your recommendations. Thank you for the detailed description of the WaTEM/SEDEM model in your feedback. We partially included this text in the new version of the manuscript. We are also grateful for the provided references, which are also included in the new version of the manuscript. We corrected the name of the model to WaTEM/SEDEM, and tried to improve the English language of the manuscript. The responses to your specific comments are in italics below. Line numbers are given in the new version of the manuscript. Text changes highlighted in yellow.

Detailed comments

Title: I suggest to correct the title as it follows:

Comparative study on sediment delivery from two small catchments within the Lena River, Siberia

Agree, corrected.

Abstract: I suggest to correct the abstract as it follows.

The paper studies the possibility of using the WaTEM/SEDEM model to assess sediment yield from the catchment area within the Lena River basin. The study was carried out on the basis of a comparison of predicted data and measured data of the suspended sediment yield at the gauging stations of the state monitoring network of Russia. The study was performed within two areas with plain and mountainous relief. The first site is located within the catchment area of the river Chara with an area of 4150 km. The second site rests on the catchment area of the Lena River between Tabaginskiy and Kangalassky capes near Yakutsk city. The catchment area of this site is 15740 km. The values of sediment yield from the “Yakutsk” catchment area are in much better agreement with the values of the measured sediment yield, than in the “Chara” catchment area. The model sediment yield from the study area remained almost unchanged from the period 1986 – 2019 and amounts to 3.5 t/km², while the suspended sediment yield in the Lena at the Tabaga gauging station slightly increased from 7 to 9.45 t/km² per year.

Agree with your version of abstract.

Introduction:

In the introduction, the authors must mention a broader spectrum of references discussing the application of the WaTEM/SEDEM for the studies on sediment transport. In its present-day shape, the introduction is quite poor and more references need to be quoted and discussed.

I agree, added more references both to the introduction and to the discussion (lines 66-78 and 376-400.).

Among the whole complex of the listed changes taking place in the Lena River basin, the dynamics of sediment yield from the catchment area deserves special attention, which affects many processes: the rate of siltation of reservoirs, the mass of incoming pollutants.

This sentence needs to be re-written, since the English form is wrong.

Agreed, we have restructured the sentence in the new version of the manuscript (lines 28-30.).

Most researchers distinguish: conceptual, empirical, physically based models [2]. Examples of empirical models are: USLE [3], RUSLE [4], MUSLE [5]. Examples of physically based models are: WEPP [6]; LISEM [7]; EROSION 3D [8]; EUROSEM [9]. Conceptual models are a combination of empirical and physical based models [10], examples are: WATEM/SEDEM [11,12]; RUSLE2 [13]; MMF [14]; SWAT [15,16].

Here the authors speak about three types of models:  conceptual models, empirical models, physically based models, but the definition of the physically based models has been omitted and needs to be added. Alternatively, the authors must speak only about conceptual models and empirical models.

Agree. They gave a definition of all types of models (lines39-43).

2. Materials and methods

2.1 Study area

Figure 1 is quite unclear for the readers which are not familiar with the Siberian geology. It is also for the lacking of a section on the geological setting of the area, which needs to be added. Information discussed in the sub-section on the study area is not enough.

Agree. In the new version of the manuscript in figure 1, lithological maps are given for the study areas, a description of lithology, relief is given in chapter 2.1 (lines 93-98 an 99-103)

2.2 Methods

The model of the average long-term assessment erosion-accumulation budget WaTEM / SEDEM, developed at the Catholic University of Belgium [11,12] was used for the average long-term assessment of the contribution of basin erosion to sediment yield and the compilation netto erosion spatial map. The methodology consists of three stages.

I don’t absolutely understand the meaning of the sentence … compilation netto erosion spatial map. Please revise and clarify.

I agree. We tried to clarify and restructure not only this sentence(lines127-129 ), but to supplement and rework the entire section of the methodology in the new version of the manuscript.

The first step is to estimate the potential soil loss within each grid pixel based on the RUSLE equation [4].

The RUSLE equation, pertaining to the first step of modeling has not been reported. It must be added as the equation (1), while the present-day equation (1), explaining the second step of modeling, must become the equation (2).

Agree, the RUSLE equation (1) has been added in the new version of the manuscript.

The parameters of the equation (1) need to be carefully checked (rows 101-106).

Thank you for your comment. All variables present in the description are in the formulas in the new version of the manuscript.

I don’t understand the meaning of the text reported at the rows 107-109. Please revise and clarify.

Agree. We tried to rebuild the sentence and make clarifications in the new version of the manuscript (lines 133-142).

The amount of eroded soil is compared with the transport capacity at the third stage. The amount of sediment delivered from the up-slope areas is added to sediment produced by erosion for each grid cell. If the sum exceeds the transport capacity (TC) of the flow, then the sediment yield from the cell is limited to the transport capacity. If the sum of the sediment delivered to a given grid cell and the sediment formed by erosion in that cell is lower than the transport capacity of the flow, then all the sediment is transported further down the slope.

The paragraph is very hard to read in English and needs to be re-written.

I agree, we tried to edit and make it more understandable for English-speaking readers. (lines  146-152)

2.3 Initial data

It should be better to speak about input data? Do you refer to the input data of the model? Please clarify.

I agree. In the new version of the manuscript "Input data"

The following cartographic models were used for calculations: relief, soil erodibility; land use; rainfall erosivity factor; model of C-factor, which were reduced to united grid pixel size - 100 meters.

Please expand this sentence and clarify in revision.

Agreed, clarifications have been made. (Lines 158-161)

“The following cartographic models were used for calculations sediment yield and netto erosion maps within study areas: relief, soil erodibility; land use; rainfall erosivity factor; model of C-factor, which were reduced to united grid pixel size - 100 meters.”

 More details information about each model are written in 2.3.1-2.3.4.

2.3.1 Relief

Currently, there are several free available global elevation models with different resolutions from 1 -  7.5 arc seconds: SRTM C-SIR, SRTM X-SAR [31,32], ASTER GDEM v.2 [33]; ASTER GDEM v.3; ALOS3D30 [34]; ArcticDEM [35]; GMTED 2010 [36] and others.

It is possible to explain the aforementioned models? What do they consist?

Thanks for the question. These free available models present the relief of the Earth; on their basis, the LS-factor map was calculated.

Additionally, there is a DEM that occupies an intermediate position in terms of spatial detail.

This sentence is unclear. The DEM occupies an intermediate position among what? Do you refer to the aforementioned models? Please revise and clarify.

Agree. We tried to make the sentence more understandable for readers (Line 168).

«Additionally, there is a DEM with middle spatial resolution.»

This DTM has a spatial resolution of 3” (about 100 m) and is available for download at http://viewfinderpanoramas.org [37]. This model was created by a group of authors based on several data sources: two global freely available elevation models SRTM C-SIR and ASTER GDEM, as well as topographic maps at a scale of 1:100,000, 1:200,000.

DTM or DEM ? Be consistent and try to use always the same terminology.

Thank you. We agree that the same terms should be used. Corrected(line 169)

Figures 2 and 3 lack of geographic coordinates, which need to be added. The legend is unreadable and needs to be enlarged.

Figures 2 and 3 have been corrected, geographical coordinates have been added, in addition, instead of soil erodibility, soil types are given on them. This is more informative for readers.

2.3.3 Landuse

Correct as 2.3.3 Land use

Agree. (Line 199)

Rows 161-165 must be shifted in the introduction. Omit.

Agree, This part has been shortened and moved to the introduction. (line87)

It was decided to use satellite data.

Correct as: the satellite data have been used, or alternatively, omit.

Thanks for the note, we agree with your , corrected .

Rows 166-170 can be omitted? If you don’t use these satellite data, why do you mention them? This paragraph seems unnecessary.

Agreed, this part has been removed.

Rows 171-174, correct as it follows:

High-resolution satellite images presented in Google Earth for the modern period and images from KeyHole-4B reconnaissance satellite (CORONA program) for the USSR period have been used. 8 KeyHole images covering the entire territory for the late 1960s with a spatial resolution of 1.8 m were selected for the "Yakutsk" catchment (Table 2).

Corrected according to the reviewer's recommendation(lines 204-208).

Rows 181-185: put in a better English shape. It is necessary ?

Corrected. The paragraph explains why KeyHole images of the end of the 60s are applicable for assessing cropland areas in the 66-85 period (lines 214-220).

Rows 186-193. Correct as it follows.

The Georeferencing module in ArcGIS was used to geo-reference the KeyHole images (projection WGS 84/UTM 50 Northern Hemisphere). Modern (2018-2021) very high-resolution satellite images were used as reference data. These images are available as global base maps from Google and ESRI in the QGIS module - HCMGIS. The maximum georeferencing errors for all images were less than 6 pixels [46].

Corrected according to the reviewer's recommendation(lines 214-220).

Rows 207-212 re-write. The English form needs to be improved.

«The modern crop fields were digitized manually and overlaid onto the KeyHole images. Boundaries were corrected, and the remaining fields were digitized. In some cases, crop field looks pretty similar to logging sites. Therefore, modern images were used as auxiliary data. If an unclear area is an abandoned field on the modern image, it was identified as a cultivated field for the Soviet period. Thus, vector layers of cropland for two periods were obtained (Figure 5).»

Re-written.(lines 234-239)

Rows 219-224. Correct as it follows.

The cropland area was calculated, amounting to 6718.5 ha in 1969 and 3337.2 ha in 2021, i.e. 0.42% and 0.21% of the total catchment area. It should be noted that with a general decrease in area, in some places plowing of new sites is observed. Despite the almost twofold reduction in the area of arable land, its insignificant share of the total catchment area allows us to conclude that its contribution to the suspended sediment yield formation is none.

Corrected according to the reviewer's recommendation(lines246-251)

Rows 230-231

Based on these models and C factor coefficients proposed by P.Panagos [47] data on protection factors for different types of land use, a C factor model was also created.

You need to discuss better this part or to eliminate it. You mention the C factor model, but you don’t explain it.

Agree. In the lines 258-265 gave a broader explanation of why C factor is needed and gave a more detailed description of how the model was obtained(Table 3).

Rows 234-236 it is necessary to mention the name of the publication in the text of the paper? It could be enough to quote it (references in square brackets).

Thanks for the comment, agree, removed.

Rows 236-243, correct as it follows.

The initial data on the intensity of rainfall for the period from 1961-1984 was used for the territory of Russia. Over the past few decades, starting from 1985-1990, an intensification of climate change was noted by many authors [49-50], expressed in a change in the amount of precipitation and the intensity of precipitation. Unfortunately, there is no modern data on the intensity of precipitation in the study area. Therefore, we analyzed the change in the amount of rainfall within the studied catchment area. Further, based on the obtained changes, the used model [48] was corrected for the time interval 1986-2019.

Thanks a lot. Agree with the wording of this paragraph (lines 270-277).

Rows 245-250, correct as it follows.

Analysis of data from the Russian Research Institute of Hydrometeorological Information-World Data Center website meteo.ru shows a slight increase in the average long-term annual precipitation in the catchment area of the Chara river, from 357 mm (average for 1966-1985) to 400 mm (average for 1986-2019).

Sorry, but we did not understand what is the difference between your and our version of this text?

This increase is mainly due to rainfall precipitation from 287 mm to 337 mm (an increase of 15%), the share of snow precipitation decreased from 59 to 46 mm per year. 

This sentence is hard to read in English. Please revise and clarify.

Agree. We edited the sentence to make it more understandable for English-speaking readers (lines281-282.).

Rows 251-256 correct as it follows

Analysis of data from the Russian Research Institute of Hydrometeorological Information - World Data Center website meteo.ru shows a slight increase in the average long-term annual precipitation in the catchment area of the Chara river, from 357 mm (average for 1966-1985) to 400 mm (average for 1986-2019).

The following sentence is very hard to read in English. Please revise and clarify.

Apparently this refers to the Yakutsk section? The sentence structure is the same. We tried to improve the English language in this sentence.

This increase is mainly due to rainfall precipitation from 287 mm to 337 mm (an increase of 15%), the share of snow precipitation decreased from 59 to 46 mm per year. 

Corrected, lines 284-289.

3. Results and Discussion

I suggest to separate the results from the discussion and the conclusions. This will improve the structure of the paper.

Agree. The new version of the article is divided into two sections “Results” from the “Discussion”.

REFERENCES

SHENG Meiling, FANG Haiyan (2014) Research progress in WaTEM/SEDEM model and its application prospect. Progress in Geography, 33 (1), 85-91, doi: 10.11820/dlkxjz.2014.01.010.

Van Rompaey A., Bazzoffi P., Jones R.J.A., Montanarella L. (2005) Modeling sediment yields in Italian catchments. Geomorphology, 65(1-2):157-169.

We agree, we added these references that are significant for the study.

Reviewer 2 Report

The article claims to be the first in assessment of the suspended sediment yield from the catchment into the river Lena, which is very good. The article is mostly about the application of a new modeling tool and that is nice.

However, I have quite a few concerns.

Fist of all the methodology is not clearly presented, rather short and a little messy. E.g. in line 97 "each grid pixel" is mentioned without previously mentioning or explaining what kind of grid, how to define it etc. Please explain this (there is one more mention of "the grid" in line 309 stating it was a 100m "pixel" but this really does not provide for the repeatability of the modeling. Next problem is still in methodology, in line 98 "RUSLE" equation needs to be explained. There is only one reference to look it up, which is not enough.

Further, how the values of "TC" have been acquired? What were the obtained values and how did these compare to measured values? Explaining this would add an insight into the correctness of the model even more than the details already given.

Next problem is with the periods of the study. First, in the abstract the period 1986 to 2009 is mentioned. Later, KeyHole images of 1969 and 1967 are used. In line 183: "data on sediment yield are given for the period 1966-1985".  Later, croplands are assessed in 1969 and 2021 (Figure 5). In the beginning of the Results and discussion, the (sub) period 1966-85 is mentioned, later the 1986-2019 (sub)period. Net erosion maps for 1985-2019. In line 322 period 1985-2015. Decide which is the period targeted with what kind of study, and state this explicitly at the beginning, in the methodology part.

Figures 6 and 7 (and the text related) use positive numbers for both erosion and accumulation, yet if one is positive the other shall be negative. This is confusing.

It is in general not clear what the goals were actually, apart from the testing of the applicability of the modeling tool. Somewhere inside the text it is mentioned that a comparison before/after the changes in precipitation is done as well, but in my opinion these results can not be interpreted as such, the modeling result for the mountain catchment being unacceptable and the lowland part (a) too small (b) as it is lowland, I presume there is a very small runoff even in positive temperatures and rain. It would be nice to have some general hydrological description as well about these modelled chatchments.

There is a very extensive part in the text about satellite images and their usage in order to create just an input to the model which is, on the other hand, almost not explained at all.  A much bigger emphasis shall be given to the model, to hydrological and sedimentological backround and less to the image processing issues which is just one part in the methodological chain. 

Lines 299-300 is there actually a proof of river channel erosion in the Lena river? I mean the acceleration of it or sediment imbalance? Otherwise it is not appropriate to try to explain the increase of suspended load with this. Give at least a few references.

Line 305 0.1 t/ha/y vs. 0.03 t/ha/y is by no means a difference that can be considered "small".

Lines 313-315 can you shortly explain the mentioned methodological differences and how they can yield so much different results?

In summary, I think that the work is valuable though not presented well. Thus, the article has to be rearranged and supplemented with additional descriptions and explanations.

Author Response

Many thanks for the careful editing of our manuscript and your recommendations. The responses to your comments are in italics below. Line numbers are given in the new version of the manuscript. Text changes highlighted in yellow.

Detailed comments

Fist of all the methodology is not clearly presented, rather short and a little messy. E.g. in line 97 "each grid pixel" is mentioned without previously mentioning or explaining what kind of grid, how to define it etc. Please explain this (there is one more mention of "the grid" in line 309 stating it was a 100m "pixel" but this really does not provide for the repeatability of the modeling. Next problem is still in methodology, in line 98 "RUSLE" equation needs to be explained. There is only one reference to look it up, which is not enough.

Agree. In the new version of the article, we tried to provide more detailed information about the spatial data model used, its resolution (lines 160-161). Since the study was carried out using GIS, it was necessary to choose a spatial data representation model for work. We chose a raster spatial data model because the WaTEM/SEDEM model requires a raster data model. A structural element of a raster data model is a georeferenced pixel and also gave the formula RUSLE(line 131-136).

Further, how the values of "TC" have been acquired? What were the obtained values and how did these compare to measured values? Explaining this would add an insight into the correctness of the model even more than the details already given.

Далее, как были получены значения «TC»? Каковы были полученные значения и как они соотносились с измеренными значениями? Объяснение этого добавило бы понимания правильности модели даже больше, чем уже приведенные детали.

TC values were obtained using formula (2). TC is a value indicating how much sediment can be transported downslope through each pixel. This value is compared with the value of E(erosion in the cell) + sediment coming from the higher slope. The results of the method are:

1) The sum of the masses of sediments that came from the slopes of the catchment area into the water body - it is this that is compared with the measured sediment yield;

2) Spatial model (map) of netto erosion intensity in each pixel of the catchment area.

The corresponding explanations were added in the text (lines 143-156).

Next problem is with the periods of the study. First, in the abstract the period 1986 to 2009 is mentioned. Later, KeyHole images of 1969 and 1967 are used. In line 183: "data on sediment yield are given for the period 1966-1985".  Later, croplands are assessed in 1969 and 2021 (Figure 5). In the beginning of the Results and discussion, the (sub) period 1966-85 is mentioned, later the 1986-2019 (sub)period. Net erosion maps for 1985-2019. In line 322 period 1985-2015. Decide which is the period targeted with what kind of study, and state this explicitly at the beginning, in the methodology part.

The abstract mentions the period from 1986 to 2019. KeyHole images for 1969 and 1967 were used because there are no other high-quality images for the period 1966-1985 for the Yakutsk site (for example, Landsat images could not be used due to cloudiness). In addition, we can assume that in the period from 1966-1985 land use remained unchanged, so the images used in 1967-1969 can characterize the land use of the entire period from 1966 to 1985. We write about this in the article on lines 210-212.

«In line 322 period 1985-2015».

Thanks, we corrected it. Corrected for 1986 - 2019 (line 335)

Figures 6 and 7 (and the text related) use positive numbers for both erosion and accumulation, yet if one is positive the other shall be negative. This is confusing.

Agree. Accumulations are shown using negative values in Figures 6, 7 and in the text of the new version of manuscript(lines 340, 350).

It is in general not clear what the goals were actually, apart from the testing of the applicability of the modeling tool. Somewhere inside the text it is mentioned that a comparison before/after the changes in precipitation is done as well, but in my opinion these results can not be interpreted as such, the modeling result for the mountain catchment being unacceptable and the lowland part (a) too small (b) as it is lowland, I presume there is a very small runoff even in positive temperatures and rain. It would be nice to have some general hydrological description as well about these modelled chatchments.

Yes, the main goal of the study is to assess the possibility of applying the WaTEM/SEDEM model to assess sediment yield from the catchment area and its dynamics depending on changes in precipitation intensity and land use change within two catchment areas with different topographical conditions.

Agree. In the new version of the manuscript (line 104-107) the values of the water surface runoff of atmospheric precipitation for the studied areas are given.

There is a very extensive part in the text about satellite images and their usage in order to create just an input to the model which is, on the other hand, almost not explained at all.  A much bigger emphasis shall be given to the model, to hydrological and sedimentological backround and less to the image processing issues which is just one part in the methodological chain. 

Agree. The part describing the remote sensing data is too extensive, some details have been removed in the new version of the manuscript. In addition, in chapter 2. were added: a description of the lithological structure and surface water runoff of the studied areas (lines 93-98, 99-107). 

Lines 299-300 is there actually a proof of river channel erosion in the Lena river? I mean the acceleration of it or sediment imbalance? Otherwise it is not appropriate to try to explain the increase of suspended load with this. Give at least a few references.

Agree. In the new version of the manuscript, this statement about the influence of channel erosion on the increase in sediment yield was removed.

Line 305 0.1 t/ha/y vs. 0.03 t/ha/y is by no means a difference that can be considered "small".

Agree. New wording provided in line 361.

Lines 313-315 can you shortly explain the mentioned methodological differences and how they can yield so much different results?

The assumption about the influence of methods for calculating the LS factor was removed in the new version manuscript. At the same time, in the new version of the manuscript, a more detailed explanation was given about the influence of the initial soil data on the final result (lines 371-375).

Reviewer 3 Report

I think that this article is about an interesting topic, and it can be published in Water after the following revision:

1.     The background needs to be improved. The overview of the topic being studied is minimal. A clear justification of the goals and objectives that the researchers set for themselves is necessary.

2.     Where WaTEM/SEDEM model description? Input data, calibration and sensitivity analysis?

3.     I didn't see that your work takes into account the effects of crop rotation, farmland fragmentation and land use change.

4.       The results are declarative. Figures 6 and 7 without any explanation or discussion.

5.     The conclusions should be clearer, there is no substantiation of the data (for example, where did the 90% come from).

Author Response

Many thanks for the editing of our article and your recommendations. The responses to your comments are in italics below. Line numbers are given in the new version of the manuscript. Text changes highlighted in yellow.

Detailed comments

1. The background needs to be improved. The overview of the topic being studied is minimal. A clear justification of the goals and objectives that the researchers set for themselves is necessary.

I agree. The introduction was improved, references to additional studies was added. (Lines 66-78)

2. Where WaTEM/SEDEM model description? Input data, calibration and sensitivity analysis?

A description of the WaTEM/SEDEM model is given in Section 2.2. Input data - in section 2.3. Calibration of the WaTEM/SEDEM model for the area of the studied watersheds was not carried out, since this requires much more test objects, for example, Van Rompaey, 2001 used 12 Belgian objects for calibration and 12 objects for verification(Van Rompaey, A.J.J.; Verstraeten, G.; Van Oost, K.; Govers, G.; Poesen, J. Modelling Mean Annual Sediment Yield Using a Distributed Approach. Earth Surf. Process. Landforms 2001, 26, 1221–1236, doi:10.1002/esp.275.), while we had only 2 objects.

3. I didn't see that your work takes into account the effects of crop rotation, farmland fragmentation and land use change.

An assessment of land use dynamics has been made and is presented in section 2.3.3. Since the area of arable land occupies less than 1% of the area of the Yakutsk site, and there are no arable lands within the Chara site at all, we believe that the effects of crop rotation on sediment yield in this catchments can be ignored.

4. The results are declarative. Figures 6 and 7 without any explanation or discussion.

Agree. Added to the discussion section a comparative analysis with the results of other studies (lines 356-400.).

5. The conclusions should be clearer, there is no substantiation of the data (for example, where did the 90% come from).

Agree. Clarified the conclusions in the new version of the manuscript (lines 407-409, 417, 419).

Round 2

Reviewer 1 Report

The paper can be accepted for the publication in Water, since a satisfactory major revision has been carried out. I suggest the authors to carefully read the revised manuscript for a final editing of the scientific content and of the English form. I suggest also to carefully check the equations another time. 

Reviewer 3 Report

The authors did a great job and corrected the manuscript according to my comments. I think the manuscript can be published.