Impacts of Integrated Watershed Management Interventions on Land Use/Land Cover of Yesir Watershed in Northwestern Ethiopia

Round 1

Reviewer 1 Report (Previous Reviewer 1)

Comments and Suggestions for Authors

This study mapped the land use/land cover dynamics and evaluated the impact of Integrated Watershed Management (IWM) for the Yesir watershed in Northern Ethiopia. First, the study applied supervised image classification algorithms to get a time-series of land use/land cover maps from Landsat 5 (2002) and Landsat 8 (2013 & 2022). geographic information system technology was used to examine the trends for land use/land cover classes: settlements, agricultural land, grazing land and land covered with other vegetation.

2.2.2. Data collection. The sample size is rather small. Why not use cross validation to test the classification result?

2.2.3. Data processing and analysis. How the dates of the images were determined? Were there lots of cloud cover in the images? At least the quality of the images should be provided.

Does pixel-based classification work well? Why not try object-based classification?

 

2.2.4. Normalized Difference Vegetation Index (NDVI). "NDVI values were classified into three categories: "no plant cover" (degraded land, bare soil, and barren areas of soil and rock), "weak plants" (grassland and shrub land), and "healthy plants" (natural forest trees) [22]." How to define the three categories? Even the reference is provided, it is hard to understand whether the standard from the reference satisifies the study in this area. At this should be justified.

There are lots of similar studies in the same topic. Please state clearly what the contributions are from this work.

Minor:

abstract: what does it mean "in focus group discussions the impacts of integrated watershed 21 management practices were discussed and analyzed."

Comments on the Quality of English Language

English revision is suggested

Author Response

Response to Reviewer 1 Comments

We would like to express our sincere thanks to the reviewer for the valuable comments/suggestions. Our point-by-point responses are given below in red colour.

Comments and Suggestions for Authors：

This study mapped the land use/land cover dynamics and evaluated the impact of Integrated Watershed Management (IWM) for the Yesir watershed in Northern Ethiopia. First, the study applied supervised image classification algorithms to get a time-series of land use/land cover maps from Landsat 5 (2002) and Landsat 8 (2013 & 2022). Geographic information system technology was used to examine the trends for land use/land cover classes: settlements, agricultural land, grazing land and land covered with other vegetation.

Point 1: 2.2.2. Data collection. The sample size is rather small. Why not use cross validation to test the classification result?

Response 1: Thank you for your comment. The total sample size is adequate, with 505 sample locations in total. From these, 20% (equating to 101 sample locations) were utilized to validate the classification results. For a more detailed description, see lines 204 to 208. To ensure representative sampling, 505 samples were randomly collected for the land use/land cover classification of 2022. Of these, 80% (404 Ground Truth Points or GTPs) were used for image classification and 20% (101 GTPs) for validation.

Point 2: 2.2.3. Data processing and analysis. How the dates of the images were determined? Were there lots of cloud cover in the images? At least the quality of the images should be provided.

Does pixel-based classification work well? Why not try object-based classification?

Response 2: Thank you for the comment. The image dates were selected based on the following criteria: the year 2002 was chosen as a reference year, representing conditions before the watershed intervention. The year 2013 was considered optimal for evaluating the initial impacts of the watershed intervention, which began in 2008. The year 2022 was selected to assess the current status of the watershed intervention's impact on land use/land cover dynamics.

The quality of the images was excellent. Data sets collected during the dry season, from January to February, were chosen to ensure cloud-free images and to accurately capture land surface reflectance values without the influence of agricultural activities.

For this study, pixel-based classification proved effective, as demonstrated by the accuracy assessment of the classification results. The overall accuracy achieved was 94% for 2002, 92% for 2013, and 90% for 2022. The kappa coefficients for these years were 0.88, 0.84, and 0.85, respectively. These accuracy levels indicate that the land use/land cover maps are reliable and meet the requirements for further analysis of land use/land cover changes.

Point 3: 2.2.4. Normalized Difference Vegetation Index (NDVI). "NDVI values were classified into three categories: "no plant cover" (degraded land, bare soil, and barren areas of soil and rock), "weak plants" (grassland and shrub land), and "healthy plants" (natural forest trees) [22]." How to define the three categories? Even the reference is provided; it is hard to understand whether the standard from the reference satisfies the study in this area. At this should be justified.

There are lots of similar studies in the same topic. Please state clearly, what the contributions are from this work.

Response 3: Thank you for your valuable comments and suggestions. However, there is no agreed NDVI classification threshold has been set; and it is different from the objectives of each literature. According to [60] stated that the NDVI class has been classified into 6 categories; no vegetation <0, lowest dense(0-0.15), lower dense(0.15-0.3), dense (0.3-0.45), higher dense(0.45-0.6), and highest dense >0.6. NDVI values were classified into three categories: "no plant cover" (degraded land, bare soil, and barren areas of soil and rock), "weak plants" (grassland and shrub land), and "healthy plants" (natural forest trees) [22].

The main contribution of this study to the literature, as detailed between lines 89 and 108 in the revised manuscript, lies in its exploration of integrated watershed management in the Ethiopian highlands. This approach not only focuses on the enhancement and preservation of the natural and ecological environment but also aims at the long-term advancement of Ethiopia's agricultural sector and overall economy. Despite the importance of these aspects, the effects of such management on land use/land cover changes and vegetation greenness have not been extensively investigated using geospatial technologies. The objective of this study is to evaluate how comprehensive interventions have positively influenced the watershed's biophysical characteristics. The necessity of this study is highlighted by Ethiopia's agriculture-driven economy, which relies heavily on natural resources. Therefore, there is a need to assess the effectiveness of sustainable land management interventions through satellite image trend analysis. The authors have committed to using GIS and Remote Sensing techniques to evaluate the long-term impacts of integrated watershed management practices on land use/land cover changes and landscape greenness over three time periods spanning 20 years for the Yesir watershed.

Point 4: Minor: Abstract: what does it mean, "in focus group discussions the impacts of integrated watershed 21 management practices were discussed and analyzed".

Response 4: Thank you for your comment. 21 refers to the line number and sorry for that I already deleted the number (21).

Comments on the Quality of English Language:

English revision is suggested.

Response: Thanks for your valuable comments. We revised the English language of the whole manuscript. See the revisions in the manuscript.

Author Response File: Author Response.docx

Reviewer 2 Report (Previous Reviewer 2)

Comments and Suggestions for Authors

the authors attended my commentaries 

Author Response

Response to Reviewer 2 Comments

Thank you, dear reviewer, for your valuable comments/suggestions to enrich the paper. We tried to address all issues to improve the paper. Replies to each review comment are given below in red colour.

Comments and Suggestions for Authors：

The authors attended my commentaries. 

Thank you very much for your appreciation.

 

Author Response File: Author Response.docx

Reviewer 3 Report (Previous Reviewer 3)

Comments and Suggestions for Authors

This revised manuscript is acceptable. The authors have included more detailed information and responses to improve the manuscript.

Author Response

Response to Reviewer 3 Comments

We would like to express our sincere thanks to the reviewer for the valuable comments/suggestions. Our point-by-point responses are given below in red colour.

Comments and Suggestions for Authors：

This revised manuscript is acceptable. The authors have included more detailed information and responses to improve the manuscript.

Response 1: Thank you very much for your admiration.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report (Previous Reviewer 1)

Comments and Suggestions for Authors

Thank you for the revision. I looked into the responses but 

For the question Point 2: 2.2.3. Data processing and analysis..., I suggest provide detail statistics of cloud cover percentage for the selected images in table 1.  For the year 2002, the date Jan 15 was selected but why not the date from 2002-02-01? Or are there difference in the image quality in the two dates?  By providing the detail of the selected images, at least readers will know that the selection of the images are rational.

For Point 3: 2.2.4. Normalized Difference Vegetation Index (NDVI) ..., the authors mentioned "However, there is no agreed NDVI classification threshold has been set; and it is different from the objectives of each literature." I do not disagree with the above statement but if NDVI threshold could be arbitrarily defined, what will the result from Table 8 reflect? Is it more reasonable to use the changes in NDVI (changes of the mean NDVI, or changes in NDVI histogram distribution) rather than using the area percentage of each category. So I think it is necessary to justify the standards defining the NDVI classification thresholds that were applied in this study.

 

Author Response

Response to Reviewer 1 Comments

We want to express our sincere thanks to the reviewer for the valuable comments/suggestions. Our point-by-point responses are given below in red colour.

Comments and Suggestions for Authors：

Thank you for the revision.

Thank you for your feedback.

Point 1: For the question Point 2: 2.2.3. Data processing and analysis..., I suggest provide detail statistics of cloud cover percentage for the selected images in table 1.  For the year 2002, the date Jan 15 was selected but why not the date from 2002-02-01? Or are there difference in the image quality in the two dates?  By providing the detail of the selected images, at least readers will know that the selection of the images are rational.

Response 1: Thank you for the comment. We incorporated your suggestion for data quality issues of the satellite images.  For this see in line, 143 we added one column to show the detailed statistics of cloud cover in percentage. The data acquisition year of the image in the table clearly showed 2002-01-15 not 2002-02-01.

Point 2: For Point 3: 2.2.4. Normalized Difference Vegetation Index (NDVI) ..., the authors mentioned "However, there is no agreed NDVI classification threshold has been set; and it is different from the objectives of each literature." I do not disagree with the above statement but if NDVI threshold could be arbitrarily defined, what will the result from Table 8 reflect? Is it more reasonable to use the changes in NDVI (changes of the mean NDVI, or changes in NDVI histogram distribution) rather than using the area percentage of each category. So I think it is necessary to justify the standards defining the NDVI classification thresholds that were applied in this study.

 Response 2: Thank you for your valuable comments and suggestions. We have removed from the manuscript the statement: “However, there is no agreed NDVI classification threshold, and it varies according to the objectives of each study. According to [60], the NDVI class has been divided into 6 categories: no vegetation <0, lowest density (0-0.15), lower density (0.15-0.3), medium density (0.3-0.45), higher density (0.45-0.6), and highest density >0.6.”

According to literature cited in chapter methodology (see lines 257 – 260), the NDVI values applied in this study were categorized into four basic categories (see Table 8) in line 342, which correspond approximately to the land use/cover classes: No vegetation, grass/fellow land, weak plant cover, and health plant cover. We already showed in line 337, Table 7 changes of the mean NDVI for 2002, 2013 and 2022.

Author Response File: Author Response.docx

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study applied remote sensing approach to map and evaluate land use changes and vegetation growth patterns over the study time window (2002-2022) in  the Amhara region's Yesir micro-watershed. The work demonstrated successfully the capacity of remote sensing in identifying land cover changes and vegetation growth. My main concern is that what contributions in terms of the method and application have been made from this study to literature. Probably the authors can provide at least a section in discussion part talking about how the current work is different from others and what contributions have been added to this study topic. Some other specific points:

2.2.1. Data sources: as quality of Landsat images can be significantly affected by raining/cloudy weather, how those scenes were decided? Also the seaon for the year 2013 was in November and the other two scenses were in Jan or Feb, what the value of comparing NDVI at different seasons?

2.2.2. Data collection: How ground truth points were collected? How the sample locations were determined? How were the 505 samples distributed among different land cover types?

Line 156 page 5, abbreviations (e.g., GTPs) should be given when they appeared for the first time.

There are no legend showing the spatial and temporal changes in Fig. 3: NDVI maps of Yesir watershed for 2002, 2013, and 2022. 

Comments on the Quality of English Language

Extensive language revision is suggested

Reviewer 2 Report

Comments and Suggestions for Authors

This paper describe the dynamics of a watershed in Ethipia. It is interesting the comparison between land use and cover. The description of the processes among all uses is a little bit confusing because some land uses expanded in one year and constrain in another. The discussion part lacks of an interpretation of main causes ol land use and land cover change. It is interesting that the autjros asked the local people their opinion on land use/cover changes, but there must be administative/political decions that provoque que change. As well some changes are due to climate emergences whicg are not mentioned. Most scientific literature in this topic refer the results to the agricultural and migration sectors, for instance because in many places in the Global South people of the rural ares migrate to the urban areas and there is a lot of literature on abandoned lands and recuperation of secondary vegetation many times with many native species and in the long time there is natural vegetation recuperation. I miss a figure where the changes are explained by its political/administrative/climate causes, and contrasted with the interviewed people opinions. 

Reviewer 3 Report

Comments and Suggestions for Authors

1.      Line141: Please check the reference again. (Error! Reference source not found.). Are there some sentences missing in this manuscript?

2.      Line 142: Table 2: Land use/land cover classification scheme and description of land use/land cover types applied for the study. Is there water in this classification?

3.      Line 176: The key informants from each community watershed were selected considering their age (60 and above). Is it suitable? The young generation opinions couldn’t be involved in this study. How to overcome this problem?

4.      Line 256: “Error! Reference source not found.” Please check this missing again.

5.      Line 280: Table 4: Accuracy assessment (%) of land use/land cover maps (2002, 2013 and 2022). The Producer’s accuracy of vegetation in 2013 was lowest with a value 71%. Why? Please explain this characteristic in detail.

6.      Line 371: The current study found a high rate of land use/land cover change to vegetation (771.6 ha), grazing land (556 ha), and settlement (527 ha) with a slight annual increment rate. On the contrary, agricultural land (1855 ha) showed a declining trend. It is better to explain the important reasons for this change. What are the fundamental driving forces for this trends?

7.      Line 253: “The practice of IWM is to restore ecological balance by utilizing, conserving, and developing degraded natural resources such as soil, vegetative cover, and water.” How to prove the ecological balance based on the IWM? Are there some evidences or references to support this statement?

Reviewer 4 Report

Comments and Suggestions for Authors

This paper uses Landsat satellite images to analyze the land use/land cover changes resulted from the Integrated Watershed Management (IWM) for the Yesir watershed in Northern Ethiopia. The method used for land classification is appropriate and correct. The results confirm the positive contribution of IWM in land use/land cover in the study area over the past 20 years. It is an useful information to share with certain group of readers. However, the analysis used only three selected years images, the data are not enough to reflect the changing dynamics over a time span of 23 years as the authors claimed. Since the authors have developed an elaborated procedure to generate the land classification from the satellite images, generating additional results for other years should not be too much work. The reviewer strongly suggests the authors to add more years in between to show the changing dynamics. The review recommends for publish after the comments are addressed.   Additional specific comments are below:

1) In 2.2.1, Data sources,  three years of images is not enough to reflect the changing dynamics,  more data in between should be added.

2) In table 2， four types of land cover are exhaustively used.  There may be other type existent.  If so,  the classification may result in errors.  Author should explain.

3) in 2.2.3. Data processing and analysis, the description is too wordy, adding some pictures will help the readability.

4) in 2.2.4 Normalized Difference Vegetation Index (NDVI),  the NDVI formula needs more explanation.  Give a short description on how to  relate variables NIR and RED to the images?  RED is not defined.  

5) in the results section,  if more years of images are added,  line charts may be more convenient to refect the changing trend for the land cover.

6) In table 4,  "User's" and "Producer's" are not defined.   The error analysis is not clear.