Impact of Differences in Land Management on Natural Vegetation in Semi-Dry Areas: The Case Study of the Adi Zaboy Watershed in the Kilite Awlaelo District, Eastern Tigray Region, Ethiopia

Round  1

Reviewer 1 Report

In general terms, the research is well developed. Land use-land cover changes, the compatibility of environmental conservation and food security as well as its relationship with the changing administrative divisions of the territory according to political regimes are topics of undoubted interest for Environments readers. It connects with one or more of the following aims of the journal: environmental conservation, governance, monitoring and modelling of environmental systems, stakeholder engagement and decision support.

Nevertheless, a major review is essential in order to improve both the methodology and the results sections. In addition, in the manuscript version, downloaded from the Environments platform, the Discussion and Conclusions sections are missing. The manuscript is interrupted on line 353 “and 3) using satellites images, GIS, ………”

Below, a series of suggestions are proposed as a guide for the review:

Manuscript: Throughout the manuscript, the term "land use" is used confusingly. Sometimes, the term is used correctly although, it is suggested to use "land use-land cover" when it refers to cropland or grassland, for example. However, in other occasions the term "land use" is used to refer to other concepts (see title of figure 2: Change of land uses in Adizaboy watershed during imperial, socialist and democratic eras.). In my opinion, on these occasions the term is used incorrectly because it refers to administrative divisions. Depending on their characteristics and the political decisions of the administrators in each of them, some land uses are allowed, limited or prohibited, but they are not land uses in themselves but administrative areas or land tenure regimes.

Abstract: It is recommended to provide more specific details about the methodology, for example: what satellite images have been used? what period of time has been used to determine the land use-land cover changes? and what other technologies have been used in the study?

Introduction: The authors contextualize well the topic addressed in the paper, reference numerous previous works of the study area and present the objectives.

Methods: In general terms, the methods employed don’t provide great originality. In this context, it is remarkable the combination of quantitative methods (NDVI values extracted from the satellite image) and field work (capture of boundaries of administrative areas with different land tenure) and qualitative methods based on interviews with key local informants.

However, in order to better understand the work, some methodological details are missing. The authors declare that the "change of land uses" have been studied (figure 2) but only one WorldView-2 image of 2016 has been used to obtain the single land use maps and to extract the NDVI values from that date (cropland). Why not repeating the process for the other land uses? In the paper by Tilahun [29] the NDVI values of 1984, 2000 and 2014 were mapped in Kilite Awulalo. It is assumed that this area of study encompasses Adizaboy Watershed. If correct, what is the relationship between the evolution of NDVI observed by Tilahun and those observed in 2016 by the authors of this article?

The source of Figure 4 is missing. Are the NDVI values derived from the 2016 WorldView-2 image?

Results:  Figure 1: It is convenient to revise your edition to improve readability. It is recommended to include a graphic scale and a color range in slope map.

Figure 2: I suggest substituting the title "Change of land uses in Adizaboy watershed during imperial, socialist and democratic eras" by "Change of administrative areas / land tenure in Adizaboy watershed during imperial, socialist and democratic eras". In addition, I recommend removing the letters and labels of each shape (eg "I-Free grazing"). They are unnecessary and make reading difficult.

I recommend replacing the title of section 3.4 "Changes in the Use of Croplands" with "Spatial distribution of crop land use-land cover". Actually, there is no analysis of temporal change.

The text between lines 297 and 312 is redundant with respect to figure 5 and does not provide additional explanations to the reader (causes of processes, for example). It is convenient to insist more on the reasons that explain the differences of the NDVI in each administrative area.

According to figure 5, the percentages expressed in line 312 "seasonally-closed grazing land was 66.6% and the free grazing land was 43.2%" are reversed.

In figure 5 (top left), the perimeter of "free grazing land" is missing. Are the white patches (croplands) related with 0 NDVI value? It seems unclear. Replace "Buildig" with "Building" in the legend box. In the lower part it is recommended to use a semi-logarithmic scale to represent the “y” axis. Currently, the bars of "Water, Building" and "Small vegetation" classes are imperceptible.

Discussion and Conclusions: As mentioned above, these sections are missing.

Other formal issues: "Hectors" is the right word (lines 67 and 68) or does it refer to "hectares"?

Tilahun's paper [29] is not published in 2000 but in 2015.

Author Response

Thank you for the thoughtful and constructive feedback you provided regarding our manuscript.

Title: Impact Assessment of Difference in Land Uses Managements on Natural Vegetation in a Semi-Dry Area.

I attached the revised paper as the following. 

Responses to the comments of Reviewer #1

Manuscript:

1.       Throughout the manuscript, the term "land use" is used confusingly.

→Response: As I pointed out, the vocabulary of land use is misleading. So I change the term “land use” to “land management”.However I used the word “land use” when I literally use resources that exist in the land, such as cultivation and grazing.

Abstract:

1.       It is recommended to provide more specific details about the methodology, for example: what satellite images have been used? What period of time has been used to determine the land use-land cover changes? And what other technologies have been used in the study?

→Response: According to your comments, we added the sentences about the information of the satellite image used for the analysis in this study. We used the one time period of satellite image used. Since we focused on the evaluation how the differences of past land managements affected on the present natural vegetation, not the current change of LULC, we used only the present period of satellite image.

Methods:

1.       The authors declare that the "change of land uses" have been studied (figure 2) but only one WorldView-2 image of 2016 has been used to obtain the single land use maps and to extract the NDVI values from that date (cropland). Why not repeating the process for the other land uses?

→Response: First, we changed the term ‘land use’ to ‘land management’. We were interesting in the impact of past land management to current natural vegetation, not time-series changes of vegetation and/or land uses. As we already explained above, since we focused on the evaluation how the differences of past land managements affected on the present natural vegetation, we used only the present period of satellite image.

2.       In the paper by Tilahun [29] the NDVI values of 1984, 2000 and 2014 were mapped in Kilite Awulalo. It is assumed that this area of study encompasses Adizaboy Watershed. If correct, what is the relationship between the evolution of NDVI observed by Tilahun and those observed in 2016 by the authors of this article?

→Response: According to your comments, the Adi Zaboy watershed, which is the research site of this study, is included in Kilite Awulaelo district.

 The NDVI relates to the photosynthetic activity of living plants. Since we tried to classify the vegetation of the study area using some threshold values, we consider that it is acceptable to adopt the Tilahum’s threshold which was tested in the same area to this study.

3.       The source of Figure 4 is missing. Are the NDVI values derived from the 2016 WorldView-2 image?

→Response: The source in Figure 4 is described in 2.3. Satellite Image Processing, lines 178 to 185. When categorizing cropland, I don’t use NDVI because I use unsupervised ISO classification and field survey when categorizing croplands.

Results:

1.       Figure 1: It is convenient to revise your edition to improve readability. It is recommended to include a graphic scale and a color range in slope map.

→Response: I corrected this part according to your comment.

2.       Figure 2: I suggest substituting the title "Change of land uses in Adizaboy watershed during imperial, socialist and democratic eras" by "Change of administrative areas / land tenure in Adizaboy watershed during imperial, socialist and democratic eras". In addition, I recommend removing the letters and labels of each shape (eg "I-Free grazing"). They are unnecessary and make reading difficult.

→Response: The word “land use” has a misunderstanding. For this reason we changed the title in Figure 2 to "Change of land managements in Adi Zaboy watershed during imperial, socialist and democratic eras". Your advice is very much appreciated, but I believe that the letters and labels of each shape are necessary to clarify the adaptation points in the figure indicated by the text.

3.       I recommend replacing the title of section 3.4 "Changes in the Use of Croplands" with"Spatial distribution of crop land use-land cover". Actually, there is no analysis of temporal change.

→Response: I corrected this part according to your comment.

4.       The text between lines 297 and 312 is redundant with respect to figure 5 and does not provide additional explanations to the reader (causes of processes, for example). It is convenient to insist more on the reasons that explain the differences of the NDVI in each administrative area.

→Response: According to your suggestion, I modified these sentences as the following;

‘As the results of analyzing the ratio of NDVI value in the natural vegetation according to indicators of small, medium and high, the water or building was 0.6%, 1.1% and 0.3%, the small vegetation was 0.6%, 0.2% and 2.6%, the medium vegetation was 25.5%, 32.1% and 53.9%, and the high vegetation was 73.3%, 66.6% and 43.2%, respectively in the prohibited grazing and protected forest land classification, the seasonally-closed grazing land classification and the free grazing land (Figure 5). ‘

The following sentence was added to line 321 and line 330 in order to explain the differences of the NDVI in each administrative area.

Line 321

In addition, many conservation activities (SWC) and afforestation activities have been carried out within the prohibited grazing and protected forest land by local farmers. Therefore, stone bunds and planted woody plants can keep rain water flowing on the slope. It was considered that vegetation growth was positively promoted by the increase of soil moisture content increase (Nyssen 2008).

Nyssen, J.; Poesen, J.; Descheemaeker, K.; Haregeweyn, N.; Haile, M.; Moeyersons, J.; Deckers, J. Effects of region-wide soil and water conservation in semi-arid areas: the case of northern Ethiopia. Zeitschrift für Geomorphologie 2008, 52, 291-315, doi:10.1127/0372-8854/2008/0052-0291.

Line 330.

Also, the spray of livestock manure during grazing period which results is supplying organic matter and nutrients into the soil would positively promote the growth of wild herbaceous plants and wild wood plants in the seasonally-closed grazing land.

5.       According to figure 5, the percentages expressed in line 312 "seasonally-closed grazing land was 66.6% and the free grazing land was 43.2%" are reversed.

→Response: Since I described different numbers, I corrected it.

6.       In figure 5 (top left), the perimeter of "free grazing land" is missing. Are the white patches (croplands) related with 0 NDVI value?

→Response: For the surroundings of free grazing lands, the image was misaligned and fixed.

→Response: The white patches don’t have relation with NDVI value. We are focusing on natural vegetation in this study. So, we excluded cropland from the survey area when assessing vegetation by NDVI. We created items for exclusion (crop land) in the legend.

7.       It seems unclear. Replace "Buildig" with "Building" in the legend box.

→Response: I corrected this word according to your comment.

8.       In the lower part it is recommended to use a semi-logarithmic scale to represent the “y” axis. Currently, the bars of "Water, Building" and "Small vegetation" classes are imperceptible.

→Response: We created figure 5 of a semi-logarithmic scale to represent the “y” axis. However, differences in land managements in middle vegetation and high vegetation, which we want to focus in this discussion, will be difficult to see. So, we use original linear graph in this discussion.

Discussion and Conclusion:

As mentioned above, these sections are missing.

→Response: About Discussion, I describe it in Results and Discussion of line 186. For Conclusion, it is added from line 364.

Other formal issues:

1.       Hectors" is the right word (lines 67 and 68) or does it refer to "hectares"?

→Response: I corrected this words according to your comment.

2.       Tilahun's paper [29] is not published in 2000 but in 2015

→Response: I corrected this point according to your comment.

Thank you for giving us the opportunity to strengthen our manuscript with your valuable comments and queries. We have worked hard to incorporate your feedback and hope that these revisions persuade you to accept our submission.

Sincerely,

Lead Author/ Ryunosuke Ogawa

Corresponding Author/ Masahiro Hirata

Institution/ Obihiro University of Agriculture and Veterinary Medicine

Inada-cho, Obihiro, Hokkaido, 080-8555.

E-mail.: [email protected]  Tel.: +81-155-49-5485

Author Response File: Author Response.docx

Reviewer 2 Report

The theme addressed, that of changes in land use, is interesting and current. The case study is also interesting considering its political phases and management models.

However, the methodology used must be better explained and some information must be reported because it is incomplete in this form.

line 161. insert (with a table) the frequencies of the multispectral images. Furthermore, are not the time scenarios analyzed, at what period do they refer?

line 165. define: small, medium and large vegetation. What kind of forest environment do you refer to?

- How did you calibrate the NDVI index for the analyzed environments? When you say "small vegetation 0.1 - 0.2 NDVI", how did you determine it? You have to explain better.

line 173. possibly increase the bibliographic references for this topic.

line 296. this period repeats itself, it is redundant.

- in the results you do not mention the accuracy of the unsupervised classification. You could use the KIA coefficient to quantify accuracy.

Author Response

Thank you for the thoughtful and constructive feedback you provided regarding our manuscript.

Title: Impact Assessment of Difference in Land Uses Managements on Natural Vegetation in a Semi-Dry Area.

Responses to the comments of Reviewer #2

1.       line 161. insert (with a table) the frequencies of the multispectral images. Furthermore, are not the time scenarios analyzed, at what period do they refer?

→Response: I can not catch your point adequately. Is it the multispectral images the 8 band images taken by the Worldview - 2 satellite used in this study? Or is it the NDVI image? Also, for what reason do you need to insert the frequencies of the multispectral images? Since we focused on the evaluation how the differences of past land managements affected on the present natural vegetation, we used only the present period of satellite image taken on September 2016.

  If you never mind, could you tell me a little bite more explanation of your comment?

2.       line 165. define: small, medium and large vegetation. What kind of forest environment do you refer to?

→Response:  According to your suggestion, I added the following sentence on line 175.

The small, medium and high vegetation represent coverage rate of natural vegetation in Adi Zaboy watershed less than 20%, 20% - 45% and more than 45%, respectively.

3.       - How did you calibrate the NDVI index for the analyzed environments? When you say "small vegetation 0.1 - 0.2 NDVI", how did you determine it? You have to explain better.

→Response:  The NDVI relates to the photosynthetic activity of living plants. Since we tried to classify the vegetation of the study area using some threshold values, we consider that it is acceptable to adopt the Tilahum’s threshold which was tested in the same area to this study.

According to John Weier and David Herring, 2000, when evaluating the NDVI of the whole earth, 0.1 or less is related to barren areas of rock, sand, or snow. 0.2 to 0.3 is described as representing shrub and grassland. Therefore, we decided that the value of 0.1 to 0.2 which is between them is a minute vegetation which is difficult to classify. Tilahum positions that value as very small vegetation.

With reference to two papers, I described the features of NDVI in 2.3. Satellite Image Processing in lines 169 to 171.

4.       line 173. possibly increase the bibliographic references for this topic.

→Response: A sentence was added by referring to the following documents on line 182.

ISO cluster unsupervised classification is commonly used as one method of estimating cropland using

satellite images (Kulkarni 2017).

Namita M Kulkarni. Crop Identification Using Unsuperviesd ISODATA and K-Means from Multispectral Remote Sensing Imagery. Journal of Engineering Research and Application 2017, 7, 45-49, doi:10.9790/9622-0704014549.

5.       line 296. this period repeats itself, it is redundant line 296. this period repeats itself, it is redundant.

→Response: According to your suggestion, I modified these sentences as the following;

‘As the results of analyzing the ratio of NDVI value in the natural vegetation according to indicators of small, medium and high, the water or building was 0.6%, 1.1% and 0.3%, the small vegetation was 0.6%, 0.2% and 2.6%, the medium vegetation was 25.5%, 32.1% and 53.9%, and the high vegetation was 73.3%, 66.6% and 43.2%, respectively in the prohibited grazing and protected forest land classification, the seasonally-closed grazing land classification and the free grazing land (Figure 5). ‘

6.       - In the results you do not mention the accuracy of the unsupervised classification. You could use the KIA coefficient to quantify accuracy.

→Response: 2.3. Satellite Image Processing I changed the sentence of line 178 as follows.

ISO cluster unsupervised classification with the WorldView-2 multispectral orthography images was used to categorize the croplands in Adi Zaboy watershed, referenced the ground-truth data, and then its accuracy was evaluated by the field survey data at 101 points in croplands.

And 3.4. Spacial distribution of crop land use-land cover. The following sentence was added to line 291.

Figure 4 shows the results of spatial distribution of classified croplands in Adi Zaboy watershed. Its accuracy was 83% from the field survey data at 101 points..

Thank you for giving us the opportunity to strengthen our manuscript with your valuable comments and queries. We have worked hard to incorporate your feedback and hope that these revisions persuade you to accept our submission.

Sincerely,

Lead Author/ Ryunosuke Ogawa

Corresponding Author/ Masahiro Hirata

Institution/ Obihiro University of Agriculture and Veterinary Medicine

Inada-cho, Obihiro, Hokkaido, 080-8555.

E-mail.: [email protected]  Tel.: +81-155-49-5485

I attached the revised paper as the following.

Author Response File: Author Response.pdf

Round  2

Reviewer 1 Report

The revised version has substantially improved the original manuscript both formally and from the point of view of the contents.

Author Response

Thank you for giving us the opportunity to strengthen our manuscript with your valuable comments and queries.

Sincerely,

Lead Author/ Ryunosuke Ogawa

Corresponding Author/ Masahiro Hirata

Institution/ Obihiro University of Agriculture and Veterinary Medicine

Inada-cho, Obihiro, Hokkaido, 080-8555.

E-mail.: [email protected]  Tel.: +81-155-49-5485

Reviewer 2 Report

I am not fully satisfied with the review of the paper.

1 - The frequency of the spectral bands used to determine the NDVI index in your work is important for comparing your results with other multi-spectral sensors. Each multi-spectral sensor uses bands with different wave frequencies If, for example, you want to use the multi-spectral images of the Sentinel-2 satellite or the Landsat satellite on the same study area, you would observe different NDVI values. This is why it is important to specify the wavelengths of the multi-spectral bands that you have used in your work.

2 - land cover, expressed as a percentage, and the relative NDVI value are not always correlated. In fact, identical NDVI values can express different types of use of plant soil. Specify the main species, trees or shrubs, that characterize the thresholds you have identified.

3 - You choose NDVI value thresholds without providing adequate information in this regard. You refer to "John Weier and David Herring, 2000" which gives general indications on a worldwide scale. Describe better the thresholds you have identified.

4 - revision accepted.

5 - revision accepted.

6 - revision accepted.

Author Response

Responses to the comments of Reviewer #2

Thank you for the thoughtful and constructive feedback you provided regarding our manuscript.

1 - The frequency of the spectral bands used to determine the NDVI index in your work is important for comparing your results with other multi-spectral sensors. Each multi-spectral sensor uses bands with different wave frequencies If, for example, you want to use the multi-spectral images of the Sentinel-2 satellite or the Landsat satellite on the same study area, you would observe different NDVI values. This is why it is important to specify the wavelengths of the multi-spectral bands that you have used in your work.

→Response: According to your suggestion, a sentence was added the following documents on line 161.

The satellite image of World view-2 consists of 8 bands. The red Band 5 (630 – 690 nm) and infrared Band 7 (770 – 895 nm) were used to calculate NDVI in this research.

2 - Land cover, expressed as a percentage, and the relative NDVI value are not always correlated. In fact, identical NDVI values can express different types of use of plant soil. Specify the main species, trees or shrubs, that characterize the thresholds you have identified.

→Response: Our focus was not on the identification of plant name, but on the identification of the coverage of total grasses, total shrubs and total trees per unit area. We do not have enough data of plants’ names. That is why we can not mention the name of local plants. The important data in this point is the threshold values of total plant coverage.

3 - You choose NDVI value thresholds without providing adequate information in this regard. You refer to "John Weier and David Herring, 2000" which gives general indications on a worldwide scale. Describe better the thresholds you have identified.

→Response: "John Weier and David Herring, 2000" provides as; very low values of NDVI (0.1 and below) correspond to barren areas of rock, sand, or snow, moderate values represent shrub and grassland (0.2 to 0.3), while high values indicate temperate and tropical rainforests (0.6 to 0.8). As this definition shows, the values of NDVI from 0.1 to 0.2 and 0.3 to 0.6 are not clear. That is why we could not adapt "John Weier and David Herring, 2000" to our study. We would like to use the thresholds (Tilahun’s value) fit to our study.

Thank you for giving us the opportunity to strengthen our manuscript with your valuable comments and queries. We have worked hard to incorporate your feedback and hope that these revisions persuade you to accept our submission.

Sincerely,

Lead Author/ Ryunosuke Ogawa

Corresponding Author/ Masahiro Hirata

Institution/ Obihiro University of Agriculture and Veterinary Medicine

Inada-cho, Obihiro, Hokkaido, 080-8555.

E-mail.: [email protected]  Tel.: +81-155-49-5485