Landcover Change in Tigray’s Semi-Arid Highlands (1935–2020): Implications for Runoff and Channel Morphology

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The topic of the article is suitable for the journal. The article is generally well written and well referenced. However, I have concerns about the methodology, particularly regarding the ability to measure high resolution changes in land surface morphology on low resolution images, on model calibration, and on the effects of hydrologic variability on river channel width.  

General comments

30 In the Abstract you call for land management strategies that “reduce hydrological connectivity”; you might flesh this out in the discussion (e.g. lines 419-422). What sorts of land management strategies?

79-85 I agree that this topic is under-researched due to lack of pre-satellite data and lack of older aerial photos of high resolution. This would be a good place to note how your data set differs. Do you have high resolution aerial images from earlier decades, or do you use a technique that allows you to measure with high resolution on a low-resolution image?   

93 Your first research question seems like something that is already known (in fact you describe it in lines 47-50). Perhaps restate as an objective to quantify these land use changes?

97 Your third objective is to “examine whether observed changes in river channel width can serve as a proxy for hydrological responses in data-scarce, semi-arid environments.” You do not clearly return to this thought in the discussion or conclusion.

143-167 In this section you need to describe the resolution of the images you use. These images are shown in figure 4, and I’m not convinced that you have high enough resolution in the historic photos to accurately delineate land uses or (especially) to measure river width to the tenth of a meter, as shown in Figures 6 and 7 and line275-278. Lines 169-174 state that the widths were measured every 10 meters (and you even measured tributary gullies, which are even smaller). I’m curious to know how you were able to manage this with any degree of accuracy on the 1935 image.

Also, measuring river channel width is somewhat ambiguous to begin with, even using high resolution images. Are you measuring the width of existing surface water or bankfull width, and are these even discernable on your images? If it is surface water width you are measuring, how are you addressing the seasonal variability that must occur even in this dry climate? (For example, did you ensure that each image was taken during a time of relatively similar hydrologic regime?)

Section 2.4 I understand the reasons why you used the same meteorological forcing for the model. A problem with this approach is that it does not catch changes in hydrologic variability over time (you note this on lines 351-352). For example, if the runoff variability (or the frequency of extreme weather events, lines 417-419) has changed significantly over the period of your study, you may be erroneously attributing changes in river channel width to land use changes. Can you account for this possibility? (It strikes me that you could re-run your model using four time periods of actual meteorological data to account for this and maybe identify the relative contribution of land use changes.)     

Also, you note that your model is not calibrated or validated (400-401), for lack of hydrologic data. It seems to me that you could calibrate it using present-day data on regional landscapes of differing land cover for which there are hydrologic data available. In other words, can you swap space for time to achieve some sort of model calibration?  

 

Minor notes:

28 I suggest “naturally vegetated land”

76 I suggest “hydrologic regimes” instead of “hydrology”

Figure 1You have the same color ramp depicting different elevations in the different maps. This is confusing. I suggest one color ramp in the legend that applies to all 3 maps.

Author Response

Reviewer 1

Comment 1: In the Abstract you call for land management strategies that “reduce hydrological connectivity”; you might flesh this out in the discussion (e.g. lines 419-422). What sorts of land management strategies?
Response 1: Thanks for the suggestion. We expanded on this in the discussion to cover specific land management strategies in relation to hydrologic connectivity and land degradation as a last paragraph in section 4.2.

Comment 2: Lines 79-85: Note how your data set differs. Do you have high-resolution aerial images from earlier decades, or do you use a technique that allows you to measure with high resolution on a low-resolution image?
Response 2: We have high-resolution scans (400 to 1200 dpi) of historical aerial photographs from Ethiopia dating from the mid-1930s and the 1960s. The georeferenced images have a spatial resolution of 1–1.6 m, now better explained and referenced in section 2.2. An appendix illustrates visual interpretation of the images.

Comment 3: Line 93: Your first research question seems like something already known. Perhaps restate as an objective to quantify these land use changes?
Response 3: Thanks for the suggestion; wording has been adjusted accordingly.

Comment 4: Line 97: Your third objective is to examine whether observed changes in river channel width can serve as a proxy for hydrological responses. You do not clearly return to this in discussion/conclusion.
Response 4: We explicitly addressed this in the discussion. While direct SWAT model calibration was not possible, observed relationships between landcover change, river channel width, and drainage network characteristics provide indirect validation. Progressive widening of channels alongside increased simulated runoff supports using river width as a proxy in data-scarce, semi-arid environments (Section 4.3, 3rd paragraph).

Comment 5: Lines 143-167: Describe the resolution of images used. Concerned that historic photos may not allow accurate measurement of river width to the tenth of a meter.
Response 5: Resolution and precision are now explained in section 2.2, with illustrations in the Appendix. Riverbank edges were measured, not surface water width. With 1–2 m resolution, visual interpolation yields ~1 m precision, with average width SE of ~0.7 m.

Comment 6: Section 2.4: Using a single meteorological scenario may not capture changes in hydrologic variability over time. Could you account for this?
Response 6: We acknowledge the limitation. Using a single 10-year scenario isolates the effect of landcover changes. Time-specific meteorological data could provide additional insights, but high-quality historical observations are limited. Consideration of ISIMIP3a datasets for follow-up studies is now noted in section 4.3.

Comment 7: Model is not calibrated or validated. Could you use present-day data from similar landscapes for calibration?
Response 7: We could not identify reliable proxy datasets. Correlating channel width with comparable streams was considered too speculative, and we acknowledge this as a study limitation.

Comment 8 (Minor): Line 28: Suggest “naturally vegetated land.”
Response 8: Added.

Comment 9 (Minor): Line 76: Suggest “hydrologic regimes” instead of “hydrology.”
Response 9: Added.

Comment 10 (Minor): Figure 1: Use a single colour ramp for all maps.
Response 10: Colour ramp of small-scale maps changed to grey scale

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

Your work is interesting because it presents issues related to current problems in this part of the world.
The work submitted for review presents issues related to these changes. The work is well-organized (from introduction to conclusions), with references to literature and figures. However, I have a few comments regarding its content. My comments are below:
1. Title: Eighty-five years of landcover change in Tigray's semi-arid highlands (1935–2020): implications for runoff and channel morphology - please consider changing the title, as 85 years should not be taken into account, as you only have data for four specific years. This is my suggestion.
2. Line 110; "Location of Figure 1." - a small error in the sentence. 3. Line 153: Table 2 - I think there's a bit little material to assess the changes; I suggest changing the title.
4. Line 189 - Please correct the citation.
5. Lines 275-277 - What were the reasons for the changes to the river channels? I hope I'll find the answer.
6. Line 194: Table 4 - Numbers should be first, then percentages. Table 4 needs to be corrected, and "133 mm" is a lot for pluses and minuses.
7. Line 317: Figure 6 - Please arrange Figure 6 like Figure 7 to better show the differences.
8. Line 324: 4. Discussion - I suggest adding papers from other regions of the world to the discussion, if available. There's a lot of speculation in the discussion, and this should be changed.
9. Line 335-337 - Are there any examples from other regions of the world regarding livestock farming?
10. Lines 338-348: The first paragraph says something similar, but without reference to specific catchments. Please consider this and choose one option – either in general or specifically for each catchment.
11. Lines 400-404: Isn't this a description that should be included in the Research Methods?
12. Lines 423-434: Has any evidence been found in the catchment area related to changes in the catchment area due to armed conflict?
13. Line 436: 5. Conclusions – I suggest further research, as the predicted rainfall will have a significant impact on changes in the catchment's relief. Perhaps it would be worth considering deploying measurement devices for weather data?
I think that's all.

Best regards,

Reviewer

Author Response

Reviewer 2

Comment 1: Title: “Eighty-five years of landcover change …” – consider changing as you only have data for four specific years.
Response 1: Suggestion taken; “the 85 years” removed from the title.

Comment 2: Line 110: "Location of Figure 1." – small error.
Response 2: Corrected.

Comment 3: Line 153: Table 2 – insufficient material to assess changes; suggest changing the title.
Response 3: Corrected.

Comment 4: Line 189: Correct the citation.
Response 4: Corrected.

Comment 5: Lines 275-277: Reasons for changes to river channels?
Response 5: River channels widened, implying higher peak discharges. While changes in rainfall regimes may contribute, modeling shows landcover changes alone increase surface runoff and peak discharges. Added details in discussion section 4.2.

Comment 6: Line 194: Table 4 – numbers first, then percentages; “133 mm” seems large.
Response 6: Section reorganized; simulated runoff figures first, followed by runoff coefficients.

Comment 7: Line 317: Figure 6 – arrange like Figure 7 to better show differences.
Response 7: Figures merged into one for clarity.

Comment 8: Line 324: 4. Discussion – add papers from other regions to reduce speculation.
Response 8: Discussion adjusted; references to studies in West African Sahel and Eastern Africa added.

Comment 9: Line 335-337: Examples of livestock farming from other regions?
Response 9: Added references to West African Sahel; note that “exclosures” policy is specific to Ethiopia.

Comment 10: Lines 338-348: Paragraphs general vs. specific; choose one.
Response 10: Section reworded for clarity.

Comment 11: Lines 400-404: Description seems like Methods rather than Discussion.
Response 11: Reworded; methodological choices clarified in M&M section, discussion reflects consequences.

Comment 12: Lines 423-434: Evidence of changes due to armed conflict?
Response 12: Aerial imagery predates the 2020–2022 conflict; limitation clarified in discussion.

Comment 13: Line 436: 5. Conclusions – suggest future research, e.g., weather measurements.
Response 13: Added recommendation: future research should integrate time-specific climate data or local weather measurements to better disentangle landcover and climate change impacts.