Peatland Transformation: Land Cover Changes and Driving Factors in the Kampar Peninsula (1990–2020)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This is an excellent paper on an important subject which many people will find of interest. My only comments are:

1: There are places where you have started a sentence by saying that (23) said. If you want to say who said it you should formulate it as Smith et al (23) said. I've indicated in the text some places where this needs to be corrected - there are more places.

2: there are a few places - I marked the first few - where you use the future tense about something which I think you have already done - I've marked a couple of places. It looks as though you have pasted in a paragraph from your funding application, You have to correct this.

3: In the conclusions you might say something about the fact that we think that deforestation has now ceased. You might say that things have now reached a steady state.

4: I'm not sure but I did think that maybe the methods part was too detailed and some of that material might be a bit repetitive and some might be submitted as supporting material - but I would not insist on this

Otherwise, I think this is an excellent paper that should attract a lot of interest and lots of citations. Congratulations

Author Response

Please see the attachment. Thank you.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

Review of “Peatland Transformation: Land Cover Changes and Driving Factors in the Kampar Peninsula (1990-2020)” Novarina et al

 The authors review changes in land cover, focussing on a forest cover change measure, over the Kampar Peninsula. They consider quantitative measures of change based on Landsat satellite data between 1990 to 2020, and compare this with historical information dating prior to this period, and to policy and industry changes also covering this period. Tropical peatlands are increasingly recognised as an important part of the global carbon cycle, as a carbon sink and store, so this research is a valuable contribution in this area.

The paper is well written, and was enjoyable to read.

Really nice introduction section, sets the scene well for the reader. I like the approach, how the effects of policy can be seen on the ground quantitatively.

Methods: they use NDFI and apply it to images from Landsat 4,5,7 and 8. How does this method take in to account the different sensors used in these satellites? For example, Landsat 7 and 8 sample slightly different parts of the spectrum for each of the bands – which is going to be very important in deriving the values of vegetation-linked indices. If these differences are not implicitly taken into account, then changes between these periods could be due to sensor differences, not due to on-the-ground differences. For time0 and time1, you only use one satellite’s dataset? Or, are they mixed? This is an important question when considering the most recent changes.

The Results present only the satellite based data, and the Policy (qualitative side) is only presented in the Discussion. I think it makes sense to introduce a summary of the qualitative data at least as a table in the results section. Especially as it appears as part of the methods section, so it is arguably results. The table could be, column 1 time-period, column 2 description of historical events/developments/policy implementation.

The conclusion I feel doesn’t reflect very well the aims of the research – or, the research aims/question are not clearly set out. If recommendations for means of balancing the needs of economic prosperity, and conserving the natural state of the peatland are the main output of this research, that needs to be clearly stated earlier, and in the introduction. If the conclusion is to stay in the current form, more effort needs to be made in the introduction and background to emphasise the value of these tropical peatlands and of their importance as carbon sinks (and other ecosystem services), and clearly state you want to make recommendations for how to go about protecting/developing tropical peatlands. Otherwise, the conclusion needs to be scaled back, these recommendations currently read as a kind of wish-list.

The key methodology point is to check whether Landsat 8 gives different results using NDFI compared to Landsat 7 (or 5). If the results are different, the most recent time period should probably be removed, or re-calculated excluding Landsat 8. 

The key research point is to clarify what are the aims of the research. If they are to make recommendations, the arguments need to be strengthened. For example in the conclusion "To address the complex challenges identified in this study" - but the challenges have not be explicitly laid out, the information has been presented about changes and policy, but no judgment is made about what is "good" or "bad". 

 

 Line 110 to 113: can you show the formula, and describe it here, as it is the basis of all the quantitative part of the work.

Line 130: typo in vegetation

Line 294: Sentence reads strangely, the calculation of these areas was performed?

Line 381: missing word for plantations colour

Lines 535 to 538: regarding conflict, is this an opinion or do you have some (at least anecdotal) evidence? It would strengthen the argument.

Lines 546 to 549: Repetition of the same key info in these 2 sentences, perhaps combine

 Fig 1, no legend for the colours shown in panel a

Table 2, typos in row headings?

Fig 6, I assume a final version will have higher resolution figures for the maps. The Acacia plantations are all the areas outlined in black, but they are classified Natural Forest in the earliest periods, is this correct? They were planted after this perhaps. And in panels b, the black areas are showing forest disturbance, but the classification is “regrowth” in 1990-1993. These are areas of regrowth, or of disturbance? In the period 2018-2020, could the inclusion of Landsat 8 data explain this reduction in disturbance?

Fig 12, needs a legend to explain features shown, same for fig 13 and 11. Or in the caption point to fig 6 for the legend

Comments on the Quality of English Language

The quality of English is very good.

Author Response

Please see the attachment. Thank you

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Manuscript: Land-3151690

Title: Peatland Transformation: Land Cover Changes and Driving Factors in the Kampar Peninsula (1990-2020)

Authors: Dian Novarina, Jatna Supriatna, Iman Santoso, and Mahawan Karuniasa

 

General comments:

 

In this paper, land cover changes from 1990-2020 and driving factors of tropical peatlands in Kampar Peninsula, Indonesia were mapped and investigated. This paper stated that the major driver of the land use changes were access road construction and government policy. While the dataset and findings are important and are current research interests in the topic of peat land use change, the presentation of the manuscript requires some improvement.

My main concerns are around the clarity and conciseness of the writing. Also, I found some repeat sentences in the paper from the introduction to conclusion. My other concerns regarding the findings can be found in the specific comment below.

 

 

Some comments and suggestions were:

Title

Peatland transformation: is the land use change was radical in this region? Maybe better to revise the title and increase the words “ Indonesia” in the title.

 

Abstract.

·        I suggest revising or adding the introduction of the abstract why land use change in Kampar Peninsula needs to studied.

·        Maybe it would be better to show the result of the land cover changes from 1990 and 2020 (not only data on land cover in 2020).

·        I suggest adding the methods: the combination of quantitative and qualitative analysis.

 

Introduction

Please make it more concise but clear. The introduction is too long and maybe some sentences need to be deleted.

Line 30-32. This sentence is similar to line 26-28.

Line 46-48. Please add a reference.

Line 48-52. Is this case in Indonesia? or the world?

Line 53-72. Please add references.

Line 78-88. Please add references.

Line 76-88. I think it would be better to revise this paragraph and make it more concise.

Figure 1. Is this figure important? Better to merge with figure 2 or delete it.

Line 95. Maybe better to be consistent with the word “Kampar Peninsula (Siak and Pelalawan)” rather than Riau, as maybe some readers don’t know where Riau is. The authors better explain where is the location of Kampar Peninsula in the introduction.

Line 95-103. I suggest revising this paragraph to make it more concise.

Line 114. Please use Kampar Peninsula rather than Riau.

Line 119-122. Is there any relationship between this sentence and your research?

Line 122-124. Please explain this sentence.

Line 146-151. I think it would be better to move it in materials and methods.

Line 146. Why will employ?

 

 

Materials and Methods

Line 157-165. Please add references.

Figure 2. Where is the map of peat? Are all areas of Kampar Peninsula peat soil? Please explain the map legend of Teluk Lanus etc.

Line 182-184. This sentence is similar or repeated in Data Analysis.

Line 208-215. These sentences are similar or repeated in line 216-223 of Data Analysis.

Figure 3. Please make a flow chart or workflow of all your methods so the reader can easily understand the flow of study.

Figure 4 and 5. Are these figures important for this paper?

Line 244-251. Is there any other agricultural use (such as seasonal agriculture).

Line 258-263. How about the data analysis of qualitative method?

Line 267-269. I think this sentence is similar to or a repetition from Materials and Methods.

Line 270-27. Please make it shorter.

Line 276-277. The reference is not a number.

Line 286-288. Is this sentence important? Maybe better to move it in introduction.

 

Results

Line 311-327. Is this from your study? If not, please add references.

Line 329-335. The sentences were similar or repetitions of materials and methods.

Line 336-353. This paragraph is likely part of materials and methods.

Figure 6,7, 8, and 9. Please increase the size or resolution of the figures.

Line 381. Maybe typos.

Line 395-408. Better to explain in percentages. Or maybe better to add the percentage of land use change in Table 5.

Line 411-412. I think better to move it to Discussion.

Maybe I missed it, but I can’t find the results of quantitative analysis in this section.

 

Discussion

Line 447-457. Please add a reference.

Line 470. Please add a reference.

Line 472. PT. CPI (abbreviation?) the same as Caltex?

Figure 10 and 11. Maybe better to be merged?

Line 533. The reference is not a number.

Line 546. Please expand the abbreviation of HTI for the first writing before using the abbreviation thereafter.

Line 546-556. This paragraph is also explained in Fig. 13?

Line 559-566. Which figure is related to this paragraph?

Line 575-576. Typos. Not number?

Line 615-643. No references?

 

Conclusions

Line 645-646. Are there any results or discussions for this conclusion?

Line 652-663. The sentences are likely similar to or a repetition of results and discussion.

 

 

 

 

Author Response

Please see the attachment and the full version of the revised manuscript. Thank you. 

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have implemented several changes to the text and layout that make the manuscript stronger. 

I still have some concerns about the mixing of Landsat sensors, although the authors now state they applied "cross-calibration methods to adjust for any discrepancies in spectral response between sensors", and later they state "To minimize discrepancies due to sensor difference, we primarily used data from a single Landsat satellite". This is reassuring, but a bit vague. What cross-calibration did you apply? In which cases have you mixed satellites in the  time0 to time1 comparison? What impact may mixing these have?

If the authors could provide some citations of studies that consider the impact of mixing Landsat 5 and 7 with Landsat 8 data, it would benefit the manuscript. If only to clearly state that the Thematic Mapper based sensor, and the OLI sensor sample different bandwidths of the electromagnetic spectrum, and this cause issues for creating unbroken timeseries analyses. This is quite important as it will surely affect the NDFI values.

Apart from this point, I think the manuscript is in a good state for publication.

Author Response

Please find the revised part highlighted in uploaded file.

Response:

For selecting and processing Landsat imagery from satellites 4, 5, 7, and 8, the workflow involves several steps that leverage cloud data and computing resources available through the open-source application Google Earth Engine using JavaScript. The data processing begins with the use of cloud data to access available Landsat images. These images are then organized into an image stack by the cloud engine. Specific years are selected for analysis, and the NDFI is calculated for these years to detect natural forest disturbances over time [28,29].

Before applying the NDFI, we calibrated and harmonized the spectral data from the different Landsat sensors. This involved using standard radiometric correction techniques and cross-calibration methods, specifically employing the Landsat Ecosystem Disturbance Adaptive Processing System (LEDAPS) and the Landsat Surface Reflectance Code (LaSRC) to achieve consistent surface reflectance values across all datasets. The cross-calibration accounted for differences in spectral response, particularly between Landsat 5 and 7 (Thematic Mapper sensors) and Landsat 8 (OLI sensor). Studies such as Roy et al. (2016) and Li et al. (2018) have documented the challenges and methods to mitigate discrepancies when mixing data from these sensors. For example, differences in bandwidths between the Thematic Mapper and OLI sensors can lead to inconsistencies in surface reflectance and spectral indices, including NDFI, which must be carefully addressed through calibration methods [30,31].

In cases where data from multiple Landsat sensors were mixed, we ensured that the temporal overlap between sensors was minimal to reduce the potential impact on time-series analysis. We prioritized the use of Landsat 8 data where possible due to its improved radiometric resolution, only supplementing it with Landsat 5 and 7 data in gaps where no Landsat 8 data were available. This approach minimizes sensor discrepancies, though mixing sensors may still introduce slight variations in NDFI values, particularly due to differences in spectral resolution and sensitivity across the visible and near-infrared bands.

Author Response File: Author Response.docx