Review Reports
- Federica Isola 1,
- Bilge Kobak 1,2 and
- Corrado Zoppi 1,*
- et al.
Reviewer 1: Carmela Mariano Reviewer 2: Matteo Vizzarri Reviewer 3: Anwar Eziz
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsReview
The Influence of Ecosystem Services on Carbon Sequestration: A Comparative Study Concerning Three Italian Regions
Authors
Federica Isola , Bilge Kobak , Sabrina Lai , Francesca Leccis , Federica Leone , Corrado Zoppi
- Main question of the Research
The research aims to determine how the multifunctional characteristics of Ecosystem Services (ESS) supply influence the capacity of ESS to sequester carbon (Carbon Sequestration, CS). It specifically seeks to understand these spatial relationships to support climate neutrality through informed spatial planning strategies in three Italian regions: Basilicata, Campania, and Sardinia.
- Originality
The topic is relevant for climate neutrality targets established by national and international policies on this topic.
The paper addresses the complexity of multifunctional ecosystem interactions and moves beyond simple carbon mapping by using Geographically Weighted Regression (GWR) to explore how these relationships vary across different geographic and environmental contexts.
- Contribution to the subject area
Compared to other studies, this study adds a regional comparative analysis, detailing how the same ecosystem services can have different impacts on carbon sequestration across three Italian regions, highlighting how local context can shape these impacts.
This comparative analysis is conducted through an integrated methodological framework that combines established tools like InVEST (for carbon and habitat quality) and ESTIMAP (for recreation) with spatial regression to provide a comprehensive planning-oriented metric.
- Specific comments for methodology
Authors might consider the following improvements:
- The authors assumed a linear decay for threats to habitat quality because no available data support alternative functions. Future research could incorporate non-linear models if ecological data on specific threat distances becomes available.
- For the Land Surface Temperature (LST) analysis, images were selected from different dates in 2023 across regions. While intended to capture peak summer temperatures, this temporal variation could introduce minor inconsistencies compared to a strictly simultaneous snapshot.
- Conclusions
The conclusions are consistent with the evidence presented. The study successfully identified that Habitat Quality (QOHA) is the strongest and most consistent positive predictor of carbon sequestration across all regions. The evidence showed that agricultural production had a "negligible impact" on carbon sequestration, which validates the authors' use of a spatially varied model rather than assuming all green areas contribute equally.
- References
The references appear appropriate and comprehensive. They include foundational global frameworks and utilise standard technical manuals for the tools employed. They cite recent, peer-reviewed studies to ground their results in current scientific literature.
- Comments on Tables and Figures
- Figure 1: Effectively establishes the study's geographic scope by showing the three distinct regional panels.
- Table 1: Provides transparency regarding the specific satellite data (Landsat codes) used for climate analysis.
- Table 2 & 3: These are crucial as they provide the statistical rigor (Adjusted $R^2$, p-values, and coefficients) needed to support the study's claims about spatial variability.
- Figure 2-6: These maps are essential for the "spatial" aspect of the research, clearly illustrating the hotspots and coldspots of ecosystem service provision and their regression coefficients.
Author Response
We are grateful to the reviewer for their appreciation of the content proposed and discussed in the submitted manuscript.
With regard to the recommendation for improvement, for which we are also grateful to the reviewer, we would like to point out the following.
The recommendation is as follows.
Authors might consider the following improvements:
The authors assumed a linear decay for threats to habitat quality because no available data support alternative functions. Future research could incorporate non-linear models if ecological data on specific threat distances becomes available.
For the Land Surface Temperature (LST) analysis, images were selected from different dates in 2023 across regions. While intended to capture peak summer temperatures, this temporal variation could introduce minor inconsistencies compared to a strictly simultaneous snapshot.
This is how we address the Reviewer’s recommendation.
Following the Reviewer’s recommendations: i. in the Conclusions, we point out that future research developments should include the study and analysis of alternative, non-linear forms of decay functions concerning threats to habitat quality (lines 959-961 and 970-972, Section 5.); ii. in Section 2.4. Regulating local climate: land surface temperature mitigation, we include the observation suggested by the Reviewer, who points out that the temporal variation in temperature measurement may introduce minimal inconsistencies with respect to a simultaneous measurement, which, for Campania and Sardinia, is not possible as complete coverage of the regional territories is guaranteed by different LANDSAT images (lines 268-272, Section 2.4).
Author Response File:
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThe authors show the variation of carbon stock in three regions in Italy depending on other ecosystem services. Findings are timely and can be important to improve spatial planning at regional scale. However, the manuscript lacks of clarity on some important aspects. I therefore suggest authors to improve the contents by considering the following guiding questions.
- I did not get the rationale of the study. What’s the reason why CCS is driven by other functions? What are the main knowledge gaps? To what extent does the InVEST model provide a cost-benefit way to bridge this gap? And as a consequence, what’s the justification of having used GWR approach?
- The study might have several limitations that must be made explicit in the text. Some examples: (i) climate change mitigation is a human-made action to reduce GHG conservation in the atmosphere, so it depends on land use activities that have implications on land functionality. How is this addressed / considered by the authors? (ii) settlements and human infrastructures seem to be included in the maps. How are this included (?) in the analysis? How do they perform when it comes to trade-offs among multiple services? (iii) InVEST is a scenario model, so how do biophysical characteristics and processes of ecosystems contribute to ES outcomes and trade-offs?
- Discussions are weak. What’s the pattern of regional geomorphology and land uses compared to ES outcomes? What are the implications for regional planning?
I strongly encourage authors to address these concerns to improve the clarity, scientific soundness and readability of the manuscript.
Minor comments:
- Spelling of CCAS and difference from CCS?
- Can the authors express the C-related outcomes (at least in the text) per hectare than 900 m-2 ?
Author Response
1.
I did not get the rationale of the study. What’s the reason why CCS is driven by other functions? What are the main knowledge gaps? To what extent does the InVEST model provide a cost-benefit way to bridge this gap? And as a consequence, what’s the justification of having used GWR approach?
Following the Reviewer’s recommendations, we have added a paragraph, explaining the knowledge gap and the reasons of having used a GWR-based approach (see lines 83-106, Section 1). In addition, we have added a paragraph explaining that the InVEST Suite has been used to map two ecosystem services, carbon capture and storage and habitat quality (see lines 122-125, Section 1). As for the reasons why carbon capture and storage is driven by other functions, we have explained that the correlations have been studied in heuristic terms, without taking into account causal links which, although evident from the results obtained with the GWR-based approach, may be the subject of future studies (see lines 106-112, Section 1).
2.
The study might have several limitations that must be made explicit in the text. Some examples: (i) climate change mitigation is a human-made action to reduce GHG conservation in the atmosphere, so it depends on land use activities that have implications on land functionality. How is this addressed / considered by the authors?
Following the Reviewer’s recommendations, we have added two statements in section 5 to recognize this limitation and to clarify that this profile represents a promising direction for further research (lines 961-963 and 972-975).
(ii) settlements and human infrastructures seem to be included in the maps. How are this included (?) in the analysis? How do they perform when it comes to trade-offs among multiple services?
Following the Reviewer’s recommendation, in Section 3.3 we have specified that the outcomes of the GWR estimates related to the three regional contexts covered by the study show the highest negative correlations between CCAS and LST in densely built-up urban areas and rural areas characterized by intensive agricultural production (lines 541-549, Section 3.3).
In Section 3.2 we have pointed out that human infrastructures are recognized as major sources of degradation and that they reduce HQ scores according to their proximity and their threat level (lines 479-481).
The description of the results from the GWR provided in Section 3.6 (lines 684-748) has been expanded to explain whether, and how, distinctive patterns characterize values and spatial distribution of the coefficients for LOHM, QOHA, AFPL, and NOAP in urban settlements, also citing, when possible, patterns concerning linear infrastructures. Moreover, Figures 4-6 were revised by adding a basemap that shows the location of the main settlements, as a complement to the textual addition.
In section 5 we have recognized that the study does not explicitly assess how settlements and human infrastructures influence trade-offs among ESs and that this is a promising direction for future research (lines 963-965 and 972-975).
4.
(iii) InVEST is a scenario model, so how do biophysical characteristics and processes of ecosystems contribute to ES outcomes and trade-offs?
Following the Reviewer’s recommendation, in Sections 3.1, 3.2, 3.3, 3.4, and 3.5, we have added some paragraphs that relate the spatial taxonomies of the various ESs to the biophysical characteristics of the territory, with reference to the three regional contexts covered by the study proposed here (lines 454-459, Section 3.1; lines 495-508, Section 3.2; lines 535-549, Section 3.3; lines 570-579, Section 3.4; lines 598-608, Section 3.5).
The trade-off issue is dealt with in the Conclusions.
In section 5 we have recognized that the study does not explicitly assess how settlements and human infrastructures influence trade-offs among ESs and that this is a promising direction for future research (lines 963-965 and 972-975).
5.
Discussions are weak. What’s the pattern of regional geomorphology and land uses compared to ES outcomes?
The discussions have thoroughly been revised to address a comment from Reviewer 3.
We addressed the Reviewer’s comment in Section 3.6, where the description of the results from the GWR has been expanded to explain, when possible, distinctive patterns concerning the relationship between, on the one hand, geomorphology and land covers, and, on the other hand, the coefficients expressing the relationship between carbon and sequestration and the other ES (lines 684-748).
Moreover, Figures 4-6 have been revised to include a terrain basemap that shows not only the location of the main settlements, already mentioned in comment 3, but also the regional geomorphology. This addition is meant to clarify sentences that relate, for instance, to high or low elevations (for example: “Elevated areas characterized by natural land cover, such as the Appennines in Campania and Basilicata, and the Gennargentu-Supramonte massifs in Sardinia tend to display darker shades of red in Figures 4–6.”, or “Conversely, in Sardinia (Figure 6) the darker shades are found in a flat, low-elevation area in the south-eastern part of the island, corresponding to the Santa Gilla wetland and Natura 2000 site west of Cagliari”).
6.
What are the implications for regional planning?
In order to address the Reviewer's question, a paragraph has been added in Section 5 that emphasizes the transition of GWR outputs from descriptive maps into actionable planning instruments, thus reinforcing the practical relevance of the study for spatial planning (lines 941-952).
Author Response File:
Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for Authorssee the attached review report.
Comments for author File:
Comments.pdf
see the attached review report.
Author Response
2.1 Title
I think the current title gets the main subject right but feels a bit passive and misses the chance to immediately signal your novel methodological approach. It reads more like a descriptor than an engaging headline. I suggest refining it to more actively capture the spatial analysis at the heart of your paper. Something like “Mapping Spatial Synergies and Trade-offs: A Geographically Weighted Analysis of Ecosystem Services and Carbon Sequestration in Southern Italy” would immediately tell readers about your contribution.
Following the Reviewer’s recommendations, the title is changed into “Mapping Spatial Synergies and Trade-offs: A Geographically Weighted Analysis of Ecosystem Services and Carbon Sequestration in Southern Italy.”
2.2 Keywords
Looking at your keywords, I feel they are a touch too generic. “Carbon sequestration” and “ecosystem services” are essential but extremely broad, making it hard for your paper to stand out in searches. To attract the right readers, I recommend incorporating more specific terms that reflect your methods and focus. Adding “geographically weighted regression (GWR)”, “spatial planning”, and “InVEST model” would make your keyword profile much more precise and effective.
Following the Reviewer’s recommendations, the list of the keywords was integrated by adding “geographically weighted regression (GWR),” “spatial planning,” and “InVEST model.”
2.3 Abstract
I believe the abstract does a fair job of outlining the what and how, but it lacks a compelling narrative why. It launches directly into the methodological framework without first anchoring the study in a clear problem or research gap. Furthermore, the most intriguing finding—that the relationships are not uniform but change across the landscape—is buried. To strengthen it, I suggest starting with a brief statement on the need for spatially explicit planning for climate neutrality, then clearly stating your research question. Crucially, include one or two of your most striking quantitative results (e.g., the strong positive elasticity for habitat quality versus the negligible effect of agricultural value) to give readers a concrete reason to delve into the full paper.
Following the Reviewer’s recommendations, at the beginning of the Abstract, we added two sentences that describe the scientific and technical context of the study and explicitly state the research question. At the end of the abstract, we added a sentence that highlights some of the most striking outcomes achieved.
2.4 Introduction
I have a major concern here. As it stands, the introduction on pages 1-2 is not functional. It is truncated midthought at line 43-44 (“The evident”) and followed by a page of line numbers. This means the entire logical buildup—from the global challenge to the specific literature gap that justifies your study—is absent. There is no stated hypothesis or research question. This section must be completely rewritten. You need to construct a clear narrative: from climate neutrality goals, to the role of ES and CS, to the knowledge gap concerning spatially nonstationary relationships in Mediterranean landscapes, culminating in your specific objectives. This is nonnegotiable for establishing the scholarly contribution of your work.
Following the Reviewer’s recommendations, the introduction has been reorganized, clarifying the narrative from climate neutrality goals to the role of ecosystem services and carbon capture and storage, to the knowledge gap, culminating in the objective of the study (lines 83-106, Section 1).
2.5 Materials and Methods
Study Area (2.1): The textual description of the regions is quite good. However, I’m puzzled that Figure 1 is referenced but not included in the manuscript file I reviewed. A map is absolutely essential for readers to contextualize the geography you’re analyzing. Please ensure a clear, labeled map of the three regions is present in the final version.
We agree with the Reviewer’s comment, as a map showing the study area is absolutely essential, and indeed it is included in the manuscript.
Carbon Storage (2.2): Your description of the InVEST model is fine, but the procedure for integrating the belowground biomass (BGB) data strikes me as unclear. On page 4, you state the BGB map was “overlayed” in a separate step “to account for potential differences in calculation methods.” This vague phrasing leaves me unsure of the actual mathematical operation. Was it a simple raster sum? If so, why does that account for methodological differences? I suggest you explicitly state the exact GIS operation used (e.g., “the final CCS raster was calculated using the Cell Statistics ‘Sum’ tool in ArcGIS”) and provide a clearer rationale for this two-stage approach to ensure full reproducibility.
Following the Reviewer’s recommendations, we have added a paragraph, at the end of Section 2.2, where we explain the tool used and why (lines 200-206, Section 2.2).
Habitat Quality (2.3): I am of the opinion that the calibration of the half-saturation constant K needs better justification. On page 5, you describe setting K to half of the maximum degradation value from a preliminary run to “maximize visual contrast.” While pragmatic, this feels somewhat arbitrary from a scientific standpoint. The choice of K directly influences the absolute HQ scores, which are key inputs to your GWR. I recommend you briefly discuss how this calibration choice might affect the model’s output, or cite a precedent for this method. At minimum, a sentence acknowledging this as a model parameter choice that influences output scale would be helpful.
Following the Reviewer’s recommendations, in Section 2.3 (lines 249-253) we cited the InVEST model’s user guide and previous studies that applied this calibration. We also acknowledged that K selection directly affects the absolute HQ scores, which serve as key inputs for the GWR analysis and therefore influence the scale of the resulting outputs
Land Surface Temperature (2.4): I have a concern about your LST data selection. You mention selecting the single image with the highest average LST from the summer period for each region. While this captures a peak heat moment, I wonder if it might introduce bias by representing an extreme rather than a typical summer condition. Using a composite (like the median) of all suitable images might provide a more robust indicator of the land surface’s thermal regulation capacity. Please consider this alternative or justify why the peak temperature snapshot is the most relevant metric for your analysis of climate regulation ES.
Following the Reviewer’s recommendation, we have added a paragraph in Section 2.4, where we explain the reason for considering mean temperature as the metric for our analysis (lines 260-263, Section 2.4: the images associated with the highest mean and median do coincide).
Geographically Weighted Regressions (2.7): This section is methodologically sound but, in my view, misses crucial practical details for replication. You thoroughly explain the theory of bandwidth selection (pages 8-9) but omit the simple software commands and settings. Which specific tool in ArcGIS Pro was used? What were the lower and upper bounds for the bandwidth search using the golden section method? Stating something like “the bandwidth was optimized between a lower bound of 1 km and an upper bound of 50 km” is vital. Without these details, another researcher cannot faithfully repeat your analysis.
The questions raised by the Reviewer were addressed in Section 3.6, where the brief text earlier provided, only citing the golden search selection and the Gaussian local weighting scheme, has now been replaced by a detailed explanation that provides information and on the chosen parameters: i, model type; ii., neighborhood type; iii, neighborhood selection method; and, iv., local weighting scheme. See lines 623-641 (“This tool requires, as input, a shapefile … as the distance from the target feature increase”). We believe that we have now provided all the practical details needed to allow for replication.
2.6 Results
I feel the presentation of the GWR results is clear but could be more impactful. My main issue is with Table 3 (pages 14-15), which is overwhelmingly dense and difficult to parse quickly. Presenting the descriptive statistics for coefficients for three regions and four variables in one table is challenging for the reader. I suggest splitting this into three simpler tables, one per region, to improve readability. Furthermore, when you describe the spatial patterns in the text (e.g., lines ~540-560), try to more directly link the description of the maps (Figs 4-6) to the summary statistics in the table. For instance, explicitly note that the “dominance of blue shades” for AFPL corresponds to the positive mean coefficient reported.
As suggested by the Reviewer, the original Table 3 has been split into three tables (Table 3, Table 4 and Table 5 in the revised version).
2.7 Discussion
My belief is that this section is competent but stays in a safe zone. It does a good job of linking each of your findings to supporting literature, which is crucial. However, it reads more like a series of independent confirmations rather than a synthesized, critical interpretation. For example, in Section 4.3, you correctly note the weak AFPL signal and cite literature on trade-offs and synergies. But the profound implication—that your GWR reveals no strong, consistent regional signal, underscoring that management is everything—could be teased out more forcefully. Similarly, for NOAP (4.4), the coastal-inland dichotomy is well explained, but I think you could push further to articulate the clear spatial trade-off this reveals for planners: prioritizing high-recreation coasts may come at the cost of optimizing CS, and vice-versa. I encourage you to reframe the discussion to first synthesize the key spatial patterns, then explain their drivers, and most importantly, elaborate on their combined implications for integrated landscape planning in a way that transcends simply validating individual results.
Following the Reviewer’s recommendations, with regard to HQ, for which the representative variable, QOHA, shows completely analogous territorial patterns in the three regional cases, it was highlighted that the spatial characterization is almost entirely consistent with that of LST and that, therefore, HQ is positively correlated with the decrease in LST (lines 782-798, Section 4.1).
Some interpretative profiles of the discrepancy in the territorial patterns of the correlations resulting from the GWR estimates in the three regional cases were examined in depth, with particular reference to the significant difference between the elasticities of the CCAS and the LST (lines 806-827, Section 4.2).
As for AFPL, the significance of management, already mentioned in the original submission, has been reinforced throughout Section 4.3, and as an example of management practices, a paragraph has been added concerning open-field extensive grazing in the study areas and how such traditional practice impacts on carbon storage (lines 873-884, Section 4.3 “Moreover, the inner … Natura 2000 sites”).
Finally, concerning NOAP, the suggestion provided by the reviewer has been incorporated into lines 909-911 (“The key takeway … of the other”).
2.8 Conclusion
I think the conclusion accurately summarizes the paper’s workflow and findings. However, it ends somewhat softly. It successfully states what you did and found but could better articulate the so what for spatial planning practice. To strengthen it, consider ending with a forward-looking, actionable insight. For instance, emphasize that your GWR maps don’t just show correlations but actually identify spatial leverage points (e.g., zones where improving habitat quality yields the greatest CS co-benefit) and zones of inevitable trade-off (e.g., coastal recreation vs. carbon stocks), providing a direct science-based tool for differentiated regional planning strategies.
Following the Reviewer’s recommendations, a paragraph has been added in Section 5 that emphasizes the transition of GWR outputs from descriptive maps into actionable planning instruments, thus reinforcing the practical relevance of the study for spatial planning (lines 941-952).
2.9 References
From my spot-check, the reference list seems generally robust and includes recent work. However, given the centrality of the GWR method to your novel contribution, I am of the opinion that the methodological grounding could be slightly strengthened. The seminal Fotheringham et al. (1998) is cited, but adding one or two more recent key papers applying or advancing GWR in the context of ES or landscape ecology (e.g., from the last 5 years) would help position your work within the most current methodological discourse.
Following the Reviewer’s recommendations, two more references related to two recent relevant papers applying the GWR approach in the context of ESs were added, at the end of Section 2.7 Geographically-weighted regressions (lines 442-451).
2.10 Language & Clarity
My impression is that the language is generally clear and professional, though it tends towards a formal, passive style that can make for dry reading in some sections. I noticed occasional minor grammatical hiccups and a few instances where acronyms like SOC (Soil Organic Carbon) appear in the discussion before being defined. I suggest a thorough proofread to activate the voice where possible (e.g., “we calculated” instead of “it was calculated”) and to ensure all acronyms are spelled out upon first use, even if they were introduced earlier in the manuscript. This will enhance the overall flow and readability.
Following the Reviewer’s recommendations, a thorough proofread was implemented, the voice was activated where possible, and hiccups addressed.
Author Response File:
Author Response.pdf
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsI appreciate the authors' efforts in addressing my comments adequately. I have no further issues to disclose. Thank you
Reviewer 3 Report
Comments and Suggestions for AuthorsThe authors have responded to the first-round review with exceptional thoroughness and scholarly care. Every critical comment has been addressed, and the manuscript has been significantly improved in clarity, depth, and rigor. The Introduction now provides a proper foundation, the Methods are fully replicable, the Results are clearly presented, the Discussion offers genuine synthesis and insight, and the Conclusion articulates a powerful practical message.
In my view, the manuscript is now suitable for publication. I commend the authors for their constructive and diligent response to the review process.