Review Reports - Land Degradation-Based Approach for Assessing the Performance of the Natura 2000 Network in Mainland Spain

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript presents the result of a work aimed to assess the performance of a network of protected areas, the European Natura 2000, in Spain through an index based on the land degradation status. While the linkages between land degradation and climate change have gained attention in recent year, how land degradation is linked to the ecological status of the territory and its biodiversity is still far less well-known. The subject of the paper is therefore compelling and quite interesting for the readers of the journal.

The manuscript has, however, some drawbacks that should be addressed by the authors when submitting a revised version for considering for publication. Main issues to be considered are as follows:

Despite the effort made by the authors in laying the foundation and explaining the development of the new Relative Severity Index indicator (SER), lines 215-241, it may be difficult for readers who are not familiar with the D2RUE methodology to follow the logical thread without referring to the original papers. This issue could perhaps be alleviated by incorporating Table 1 and Figure 1 from the supplementary materials directly into the main text. Besides that, the authors may consider illustrating some examples of specific land conditions with some picture in as part of the supplementary materials.

On the methods, there is a need for clarification about some aspects:

How were the frequency distributions of the SER´s scores compared? (line 250-252). By the Mann-Whitney-Wilcoxon test (line 285)?

When authors refer to the calculation of the SER index across the entire study area (line 263, does it mean that they sampled an equal number (from 10 to 1000) of SER in both protected and non-protected areas?. Please clarify it.

Could you explain where the number of 999 SER calculated indices comes from? What do the authors mean by statistical rigor (line 274)?

I would suggest that the authors indicate whether the values shown in Table 1 and figure 1 are statistically different for SACs and UN2K areas?

On the discussion of the paper

Based on the correlation between SER and tree cover density, the authors argue for the importance of well-managed forest and woodland areas in achieving and maintaining good ecological status (lines 434–435). However, there are non-tree-covered ecosystems that represent good and integral, 'optimum' ecological status under specific biophysical circumstances (e.g., high altitudes, arid zones). How are these types of ecosystems captured by the SER? Could the correlation between tree cover density and SER be somewhat spurious due to the primary definition of land condition status?

Including additional indicators that reflect the degree of ecological integrity of the territory (both in SACs and UN2K areas), such as biodiversity or naturalness indexes as ancillary or explanatory variables, would provide deeper insight into the linkages between land degradation and ecological status.

Finally, the authors state that 'this study shows that degradation is not closely related to the current human footprint' (lines 476–477). Could you further elaborate on how the findings of this paper support this statement?

Other minor comments are:

Line 185. Which is the minimum area of analysed SACs? I think it is 10 km2 instead of 10 km

Author Response

Response to Revision (Manuscript ID: land-3424337)

Title: Land degradation-based approach for assessing the performance of the Natura 2000 Network in mainland Spain

Response to reviewer 1

Authors: First and foremost, I would like to express my gratitude for dedicating your time to this review. Thank you for this thoughtful report and your comments.

The manuscript has, however, some drawbacks that should be addressed by the authors when submitting a revised version for considering for publication. Main issues to be considered are as follows:

Authors: We appreciate this suggestion. In this revised version, Figure S1 and Table S1 have been incorporated into the main document to facilitate a clearer understanding of the 2dRUE product.

Besides that, the authors may consider illustrating some examples of specific land conditions with some picture in as part of the supplementary materials.

Authors: With respect to pictures to illustrate land condition, we believe the intended benefit could backfire. Land condition is assessed in terms of NPP or biomass for a given aridity level. Hence a series of pictures illustrating land condition classes would have to be taken within a certain aridity level, and would not be valid for a different one. For example, the amount of biomass associated with Reference performance in arid SE Spain would be considered as Productive with low biomass or even Degraded in more humid areas. Dealing with this properly would lengthen considerably the Supplementary materials, and is beyond the scope of this article. We consider that the interested reader might resort to the published references instead.

Methods, there is a need for clarification about some aspects:

How were the frequency distributions of the SER´s scores compared? (line 250-252). By the Mann-Whitney-Wilcoxon test (line 285)?

Authors: Yes, line 250 introduces how the analyses will be conducted. As the reviewer correctly points out, it is later stated that this comparison is performed using the Mann-Whitney-Wilcoxon test. To avoid altering all references, and since it is clearly indicated a few lines below, we have chosen to retain it.

Authors: Yes. In this new version, we have clarified the statement by adding more information to the sentence. Please, here is the new sentence:

<< The procedure started by calculating the SER index, first for the entire study area under SAC protection and then for the unprotected area. This process began with the computation for 10 cells and progressively increased to include up to 1000 cells (see Figure S1).>>

Could you explain where the number of 999 SER calculated indices comes from? What do the authors mean by statistical rigor (line 274)?

Authors: The choice to calculate 999 SER indices is based on the need to obtain a robust and reliable estimate of the index's variability for each study unit. This number is not arbitrary, but was selected as a random value large enough (~1000 calculations) to capture the variability of the sample size effect on the calculation of the SER index. The term "statistical rigor" refers to ensuring that the calculations performed are sufficiently representative and precise to provide reliable results. By performing 999 calculations, the goal is to obtain a wide enough distribution of values to allow for a reliable estimate of the median, minimizing the risk of results being influenced by an insufficient number of samples.

In this new version, a new sentence has been modified to facilitate understanding:

<< The number of 999 calculations was selected as a sufficiently large random value to capture the variability of the sample size effect on the SER index, ensuring statistical rigor without concerns about oversampling at this stage.>>

I would suggest that the authors indicate whether the values shown in Table 2 and figure 1 are statistically different for SACs and UN2K areas?

Authors: We apologize. The differences corresponding to the Basque Country in the SER index were not marked with an asterisk. In this revised version, we have modified Table 2 (now Table 3). Thank you for pointing this out.

Discussion:

Authors: Your point is correct and 2dRUE was developed to assess land condition accounting for differences in vegetation cover due to aridity. You may observe in the map (new Figure 1) that there are significant patches in arid SE Iberia that are correctly detected as Reference performance or Mature. Moreover, the formulation of RUE (NPP/P) makes and overcorrection, and tends to overestimate the condition of dry locations due to the use of low P as denominator. This is why 2dRUE further detrends the raw RUE scores for aridity, so that the resulting states can be compared irrespective of aridity variations in large study areas. Therefore, you are right at saying that the correlation between SER and TCD is somewhat spurious. It holds statistically for the overall population of sampling points due to the relative prevalence of semiarid and more humid zones in mainland Spain, but fails to explain arid zones. Because discussing such correlation contributes very little to the outcome of this study, we have deleted any mention to it.

Authors: As stated in the Introduction, we wanted to focus in the role of conservation networks to prevent negative effects (such as land degradation) rather than to favour positive ones (such as biodiversity or naturalness). Further, the way land degradation is approached here, through a full scale of ecological maturity, has appropriate explanatory power on the function of the involved ecosystems. We do not think that biodiversity data at a level of detail comparable to that used here for land degradation would be available for whole countries, as it is the scope of this study. Feedbacks between land degradation and biodiversity are very interesting themselves and deserve a dedicated study with a different experimental setup.

Authors: This phrase is simply wrong and we thank you for having detected it. In the context of this study, which is the same of 2dRUE and the initiatives UNCCD and SDG for which it was developed, land degradation is always attributed to human action, caused by desertification in terms of abuse or overexploitation of ecosystems. The phrase in question has remained undetected from early versions of the manuscript, trying to distinguish between ongoing land degradation an inherited desertification. Please find below the new wording referring to this aspect:

<<Land degradation occurs when soils are improperly managed, misused, or inefficiently utilized. This study shows that degradation is not necessarily related to the current human footprint, but it is complex to find a place in the Iberian Peninsula where various ecosystem compositions have not been altered, suggesting that the driving force of land degradation has been in operation for several centuries [23,75] and appears now as historical desertification []>>

Other minor comments are:

Line 185. Which is the minimum area of analysed SACs? I think it is 10 km2 instead of 10 km

Authors: Yes, we apologize for this error. In this revised version, it has been corrected.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript evaluates the effectiveness of the Natura 2000 (N2000) network in mainland Spain by utilizing the Normalized Relative Severity Index (SER) to assess land degradation and conservation. This manuscript provides a quantitative method for assessing the land status of protected areas. I recommend its publication after the following minor revision.

1. Methods:

It would be clearer if the authors could explain the reason in selecting variables like elevation, slope, and tree cover density in detail, as well as their ecological relevance to land degradation and conservation.

2. Results:

Table 1 does not show much information and should be able to be combined with Table 2.

3. Discussion:

For regions where unprotected areas outperform SACs (e.g., Cantabria, Galicia), please provide more context. For instance, are socio-economic factors, forest management practices, or land use histories influencing the outcomes?

Author Response

Response to Revision (Manuscript ID: land-3424337)

Title: Land degradation-based approach for assessing the performance of the Natura 2000 Network in mainland Spain

Response to reviewer 2

Authors: Thank you for kindly accepting this review. I deeply appreciate your effort in this assessment.

Methods:

Authors: This is described in the paragraph above. In this revised version, we have modified the sentence for better clarity. Below is the new sentence:

<<It was assumed that areas with higher elevation, relatively steeper slopes, minimal human impact, and greater tree density are likely to have undergone less transformation [49,50]. Consequently, these variables are expected to have an effect on the SER index.>>

Results:

Table 1 does not show much information and should be able to be combined with Table 2.

Authors: We appreciate this suggestion; however, we believe that, despite containing limited information, Table 1 is appropriately placed in its section (mainland Spain). Combining it with Table 2 would compromise the reader’s understanding. Currently, Table 1 (renamed Table 2 in this version) focuses on mainland Spain, allowing readers to identify differences in variables between areas under N2000 protection and those that are not protected. When readers proceed to the next section, where the analysis is broken down by autonomous communities, they will be able to perceive the effects of regionalization, having formed a preliminary understanding from the current Table 2. We hope this provides a solid argument for maintaining these tables separately.

Discussion:

Authors: A new, improved explanation of these results is now provided:

<< This result, contrary to the general trend, is likely related to the fact that these three ACs are located in the Atlantic region, where water is not a limiting factor for vegetation growth. The RUE paradigm was developed originally for drylands. In humid or wet sub-humid zones, 2dRUE correctly detects areas where tree cover is not as its maximum potential as a form of degradation, but struggles to tell condition differences in areas with a high vegetation cover. Such areas are frequent in UN2K land of that region, resulting from vigorous growth of vegetation, for instance, in marginal areas or forest crops. Compounding this, some relevant SACs in the Atlantic zone are designated on areas where the traditional management involves keeping grasslands for grazing at the expense of forests that grow naturally there. These areas are allocated a lower state than their forested surroundings.>>

Author Response File: Author Response.pdf