Evaluating and Improving the Effectiveness of Protected Areas to Conserve Plant Diversity Under Climate and Land-Use Changes

Abstract

Biodiversity is declining globally principally because of land degradation and more because of climate change. Its effective conservation is vital for species and habitats, but also to maintain the related ecosystem services they provide for human well-being. In this context, evaluating the ability of Protected Areas (PAs) to cover species distribution under current and future environmental conditions is highly valuable. Considering the distributions of 1692 species of plants in the cross-border region of Grand Genève, located between France and Switzerland, the effectiveness of existing PAs in preserving plant diversity through local hotspots and priority areas for rare and vulnerable species was evaluated. The results show that PAs are moderately effective in conserving plant diversity, but are not expected to lose effectiveness in future conditions because important areas for plant diversity conservation will remain at similar locations. To address this gap, a spatial conservation network combining hotspots and priority areas was identified to cover 30% of the study area. It captures a significantly higher proportion of species distributions under both current and future conditions, and covers a greater representation of rare and ecologically important habitats, such as subalpine meadows and wetlands. The proposed solution aims to inform local stakeholders about areas of high ecological value that could be used to identify the Blue-Green Infrastructure, supporting the expansion of PAs and the improvement of conservation strategies in the face of environmental change.

1. Introduction

Biodiversity is declining worldwide due to several causes amongst which are changes in Land-Use-Land-Cover (LULC) and climatic conditions [1,2]. This decline has tremendous impacts on ecosystem functions and processes, ultimately threatening our societies, jeopardizing the benefits we derive from a functioning and diverse nature [3]. Climate and LULC changes are expected to continue, intensifying their impacts on species distribution despite protection efforts [4,5]. To counter biodiversity’s decline, several protective measures were created. The IUCN Red List assesses the vulnerability of species to extinction by following a protocol measuring the evolution of species’ populations over the years. It plays a key role in informing the stakeholders about the prioritization of conservation efforts on the most threatened taxa. Furthermore, it is estimated that 16.8% of terrestrial areas and 8.01% of marine areas have protective or conservation measures for their environment according to the United Nations Environment Program World Conservation Monitoring Centre [6], with mixed results. Protected areas are globally increasing although there are some examples of decline [7,8]. At the landscape levels, the Green Infrastructure (GI) aims at identifying and conserving a network of (semi-)natural areas. GI is designed and managed for the preservation of biodiversity, ecosystem services and connectivity in a more flexible way than strictly protected areas [9,10]. It could serve as a link to connect core areas, allowing species movements and dispersion, and as a network of areas with different conservation statuses able to sustain a high biodiversity level. The distribution of species and more globally of biodiversity across a landscape serves as an interesting input to identify areas to be included in the GI or in the freshwater system-related concept named Blue Infrastructure.

Most ecosystems are expected to face changes in their climatic conditions by the end of the century [11]. Consequently, questions are raised about the ability of the current network of Protected Areas (PAs) to maintain their effectiveness in conserving species due to the loss of the landscape’s suitability or range shifts [12,13]. A literature review concluded that PAs’ efficiency under climate change is highly dependent on the context, such as their location and characteristics, as well as the taxa considered. But overall, it concluded that their ability to act as a refuge is becoming less certain [14]. Similarly, the IUCN statuses have shown limits in identifying species threatened by climate change [15,16,17], and a previous study concluded that they were not able to predict species’ vulnerability to global changes [18]. As a consequence, the tools at our disposal to protect species should be tested to ensure their relevance in the face of global changes.

Several methods allow for the identification of areas that are valuable for current and future conservation. For example, the spatial arrangement of specific richness (such as local hotspots) or the results of a spatial conservation prioritization. This last method is widely used in systematic conservation planning because it allows us to find the right compromise between conservation targets and a spatially optimized priority area network [19]. The software Zonation [20] produces a hierarchical prioritization map across the landscape, iteratively removing pixels with the lowest aggregation value according to the desired result. Therefore, it provides maps identifying priority areas depending on the settings and weights attributed to the inputs [21,22].

The study area is a dynamic transborder region between France and Switzerland, constituted of a plateau at low elevation, where most urban areas are located, surrounded by mountain ranges peaking at around 2000 m.a.s.l. These mountain ranges are mostly covered by forests with diverse degrees of management, while subalpine meadows and pastures are found approaching the summits. This complex organization allows for a wide range of (semi-)natural habitats to coexist on the same territory together with human infrastructures such as cities and agriculture. This combination of natural and anthropic areas has led to substantial assessments and identification of the regional GI in Geneva (Switzerland) and in its surroundings [23,24]. However, past studies focused only on current biodiversity features, and the information regarding the effectiveness of existing PAs to actually protect species distributions through time is lacking.

In this article, the distribution maps of 1692 plant species under current conditions and future conditions for 2050 based on two scenarios of climate change (optimistic and pessimistic) were used to assess the capacity of PAs to incorporate species distribution according to their IUCN and native statuses across a wide range of future conditions. All plant species for which sufficient data were found were considered, including exotic and archaeophyte plants (alien species introduced more than 500 years ago). To be aligned with the biodiversity strategy of the Canton of Geneva and the European Commission strategy of the United Nations Convention on Biological Diversity, an informative Plant Conservation Network (PCN) covering 30% of the territory was identified. It was computed by identifying and combining the distribution of local hotspots of species richness with the priority areas for rare and vulnerable species derived from a prioritization exercise [25,26,27]. The effectiveness of this optimal theoretical network of areas was compared to current PAs’, and could serve as a new input for identifying GI integrating not only the current distribution of biodiversity but also future projections. It also aims to inform regional stakeholders on the distribution of plant diversity in the landscape through time and even serves as a reservoir of areas to improve current PAs. This work contributes to (1) identifying the spatial distribution and characteristics of areas with high conservation value for the long-term preservation of plant diversity; (2) assessing the current and future effectiveness of PAs in protecting species distributions with a focus on native plants. The discussion further develops the utility and limitations of developing an informative PCN for stakeholders.

2. Materials and Methods

2.1. Protected Areas

Focusing on the strict perimeter of the cross-border territory of Grand Genève (France–Switzerland), all areas with a strict protection status were identified (Figure 1). These PAs represent protected areas and natural reserves designated by Swiss and French local, cantonal, regional or national authorities, as well as priority habitats recognized for their conservation value. They are distributed across the whole study area and represent 13.69% of the perimeter, which is similar to the proportion of protected surface in Switzerland [28]. Areas with too permissive legislation such as the Natura2000 network, natural areas of ecological interest, as well as the natural regional park of the Haut-Jura, were not included in PAs. Even if they also participate in nature conservation, they cover more than half of the territory and allow for intensive human activities. Thus, they were not considered as compatible with a strict protection system.

2.2. Plant Diversity Distribution

To test the effectiveness of PAs, we have to model and map the distribution of biodiversity. This was made in two steps. First, current and future species distributions were modelled. Then, these maps were aggregated to create two indexes: the distribution of specific richness identifying local hotspots of suitability, and priority areas resulting from a prioritization network focusing on the identification of important areas for rare and vulnerable species conservation.

2.2.1. Species Distribution Models

The distribution of 1692 plant species was modelled for current conditions and two future scenarios in 2050. Two separate models were designed at two different scales (regional vs. European continental). After data cleaning, 755,259 georeferenced observations were selected from local naturalist institutions to create a regional model at 25 m resolution using the current distribution of LULC categories as well as topographical and pedological predictors. LULC categories were projected in 2050 following a business-as-usual scenario, including already planned new infrastructures to create a set of future predictors. Then, approximately 22 million occurrences were compiled at the continental scale, retrieved from the Global Biodiversity Information Facility (GBIF, accessed in August 2021) and complemented with the regional ones. A table summarizing the species, number and origin of occurrences is available in Supplementary Materials (Table S1). Occurrences were corrected for potential spatial and sampling biases and were used to calibrate a model at the continental scale using climatic predictors at 30 arc-second resolution (~1 km in Europe). The models were projected in and downscaled to the study area using bilinear interpolation for current and future conditions in 2050 for six different general circulation models and two Representative Concentration Pathways (RCP) that represent optimistic and pessimistic scenarios of climate change (RCP26 and RCP85 from the fifth Coupled Model Intercomparison Project respectively) [29]. Average evaluation metrics from these two models are available in Appendix A (Figure A1). The resulting maps from these two models were then merged, resulting in three different time-step distribution maps at 25 m resolution for each species: current situation, optimistic future conditions in 2050 and pessimistic future conditions in 2050. A full description of the modelling method, algorithm, calibration, projection, performance metrics and a detailed explanation of the outputs and the methodology used to identify vulnerable species to global changes are available in Sanguet et al. (2025) [18], where the distribution maps were retrieved. The resulting maps have values ranging between 0 and 1, according to the suitability of the landscape, and were used to create two indexes representing the spatial arrangement of plant diversity.

2.2.2. Local Hotspots

Once modelled, the distribution maps were aggregated into two different informative indexes representing the distribution of plant diversity. Species distribution maps were summed for each time-step to create a map highlighting areas suitable for many species, named hereafter “local hotspots”. The term “hotspot” used here does not refer to the global hotspots concept developed by Myers et al. (2000) [30], but rather to a spatial area where many species have the possibility to grow compared to the rest of the study area [31,32]. These maps do not show the exact number of species able to grow in each pixel but rather depict a relative suitability index of the landscape.

2.2.3. Priority Areas

The software Zonation 5 was used to run a spatial prioritization analysis, using the Core-Area Zonation (CAZ) function which allows the selection of priority areas for single rare features. This setting was selected over the Additive-Benefit Function (ABF) because the latter gives more importance to pixels with many features, which would have been redundant with the local hotspots described in the previous section (Section 2.2.2), neglecting the importance of rare and localized species in conservation goals [21,33,34]. The CAZ setting thus offers an interesting complementary input to inform about the spatial arrangement of plant diversity, focusing on rare features instead of on areas incorporating many features.

The Zonation software was run using current and future species distributions, giving twice as much weight to the species identified as vulnerable to global changes in a previous study [18]. The vulnerability was measured according to their expected changes in suitability, habitat fragmentation, size of suitable surfaces and migration potential in future conditions. A full description of the method used is available in Sanguet et al., 2025 [18]. The continuous maps of habitat suitability were used in order to avoid biases [35]. Among the 1692 species assessed, 621 are considered vulnerable for optimistic scenarios and 753 for pessimistic ones, and they are almost entirely composed of native species. The resulting map represents a prioritized ranking of pixels where the highest values determine the most valuable areas for species conservation, named hereafter “priority areas”.

2.3. Relevance of Protected Areas and Plant Conservation Network Assessment

The relevance of PAs in conserving plant diversity was assessed by measuring their ability to encompass the distributions of species and of the local hotspots and priority areas. A focus was placed on the effectiveness of PAs to protect endangered species based on their IUCN status, as well as native plants, which have been demonstrated to be more vulnerable to global changes in the region [18]. This was done for current and future conditions, and a Wilcoxon mean test was applied to measure the effectiveness of the current network of PAs to conserve long-term plant diversity.

Then, to better understand the distribution of plant diversity in the territory, the most valuable areas from the local hotspots and priority areas were extracted and analyzed. To align with European and Swiss objectives in conserving 30% of their territory, these maps were normalized and merged together to extract the 30% areas with the highest ecological interest for plant diversity. This resulting map was named “Plant Conservation Network” (PCN). This was done for current species distributions, as well as in optimistic and pessimistic future distributions, and the spatial dissimilarities between scenarios were studied. The PCN was then compared to the current network of PAs in terms of habitats included, as well as integration of species distribution depending on their IUCN and native statuses. Finally, the improvement of PAs’ effectiveness and the utility of the PCN for GI identification were discussed.

3. Results

3.1. Distribution and Evolution of Plant Diversity

The local hotspots were distributed similarly in the current time and in the two future scenarios, which was surprising considering the changes in climatic conditions and the expected shifts in species distribution (Figure A2 in Appendix A). Indeed, they showed a great spatial overlap of 93.70% and 88.30% between current and future optimistic and pessimistic scenarios, respectively. The most suitable areas were located in the mountainous habitats, corresponding to subalpine meadows and pastures, as well as further South in the lowlands, corresponding to forested habitats. The magnitude of change was similar when comparing the current distribution with both future ones, although the trends were accentuated in the pessimistic scenario (Figure A3 in Appendix A). However, although the location of local hotspots remained similar across time, they were found to lose average suitability in the future. On the contrary, current areas with poor suitability in the lowlands tended to show a slight gain of suitability in future scenarios. The main difference observed from the optimistic and pessimistic future scenarios was an important decrease in suitability in the highlands, indicating that these areas are more vulnerable to global changes and to warming temperatures in particular.

The output maps from the prioritization process showed similar patterns across the three treatments (Figure A4 in Appendix A). Areas located in the highlands, and especially subalpine meadows and pastures, were classified as high priority for rare and vulnerable species. Some disparate patches located in the plateau were also integrated into these priority areas, but mid-elevation forests were not selected. These patterns remained similar in future conditions, and the spatial overlap between priority areas was 80.81% and 77.61% between current and optimistic future scenarios, and current and pessimistic future scenarios, respectively.

The local hotspots and the priority areas are presented in Figure 2. Interestingly, while priority areas tended to be located in the highlands, the local hotspots showed a wide distribution in the lowlands. The overlapping surfaces combining both indexes were located all over the study area and are represented by large patches of subalpine meadows and pastures, and some deciduous forests in the lowlands. The evolution of the local hotspots, the priority areas and their overlap for the three treatments are presented in Figure A5 in Appendix A.

3.2. Effectiveness and Relevance of Protected Areas

Selected PAs represented 13.69% of the study area and included a tremendous disparity in their habitat composition. Indeed, they were mostly composed of closed forest (68.55%), followed by meadows (14.99%) and crops (6.38%), while open forests and wetlands only represented a small proportion, respectively, of 1.74% and 2.01% of their surface. The remaining proportion was made up of disturbed habitats. However, habitats’ representativity was more constant with the integration of 18.42% of closed forests, 17.87% of open forests, 14.95% of wetlands, 11.30% of meadows and only 3.60% of crops.

Interestingly, PAs only included 14.84% of the identified local hotspots for the current time, and this proportion stayed relatively similar in future conditions (14.23% for optimistic future scenarios and 13.53% for pessimistic ones). This result was surprisingly low, especially considering that these hotspots only represented 30% of the territory. PAs included a slightly higher proportion of priority areas of 17.98% for the current time, 17.59% and 17.23% in future optimistic conditions and pessimistic ones, respectively. In other words, current PAs do not protect most of the local hotspots and priority areas, but their effectiveness is expected to remain broadly constant in future conditions.

Four species were not protected by PAs at the current time. This number increased to 18 species in optimistic future scenarios and 26 in pessimistic ones. They were only native species, mostly alpine plants, that were not threatened, with the exception of two critically endangered species for future scenarios. Table S2 in the Supplementary Materials shows detailed values of the proportion of species distribution included in PAs and the proportion of PAs suitable for species. On average, 12.30% of species distribution was integrated into PAs in the current situation, and this number stayed relatively constant in future conditions (see

Table 1. Average proportion of species distribution covered by PAs and by the PCN. The numbers are expressed in % of the total distribution.

Group of Species	Treatment	Protected Areas	Plant Conservation Network
All	Current	12.30%	34.89%
All	Optimistic	12.61%	35.82%
All	Pessimistic	12.85%	36.71%
	Current	13.77%	37.78%
Native	Optimistic	14.22%	38.86%
	Pessimistic	14.52%	39.93%
Archaeophyte	Current	4.57%	19.45%
	Optimistic	4.59%	20.27%
	Pessimistic	4.62%	20.42%
	Current	5.21%	19.26%
Exotic	Optimistic	5.06%	19.49%
	Pessimistic	5.08%	19.61%
Not endangered	Current	12.90%	36.78%
	Optimistic	13.04%	37.56%
	Pessimistic	13.18%	38.36%
	Current	10.54%	34.40%
Endangered	Optimistic	11.13%	33.89%
	Pessimistic	11.58%	34.13%

for details). A greater proportion of native species distributions was integrated in PAs compared to the other native statuses for current and future scenarios (p-value < 0.001; Figure 3).

3.3. Characteristics of the Plant Conservation Network and Comparison with Protected Areas

The PCN covered 30% of the territory and was identified by merging the local hotspots and priority areas (Figure 4). It is composed of closed forests (49.40%), meadows (24.18%), crops (10.99%), wetlands (3.51%), open forests (3.25%) and disturbed vegetation (1.91%), the rest being urban areas and communication roads. However, its habitats’ representativity was very different from PAs and included 29.97% of closed forests, 75.56% of open forests, 58.91% of wetlands, 41.15% of natural meadows, 14.20% of crops, and 10.68% of disturbed vegetation. Although its area was more than twice the surface of PAs, it is interesting to note that it integrated a lower proportion of closed forest than expected, but a higher proportion of open forests, wetlands, crops and meadows, suggesting that these specific habitats are important for plant diversity and should be further integrated into conservation planning. It integrated 83.78% of the local hotspots and 72.73% of the priority areas for the current situation.

In future conditions, the PCN integrated up to 80.00% of suitability hotspots and 67.86% of priority areas in optimistic scenarios and 77.15% and 66.20% in pessimistic ones. The overlap between current and future PCN is of 89.54% for optimistic scenarios and 84.99% for pessimistic ones (Figure A6 in Appendix A). This signifies that today’s most valuable areas for the conservation of plant diversity will mostly stay valuable under future conditions, which is an encouraging element from a conservation perspective. Finally, around 40% of PAs were integrated into the PCN, supporting the previous finding that most of the PAs currently cover areas that are not especially valuable for plant diversity, and suggesting that there is still room for improvement.

Overall, native species had a larger proportion of their distributions included in the PCN compared to archaeophyte and exotic species, for the three time-steps (p-value < 0.001; Figure 5). In addition, the proportion of native species distributions included was significantly higher in future pessimistic scenarios compared to the current situation (p-value < 0.01). This observation should be interpreted with care, as native species are expected to experience a greater reduction in their range, which could lead to a higher proportion of their distribution included in the PCN. However, it still implies that areas within the PCN were more suitable than those located outside of it. A comparison of PAs’ and PCN’s effectiveness depending on species’ IUCN and native statuses can be found in Table 1, and a detailed document showing the proportion of species distribution included in the PCN can be found in Table S3 in the Supplementary Materials.

4. Discussion

4.1. Valuable Habitats for Conservation

Areas located at altitude appeared to be of great interest for plant diversity, showing a high conservation priority for all treatments but a loss of average suitability in the future. These observations could be explained by the expected altitudinal shift in species distributions as they follow their optimal climatic conditions. This phenomenon is already observed in similar taxa across the Alps [36]. Mountain plants living in subalpine open habitats do not have many possibilities to migrate higher in elevation in the study area. Mountain summits reach around 2000 m.a.s.l., which corresponds to the beginning of the altitudinal range of their habitat and the upper limit of montane forests. The rarity, specific diversity and vulnerability to global changes in these habitats explain why they were particularly selected during the prioritization ranking process. They represent a relatively small area, concentrating a very unique diversity. Although most of the alpine species will still find suitable conditions by 2050, increased competition with forest species shifting higher in elevation combined with less suitable climatic conditions might contribute to increasing the local risks of extinction. Additionally, these subalpine habitats might also be vulnerable to forest encroachment due to the combined effects of the loss of traditional pastures that maintain habitats open already observed in the Alps, and the rising temperatures allowing more species of trees to grow higher in elevation [37,38]. While these habitats were found to be highly diverse, mid-elevation forests, on the contrary, showed a low priority and suitability. Thus, a strengthened management of these areas might be necessary in the future to maintain their quality. Higher mountain ranges are found outside the study area and species might still be able to migrate to these locations, despite the expected difficulties due to landscape fragmentation and the unprecedented rapidity of change [39,40,41,42].

Average suitability was found to slightly increase in the lowlands. This could be explained by the rising temperatures leading to an improved suitability for species usually growing in warmer urban environments found in the lowlands due to the urban heat island phenomenon [42,43]. Furthermore, climate change would allow more Mediterranean species to expand their current range within the study area or to colonize it in the near future. Latitudinal shifts are already observed in neighbouring Mediterranean habitats and are expected to continue as temperatures rise [44,45,46,47]. The results of this study thus possibly underestimate the increase in suitability in the lowlands because the models were developed only for species that are currently growing in the study area without considering potential future arrivals.

4.2. Ecological Characteristics of the Plant Conservation Network and Protected Areas

PAs are not always spatially congruent with species distributions, nor necessarily relevant when considering distributional shifts expected in the future. This finding alone supports the necessity of assessing their long-term effectiveness [12,48]. In this study, PAs are not expected to lose relevance in the future but they struggle to protect the local hotspots and the priority areas.

On the other hand, all indicators seem to point to the conclusion that subalpine meadows and pastures are highly valuable for plant diversity conservation, especially in the context of climate change. They represent relevant candidates for the expansion of natural reserves to better cover this overlooked habitat in the current network of PAs. In addition, most species whose distribution is not currently covered by PAs are alpine plants, and more are expected in the future due to climate change. Therefore, including more of these subalpine habitats in the network of PAs would largely benefit the most vulnerable species and the resilience of plant diversity. The areas included in the PCN could generally serve as a reservoir of candidates for increasing the surface of PAs or for identifying GI.

An interesting finding of this study is that the location of valuable areas for plant conservation does not fundamentally shift in future projections. Local hotspots, priority areas and the PCN calculated for current and future scenarios show a relatively high similarity, although species distribution is expected to change and shift higher in elevation. Several factors could explain this counter-intuitive result. First, the relatively small amount of time between the two time-steps (current situation and 2050) does not allow for major transformations in the study area. Second, only small changes were found in the future LULC map used as predictors for the regional species distribution models. This is also explained by the short timeframe of this projection (greater changes would be unrealistic), but also by the relatively high management of the environment in the territory. Indeed, applying the same methodology in a different context and on a larger scale would probably result in more visible changes in LULC. Consequently, the distribution of habitats remains similar between the current and future conditions, and the intrinsic ecological value of PAs covering them also remains relatively similar across time. Third, habitats play a major role in defining the ecological niche of a species. While climate is shifting, habitats might allow species to survive longer at the margins of their climatic niche, increasing their overall resilience to environmental changes. Fourth and finally, the areas identified as ecologically relevant for plant diversity conservation might be found in the core area of species distribution, which could explain why they stay mostly relevant in the future, even though peripheral areas are expected to shift.

4.3. Limits of the Methodology and Perspectives

The observations made regarding PAs must be nuanced. Although plants are at the bottom of all trophic chains and constitute the habitats upon which all animals depend, a similar assessment using animal species would be necessary. Indeed, while some habitats are not essential for the conservation of plant diversity, they might be highly relevant for animal movements, migrations, feeding habits or reproduction. In addition, PAs should not be identified based on the distribution of biodiversity only, but also considering landscape and species connectivity as well as ecosystem services supplies, as supported by the methodology to identify GIs [10,23,24].

In this work, it has been decided to consider only the areas with a strict protection level but areas with a more flexible legislation also participate in nature’s conservation. Even if human activities are allowed in these more flexible areas, they could be managed to deliver interesting conservation outputs such as connectivity, the supply of ecosystem services or the preservation of habitats of interest for some species. Consequently, the PAs considered here are not the only areas participating in species conservation. Finally, freshwater ecosystems also play an important role in the study area due to the combined presence of rivers (amongst which the Rhône) and Lake Geneva. Including freshwater species would be an interesting complement to this terrestrial ecosystem-based study, and would participate in the identification of the Blue Infrastructure in the territory.

This paper presented a workflow to identify relevant and valuable areas for the conservation of plant diversity, but the actual integration of these newly identified areas is complex and depends on other factors independent of their ecological interest. Indeed, imposing new legislation on private lands or public acquisitions is a very complex process, especially in a human-managed transboundary territory such as the one studied here. Dedicating new areas to species conservation would inevitably generate conflicts with other uses in the region. However, identifying highly valuable areas in a theoretical exercise and creating informative outputs designed for stakeholders to participate in improving the knowledge of the territory and, ultimately, in making better decisions. PAs implementation thus depends on complex trade-offs between many aspects, occasionally in conflict with each other. The aim of this work is not to criticize the current network of PAs, but to stress-test its effectiveness in future conditions and propose adaptive solutions.

5. Conclusions

This work has led to the identification of areas with high ecological values in the transboundary territory of the Grand-Genève. We found that local hotspots and priority areas were partially spatially congruent; they respond to different ecological processes, but they also show a great spatial overlap, mostly located in the subalpine meadows and pastures. This indicates that these habitats are highly important for the conservation of plant diversity despite their lack of protection. Overall, the two indexes calculated (local hotspots of specific richness and priority areas for rare and vulnerable species) have shown an interesting complementarity to identify the most relevant areas for the long-term conservation of plant diversity. These indexes could be used more broadly, including more species and more realms, to help identify the Blue-Green Infrastructure of a territory. PAs showed moderate but constant effectiveness in protecting plant diversity, but tended to lose their ability to cover species distribution in future conditions. On the contrary, the areas identified by the PCN stay relevant in future conditions, supporting the interest and the value of the method developed in this article. We encourage other researchers and conservationists to consider several aspects of biodiversity’s distribution prior to the identification of potentially new areas for conservation, and to test their relevance in future conditions on their own territory.