Hainan Island (108°37′–111°03′ E, 18°10′–20°10′ N) is located in the southern part of China and harbors the most concentrated and well-preserved continental island tropical rainforest in the country ( Figure 1 ). This rainforest ecosystem, which is located in the central mountainous region of the island, is a harbor for diverse species and is designated biodiversity conservation area, recognized as one of the nine hot spots for biodiversity conservation in China (Wang et al., 2021). Furthermore, these central mountains serve as a vital source of ES for local communities, providing ES like forestry products, water provision, and flood mitigation (Zheng et al., 2019). A significant portion (8.82%) of Hainan Island has already been designated as natural reserves in recognition of its ecological treasures. Such nature reserves aim to protect the representative rainforest and its diverse inhabitants. However, this existing nature reserve network, called exsiting PAs in this paper, falls short of achieving comprehensive conservation goals for both biodiversity and the vital ES it provides (Wang et al., 2021). Therefore, Hainan Island presents a compelling case study area for comparing the locking and unlocking methods for PAs expansion. These contrasting approaches can help identify the most effective strategy for expanding PAs and provide valuable insights for other areas facing similar challenges.

We employed a biodiversity importance index, which is the sum of potential habitat suitability for all species across five taxonomic groups: plants, mammals, birds, reptiles, and amphibians, to quantify biodiversity (Wang et al., 2021; Xu et al., 2017). Habitat suitability models were developed by correlating species distributions with environmental factors, such as elevation, slope, and ecosystem types (Zhang et al., 2011). Refer to Wang et al. (2021) for more details.

We focused on five ES (e.g., water yield, soil retention, water quality, flood mitigation, and carbon sequestration) to evaluate the ES provided by PAs in this region. These ESs are crucial for the local community, downstream populations, and global efforts to address environmental challenges. We utilized the Integrated Valuation of Ecosystem Services and Trade-offs (InVEST) model to generate spatial distribution maps for each ES (Sharp et al., 2019). A detailed methodology and data description can be found in Wang et al. (2021).

We used a marine reserve design, applying the spatial explicit annealing algorithm in Marxan tool to identify priority areas for expanded PAs. This tool is a widely used decision-support tool in systematic conservation planning (Watts et al., 2017; Zhang and Li, 2022). Marxan prioritizes PAs for biodiversity and ES by minimizing a composite score that balances three factors (Equation 1) (Watts et al., 2017). 1) Cost: represents the cost of protecting each planning unit. 2) Boundary length: penalizes long and complex boundaries (measured by the boundary length modifier, BLM) to create compact, manageable PAs. 3) Species penalty: penalizes solutions that inadequately protect target species and habitats (assessed by the Species Penalty Factor, SPF). A high SPF indicates a greater shortfall of that conservation feature in protection.

This study identified one biodiversity element and five ESs as protection targets to safeguard 40% of the biodiversity and each ES throughout Hainan Island. These targets were established based on the research specificity of Hainan Island. Watersheds across Hainan Island were used as planning units, with their area determining the cost of protection. Adjacent planning unit lengths were incorporated as boundary data to enhance calculation accuracy. All protection targets were assigned the same weights during the Marxan model execution to ensure equal consideration. An SPF of 1 was set to meet all protection goals. We employed an iterative approach to determine the BLM parameter. We determined an appropriate BLM value of 0.72 by gradually increasing the BLM value and evaluating the resulting spatial compaction of multiple scenarios (Zhang et al., 2014). We used Marxan to run two distinct scenarios and investigated various strategies for expanding PAs: (1) locking existing PAs during expansion. In this scenario, we prioritized extending PAs beyond the current network while keeping the boundaries of existing reserves intact. We designated the planning units as priority conservation areas, where over 5% of their total area overlapped with existing PAs (Yang et al., 2019). This procedure was made by “locking in” these units, thus ensuring their inclusion in conservation planning. (2) Unlocking existing PAs: this approach allowed Marxan to consider the entire landscape, without being limited by existing reserve boundaries, for potential expansion opportunities.

We ran Marxan simulations 1,000 times to calculate the irreplaceability index for each planning unit. This index reflects the number of times a unit is selected within the optimal solutions generated by Marxan, indicating its relative importance for achieving conservation goals. The higher the irreplaceability index, the greater the priority. We identified the 15% high-ranking areas of Hainan Island as priority expansion areas for protected zones, effectively doubling the size of the existing PAs. This selection aligns with the area designated by the national park pilot initiative, ensuring consistency with broader conservation efforts (Li et al., 2022).

Independent samples t-tests were conducted to assess the differential impacts of locking and unlocking existing PAs on biodiversity and ESs. The protection target achievement rate was calculated as the percentage of protection objectives met within PAs relative to the desired targets. For instance, if our goal for PAs on Hainan Island was to protect 40% of biodiversity, and under a locking strategy, the selected PAs could only protect 30%. The protection target achievement rate would be 30% divided by 40%, equaling 75%. PAs fragmentation was simply represented by the number of patches (Santiago-Ramos and Feria-Toribio, 2021).

Existing PAs in Hainan Island, covering only 8.82% of the total area of the island, are insufficient for comprehensive protection of biodiversity and ES. Although the existing PAs provide partial protection for some ES—including water quality (98.50%), and soil retention (81.17%)—significant gaps exist in safeguarding biodiversity (55.67%), carbon sequestration (55.78%), water yield (35.66%), and flood mitigation (47.42%) ( Figure 2 ). This limited coverage necessitates strategic expansion of PAs to ensure the robust conservation of biodiversity and the vital ES they support.

According to the results of the independent-samples t-test ( Figure 3 ), the locking and unlocking of existing PAs strategies significantly influenced the protection of biodiversity, soil retention, water quality, and flood mitigation services (p<0.05). However, no significant influence on water yield or carbon sequestration was observed (p>0.05).

Comparing the locking and unlocking methods for expanding the PAs of Hainan to 15% of the island revealed distinct outcomes for biodiversity and ES protection outcomes. The unlocking approach, which considered the entire island for potential expansion, was more effective in protecting species habitats than the locking method (86.84% vs. 66.49%, respectively). However, it showed slightly lower performance in safeguarding the overall ES. Interestingly, both methods achieved near-complete protection for soil retention (locking: 112.82%, unlocking: 103.81%) and water quality (locking: 120.01%, unlocking: 115.52%) ( Figure 4 ). This result highlights the complex trade-off between biodiversity and ES goals when selecting a PAs expansion strategy with limited resources.

The locking method, focusing on expanding existing PAs, resulted in a more concentrated expansion (41.19% increase in adjacent land) and few protected patches (12). This approach may offer increased efficiency by building upon existing conservation efforts and potentially minimizing habitat fragmentation. Conversely, the unlocking method yielded a broader expansion (56.32% increase outside existing areas) but with a significantly high number of protected patches (102) ( Figure 5 ), raising concerns about potential fragmentation within these new areas.

Our study examined the effectiveness of different strategies for expanding PAs to enhance biodiversity and ES conservation. By comparing the “locking” and “unlocking” approaches on Hainan Island, we identified critical trade-offs and implications for PAs management. Our protection-efficiency analysis revealed that existing PAs on Hainan Island demonstrate a insufficient protection for both biodiversity and ES, a common trend in many other locations (Neugarten et al., 2020; Watson et al., 2020). This result may be attributed to two reasons. First, the initial aim of the existing nature reserve was to protect specific species (e.g., Hainan black crested gibbon) or rainforest ecosystems, which did not include ES (e.g., water yield) and some other species as protection objectives at that time. Second, there exists a spatial mismatch among different protection objectives, which is influenced by biophysical conditions (e.g., soil and climate), a common phenomenon in the spatial protection of ES and biodiversity (Wang et al., 2021; Zheng et al., 2016). These findings highlight the urgent need for strategic PAs expansion to address these gaps.

By comparing two PAs expansion strategies, we found that the locking approach, which focuses on expanding adjacent to existing nature reserves, demonstrated a preference for ES protection rather than biodiversity. In contrast, the unlocking approach, which considers the entire landscape, achieved broader biodiversity coverage but at the cost of increased fragmentation and potential management challenges. These differences may be attributed to several factors. The spatial inconsistency between ES and biodiversity on Hainan Island (Wang et al., 2021), coupled with the greater spatial flexibility afforded by the unlocking strategy, likely played a significant role. Additionally, small PAs, which are more common in the unlocking approach, may be more effective at protecting biodiversity than ES (Volenec and Dobson, 2020). Consequently, these findings emphasize the complexity of balancing biodiversity and ES conservation in PA expansion (Fastré et al., 2020).

Although the locking and unlocking methods employed in this study offer valuable insights into potential strategies for expanding existing PAs, it is important to acknowledge that they represent only one perspective. Real-world PA expansion is influenced by a complex interplay of factors beyond biodiversity and ES (Yang et al., 2020). Challenges such as conflicting economic development interests, cultural factors, and the associated costs and benefits must be carefully considered (Hoffmann, 2022; Waldron et al., 2022). Effective PA expansion strategies should be tailored to specific site conditions and involve active engagement with local communities. Moreover, successful PA expansion requires robust management practices, particularly legislation. Implementing clear and enforceable regulations is essential to ensure the long-term sustainability of expanded PAs and their contribution to biodiversity conservation (Zabala et al., 2024).

Decision-makers must carefully weigh the costs and benefits of different expansion strategies based on specific conservation goals and available resources. We recommend developing robust decision-support tools to quantify trade-offs and identify optimal expansion scenarios to address these challenges. Additionally, exploring innovative financing mechanisms to support PA expansion and management is crucial (Huang et al., 2024; Zafra-Calvo and Geldmann, 2020). This scenario includes diversifying revenue sources beyond government funding, such as payments for ES and private conservation investments. Moreover, integrating PAs expansion with broader landscape-scale conservation initiatives can enhance overall conservation effectiveness. Finally, collaborative governance approaches involving local communities, Indigenous peoples, and other stakeholders can facilitate the successful implementation of PAs expansion plans (Poppenborg and Koellner, 2013).

Although this study provides valuable insights into the trade-offs between biodiversity and ES conservation in PAs expansion, it is essential to acknowledge its limitations. First, the expansion of PAs did not consider local livelihood and other human impacts or costs, which substantially influence PAs expansion (Geldmann, 2023). Second, the study primarily considered biodiversity and ES as static variables, neglecting potential changes in their distribution due to future climate change or other environmental factors and ES flow (Lu et al., 2022; Wang et al., 2022). Third, the lack of cost-benefits analysis of the two methods may weaken the robustness of the findings.

Future research should explore the dynamics of biodiversity and ES over time, incorporating climate change projections and other relevant drivers. Furthermore, conducting comparative studies across different regions with varying ecological characteristics and governance systems would enhance the understanding of the generalizability of our findings. Developing more sophisticated spatial modeling tools that can incorporate multiple, often conflicting, conservation objectives is crucial to refining the decision-making process. In conclusion, our study provides a possible expansion of PAs, which is important for realizing the ambition of Kunming–Montreal biodiversity initiative.