Protected areas (PAs) are crucial for mitigating global change impacts on biodiversity and human well-being. Despite challenges like weak regulations and conflicts, PA coverage has significantly increased, driven by the Kunming-Montreal Global Biodiversity Framework's 30% target by 2030. However, this expansion isn't uniform; PAs are often biased towards higher-income countries and less valuable lands. While the benefits of PA networks are well-studied, the socioeconomic and environmental factors enabling PA establishment are less understood. This research aims to quantitatively assess the global niche of terrestrial and marine PAs using a multidimensional social-environmental framework, predicting PA occurrence and identifying enabling conditions. This will help pinpoint areas with 'potential' versus 'unrealistic' conservation gains, guiding strategic interventions for effective long-term conservation outcomes.

Previous research highlights the heterogeneous distribution of PAs globally, influenced by factors like geopolitical conflicts, human development, land use, cultural contexts, and elevation. However, other potential factors like resource dependency and NGO presence remain understudied. Studies have shown biases towards higher-income countries and remote, less economically valuable areas. While studies evaluate the potential benefits of PAs, the combined socioeconomic and environmental conditions enabling establishment require further investigation to inform strategies for achieving the 30% coverage target.

A global 10x10 km grid was used, with cells overlapping PAs (IUCN categories I-VI) and an equal number of randomly selected unprotected cells. Fourteen socioeconomic and environmental factors were extracted for each cell, including GDP, accessibility, human footprint, HDI, conflict levels, NGO numbers, resource dependency, temperature, primary productivity, elevation/depth, distance to coast/seamounts, and salinity/precipitation. Missing data were imputed using a machine learning algorithm. Ecological Niche Factor Analysis (ENFA) visualized PA distribution in the multidimensional space. Random Forest (RF) models predicted the relative likelihood of PA occurrence using the 14 factors, employing 10-fold spatial cross-validation. Model performance was assessed using accuracy and true skill statistic (TSS). Factor importance was determined using permutation-based variable importance. Unprotected areas were then classified as 'potential' or 'unrealistic' conservation gains based on vertebrate diversity and the model's predictions. Zonation software prioritized areas for maximizing species range size coverage.

PA distribution is highly heterogeneous but predictable from social-environmental factors. PAs, particularly restrictive ones (IUCN I), over-aggregate where HDI and NGO numbers are high, but conflicts, human footprint, and resource dependency are low. The global PA network fails to cover all socioeconomic and environmental conditions. RF models accurately predicted PA occurrence, with socioeconomic factors being more important than environmental ones, especially at the national level. HDI and NGO numbers positively correlate with PA likelihood, while conflicts and resource dependency negatively correlate. Accessibility is a key factor for marine PAs. Potential high conservation gains are clustered, mainly in regions with favorable socioeconomic conditions and high vertebrate diversity but largely absent in Africa, Europe, and South Asia. Unrealistic high conservation gains are widespread in areas with unfavorable conditions despite high biodiversity. Mean probabilities of protecting unprotected areas vary significantly among countries, with low probabilities for most countries, particularly for marine areas.

The study's findings highlight the strong influence of socioeconomic factors on PA establishment. The positive correlation with HDI and NGOs suggests potential levers for increasing PA coverage, but the non-linear relationship with HDI reveals that simple development isn't a sufficient solution. The negative correlation with conflicts and resource dependency identifies key barriers. Spatial heterogeneity within countries underscores the importance of local factors, suggesting the need for tailored interventions. The identification of 'potential' vs. 'unrealistic' gains helps prioritize conservation efforts, guiding the allocation of resources and highlighting the need for alternative strategies like OECMs and PPAs in areas with unfavorable conditions. The study emphasizes that achieving the 30% PA coverage target requires addressing socioeconomic barriers and utilizing diverse conservation approaches.

This research provides a comprehensive assessment of the socioeconomic and environmental factors influencing PA establishment globally. The findings highlight the critical role of socioeconomic conditions and the need for targeted strategies beyond simply expanding PA coverage. Future research should explore the effectiveness of alternative conservation measures, investigate the dynamics of PA establishment over time, and develop more inclusive and equitable conservation strategies.

The study relies on presence-background data and existing PA designations, potentially omitting unreported PAs. The analysis uses spatial cross-validation but may not fully account for all forms of spatial autocorrelation. The focus on vertebrate diversity might not capture the full extent of biodiversity. The socioeconomic and environmental factors were used as proxies and may not fully capture the complexity of causal relationships.