Animal invaders threaten protected areas worldwide

The study addresses how resistant protected areas (PAs) are to alien (non-native) animal species incursions globally, and which factors drive established alien species richness within PAs. Although terrestrial PAs cover about 15% of Earth’s land and are central to conserving biodiversity, human-driven environmental change and globalization have increased rates of alien species introductions and establishment. Prior work suggests human pressure is higher in less strictly protected PAs and that invasive species threaten conservation outcomes, but comprehensive, global, multi-taxon analyses across the full range of IUCN PA categories have been lacking. The authors hypothesize that alien establishment should be higher in less strictly protected PAs, and that PA characteristics (area, age/designation year), human footprint, native biodiversity, biomes, and insularity influence alien richness. The purpose is to quantify current alien animal establishments in and around nearly 200,000 PAs worldwide, identify abiotic, biotic, and anthropogenic correlates, and project potential richness based on habitat suitability to assess future invasion risks. The work is important for informing global biosecurity and PA management under accelerating global change.

The paper synthesizes evidence that alien species are among the top five direct drivers of biodiversity loss and are linked to declines in threatened species and historical extinctions. Globalization and increased human mobility and trade elevate propagule pressure and spread, potentially affecting PAs. Prior regional/continental studies (e.g., European Natura 2000) found more invasives in less strictly protected PAs; SCOPE case studies documented variable management effectiveness against invasions; and island systems are recognized hotspots for alien species impacts. The IUCN PA categorization sets varying management objectives but all aim to minimize invasion risks. The literature also debates biotic resistance (native richness reducing invasions) and highlights taxon- and region-specific differences in establishment success. Despite these insights, a comprehensive, multi-taxon, global assessment of alien animal establishment in PAs, including the roles of human footprint, PA attributes, and native biodiversity, has been missing.

Study scope and PA data: The authors analyzed 199,957 terrestrial PAs worldwide using the April 2019 World Database on Protected Areas (WDPA). They included sites with status designated/inscribed/established (excluding UNESCO MAB reserves), dissolved overlapping polygons per WDPA best practice (assigning the stricter IUCN category), and simplified geometries (1 km tolerance) to aid computation. PAs were classified as mainland or insular using a new 30 m global shoreline and islands database. IUCN categories I–VI and unassigned/other were retained, and 14 global biomes (aggregated from 867 terrestrial ecoregions) were used to stratify comparisons among categories.

Alien species and occurrences: A global database of 894 established terrestrial alien animal species from 11 taxa was compiled: vertebrates (98 amphibians, 178 reptiles, 391 birds, 150 mammals) and invertebrates (77 species across Arachnida, Branchiopoda, Chilopoda, Diplopoda, Malacostraca, Gastropoda, and Insecta). Sources included IUCN, BirdLife/NatureServe, GAVIA, Kraus’ compendium, Long (2001), Capellini et al., and the Global Invasive Species Database (GISD) for invertebrates. Species with imprecise distributions or non-breeding populations were excluded. Presence within PAs was determined from precise occurrence coordinates for all taxa except birds, for which established-range polygons (GAVIA) were used; cross-validation showed high consistency between point and polygon-derived bird richness (r=0.95; regression R=0.94). Alien richness per PA was calculated as the number of established alien species present.

Proximity and flows: To gauge incursion pressure on uninvaded PAs, established alien populations within 10 km and 100 km of PA boundaries were identified using ArcGIS Near analyses. Geographic flows between native and alien continents for species established in PAs were quantified via network (chord) diagrams using the R circlize package.

Predictors and statistical analysis: Four predictors were evaluated for associations with alien richness in PAs: PA surface area, designation year (age), human footprint index (HFI; 1 km grid aggregated to analysis resolution), and native vertebrate species richness (amphibians, reptiles, birds, mammals; from Biodiversity Mapping and global reptile datasets). Invertebrate native richness was not available. Analyses were conducted at 5 arc-min resolution (~10×10 km), with sensitivity checks at 2.5 and 10 arc-min. Because most PAs had zero alien species, zero-inflated negative binomial mixed-effects models (ZINB) were fitted using glmmTMB, with region/country identity as a random effect to account for spatial pseudo-replication and variable sampling effort.

Habitat suitability modeling: Species distribution models (SDMs) were built for all 894 species using Biomod2 ensemble methods (BRT, GAM, MARS, CTA, RF), trained on combined native and invaded occurrences to capture realized niche shifts. Predictors included climate (taxon-specific bioclimatic variables from WorldClim), vegetation (annual NDVI, 2001–2005), and water availability (global lakes and wetlands, excluding saline lakes). Datasets were projected to a Behrmann equal-area grid (~10×10 arc-min). A target-group background approach used GBIF occurrence pools per taxon to mitigate sampling bias; background points: 30k (amphibians/reptiles), 70k (mammals), 100k (birds), 80k (invertebrates); prevalence set to 0.5. Models with AUC<0.8 or TSS<0.6 were excluded; ensemble predictions were TSS-weighted. Continuous outputs were thresholded using TSS-maximizing thresholds to obtain binary predictions, summed to estimate potential alien richness per grid and intersected with PA polygons. Projections were restricted to analogous climates to reduce extrapolation uncertainty.

Comparisons across IUCN categories and biomes examined whether categories reflect alien incursion states. Insular vs mainland PAs were contrasted in alien richness patterns.

• Extent of establishment: 58% (520/894) of alien animal species have established in at least one PA, but only 9.1% (18,110 of 199,957) of PAs host any of these aliens.
• Environmental suitability vs observed richness: >95% of PAs are environmentally suitable for establishment by at least some alien species. Predicted alien richness per PA based on SDMs (mean 53.34 ± 0.070 species) is 194 times higher (95% CI: 189–199) than observed richness (Wilcoxon signed-rank test, P<0.0001), indicating high latent invasion risk.
• Proximity risk: 89.4% of PAs have established alien populations within 10 km of their boundaries (encompassing 58% of the alien species), and 99.0% have established aliens within 100 km (79.6% of species).
• Taxonomic patterns in invaded PAs: The largest proportion of invaded PAs hosts alien birds (4.7% of PAs; 252 species), followed by mammals (3.7%; 91 species), invertebrates (2.2%; 63 species), amphibians (0.5%; 48 species), and reptiles (0.4%; 66 species). Common aliens in PAs include birds (Columba livia; Phasianus colchicus; Passer domesticus), mammals (Oryctolagus cuniculus; Neovison vison; Mus musculus), reptiles (Hemidactylus turcicus; Trachemys scripta), amphibians (Rhinella marina; Lithobates catesbeianus), and invertebrates (Harmonia axyridis; Apis mellifera scutellata; Aedes aegypti).
• Spatial patterns: Alien richness hotspots in PAs are concentrated on islands (mean insular PA alien richness 0.407 ± 0.003 vs mainland 0.252 ± 0.003), with Hawaiian PAs (e.g., Volcanoes National Park, 80 species) among the highest. High-richness regions include the tropical/subtropical Pacific and Caribbean islands, New Zealand, Florida flooded grasslands/savannas, and temperate regions in western Europe, Australasia, and the US west.
• Donor/recipient flows: Flows are dominated by intra-continental exchanges and movements between South/East Asia and Europe; South/East Asia and Africa are major donor regions, Europe and North America major recipients.
• IUCN categories: Contrary to expectations, IUCN Category II (National Parks) show the highest average alien richness across many biomes, whereas less stringent categories show moderate richness except in certain human-pressured biomes (e.g., mangroves).
• Drivers (ZINB mixed models): Larger PA area, higher human footprint index, and more recent designation year are significant positive correlates of alien richness across taxa. Relationships between native and alien richness vary: negative for reptiles and birds; positive for mammals and amphibians. Results are robust across spatial resolutions tested (2.5, 5, 10 arc-min).

The findings indicate that while most PAs currently lack established terrestrial alien animal populations from the studied set (over 90% uninvaded), they are broadly suitable for alien establishment and are surrounded by numerous nearby established populations, implying substantial future incursion risk. This helps address the core question of PA resistance to alien invasions: PAs have so far limited incursions, potentially due to lower human activities in many PAs and management actions to prevent and eradicate invaders, but their vulnerability is high given environmental suitability and proximity pressures.

Older PAs harbor fewer alien species, consistent with earlier designation facilitating prevention and with older PAs often being more remote. Human footprint is a strong predictor—especially for taxa primarily transported by people (herptiles, mammals, invertebrates)—highlighting anthropogenic pathways as key drivers. For birds, PA area is the dominant predictor, consistent with high natural dispersal and challenges in excluding volant species.

The expected protective effect of native biodiversity (biotic resistance) is not consistent: negative associations for reptiles and birds support some evidence of resistance, while mammals and amphibians show positive associations, reflecting the debated invasion paradox.

Islands, typically species-poor and highly impacted by introductions, show greater alien richness in PAs, aligning with global patterns of invasion-driven extinctions and risk concentration. The paradoxical higher richness in Category II National Parks suggests that current IUCN management categories do not consistently reflect invasion status; large park sizes and tourism not fully captured by the human footprint metric may contribute.

Given PAs’ role as refuges under global change, integrating alien species risk assessment and management into PA planning is urgent. The results underscore the need for proactive biosecurity, surveillance, and control in and around PAs, particularly where human access is high and along dispersal corridors (rivers, roads).

This study delivers a comprehensive, global, multi-taxon assessment of terrestrial alien animal establishments in protected areas. It shows that PAs have thus far limited alien animal incursions, yet the vast majority are environmentally suitable and adjacent to established alien populations, indicating high future invasion risk. Alien richness is higher in larger, more recently designated PAs with greater human footprint, and unexpectedly highest in IUCN Category II National Parks across many biomes. Islands and certain regions (e.g., Hawaii) are hotspots. The work highlights inconsistencies between PA categories and invasion status and variable roles of native biodiversity among taxa.

Recommendations include: strengthening biosecurity at PA access points; routine monitoring of visitors and vehicles; early detection and rapid response within PAs and surrounding areas; targeted control of nearby alien populations; and coordinated action among agencies, scientists, and stakeholders. Future research should incorporate aquatic animals and plants, improve global data coverage (especially in under-sampled regions), develop higher-resolution predictive models, and include metrics capturing tourism and other human activities not reflected in current footprint indices.

• Spatial resolution: SDMs and some analyses were conducted at relatively coarse resolutions (e.g., 10 arc-min), potentially missing fine-scale habitat heterogeneity; finer-scale modeling is desirable in future work.
• Taxonomic scope: Focused on terrestrial animals; excluding aquatic animals and plants likely underestimates the true extent of alien incursions in PAs.
• Data availability and bias: Geographical and taxonomic sampling biases exist (more data from Europe, USA, South Africa, Australia; fewer from many developing regions). Invertebrate native ranges are often poorly resolved, limiting inclusion.
• Proxy limitations: Human footprint index may not fully capture relevant activities (e.g., tourism) that influence invasion risk; IUCN categories may not reflect on-the-ground management effectiveness.
• Native diversity effects: Inconsistent relationships across taxa suggest unresolved mechanisms (biotic resistance vs. facilitation) and potential confounding factors.