Global evaluation of current and future threats to drylands and their vertebrate biodiversity

The study investigates how well global drylands—covering roughly 42% of Earth’s land and home to a third of humanity—are protected, what threats their vertebrate biodiversity faces, and how future land-use change under climate and socioeconomic pathways will affect them. Despite their ecological and societal importance, drylands have been underrepresented in global conservation agendas that have focused more on megadiverse tropical regions. Drylands are highly sensitive to anthropogenic pressures such as agricultural expansion, alternative energy development, overgrazing, invasive species, water depletion, and climate change, which can drive land degradation, fragmentation, and biodiversity loss. The research aims to quantify protection shortfalls, identify threat patterns for vertebrates (amphibians, birds, mammals, reptiles) concentrated in drylands, and project the extent of future land conversion under shared socioeconomic pathways to inform conservation targets and priorities.

The paper situates drylands within global conservation discourse, noting historic focus on tropical biomes and gaps in dryland-specific initiatives. Prior work documents extensive dryland degradation, increasing bare ground, vegetation loss, lake and groundwater declines, and vulnerability to invasive species and climate extremes. Studies highlight unique biodiversity in grasslands, savannas, deserts, and Mediterranean systems embedded within drylands, as well as sensitivity of species near physiological heat and water stress limits. The literature also underscores shortcomings in protected-area placement and effectiveness, under-reporting of community and indigenous conserved areas, and uneven enforcement. Recent assessments show reptiles and other groups remain under-assessed, with data deficiency masking true threat status, and large-scale land-use trends driven by agriculture, urbanization, and energy transitions pose escalating risks to biodiversity.

Global drylands were delineated using the UNEP-WCMC Drylands Dataset (2007), based on aridity index P/PET < 0.65, with subtypes: hyperarid (<0.05), arid (0.05–0.20), semiarid (0.20–0.50), and dry subhumid (0.50–0.65). Presumed drylands (P/PET ≥ 0.65) were excluded. Spatial processing used ArcGIS (v10.8.1) with an equal-area Behrmann projection, overlaid with world continents; Oceania and Antarctica were omitted due to low dryland proportion. WWF biomes were intersected with drylands to quantify biome representation. Protected areas were compiled from WDPA (downloaded Nov 2023). China’s 2016 WDPA data were merged with 2023 due to reporting gaps; Turkey and India were supplemented with OpenStreetMap (Nov 2023). Analyses were performed for IUCN categories I–IV (strict protection) and for all categories (I–IV, V–VI, and uncategorized: not applicable, not assigned, not reported). Polygons without boundaries and OECMs were excluded. Data were cleaned using the R package wdpar (v1.3.7). Species data: Extent of occurrence polygons were obtained for amphibians and mammals (IUCN v6.3), birds (BirdLife International v4), and reptiles (GARD v1.7). Dryland species were defined as those with ≥50% of global range in drylands; endemic species as >99% range in drylands. Final datasets included 1,001 amphibians, 2,000 birds, 1,493 mammals, and 3,589 reptiles inhabiting drylands. Range overlap with IUCN I–IV protected areas was computed and binned: 0% (<0.1% overlap), 0–10% (≥0.1%–<10%), 10–100% (≥10%). Threat status (VU, EN, CR) came from IUCN Red List; for 372 unassessed reptiles, modeled threat categories from de Oliveira Caetano et al. (2022) were used. Analyses were also repeated for species with ≥90% dryland range and at continental scales. Representation targets: Species were classified by range size within drylands: narrow (<1,000 km²; target 100% protection), intermediate (1,000–250,000 km²; target 10–100%), widespread (>250,000 km²; target 10–100%). Range protection categories: 0% (<0.1% protected), 0–10% (0.1–9.9%), 10–100% (9.9–99%), and 100% (>99%). Land-use scenarios: Current and future human land-use states in drylands were extracted from LUH2 (v2h; 2015 baseline) and CMIP6 SSP scenarios (SSP1-2.6, SSP2-4.5, SSP5-8.5). Land classes considered: cropland (C3/C4 annual/perennial and C3 N-fixing), managed pasture, rangeland, and urban land. Grids at 0.22180° (~25 km²) were classified by fraction intervals: 0, 0–20, 20–40, 40–60, 60–80, 80–100%. Total fractional land areas were computed as the area of grids with >0% fraction for each land class; additional areas under future scenarios were quantified. Analyses were repeated excluding rangelands (cropland+pasture+urban only). Gains/losses in fractions per grid relative to 2015 were calculated using zonal statistics. The approach treats land classes equally (no global weighting) due to varying regional sensitivities and non-stationarity.

- Protection shortfalls: Only ~12% of total drylands are covered by protected areas (all IUCN categories), compared to ~21% in non-drylands. Less than 5% of each dryland subtype is covered by IUCN I–IV categories. Dry subhumid regions have ~14.4% total protection versus hyperarid regions at ~6.7%. - Continental patterns: Lower IUCN I–IV coverage across all continents; e.g., Australian drylands drop from ~20% to ~6% when considering only IUCN I–IV; North American drylands <7% in IUCN I–IV. - Species protection gaps (≥50% range in drylands): Vast majority of species have <10% of ranges protected (IUCN I–IV). Gap species (0% protected): 30% of amphibians, 7% of birds, 16% of mammals, 27% of reptiles. Among 0%-protected dryland endemics: 224 amphibians, 73 birds, 144 mammals, 720 reptiles. - Reptile richness: Reptiles (n=3,589) are the most diverse dryland vertebrates; 80% have <10% of their dryland distributions protected; 53% are endemic to drylands; 13% are IUCN threatened. Australia is a reptile hotspot (726 species; 58% endemic; 75% with <10% protection; 5% threatened). - Representation targets: Narrow-ranging species are largely inadequately protected. Amphibians: narrow n=331, only 24 adequately protected, 193 with 0% protection. Reptiles: narrow n=629, only 36 adequately protected, 431 with 0% protection. Birds: narrow n=65, only 2 adequately protected. Mammals: narrow n=96, only 3 adequately protected. Intermediate/widespread species adequately protected (meeting targets) remain low: amphibians 18%, birds 18%, mammals 16%, reptiles 16%. - Threats to threatened species in drylands (proportion of species affected): • Amphibians: agriculture 78%, timber/plant harvesting 40%, infrastructure 39%, invasion/disease 36%, pollution 31%, climate change 24%, water management 20%, fire 19%, mining/energy 15%, over-exploitation 8%, human disturbance 2%. • Birds: agriculture 70%, timber/plant 30%, climate change 29%, over-exploitation 26%, invasion/disease 21%, infrastructure 16%, fire 16%, pollution 8%, human disturbance 4%, mining/energy 4%, water management 2%. • Mammals: agriculture 70%, over-exploitation 42%, timber/plant 31%, infrastructure 29%, invasion/disease 21%, climate change 23%, human disturbance 16%, mining/energy 13%, fire 10%, water management 8%, pollution 6%. • Reptiles: agriculture 63%, infrastructure 37%, timber/plant 25%, mining/energy 21%, invasion/disease 20%, over-exploitation 16%, fire 11%, pollution 10%, climate change 9%, human disturbance 6%, water management 6%. - Land-use change projections: By 2100, most natural drylands are projected to undergo some degree of conversion under all SSPs. Additional natural drylands converted (above 2015): SSP1-2.6: 9.6 × 10^6 km²; SSP2-4.5: 12.6 × 10^6 km²; SSP5-8.5: 10.4 × 10^6 km². Under SSP2-4.5, 100% of drylands become affected by some human land use fraction; even under SSP1-2.6, ~95% are affected. Trends show increases in cropland and urban areas with reductions in pasture and rangeland; alternative energy development (solar, bioenergy) is a notable driver for conversion and fragmentation. - Knowledge gaps: 38% of reptile species in drylands are unassessed or data deficient, higher than elsewhere (24%), suggesting threats may be underestimated.

Findings reveal a marked disparity in protection between drylands and non-drylands, with the most arid and biodiversity-unique regions being least protected. Strictly protected areas (IUCN I–IV) are particularly scarce across drylands, undermining the capacity to maintain viable populations, especially for endemic and narrow-ranging species that face multiple, interacting threats. Agriculture emerges as the dominant threat across taxa, but threat profiles differ: water management for amphibians, climate change for birds, over-exploitation for mammals, and infrastructure/mining for reptiles. Projections indicate that anthropogenic land uses will pervade nearly all drylands by 2100 under all scenarios, including optimistic pathways, due to cropland and urban expansion and alternative energy deployment. These results underscore the urgency to expand and strengthen dryland protected-area networks (favoring stricter categories where appropriate), integrate community and indigenous conservation where effective, and proactively plan to mitigate impacts from agriculture, infrastructure, and renewable energy development. Addressing data deficiencies—especially for reptiles—and improving threat assessments are critical to refine priorities and avoid overlooking at-risk species.

Drylands host high levels of unique vertebrate biodiversity yet are grossly under-protected and face accelerating human pressures. The study shows extensive protection gaps (particularly for endemic and narrow-ranging species), dominant and taxon-specific threat profiles, and near-ubiquitous future land conversion under all socioeconomic pathways, including green-growth scenarios. Meeting global biodiversity targets will require prioritizing drylands within conservation and development frameworks, expanding protected-area coverage—especially IUCN I–IV—enhancing effectiveness and enforcement, recognizing and supporting community-based conservation, and planning land uses (agriculture, infrastructure, and renewable energy) to minimize habitat loss and fragmentation. Future work should focus on improving species assessments (notably reptiles), refining land-use projections at finer scales, and developing strategies that balance climate mitigation with biodiversity conservation in drylands.

- Protected area data gaps: WDPA reporting incompleteness in some regions (e.g., China, Turkey, India) required merging older datasets and supplementing with OpenStreetMap; community and indigenous conserved areas may be under-reported; some designated PAs may be ineffective due to poor enforcement. - Species data uncertainties: Many dryland species—especially reptiles—are unassessed or data deficient, likely underestimating true threat levels; use of modeled threat categories introduces uncertainty. - Land-use modeling constraints: LUH2 operates at coarse spatial resolution (~0.25°), limiting local inference; regional differences in historical data introduce variability; alternative higher-resolution datasets have limitations and do not distinguish grazing categories critical to drylands; equal weighting of land-use classes does not account for region- and species-specific sensitivities; rangeland definitions and extent may disproportionately influence patterns.