This is a Preprint and has not been peer reviewed. The published version of this Preprint is available: https://doi.org/10.1016/j.jnc.2026.127380. This is version 1 of this Preprint.
The cost of tree cover loss: quantifying global extinction rates per million hectares
Downloads
Supplementary Files
Authors
Abstract
Aim:
To quantify the relationship between annual tree cover loss and changes in extinction risk across terrestrial vertebrates, expressing responses as both absolute numbers of species associated with each IUCN threat category per million hectares of tree cover loss and relative odds of assignment to higher-risk categories.
Location: Global.
Time period: Contemporary period.
Major taxa: Mammals, birds, reptiles, and amphibians.
Methods:
An ensemble of machine learning and statistical models was used to estimate associations between annual tree cover loss and IUCN threat categories—Critically Endangered (CR), Endangered (EN), and Vulnerable (VU). Responses were quantified as absolute estimates per million hectares of annual tree cover loss and as relative associations using odds ratios (ORs). Analyses further examined variation among biomes and major drivers of tree cover loss, including agricultural expansion, settlement expansion, timber extraction, and wildfire.
Key results:
Marked differences emerged among vertebrate groups. For each million hectares of annual tree cover loss, reptiles showed the highest estimated absolute numbers of species across threat categories, with 19.9 CR, 34.7 EN, and 20.6 VU species, followed by amphibians (15.3 CR, 26.3 EN, and 7.3 VU). Estimates were lower for mammals (2.4 CR, 5.4 EN, and 4.2 VU) and birds (2.8 CR, 5.1 EN, and 6.2 VU). Relative associations showed a contrasting pattern: mammals had the strongest increase in the odds of assignment to a higher-risk IUCN category (OR = 6.86), followed by birds (OR = 2.83), whereas reptiles showed little evidence of an overall association (OR ≈ 1). Amphibians showed a stronger positive association in subtropical biomes (OR = 3.08), while associations were below 1 in other biome contexts. Associations also differed among major drivers: mammalian threat status was most strongly associated with settlement expansion and shifting agriculture; reptile and amphibian responses were more strongly associated with wildfire; and bird responses were strongest in relation to logging and agricultural expansion.
Main conclusions:
The relationship between tree cover loss and extinction risk varies substantially among vertebrate groups in both magnitude and associated drivers. Absolute estimates indicate a greater burden among reptiles and amphibians, whereas mammals show the strongest relative association with higher-risk IUCN categories. Expressing biodiversity responses per unit of tree cover loss provides a common metric for comparing vulnerability among taxa and can support taxon- and context-specific prioritization of conservation interventions.
DOI
https://doi.org/10.31223/X5GJ55
Subjects
Life Sciences
Keywords
biodiversity crisis, defaunation, land use change, machine learning models, threatened vertebrates, tree cover loss.
Dates
Published: 2026-09-03 06:02
Last Updated: 2026-09-03 06:02
License
CC-BY Attribution-NonCommercial 4.0 International
Additional Metadata
Conflict of interest statement:
The author declares no conflict of interest.
Data Availability:
The data are available as supplementary material in the published article
Metrics
Views: 32
Downloads: 4
There are no comments or no comments have been made public for this article.