Mapping deforestation frontiers in Central America’s tropical moist forests

Deforestation assessments often treat tropical moist forest loss as a uniform process, obscuring the frontier dynamics that shape where and how forests are lost and the spatial correlates associated with them. Here, we show that Central America’s tropical moist forest frontiers form a heterogeneous regional mosaic by combining annual 30-m forest-change data from 1990 to 2023 with a frontier-metrics framework aggregated to 1.5-km cells. We derive three frontier typologies and five archetypes from six metrics and use a Bayesian multinomial logit model to quantify the covariate associations for each archetype. Region-wide, the Other accounts for the largest share of mapped area (34.7%), followed by Consolidated (16.1%), Fragmented (15.1%), Critical (13.3%), Developing (11.5%), and Dormant (9.3%). Critical, Developing and Dormant frontiers occur in more remote, undisturbed forest contexts than Consolidated and Fragmented frontiers. Protected-area associations differ by archetype, with critical frontiers less associated with protected areas and Dormant frontiers more associated with them. These findings support regional archetype-specific conservation targeting strategies.