Logo image
From dry forest to Anthropocene mosaic: Divergent land conversion pathways and erosion dynamics in the Brazilian Caatinga (1985–2023)
Article de revue scientifique   Open Access

From dry forest to Anthropocene mosaic: Divergent land conversion pathways and erosion dynamics in the Brazilian Caatinga (1985–2023)

Lionel Louis Siame, José Carlos de Araújo, Gabriela Goudard, Marie-Pierre Ledru, Eduardo Sávio P.R. Martins, Francisca Soares de Araújo, Icaro Breno da Silva et Frederico de Holanda Bastos
Global and Earth Surface Processes Change, Vol.7
12/2026

Résumé

Brazilian Northeast Semi-arid Land use change Vegetation dynamics Deforestation trajectories Agricultural conversion
The Brazilian Caatinga, the largest seasonally dry tropical forest in the Americas, sustaining 28 million people and harboring exceptional endemic biodiversity, is simultaneously the least legally protected of all Brazilian biomes and a landscape in accelerating Anthropocene transition. Whether the degradation trajectories reshaping it are approaching ecological thresholds beyond which autonomous recovery becomes compromised remains a critical open question for conservation science and Earth surface process dynamics. We address this question through an integrated, spatially explicit analysis of nearly four decades of landscape transformation (1985–2023) across the Central Ceará region (39,000 km2), combining satellite-derived land-cover trajectories, independent deforestation and fire records, natural erosion factors, and RUSLE-based soil loss estimates within a common hexagonal analytical framework. These absolute soil loss estimates are deliberately conservative: realistic crop cover raises them by a factor of 2.4, and the combined cover and rainfall-erosivity biases could approach4.More than half the study area underwent net agricultural conversion over the study period, with the pace of change apparently accelerating after 2013, a pattern consistent with genuine intensification of land conversion but potentially amplified by the improved radiometric sensitivity of Landsat 8. Our landscape typology identifies three mechanistically distinct disturbance regimes across contrasting geomorphic settings: direct mechanized clearing in the accessible crystalline lowlands, fire-driven conversion along piedmont slopes, and recurrent fire without permanent conversion on the elevated Araripe Plateau. The spatial segregation of these regimes follows topoclimatic gradients rather than rainfall alone, highlighting the dominant role of anthropogenic disturbance in shaping contemporary landscape trajectories. Interpreted through the lens of alternative stable states theory, the most degraded landscape cluster exhibits hallmarks of a system that has transitioned beyond its local resilience boundary, while the Araripe Plateau constitutes an empirical demonstration of fire-resilient edaphic buffering against irreversible conversion. Conversely, natural vegetation recovery signals, detectable across about 3% of the study area, are consistent with a measurable reversal of this trend, though phenological variability in satellite-derived land-cover products precludes firm attribution to structural regeneration. Where genuine, these recovery dynamics provide a quantitative basis for targeting payment-for-ecosystem-services mechanisms in a biome where the window for effective conservation action is narrowing.

Fichiers et liens (2)

url
Find in HALAfficher
url
https://doi.org/10.1016/j.gespch.2026.100024Afficher
Publié (version de la notice) Ouvrir

Indicateurs

1 Consultations de la notice

Détails

Logo image