Forest Fragment Proximity as an Important Predictor of Restoration Success in Mining‐Degraded Areas
Carlos Vinicius Gonçalves Ribeiro, Edivaldo Barbosa de Almeida Junior, Artur Ferro de Souza, Cléber Rodrigo de Souza, Denise Moura Madeira, Mariana Miranda Andrade, Ana Clara Gerônimo Silva, Guilherme Ribeiro Aguiar, Pedro Machado Neto, Lucas Amaral de Melo, Soraya Alvarenga BotelhoABSTRACT
Ecological restoration is essential for achieving global climate and biodiversity targets, particularly in tropical landscapes affected by large‐scale disturbances. Following the 2015 Fundão tailings dam collapse in Mariana, Brazil, extensive areas of the Atlantic Forest were buried by mining waste, creating barriers to forest recovery. This study evaluated how vegetation cover, soil surface conditions, landscape context, and restoration strategies were associated with natural regeneration and soil seed bank composition in tailings‐affected environments. Ten sites were sampled along the Doce River Basin, including areas under active restoration, passive restoration, and reference ecosystems, 6 years after disturbance. Natural regeneration and the soil seed bank were evaluated through field sampling and data collection. Generalized linear mixed models (GLMMs), non‐metric multidimensional scaling (NMDS), and permutational multivariate analysis of variance (PERMANOVA) were used to analyze ecological patterns and environmental variables. Natural regeneration in degraded areas was dominated by pioneer and shrub species, whereas passively restored areas showed regeneration patterns more similar to reference ecosystems, with higher proportions of tree and zoochorous species than actively restored areas. In contrast, actively restored areas were characterized by a greater dominance of anemochorous species. Distance to forest fragments was associated with regeneration and seed bank attributes, with greater richness and diversity observed in sites closer to forest remnants. Canopy cover was positively associated with regeneration density and diversity, whereas grass biomass was negatively associated with regenerant growth. Overall, the results suggest that landscape connectivity and local environmental filters influence forest recovery trajectories after mining disasters, while passive restoration may favor natural regeneration under suitable landscape conditions.