RESUMO
Forests provides major ecosystem services worldwide. The Brazilian Atlantic Forest (BAF) has been dramatically devastated, with fragmentation processes jeopardizing its long-term sustainability. This study investigated the structure and successional dynamics in BAF natural regeneration along an anthroposequence characterized by increasing human disturbance histories as: secondary (SF) > disturbed (DF) > late forest (LF). We aimed to understand how and the degree to which BAF fragmentation and human disturbance affected plants, soils, and the whole soil-plant relationships and feedbacks. We investigated the natural regeneration conditions of plants (using plant classification and quali-quantitative analyses) and soil chemistry (including pH-CaCl2, H + Al, C, N, Pt, cation-exchange capacity (CEC), exchangeable cations, Al, B, Cu, Fe, K, Mn, and Zn) at twelve permanent, 2000 m2 plots, distributed across LF, DF, and SF forests. Significant differences were determined by ANOVA. Correlation matrix (CM) and factor analysis (FA) were used for understanding correlations and feedbacks/variability among investigated parameters, respectively. Most of investigated plant and soil parameters showed significant differences (p < 0.05) between more developed plant formations (LF) vs less developed ones (SF), with differences mainly due to soil's development stage. All investigated forest formations are featured by a great influence of the soil-plant relationships and feedbacks, with a decreasing magnitude as LF â DF â SF. Thus, there is a direct, statistically recognizable impact of both "recent" as well as "ancient" human disturbance on investigated soil-plant formations. The anthropogenic influence clearly affected not only plant and soil as "separate" systems but the whole complex of interactions and feedbacks among ecosystem components. A decreasing quality in soil and plant parameters was observed as human disturbance increased. We demonstrated that BAF plant and soil require decades for their recovery after human disturbances, with complex mechanisms and behaviors in the relationships among ecosystem components. The results can be useful for managing future recovery in an ecosystem of worldwide strategic importance.