Multidimensional Recovery of Young Secondary Forests in Human-Modified Tropical Landscapes

Abstract

Secondary succession is a widespread phenomenon in the Anthropocene due to global land-use and climate change. Our ability to predict successional trajectories remains limited due to key knowledge gaps related to early secondary succession and how successional trajectories vary across socio-ecological systems and multiple forest attributes. Therefore, we analyzed the first 5 years of secondary forest succession across six tropical landscapes (i.e., socio-ecological systems) in three countries (Australia, Ghana, and Mexico) that differ in land-use intensity and two main forest types (dry and wet). We established 122 permanent plots in recently abandoned agricultural fields, monitored them annually for up to 5 years, and quantified 12 forest attributes related to structure, diversity, functional composition, and biotic interactions. We found that a large variation in successional trajectories was explained by the six landscapes (average r2 across 12 attributes is 54%; range: 18%–78%), indicating that succession is the result of a socio-ecological system. An additional 39% of the variation (range: 19%–70%) was explained by plots occuring within landscapes, which reflects variation in landscape context and local land use intensity. Countries had a stronger impact on succession than forest type, indicating that the social component is more important early in succession, whereas the ecological component may become more important later in succession. Countries with lower land use intensity (e.g., subsistence agriculture, shorter duration of use, no mechanization) showed a higher start and speed of succession, as vegetation legacies can kickstart succession. Forest attributes followed distinct successional trajectories: forest structure and diversity increased over time, reflecting a deterministic component of succession, whereas functional composition and biotic interactions varied more with forest type, reflecting environmental filtering. These findings highlight the importance of integrating early succession, socio-ecological systems, and multiple dimensions of forest recovery to better understand and predict forest succession in human-modified tropical landscapes.