The importance of connected landscapes for forest conservation and restoration of plant-frugivore interactions

Forests are among the most biodiverse ecosystems on Earth. Yet they face ongoing threats from deforestation and fragmentation, driving biodiversity loss and disrupting key ecological processes. Ensuring their long-term persistence requires a dual approach: protecting remaining fragments and restoring degraded areas through regeneration. This thesis investigates how these complementary strategies contribute to biodiversity conservation and ecological recovery in fragmented (tropical) landscapes. The first part evaluates the effectiveness of protected areas in maintaining bird species in fragmented forests. While protected areas are a cornerstone of global conservation, their performance in highly fragmented landscapes remains uncertain. Through a global meta-analysis of 46 forest landscapes across six continents, we assessed whether protected fragments retain more bird species than unprotected sites. Both moderate (IUCN categories V–VI) and strict protected areas (cat. I–IV) were associated with higher bird occurrence, especially in larger fragments. In large areas, moderate and strict protection showed similar outcomes, but in smaller fragments (~50–175 ha), strict protected areas retained more species—particularly forest-dependent ones. The strong link between threatened bird species and large, strictly protected areas highlights their importance for conservation. However, the overall decline in bird species across all protected areas as fragments shrink shows that protection alone is insufficient in heavily fragmented landscapes. The second part draws on extensive fieldwork in Brazil’s Atlantic Forest, one of the world's most threatened biodiversity hotspots. We studied forest fragments naturally regenerating for different lengths of time to understand how forests recover after clearing. Recovery was assessed through three lenses: interactions between fruit-eating animals and plants (Chapter 3), the reassembly of plant and animal traits (Chapter 4), and the relationship between soil microbial diversity and the recovery of plants and animals (Chapter 5). This multi-trophic approach combined interaction network analysis, trait-based methods, and microbial profiling to provide an integrated view of recovery. A central finding emerges: landscape connectivity shows a stronger and more consistent association with ecological recovery than forest age. Fragments embedded in more forested landscapes recovered faster and more completely. They supported a greater variety of frugivores, more plant–animal interactions, and a stronger tendency for these interactions to cluster into distinct modules (network modularity). They also showed higher specialization, with species more likely to have unique fruit preferences. More connected fragments supported greater functional diversity among plants and animals and higher soil fungal diversity, all signs of healthier, more resilient ecosystems. By contrast, forest age showed weaker and inconsistent effects—particularly on plant–frugivore networks—or even ran counter to expectations, such as declining trait diversity among interacting species. Importantly, some key species, especially large specialized frugivores, remained absent from regenerating forests even after many years. This suggests that certain biodiversity elements and their functions may take very long to return, or may never fully re-establish if the surrounding landscape remains too degraded. These findings have clear implications for conservation and restoration. While crucial, protected areas often fall short of maintaining full species assemblages and ecosystem functions due to limited size and isolation. Natural regeneration offers a cost-effective path to recovery, but only when embedded in well-connected landscapes. In regions with low forest cover, increasing forest age alone cannot ensure full recovery, particularly for interaction-dependent functions like seed dispersal. To restore ecological functions in fragmented (tropical) landscapes, integrated strategies are needed that prioritize landscape-scale connectivity, functional diversity, and species interactions. Despite ongoing degradation, the Atlantic Forest retains strong recovery potential, especially if environmental laws are effectively implemented and restoration focuses on rebuilding ecological interactions through evidence-based approaches. By integrating global and local perspectives, and combining network, trait-based, and microbial approaches, this thesis provides a holistic view of ecosystem recovery and practical insights for forest conservation.