Tropical forest restoration is a critical component of natural climate solutions (NCS) aimed at mitigating climate change and biodiversity loss. The Intergovernmental Panel on Climate Change (IPCC) highlights the need for significant carbon removal through NCS, with forests playing a major role. Forest restoration, while ecologically responsible and cost-effective through natural regeneration, faces challenges in degraded areas with depleted soil or altered seed banks. A key challenge is determining where animal contributions to natural carbon recovery are sufficient and where active interventions are necessary. Growing evidence suggests that animals, especially seed-dispersing frugivores, are essential for carbon recovery in tropical forests. Frugivores disperse a significant portion (70-90%) of tree species, and their movement restrictions can hinder natural regeneration. However, factors like animal composition, diet, traits, and movement barriers affect their contribution to seed rain. This study focuses on quantifying the impact of frugivorous birds on seed dispersal and carbon recovery across a fragmentation gradient in the Brazilian Atlantic Forest, a biodiversity hotspot with significant fragmentation.

Previous research indicates the crucial role of animals in the carbon cycle and NCS. Studies have shown animals are determinants of carbon recovery in tropical restoration projects because of their importance in seed dispersal. Seed arrival at degraded sites is a limiting factor in ecological recovery. The composition of animal communities, their diets and traits, and movement barriers all influence seed dispersal. Large frugivores tend to consume seeds from long-lived species with high carbon storage potential, but their movement is disproportionately impacted by fragmentation. Empirical evidence from local studies provides mixed results, highlighting the need for regional-scale assessments to understand the impact of animal movement on forest carbon recovery.

This study employed a regional-scale assessment to quantify seed dispersal by birds and its influence on forest recovery across a fragmentation gradient in the Atlantic Forest, Brazil. Ten landscapes spanning a gradient of 9-61% forest cover were selected. Three hierarchical Bayesian models were built to simulate three key components of the seed dispersal process: (1) fruit ingestion (frugivory), (2) frugivore movements between forest fragments, and (3) gut passage time (GPT). These models incorporated bird traits (body mass, gape size, frugivory degree) and landscape characteristics (forest cover, isolation). Simulations estimated potential seed rain deposited by each bird species in open areas and subsequent potential aboveground biomass (AGB) and carbon stock. The models accounted for bird-plant interactions, bird movement patterns based on landscape structure, and seed passage time through the birds' digestive systems. The models used a dataset of frugivory interactions, data from bird movement tracking, and data on seed retention time. The resulting simulations generated seed rain maps and potential AGB estimations across the fragmentation gradient. Inflection points were calculated to pinpoint thresholds of forest cover and isolation, where natural regeneration ceases to be effective.

The study revealed a nonlinear relationship between forest cover, isolation, seed deposition, and potential AGB. Seed deposition and AGB increased with forest cover and decreased with forest isolation. A threshold was identified around 40% forest cover, beyond which seed deposition and AGB significantly declined. In landscapes with less than 40% forest cover, the movement of large birds was restricted, resulting in a 38% reduction in potential carbon stock compared to less fragmented landscapes. Bird body mass and gape size inversely correlated with seed dispersal, with smaller birds contributing more seeds in highly fragmented landscapes, but these seeds were primarily from species with lower carbon storage potential. The contribution of birds to seed deposition and AGB increased with their degree of frugivory, but this effect was only pronounced in landscapes with high forest cover. A strong decay in seed deposition and potential biomass restoration was observed with increasing distance from the forest edge. The effectiveness of natural regeneration is dependent on the interaction between seed-dispersing frugivores and landscape structure. The study found that natural regeneration was effective in landscapes with >40% forest cover and fragments separated by <133m.

The findings highlight the importance of landscape structure and frugivore community composition in forest restoration. Maintaining forest cover above 40% is crucial for enabling sufficient bird movement and maximizing seed dispersal and carbon sequestration potential. In more fragmented landscapes, active restoration methods are needed to compensate for the limitations of natural regeneration. The observed threshold of 30-40% forest cover aligns with previous recommendations for preserving biodiversity, supporting the importance of incorporating this threshold in restoration planning. The study supports the need to integrate animal movement and landscape connectivity into restoration strategies to effectively enhance both carbon sequestration and biodiversity.

This study demonstrates the critical role of frugivorous birds in tropical forest restoration and carbon sequestration. Maintaining landscape forest cover above 40% is vital for optimizing natural regeneration. In more fragmented areas, active restoration strategies are needed to enhance animal movement and improve carbon stock recovery. Future research should investigate the impact of resource distribution, post-dispersal processes, and other seed disperser groups to refine carbon potential estimations and improve the generalizability of these findings to global conditions. Integrating animal movement and landscape connectivity into restoration policies is necessary to maximize the effectiveness of forest restoration efforts.

This study primarily focused on birds and did not consider other seed dispersers (mammals, reptiles). Post-dispersal processes (seed predation, germination success, environmental filtering) were not explicitly modelled, potentially overestimating potential AGB. The study's conclusions might not be universally applicable across different tropical regions; thresholds for forest cover and isolation could vary with specific environmental conditions. The use of a single reference forest to inform potential plant composition might not be completely representative of the regional species pool.