Forest structure recovery around West Africa’s last great rainforest: Modelling complex dynamics in Taï national park

• Recovery rates depend on attribute: diameter > height > biomass. • Biomass recovery is slowest, needing over 100 years. • Cocoa fields show the fastest recovery across all structural attributes. • Gold mining sites show very slow recovery due to severe soil degradation. • Remnant trees are key to rapid recovery. Tropical secondary forests play a key role in restoring biodiversity and biomass, yet their recovery dynamics remain poorly understood in West Africa, even within protected areas. In Taï National Park, one of the region’s last major rainforests, we modelled forest recovery trajectories in the past agricultural areas to better understand and support effective conservation strategies. We inventoried 118 plots spanning old-growth undisturbed, old-growth disturbed, and secondary forests. Four structural attributes (aboveground biomass, Lorey’s height, quadratic mean diameter, and structural homogeneity) were modelled using a Bayesian framework. We assessed the influence of past land use, disturbance, and environmental factors on recovery rates. Structural attributes recovered at markedly different rates. Structural homogeneity and mean diameter recovered fastest (∼20–30 years), followed by height (∼30–40 years), while biomass required over a century for near-complete recovery. Recovery was strongly enhanced by the presence of remnant trees and forest connectivity, especially for biomass. In contrast, Marantaceae presence, hydromorphic soils, anthropogenic disturbance, and herbivory slowed recovery. Former cocoa fields showed the highest recovery rates across all attributes, while former gold mining sites exhibited extremely slow regeneration due to severe soil degradation. Our results suggest that forest recovery around Taï National Park will be highly heterogeneous. Cocoa fields with remnant trees offer strong potential for rapid passive recovery, while mined areas may need active restoration. Protecting remnant trees and managing disturbances will be crucial. Overall, passive regeneration holds significant promise, but realistic expectations and adaptive management are essential to support long-term forest resilience in this landscape.