Leaf litter Decomposition dynamics across a Recovering Tropical Forest in the lowland Ecuadorian Chocó

Litter decomposition by arthropods, microbes, and fungi is a key ecosystem process in tropical forests, yet its response to forest disturbance and recovery remains poorly understood. To investigate decomposition dynamics across forest succession, we conducted an experiment in the Ecuadorian lowland Chocó (Esmeraldas) using a chronosequence approach. We deployed above- (AG) and belowground (BG) litterbags in 32 plots spanning active cacao plantations and pastures (age 0), secondary forests (1–38 years), and old-growth forest. AG litterbags (5 mm mesh) allowed arthropod access, while BG litterbags restricted decomposition to microbial activity. Each contained standardized leaf litter from five common tree species: Pourouma bicolor, Brosimum utile, Compsoneura atopa, Vochysia macrophylla, and Trema micrantha. Litterbags were collected at three time points, with replacements every 45 days. We examined decomposition drivers by modeling litter mass loss (%) against forest age and environmental factors. AG decomposition was analyzed in relation to tree aboveground biomass, surface temperature, leaf litter biomass, elevation, and terrain slope, while BG decomposition was assessed with soil pH, soil C:N, terrain slope, soil moisture, and soil temperature. Additionally, we tested how small-scale disturbances and large animal exclusion affected decomposition (PREX) using four treatments: control (C), fenced (CF, exclusion of large ground-dwelling animals), perturbed (P, removal of litter and understory vegetation in 100 m²), and perturbed-fenced (PF, combined litter removal and animal exclusion). AG decomposition rates increased with forest succession but followed a U-shaped pattern in plots recovering from cacao land-use, with a mid-succession decline and higher rates in old-growth forests. Key drivers included surface temperature, elevation, and tree aboveground biomass, with temperature varying significantly depending on land-use history. BG decomposition was unaffected by forest age, decreased with C:N, and showed a bell-shaped response to soil moisture. Large animal exclusion (CF) had no effect, whereas perturbation (P, PF) significantly altered decomposition. Notably, decomposition in P plots showed dynamic recovery, whereas in PF plots, mass loss remained suppressed throughout the 135-day study, emphasizing the role of large animals in facilitating ecosystem recovery.