Relatively little information exists on neotropical secondary rain forests that have progressed beyond the pioneer stages of succession, or on the potential of natural regeneration to restore forest on large areas. We determined the structural and floristic characteristics (10 cm dbh) of a 30‐year‐old secondary forest developing on a 32.5 ha pasture on hilly terrain, abandoned after use of moderate intensity. Ten 0.24 ha sample plots covered the range of site conditions. The forest was dominated by long‐lived pioneer tree species; overall, the majority of species (70%) was vertebrate dispersed but the majority of individuals (52%) was of wind‐dispersed species. Tree species, including the dominants, were a mixture of those present in old‐growth and adventives colonizing from agricultural land. The forest was very heterogeneous. Vochysia ferruginea‐dominated stands characterized slopes with soils of high exchangeable acidity, while the adventive Cordia alliodora dominated sites with gentler topography and soils of lower acidity. Structural differences between the two forest types were slight, but Cordia forest had significantly greater species diversity and absolute and relative abundances of vertebrate‐dispersed tree species than Vochysia forest, which had significantly greater absolute and relative abundances of wind‐dispersed tree species. These latter differences between forest types, as well as the wide structural variation of the forest as a whole, were probably largely due to spatial and temporal variation in seed rains, some of it linked to the characteristics of the dominant species. Rain forest restoration on large pastures may depend greatly on wind dispersal and adventive tree species, and techniques for silvicultural diagnosis must be developed as a basis for the management of heterogeneous successional stands. Studies of early colonization of pastures should be expanded to focus on the causes of heterogeneity in older forests.

Aim We sought to determine the relationship of forest composition and alpha diversity (the species diversity of a local assemblage) to altitude, soil, and spatial factors over a 440–2,950 m a.s.l gradient. Location Altitudinal gradient on the Caribbean slope of the Talamanca Cordillera, Costa Rica. Taxon Angiosperm and gymnosperm trees, palms, and tree ferns. Methods We measured and identified all stems ≥10 cm dbh in 32 0.25‐ha undisturbed rain forest plots over the gradient. We determined compositional patterns using Non‐Metric Multidimensional Scaling (NMS) ordination, and used linear regressions to explore the relationship between four alpha diversity metrics and altitude. With variation partitioning (VARPART), we determined the compositional variation explained by altitude, soil, and spatial variables quantified using Principle Components of Neighbor matrices. Results We identified 425 species. NMS axis 1 separated a lowland zone (440–1,120 m asl) from a transitional one dominated by holarctic Oreomunnea mexicana (1,400–1,600 m asl) and Quercus‐dominated forests at altitudes >2,100 m asl. The lowland zone was separated into two clusters of plots on NMS axis 2, the first in the 430–620 m asl range and the second at 1,000–1,120 masl. Regressions showed that all alpha diversity metrics were strongly negatively related to altitude (R2 > 0.78). Overall, adjusted R2 from VARPART was 0.43, with 0.30, 0.21, and 0.17 for altitude, soil, and space respectively. The respective adjusted R2 of individual matrices, on controlling for the other two, was 0.06, 0.05 and 0.09 (p < 0.001). Main conclusions There are two well‐defined forest compositional zones on this gradient—lowlands 430–1,120 m asl and montane forests >2,150 m asl—with a transitional zone at 1,400–1,600 m asl, where lowland tropical and montane holarctic species are found together. Montane forests are very distinct in their composition and low alpha diversity. Vegetation and soil respond to altitude, and therefore temperature, as an integrated system, a model that goes beyond niche assembly as shown by the significant effect of space in the VARPART.

Multi‐year studies comparing changes in litterfall biomass and nutrient inputs in sites under different restoration practices are lacking. We evaluated litterfall dynamics and nutrient inputs at 5 yr and after a decade of recovery in four treatments (natural regeneration—no planting, plantation—entire area planted, tree islands—planting in patches, and reference forest) at multiple sites in an agricultural landscape in southern Costa Rica. We inter‐planted two native species (Terminalia amazonia and Vochysia guatemalensis) and two naturalized N‐fixing species (Inga edulis and Erythrina poeppigiana) in plantation and island treatments. Although litterfall N was higher in plantations in the first sampling period, litter production and overall inputs of C, N, Ca, Mg, P, Cu, Mn, and Fe did not differ between island, plantation, or reference forest after a decade; however, all were greater than in natural regeneration. Potassium inputs were lower in the natural regeneration, intermediate in island and plantation, and greater in reference forest. The percentage of litterfall comprised by the N‐fixing planted species declined by nearly two‐thirds in both plantations and islands between sampling periods, while the percentage of V. guatemalensis more than doubled, and the percentage from naturally regenerated species increased from 27 to 47 percent in islands. Island and plantation treatments were equally effective at restoring litterfall and nutrient inputs to levels similar to the reference system. The nutrient input changed substantially over the 7‐yr interval between measurements, reflecting shifts in vegetation composition and demonstrating how rapidly nutrient cycling dynamics can change in recovering forests.

The recovery of soil conditions is crucial for successful ecosystem restoration and, hence, for achieving the goals of the UN Decade on Ecosystem Restoration. Here, we assess how soils resist forest conversion and agricultural land use, and how soils recover during subsequent tropical forest succession on abandoned agricultural fields. Our overarching question is how soil resistance and recovery depend on local conditions such as climate, soil type and land-use history. For 300 plots in 21 sites across the Neotropics, we used a chronosequence approach in which we sampled soils from two depths in old-growth forests, agricultural fields (i.e. crop fields and pastures), and secondary forests that differ in age (1–95 years) since abandonment. We measured six soil properties using a standardized sampling design and laboratory analyses. Soil resistance strongly depended on local conditions. Croplands and sites on high-activity clay (i.e. high fertility) show strong increases in bulk density and decreases in pH, carbon (C) and nitrogen (N) during deforestation and subsequent agricultural use. Resistance is lower in such sites probably because of a sharp decline in fine root biomass in croplands in the upper soil layers, and a decline in litter input from formerly productive old-growth forest (on high-activity clays). Soil recovery also strongly depended on local conditions. During forest succession, high-activity clays and croplands decreased most strongly in bulk density and increased in C and N, possibly because of strongly compacted soils with low C and N after cropland abandonment, and because of rapid vegetation recovery in high-activity clays leading to greater fine root growth and litter input. Furthermore, sites at low precipitation decreased in pH, whereas sites at high precipitation increased in N and decreased in C : N ratio. Extractable phosphorus (P) did not recover during succession, suggesting increased P limitation as forests age. These results indicate that no single solution exists for effective soil restoration and that local site conditions should determine the restoration strategies. This article is part of the theme issue ‘Understanding forest landscape restoration: reinforcing scientific foundations for the UN Decade on Ecosystem Restoration’.

Active restoration strategies increase the production of leaf litter in tropical forests, but little is known about their effect on litter decomposition and subsequent nutrient release. We quantified changes in leaf litter stoichiometry during decomposition in former pasture sites under contrasting restoration strategies (natural regeneration, applied nucleation/islands tree planting and plantation), as well as in nearby primary forest. Litterbags were employed to evaluate decomposition. We used a leaf mixture of either the four planted tree species in the plantation and island treatments or the nearby primary forest and compared them under a factorial design. Decomposition rates were similar between restoration treatments (p > 0.5), but leaves decomposed faster in the forest mixture than in the plantation mixture (p < 0.01). The content of Ca, Mg, K, P, and the C:N ratio were higher in the forest mixture at the beginning and during decomposition (p < 0.05); the N content in the plantation mixture was higher at the beginning but lower during decomposition (p < 0.05), which meant greater mobilization of nitrogen per unit of carbon lost. K and P had a strong initial release, while Mg was released more gradually. N and Ca had an irregular pattern of initial fast release, immobilization, and re-release in the later stages. We conclude that the differences in rates of decomposition and nutrient release in these systems under restoration were at least partly determined by the floristic heterogeneity and chemical quality of the leaf litter that reaches the soil.