How forest structure varies with elevation in old growth and secondary forest in Costa Rica

Clarifying how forests vary with elevation can inform our understanding of forests and their responses to climate. Despite the prevalence of secondary forests (defined as regrowth after disturbance), few studies have examined how their structure and recovery vary with elevation. Here, in apparently the first study of its kind, we explored how basal area, stem density and community wood specific gravity of stems ≥ 10 cm diameter vary along an old growth and a 25–30 years old secondary forest transect, each spanning over 2000 m of elevation in Costa Rica. The old growth transect comprises 29 plots of 0.25 ha from 440 to 2900 m.asl. and the secondary 28 plots of 0.1 ha and 10 plots of 0.25 ha from 600 to 2750 m.asl. We characterized how stand characteristics vary with elevation using linear and generaled additive models. To our surprise basal area increased with elevation not only in the old growth (from 24 to 56 m²/ha) but also in the secondary forest (11 to 59 m²/ha). This increase reflected the greater density of larger trees (≥40 cm diameter) at higher elevations. Basal area decreased with inferred water availability (global aridity index) in the secondary, but not the old growth forest. Surprisingly, wood specific gravity revealed no clear pattern with elevation in either transect though within stand variation decreased with elevation in both. In both forests, basal area is positively correlated with mean wood specific gravity, being significantly greater in stands with less variation in wood specific gravity. We interpret these patterns as resulting from reduced mortality of larger trees at higher versus lower elevations in both old-growth and secondary forest. Our results imply that planting trees to enhance recovery would be more beneficial in lowland than in high elevation forests.