Effects of remnant trees on secondary forest regeneration in the Osa Peninsula, Costa Rica

Background/Question/Methods

Remnant trees, left when forests are cleared for agriculture or grazing, act as nuclei of tropical forest regeneration directly following field abandonment. Our study is among the first to investigate the legacy of remnant trees on surrounding forest structure and biodiversity 20-30 years after pasture abandonment. We located 10 remnant trees and 20 reference trees in two former pasture sites on the Osa Peninsula, Costa Rica and confirmed their classification using aerial photographs. We sampled along four 30 m transects around each tree. We measured the diameter at breast height (dbh) of all non-liana woody stems ≥1 cm and identified to species all stems ≥5 cm. For half of our trees, we counted all non-liana woody seedlings <1 cm dbh in eight 1 m² quadrats. We used ANOVA’s to determine if remnant tree presence or site influenced tree density, seedling density, basal area, and alpha diversity. We compared species composition using Chao-Jaccard similarity indices along with multi-dimensional scaling (MDS) and analysis of similarity (ANOSIM). We used a novel statistical method to classify old-growth and second-growth specialists and generalists, and compared their relative abundance in survey transects around remnant and reference trees. 

Results/Conclusions

Regeneration beneath remnant trees does not significantly differ in tree density, seedling density, basal area, or diversity (Shannon exponential mean) compared to reference trees in surrounding second-growth forest areas (p>0.05). However, the species composition around remnant and reference trees did vary significantly (p<0.05). Species composition around all focal trees also varied significantly between pasture sites (p<0.01). When pooled by site, species composition around remnant trees was more similar to that of old-growth forest than to reference trees. We observed significantly greater relative abundance of old growth specialists around remnant trees and significantly greater proportions of second growth specialists around reference trees. Our results suggest that by 20-30 years post-abandonment, secondary forest growth in this region has homogenized sufficiently to erase prior regeneration signatures surrounding remnant trees. Although remnant trees do not locally affect the structure of regenerating forests more than 20 years after pasture abandonment in this study area, they do exert a persistent effect on the species composition of the nearby surrounding forest, 20-30 years after forest regrowth.