Return of forest structure and diversity in tropical restoration plantings

Stepping-stone restoration plantings can reconcile conservation goals and local land use needs in highly fragmented ecosystems. We explored how initial planting composition influences recruiting plant species density, diversity, abundance, and forest structure in a 13-year-old restoration experiment in Los Tuxtlas, Veracruz, Mexico. Treatments included 8 fenced plantings with animal-dispersed species, 8 plantings with wind-dispersed species, 8 unplanted plots to favor natural succession, and 8 plots in the primary forest as reference sites. We predicted that that by attracting more seed dispersers, animal-dispersed plantings would most closely resemble the primary forest. A census of trees taller than 2 m showed that while wind-dispersed plantings had more recruits, the animal-dispersed plantings most closely resembled the primary forest in pioneer abundance, species density and abundance of biotically-dispersed and abiotically-dispersed plants, individual tree basal area (m2/ha), and vertical structure. The wind-dispersed plantings more closely approximated the forest in non-pioneer abundance and community composition. However, restoration treatments were more similar to each other than to the primary forest and did not differ in plant diversity. Animal-dispersed and wind-dispersed plantings did not differ in non-pioneer species density and matched the primary forest in total plot basal area. Higher abundance of trees in wind-plantings is explained by lower establishment limitations, seed legacy effects, and rapid reproduction of a few planted species. As the experiment continues, we expect treatment effects on seed dispersers will more strongly influence the recruiting plant community, leading the animal-dispersed plantings to more closely resemble the primary forest in diversity and forest structure.

Methods
This study was conducted in Los Tuxtlas, Veracruz, Mexico. The restoration experiment was established in June 2006 in 12 ha of actively grazed pasture adjacent to a privately owned patch of late secondary forest continuous with the biosphere reserve. The experiment consists of 24 30 x 30 m fenced plots set on an 8 x 3 grid, each separated by 35 m of pasture (central GPS point 18° 35' 43.64" N, 95° 06' 06.29" W). We recorded the circumference, identity, and origin (non-planted plants are from hereafter referred to as “recruits”) of every woody plant and palm (>2m tall) in all 24 plots in June and July 2019. Eight additional plots were randomly established in the primary undisturbed forest of Los Tuxtlas (from here referred to as “forest”) by the Los Tuxtlas Biological Station to compare with the restoration treatments; forest plots were divided into subplots in the same way as the restoration plots. Recorded species were classified by dispersal mode and life history using the available literature (Appendix S1: Table S1). Diameter and basal area values were calculated from recorded circumference values.