Planting Composition Effects on Tropical Forest Restoration

Deforestation in the Neotropics is driven primarily by the demand for pastureland for cattle. For pastures unlikely to transition back to forest, a proposed solution is establishing fenced restoration plantings throughout the pastures as stepping-stone biological corridors between forest fragments. Most tropical tree species are animal-dispersed, so in selecting species for plantings, animal-dispersed trees are favored as they attract more dispersal agents. However, whether the enhanced seed dispersal results in plantings that more closely resemble the forest has not been experimentally demonstrated. Between 2018 and 2019, I visited a restoration plantings experiment established in 2006 in Los Tuxtlas, Veracruz, Mexico to compare among and evaluate which of three treatments (plantings of animal-dispersed species, plantings of wind-dispersed species, and unplanted plots that simulate natural succession) most closely resembled the primary forest. First, I compared by forest structure by measuring the diameter and identifying every tree (> 2 m tall) in the experiment and in forest plots of equal size to the treatments. I found that animal-plantings most closely resembled the forest. Next comparison was by invasibility to the swordtail fern (Nephrolepis brownii), an invasive species seldom found in the forest. I found that the invasion was most acute (densest and longest fern fronds) in the natural succession plots and least acute in animal-plantings. Following that, I evaluated soil nutrient concentrations. Nitrate was marginally higher in animal-plantings soil, more closely approximating the forest. Fourth comparison was in foraging intensity by cricetid rodents largely absent in the forest, which I conducted using experimental food patches with millet seeds mixed with sand as substrate. I found that foraging was most intense in both animal-plantings and natural succession plots. My dissertation is capped off with a perspective piece on the differential roles that dispersal agent species play in the restoration of plantings in fragmented landscapes. The combined results of this dissertation and previous research that evaluated other key variables suggest that selecting animal-dispersed over wind-dispersed species is most effective towards establishing plantings that more closely resemble the forest. We expect stronger differences between treatments to manifest as the experiment matures.