97th ESA Annual Meeting (August 5 -- 10, 2012)

COS 39-5 - Seed dispersal limitation of tree Aextoxicon punctatum in historically isolated and fragmented rainforest in semiarid Chile

Tuesday, August 7, 2012: 9:20 AM
E143, Oregon Convention Center
Mariela C. Núñez-Ávila, Departamento de Ecología, P. Universidad Católica de Chile, Institute of Ecology and Biodiversity, Santiago, Chile, Maria Uriarte, Ecology, Evolution, and Environmental Biology, Columbia University, New York, NY, Pablo A. Marquet, Department of Ecology, Pontificia Universidad Catolica de Chile, Santiago, Chile and Juan J. Armesto, Ecology, Institute of Ecology and Biodiversity, Santiago, Chile
Background/Question/Methods

The processes determining where seeds fall are critical to maintain tree populations in fragmented habitats, particularly for fleshy-fruited, bird-dispersed trees. We use a spatially explicit simulation to predict seed rain and seed limitation at 1-square meter scale, for the bird-dispersed tree Aextoxicon punctatum (Aextoxicaceae) in two relict forest patches (2.4 and 0.97 ha) isolated on coastal mountaintops of semiarid Chile (30ºS). For this purpose, we estimated dispersal kernels produced by avian seed vectors and estimated the magnitude of seed dispersal limitation for three consecutive fruiting seasons. All Aextoxicon trees were mapped and reproductive female trees identified. Fruit yield was quantified using 0.25 m2 seed traps (2 or 3) below each reproductive tree. To estimate dispersal kernels produced by avian seed vectors we used microsatellite markers to genotype 65 female trees and seed endocarps from 197 avian dispersed seeds collected in seed traps (n=288) located randomly within forest patches. We used maximum likelihood estimates to parameterize functions that resulted in the best fit to empirical data, which were used in stochastic simulation. The number of 1-m2 sites within patches that did not receive seeds in each reproductive season was quantified as indicator of dispersal limitation.

Results/Conclusions

Only 10 and 5% of all Aextoxicon trees in each patch produced fruits. We unambiguously identified the source tree for 93.4 and 82.3 % of seeds collected from seed traps in each patch. In both forest patches, mean dispersal distance due to avian frugivory was less than 10 meters, with a unimodal and strongly leptokurtic lognormal probability distribution. Spatial distribution of the seed rain in each patch was markedly heterogeneous, with peaks of seed density concentrated around few reproductive females. Between 78 and 79% of 1-m2 cells within both patches did not receive seeds in a given season, suggesting strong dispersal limitation, despite Aextoxicon dominance in the forest canopy. The proportion of seeds falling inside each patch did not differ between patches over the three years, despite fluctuations in fruit production. Strong seed limitation over several years suggests that manipulations of seed availability may be necessary to assure continuous tree regeneration in this fragmented landscape. Avian frugivory concentrates in few reproductive females per patch and few seeds arrive from outside the patch, which may have important consequences for genetic diversity in isolated habitats.