Modeling seed dispersal of black cherry, an invasive forest tree: how microsatellites may help?

We used empirical models and three dispersal functions (Weibull, lognormal and 2Dt) to model seed distributions derived from the black cherry (Prunus serotina Ehrh.) understorey of a pine-dominated stand. Two different approaches were used to disentangle the overlapping seed shadows: the traditional inverse modeling approach and the genetic approach that uses microsatellite markers to assign a dispersed seed to its maternal parent. The distinction was made between the seeds passively dispersed by gravity (fruits with mesocarp) and those dispersed by birds (fruits without mesocarp). Our main objectives were to compare the three dispersal functions and assess the differences between the two approaches used. The functions performed equally well, but the lognormal function often showed a better data correlation. The best dispersal curves obtained by both traditional and genetic approaches were quite similar for the seeds dispersed by gravity, and 95% of these seeds were predicted to fall 5 and 3 m away from the parent tree for the traditional and genetic approaches, respectively. Differences were more important for the seeds dispersed by birds. The traditional approach predicted a lower number of seeds near the parent plant and a higher dispersal distance. Microsatellites provided accurate information on individual dispersal events and led to a better insight into the dispersal process.