Population matrix models of Aeschynomene virginica, a rare annual plant: implications for conservation

Matrix models are valuable tools for comparing population growth rates (λ) in different environments. Life table response experiments (LTRE) can then be used to identify those life history stages that are most important to observed population growth rate differences. We used matrix models of population growth and LTREs to analyze the effects of standing vegetation and seed dispersal on the population growth rate of Aeschynomene virginica, a rare, tidal, wetland annual. We combined seedling establishment, seedling survival to adulthood, adult fecundity, winter seed survival, probability of seeds entering a seed bank, and seed dispersal rates to calculate the matrix elements in a matrix model. Experimental removal of other plant species from field plots increased A. virginica fecundity, seedling establishment, and survival. The magnitude of these changes doubled between 1998 and 1999. Removal of interspecific competitors from population patches of A. virginica explained 54% of the variability of λ, while environmental variation between 1998 and 1999 explained only 1%. An additional 9% of the variability in λ was explained by adding the interaction of interspecific removal and year to the regression. Sensitivities of λ to changes in lower-level parameters varied over several orders of magnitude. Fecundity had the smallest sensitivity, while seedling establishment made the largest contribution to λ in 1998, and seedling survival to adulthood made the largest contribution in 1999. Simulations indicated that changes in seed dispersal would have small effects on λ. Eighty percent (1998) to 90% (1999) of the seeds produced in local patches growing without interspecific vegetation would have to disperse before patch population growth rates would become negative. Although seed dispersal had little impact on within-patch population dynamics, it may make seeds available to establish populations in nearby, open habitat patches. In population patches with interspecific vegetation, 30 seeds per plant (1998) and 80 seeds per plant (1999) would have to be added before patch population growth rates were positive. The most efficacious treatment to increase population sizes of this rare plant would be creating disturbed patches and not adding seeds.