Most studies that use demographic matrices or Integral Projection Models (IPM) to investigate population dynamics rely on following marked individuals from one time to another. While some perennial plants, and many types of animals, are readily marked with permanent tags and can be relocated, this can be difficult for clonal plants with stems that die back. We studied the population dynamics of the weedy plant Solanum carolinense over four years (three annual transitions) to parameterize density dependent matrix models and IPMs. Stems (ramets) of Solanum carolinense appear above-ground in spring and summer, grow through the summer, and senesce in summer or fall, however the underground rhizomes of a genetic individual can persist in the same general location for many years. When stems appear in spring they typically do not appear in the exact locations as the previous year, as different growth nodes are activated. New stems can appear up to tens of centimeters from the location of the previous year's stems. Thus when different genetic individuals are growing in proximity it can be difficult to assign stems to the previous seasons "parents". We asked how different methods of assigning parentage to clonal stems affected the outcome of demographic analyses.
Results/Conclusions
To construct our demographic models it was necessary to estimate the number and size of stems that a parent stem of a given size produced the following year. We applied multiple methods for assigning stems in one year to putative parents in the previous year, and evaluated the consequences of these assignments for projected population dynamics. Stems were (1) assigned wholly to the nearest stem the previous year, (2) fractionally assigned to multiple parents based on size and distance, and (3) assigned in a Bayesian analysis using genotype data from a subset of plots to inform priors on the number and spatial dispersion of offspring stems. Alternative methods of stem assignments affected predicted population growth rates and stable size distributions, although the pattern of effects varied some among the three annual transitions. Methods 2 and 3 increased the sensitivity of growth to larger stem sizes, as a larger proportion of stems were assigned to larger putative parents. Genetic data, although more difficult to collect, can improve predictions when genotypes vary widely in their life history traits. Assigning parentage to clonal plants without persistent aboveground stems remains a challenge for making demographic projections and studying selection on life history traits.