Landscape demography of Centaurea stoebe (spotted knapweed) in New York State




Green, Ashley R.

Journal Title

Journal ISSN

Volume Title



Variation in space and time is ubiquitous in populations, communities, and ecosystems, and understanding the nature and causes of spatial and temporal variation is one of the central goals of ecology. Landscape demography, the study of demographic variation among populations, focuses on documenting patterns and causes of demographic heterogeneity among populations. I compared 41 populations of Centaurea stoebe (spotted knapweed), an herbaceous plant, in two regions of New York State, many of them for surveyed 6 times over 2 annual intervals (3 years), with a total of ~ 24,500 plants. The large sample size allowed me to address demographic questions at exceptional depth and breadth. This study focused on the amount and type of demographic variation among populations in space and time, the former at two scales: within and between regions. A complete set of demographic information was collected, including recruitment rates and seed bank variables, and a separate integral projection model (IPM) for each population in each annual interval was constructed. I found substantial variation at both spatial scales, and between annual intervals, in the finite rate of population increase (λ), plant size distributions, the underlying demographic variables (survival probabilities, plant growth rates, fecundities), and the importance of the latter to population dynamics (elasticities and amounts of existing variation). I found little to no consistency in the ranking of populations between annual intervals. The results of the models differed from observed changes in population sizes. These results suggest that simpler methods, or fewer populations, do not give representative results. The large sample sizes and IPMs allowed detailed analyses that revealed other interesting findings. For example, the growth of large plants was especially important in determining λ, due to high elasticities and large amounts of natural variation. I also found an intriguing pattern of correlations among demographic variables in one region, suggesting the classic fast-slow life history gradient. I found that, contrary to the literature, the distribution of elasticity across plant sizes closely matched the predicted size distributions. The low elasticities of fecundity variables suggested that ongoing biocontrol efforts using seed eating insects are unlikely to be effective.


LCSH Subject Headings