PS 11-164 - Seven years of invasion opportunity for Microstegium vimineum in suburban forests

Abstract: Suburbanization fragments the natural areas that support most of suburban regions’ biodiversity, yet they often harbor large populations of nonindigenous plants. In suburban forests, deer, among other drivers, may facilitate nonindigenous plant invasion due to especially high densities because of abundant edge habitat, few predators, and little hunting. A facilitative effect of deer on plant species’ success may be due to a network of direct and indirect effects upon and among a set of key community variables. Structural equation modeling (SEM), done with the piecewiseSEM package in R, was used to investigate the interacting roles of white-tailed deer, the plant community, earthworms, and abiotic conditions on the initial recruitment and subsequent success of the invasive grass Microstegium vimineum (Japanese stiltgrass). Data were obtained from a field experiment in six suburban forests in New Jersey, USA, with 16-20 16 m² plots in each (112 total). Experimental stiltgrass stands were initiated by seed addition, and deer were excluded from half the plots by fencing. Data were collected over seven years in each plot, including annually for stiltgrass proportion cover, total non-stiltgrass herb layer cover (competitors), and photosynthetically active radiation. Soil pH, soil water potential, and earthworm activity were measured once in each plot. A deer pressure index was set to zero for fenced plots, and for unfenced plots had a unique value for each forest. SE modeling began with a system-wide hypothesis of how the variables affected each other directly and indirectly, and ended with a fitted model including only the important paths between variables. It indicated that initial stiltgrass recruitment increased with more sunlight, wetter soils, lower soil pH, and less activity of (mostly) nonindigenous earthworms. Competitor cover decreased with deer pressure but had no effect on stiltgrass recruitment, which itself was not affected by deer. Six years later, stiltgrass cover was most strongly influenced by initial recruitment, but also increased with the same abiotic factors that affected recruitment. Its cover also was greater with more earthworm activity and greater deer pressure, but was unaffected by competitor cover, which was less under greater deer pressure. Combining a field experiment with multivariate SEM enabled a fuller understanding of stiltgrass invasion in suburban forests, revealing facilitation by deer only in the post-recruitment phase, the ongoing importance of abiotic factors, a changing role for earthworms, and the ability for this competitive invasive plant to establish regardless of biotic resistance from the plant community.