Are weed population dynamics chaotic?
Abstract
There have been suggestions that the population dynamics of weeds may show chaotic dynamics, and that therefore it will not be possible to predict the impact of changing management regimes on weed abundance. The instability of weed populations is presumed to result either from overcompensating yield-density responses or from threshold management. Using theoretical arguments and empirical evidence we argue that this contention is likely to be incorrect. Overcompensating yield-density responses are unlikely in plant populations and this point has been extensively discussed. Such responses have only been observed in high-density artificially sown stands of weed populations. The form of chaos that results from threshold management is a consequence of high population growth resulting from the cessation of management when weed densities are lower than a threshold level. Consequently the dynamics of such populations may be argued to be extrinsically rather than intrinsically driven. There are many studies that have shown weed populations to be dynamically stable, both spatially and temporally. Here, we present an analysis of data from the Broadbalk experiment that demonstrates long-term stability of 12 species of common weeds (Ranunculus arvensis, Medicago lupulina, Vicia sativa, Tripleurospermum maritimum [Matricaria perforata], Alopecurus myosuroides, Poa annua, Cirsium arvense, Tussilago farfara, Equisetum arvense, Agrostis stolonifera, Poa trivialis and Papaver spp.) over a 12-year period (1955-67). Using parameter estimates derived from the literature we show that the stability of these populations is similar to other annual species, both weedy and non-weedy. We argue that weed population dynamics are more generally better viewed as resulting from the impacts of broad-scale types of management, as well as temporal variability in population numbers. The significance of chaotic dynamics is likely to be minimal.