Predicting the determinants of weed abundance: a model for the population dynamics of Chenopodium album in sugar beet.
Abstract
Previously published literature on the population dynamics of a common arable weed, Chenopodium album, and its interactions with an arable crop, sugarbeet, is reviewed with a view to (i) assessing the degree of variability in life-history traits, and (ii) parameterizing simple models of population dynamics to explore the factors determining weed abundance. Comparison of previously published data sets indicates that (i) the yield-density responses of C. album in monoculture are remarkably consistent across sites and years; (ii) the per plant interspecific competitive effect of sugarbeet on C. album is roughly the same as the per plant intraspecific effect of C. album; (iii) the allometric relationship between the seed production of C. album and plant biomass is invariant; and (iv) there is considerable variation in published estimates of rates of seed emergence, mortality and seed bank decay. These data are used to parameterize a simple analytical model for the population dynamics of C. album in a rotation containing sugarbeet. Sensitivity analysis indicates that the key parameters in determining changes in population size are the rates of emergence and mortality of seeds, as well as the rate of mortality of plants through control. Using inferred levels of variability to compare deterministic and stochastic implementations of the model indicates that population sizes, and variability in population sizes, will be dominated by environmentally driven variations in the rate of seed germination and the rate of control. From a practical perspective, these results indicate that (i) further information on the effects of control on plant numbers, and (ii) monitoring and prediction of emergence rates is likely to be the most successful approach to predicting weed numbers and levels of infestation. The modelling shows how it is possible to use existing published data to parameterize simple analytical models, as well as to use information on scales of parameter variability with sensitivity analysis of such models to explore population dynamics. This provides an effective basis for exploring the impact of changing management on weed numbers in variable environments.