Modeling effects of spatial patterns on the seed bank dynamics of Alopecurus myosuroides

A spatially explicit metapopulation neighborhood model was extended to encompass the seed bank dynamics of the annual weed Alopecurus myosuroides Huds. (blackgrass), growing in U.K. winter wheat crops established after noninversion tillage. The effects of the initial spatial pattern of infestation, herbicide, and combine harvesting on seed bank densities and on expected yield losses of the infested crop were studied within a 634 m(2) area. In the absence of herbicide, all seed bank populations were large and typical of values in the literature; those with patchy initial distributions spread quickly over the entire field. The effects of intraspecific competition ensured that even after 10 yr, the average seed bank density from three patterns with the same initial density consistently retained the same rank order: the initially uniform pattern consistently ranked largest; a patchily distributed moderate infestation ranked next; a more patchily distributed heavy infestation ranked least. Expected grain yield losses sometimes exceeded 40%. With the introduction of a herbicide, seed densities declined exponentially, but the rank order with regard to spatial pattern remained. Relatively dense patches occasionally persisted for longer than 10 yr. The economic threshold for treatment was achieved within 2 and 5 yr, depending on the infestation pattern, but the model predicted that it would take many more than 10 yr before infestations could be completely eliminated. The effects of harvesting by combine were modeled. In all cases, the speed of the spread of infestation along a row appeared to be largely due to the combine rather than to natural dispersal or other cultivation practices. The proportion of seeds removed by the combine was an important determinant of the number of years required to drive the population below the economic threshold and of the probability of long-term eradication of the weed metapopulation.