Spatial distribution of weeds in arable crops: are current sampling and analytical methods appropriate?

This paper reviews the literature concerning the spatial distribution of we eds; highlighting the limitations of our current sampling and analytical me thodologies, and suggesting how these inadequacies can be addressed. Most r esearch studies have used discrete sampling, i.e. weeds are counted within a quadrat, on a grid basis. Few have mapped weeds at a whole-field scale, e ither with a resolution appropriate to spraying operations or key ecologica l processes. Statistical analyses used to describe the data can be divided into two main types, spatially implicit (also at the scale of the sampling unit) or spatially explicit, in which the location of individuals is includ ed in the analyses. Spatially implicit methods can be strongly affected by quadrat size and mean density and are of doubtful benefit. More attention i s required to address sampling resolution issues for spatially explicit met hods. Our understanding of the formation and dynamics of spatial pattern, a s well as predicting the consequences of site-specific management, can be i mproved with models. Unfortunately, most models consider only newly expandi ng patches and appear incapable of predicting spatial distributions when an area has been fully invaded. More detailed biological information is requi red if models are to become more realistic and informative. We also need to ensure that we understand the spatial processes in the context of the whol e field environment, to optimize the success of site-specific weed manageme nt in the longer term.