MATHEMATICAL-MODELS IN WEED MANAGEMENT

Some of the attempts to model aspects of weed control are reviewed, together with an assessment of how successful these have been in answering the key pragmatic considerations in weed management, namely the likelihood of invasion, rate of spread, crop competition and the effectiveness and economics of control. Although mathematical modelling has contributed some useful insights, there has been a tendency to concentrate on certain aspects, such as defining threshold levels and weed-crop competition, to the exclusion of other critical areas. In particular, weed dispersion, variation in recruitment and mortality, spatial heterogeneity in weed populations and the existence of multispecies assemblages have received scant attention. The future of weed modelling would seem to lie in trying to marry these various aspects to create models more firmly anchored on physiological principles and biological variation.