Strategies for non-chemical weed control on public paved areas in Denmark.

To be proactive in minimizing pesticide use, the public authorities in Denmark agreed in 1998 to phase out the use of pesticides on publicly owned areas by the end of 2002. A part of the agreement was an increasing focus on research into and development of new methods and implements for non-chemical weed control on paved areas. Due to a large increase in the costs of non-chemical weed control, the park authorities have to put the different types and locations of paved areas in order of priority to optimize the weed control effort. The present authors divided the paved areas into five weed control levels, dependent on placement, quality and use. For the 3 years 1999-2001, experiments with different non-chemical weed control methods were conducted on pavements at six locations in Denmark. The aim was to test the reaction of the weeds to different treatments and strategies. The efficacies of the methods were evaluated by analysis of digital images to estimate the fraction of the paved area covered with green vegetation (weed coverage). The weed coverage was used as the dependent variable in the subsequent statistical analysis. The independent variables in the model were incoming radiation, wear, area of joints in the pavement, the dying process of the weeds and the number of runs/applied energy of the mechanical or thermal weed control methods, respectively. The estimated parameters from the statistical model were used to build a simulation model, which was used to optimise five weed control strategies to fulfil the suggested weed control levels. In the suggested strategy for maximum weed control, 12 thermal weeding applications at 2-week intervals are suggested. The 'clean-up' strategy is based on one weed-brushing in late spring or early autumn.