Development of a decision support system for integrated weed management

In the last years only a few new action mechanisms have been introduced in the chemical weed control sector. Consequently, when product innovation slows down, process innovation becomes the only possible solution to reduce dependence on chemical herbicides. To rationalize weed control is a priority in this process. Developing and spreading via the Internet a DSS that can advise farmers about weed management, suggesting the correct timing and appropriate mixture of active ingredients for each situation, may represent a key step for this. The combination of two existing models, GESTINF and AlertInf, makes possible the realization of this complete DSS. GESTINF is an existing DSS that evaluates the cost effectiveness of weed control, providing users with a ranking of possible technical alternatives, i.e. herbicide mixtures, according to their specific economic return and the environmental risk of ground and surface water contamination by leaching and runoff. AlertInf has recently been created to predict weed emergence dynamics of several weeds in maize fields. AlertInf provides the percentage of emergence reached by a given weed species in real time using meteorological data, such as soil temperature and soil water potential. The major cause of poor post-emergence weed control is improper application timing, which can be either too early or too late. AlertInf, estimating daily the percentage of weeds that have already emerged and the successive seedling emergence dynamics, allows to make appropriate evaluation on the timing for post-emergence applications to achieve efficient weed control. The two models are integrated in a unified DSS that is able to indicate to the users the date for the single scouting to be carried out in the field to know the quali-quantitative characteristics of the infestation and estimate the density of each weed species for the rest of the season. The predicted daily density of each species is used by the DSS as input without any further survey. The single time survey method for indicating the survey date and estimating the density for the rest of the season, uses the cumulated emergence estimated by AlertInf; the method is described in Masin et al. (2011). The DSS gives indications on post-emergence treatments in traditional tillage maize and soybean and no-tillage maize in terms of herbicide to use ranked according to economic net return. The output of the combined DSS is a more complete information about if, when and how control weeds in their specific situation. This combination of models has the objective to overcome the constraints that have so far hindered the use of DSS, such as GESTINF, by farmers. Furthermore the combined DSS available on-line should encourage farmers to adopt the criteria and methods of IWM