A SIMULATION OF HERBICIDE USE BASED ON WEED SPATIAL-DISTRIBUTION

Intensive field surveys were conducted to determine the spatial distri bution of composite broadleaf and grass weed seedlings in seven maize and five soybean fields in eastern Nebraska in 1992. Farmer fields whe re herbicides were applied in a 38-cm band over 76-cm spaced rows were chosen for this study. Weed populations were measured in the interrow and in the band-treated intrarow. Spatial maps constructed for grass or broadleaf weeds revealed that individual species as well as species assemblages are highly aggregated. On average, 30% of the sample area in the 12 fields surveyed was free of broadleaf weeds and 70% free of grass weeds in the interrow area (no herbicide). Where a pre-emergenc e herbicide was applied (intrarow), 71% of the sample area was free of broadleaf weeds and 94% free of grass weeds. Increasing the threshold to some value other than zero resulted in a larger field area not req uiring a herbicide application. The results of these distribution stud ies indicate that herbicide use could be substantially reduced if weed distribution maps or real-time plant sensing were available to provid e information for intermittent herbicide application systems or refine ment of economic thresholds.