Studies of spatial distribution of weeds are important in the agronomic and environmental context. The objective of this research was to describe the spatial distribution of weeds in two soil management systems. This study was accomplished in a Distroferric Red Latosol (Oxisol), located in Campo Mourão, Parana State, Brazil, in no-tillage (NT) and conventional tillage of soil (CT). A grid was established in coordinated x and y, with 128 sampling points. The studied variables were biomass of shoot weeds, density of weeds, Commelina benghalensis density and Bidens pilosa density. Exploratory data analysis was used to evaluate distribution of variables as well as to adapt them for the spatial analysis. The spatial variability was studied using geostatistics by analysis of semivariograms, kriging interpolation and constructions of maps. The ranges of biomass and density of Bidens pilosa, were larger in NT than in the CT. For plant density range was larger in the PC. The spatial distribution described by geostatistics allowed weed mapping and its site specific management and the establishment of criteria for future experiments.
site specific management; spatial dependence; interpolated maps; semivariograms; kriging
