Intensive field surveys of Convolvulus arvensis were conducted over the course of four growing seasons (1999)(2000)(2001)(2002) in a 1.6 ha section of a field located in Andalusia, southern Spain, cropped with wheat in 1999 and 2001 and with sunflower in 2000 and 2002. In order to characterise the spatial and temporal variability of C. arvensis density, geostatistical techniques were used. Data from the study were used to design precise herbicide spraying. The density and spatial distribution of C. arvensis varied widely with years and crops. In general, weed density was higher in wheat years (30.41 and 53.72, in 1999 and 2001, respectively) than in sunflower ones (20.67 and 24.03, in 2000 and 2002, respectively). C. arvensis density data showed strong and moderate spatial dependence. The C. arvensis density maps achieved by kriging were used to estimate the percentage of surface susceptible to be treated with a site-specific herbicide treatment based on an estimated economic threshold (ET) for each year of study. The C. arvensis infested areas varied considerably between years and crops. About 33 and 64% of the total infested area were moderately infested (density > 14 plant m −2 ) in wheat in 1999 and 2001, respectively, and about 17 and 19% in sunflower years (2000 and 2002, respectively). During the years cropped with sunflower the percentage of the total area infested with C. arvensis was reduced and therefore the area exceeding the economic threshold (AEET). So, if a given herbicide were applied just only to the areas exceeding the economic threshold, the average reduction in herbicide cost achieved in sunflower years could be around 81%. These approaches demonstrated the importance of growing different crops in weed management outcome, and could be useful to farmers and researchers in formulating a multiple-season weed management plan, especially since C. arvensis has been shown to have relatively stable patches over time and crop rotations in no-tillage systems.