The aim of this work is to assess the effect of planting patterns on the spatial distribution of total copper and other Cu fractions in vineyard soils. Both classical and geostatistical tools were used for the study. The soil of the plot had a loam texture and was strongly acid. The mean total Cu concentration (CuT) was 368 mg kg−1. The mean value of potential available fractions was 188 mg kg−1 for CuEDTA and 122 mg kg−1 for CuDPTA, whereas the mean exchangeable Cu (CuEX) was 5·2 mg kg−1. All Cu measurements exhibited a wide variation. These values are very high compared with those found in non-polluted soils, and they can affect the soil, plants and microorganisms. The best correlation for CuEX was with soil pH, whereas for CuEDTA, CuDPTA, and CuT, the best correlation was with soil organic carbon. Directional semivariograms were fitted with a spherical model (parallel to plant rows) and a periodic model (perpendicular) showing a dependence on orientation and distance. All Cu measurements were higher along plant rows than among them, finding a periodic pattern in the variance for the normal direction from plant rows. However, in site-specific management, it is crucial not only to describe the pattern of variation but also to estimate the Cu content in the soil. Copper concentration maps were estimated by kriging interpolation. These maps show a higher Cu accumulation along the cultivated rows than the uncultivated rows. Copyright © 2011 John Wiley & Sons, Ltd.