Assessing the Spatial Uncertainty of Mapping Trace Elements in Cultivated Fields

Abstract

Many of the cultivated soils in Galicia (NW Spain) consist of grassland areas and, subsequently, cattle density is also considerable. As a result, dispersion of heavy metals into the rural environment associated with manure and slurry applications has been reported. Mapping the spatial distribution of heavy metals in soils is important for soil management; currently, geostatistical methods are used to generate maps of fertilizer and heavy metal contents and also to assess the uncertainty of the predicted concentrations of these elements. The main objective of this study was to assess the spatial variability of trace elements in a small catchment (10.7 ha), where slurry application was intense. Moreover, the use of geostatistical tools for the description and modeling of metal distribution is illustrated. Fifty‐five soil samples were digested by nitric acid in a microwave (U.S.E PA‐SW‐846 3051) method to assess Mn, Zn, Cu, Ni, Pb, and Cd contents by ICP/AES. The pattern of spatial distribution of total metal contents was analyzed by means of different geostatistical techniques and by the inverse distance method. When a pattern of spatial dependence was found, kriging was used to construct contour maps after semivariogram analysis. High statistical variability was observed for the metals studied. Mean Cu content was 28.5 mg kg−1 and that of Zn was 64.7 mg kg−1, the differences between maximum and minimum being 62.3 mg kg−1 and 122.1 mg kg−1 for Cu and Zn, respectively. Also, some hot spots with high levels of Cd (>l3 mg kg−1) with respect to background values were recorded. The spatial distributions of total Zn and Cu contents were characterized using spherical semivariograms with no nugget effect. Ordinary kriging estimates were compared with results obtained by indicator kriging. Sequential simulation was also applied to analyze the spatial distribution of the metals studied. This study reveals that these methods are useful to determine spatial distribution and uncertainty of trace metal contents in cultivated fields and thus to characterize the heavy metal status at the scale studied. In contrast, EDTA‐ and CaCl2‐extractable Zn and Cu contents did not show spatial structure at all; hence, nongeostatistical techniques were used to attempt to interpolate these metal fractions.