Spatial variability of heavy metals in soil on a one-hectare scale
1988
WOPEREIS, M. C. | GASCUEL-ODOUX, C. | BOURRIE, G. | SOIGNET, G.
We studied the spatial variability of heavy metals (Ni, Cu, Zn, Cd, and Pb) in a small (1-ha) forested plot on acid brown earth soils with granite underneath. The sampling pattern consists of a nested sampling to describe the spatial variability from 0 to about 60 m. Data were analyzed by geostatistical techniques, specifically variogram computations. Analyses were replicated to separately assess spatial and analytical variances.Analytical errors are not statistically different from spatial variability of Zn and, to a lesser extent, of Pb. Spatial variances of Cu, Ni, and Cd are higher than analytical variances, but do not show a spatial structure, so that only random models—“pure nugget effect”—can be fitted to the variograms. With the exception of Cd, for which spatial variance is very high, this study shows that a limited number of samples is sufficient to assess a mean value for heavy metal contents in this plot. Moreover, in all cases, variance observed at the very center of this plot (set of nine data) is not statistically different from the variance computed from the samples collected throughout the entire plot (65 data). This means that spatial variability is due to a heterogeneity at a microscopic scale. Samples collected at close distances can be considered as independent experiments, normally distributed. Two particle-size fractions, namely coarse silt and coarse sand, show a definite spatial structure, with a linear drift, which has no bearing on other variables and particularly heavy metals. Nickel, Cu, and Zn are positively intercorrelated and are negatively correlated with organic matter. Lead and Cd are not correlated with each other or with any other variable.
显示更多 [+] 显示较少 [-]