Spatial structure analysis and kriging of dichlorodiphenyltrichloroethane residues in topsoil from Tianjin, China

Abstract

Spatial structure analysis was conducted on a previously published dataset of DDT species concentrations in topsoil samples from Tianjin, China ( p, p′-DDT, p, p′-DDD, p, p′-DDE and ∑ p, p′-DDTs). Positive nuggets were observed for all three DDT species and ∑ p, p′-DDTs in the four directions modeled. The Nugget–Sill-Ratio values, 0.62, 0.50, 0.63 and 0.58, for p, p′-DDT, p, p′-DDE, p, p′-DDD and ∑ p, p′-DDTs respectively, indicated that spatial distribution of DDTs residue in Tianjin topsoil could be characterized with high randomness, on average more than 50%, and less than 50% variance is explainable. Nevertheless, all DDT species have evident spatial structure. These spatial structures are anisotropic, showing more extended transportation along the E–W direction. Furthermore, the spatial structure of p, p'-DDE is more evident than the two other species, i.e. p, p′-DDE showed higher continuity in space than p, p′-DDT and p, p′-DDD, which signs a better mobilization of this species in the soil. Two-dimensional ordinary block kriging was applied to DDT concentrations dataset for mapping purposes. The mean errors were close to zero and the mean squared errors varied from 0.23 to 0.46 log units. The accuracy of interpolation proved to be acceptable for p, p′-DDE and ∑ p, p′-DDTs. However, for p, p′-DDT and p, p′-DDD, the interpolation results were accompanied by relatively high kriging errors, due to their high randomness in spatial distribution.