ENRICHMENT AND DISTRIBUTION OF ELEMENTS IN FLOODPLAIN SOILS OF THE BITTERFELD INDUSTRIAL-AREA - STUDIED BY NEUTRON-ACTIVATION ANALYSIS

Bitterfeld is an industrial area of central Germany which has been exp osed to extraordinarily high concentrations of anthropogenic environme ntal pollutants for many decades. High concentrations of organic (e.g. PCB and hydrocarbons) as well as inorganic (e.g. heavy metals) pollut ants are present in the soils. The major objectives of this study were the quantification of inorganic components in the soils and the verif ication of the analytical results through different analytical methods . The origin of the anthropogenic heavy metal input in the study area is twofold: first, at present, industrially produced fly ash is deposi ted upon hood-plain sediment and second, sedimentation processes durin g the flood periods of the Mulde river over the past number of years. Flood-plain soil profiles were studied and characterized by four horiz ons, designated as A(h), B-v, Gor and D. Samples were taken at 10-cm i ntervals down to a total depth of 170 cm. All components studied were analyzed using activation methods with reactor neutrons and accelerato r-produced high energy bremsstrahlung photons. Thirty-four elements an d their vertical distributions throughout the profile were examined. T he results are interpreted in relation to geochemical and soil-related physical parameters, e.g. pH, clay content, density and organic compo nents. We conclude that for a number of elements, estimations of the a nthropogenic input and their mobility behaviour can be made. For insta nce, the distribution characteristics of calcium, nickel, zinc, molybd enum, arsenic, antimony, lead and uranium indicate mainly an anthropog enic input, whereas the distribution of scandium, titanium, rubidium, yttrium, zirconium, tin, cesium and the rare earths are obviously to a large extent of geogenic origin. For the classification of the depth distribution of the elements studied, a hierarchic cluster analysis wa s carried out, the results of which were correlated subsequently with the help of canonic discriminance functions. (C) 1997 Elsevier Science B.V.