engleski

Spatial Patterns of Key Soil Properties and Trace Element Contents in Soils from the Mediterranean Agricultural Karst Catchment

The Vrana Basin, located in the central part of Croatian Adriatic Sea coast, is a typical cryptodepression formed by tectonic activities, with agricultural land and Lake Vrana situated in the very basin. The area is ecologically highly sensitive and it is characterised by intensive vegetable production in a karst environment where, due to geological and climatic conditions as well as anthropogenic influence, it is easy for harmful substances to enter the food chain or penetrate ground water or open watercourses. This threatens the largest Croatian lake, Lake Vrana, as well, which is a large potential source of water for agriculture, possibly also drinking water, as well as for tourism and leisure activities. Because it is one of the rare natural habitats of waterbirds as well as because of its fresh water springs and biodiversity, the Vrana Lake and its surroundings have been declared a nature park. To evaluate the impact of natural processes and agricultural practice on trace metals accumulation and behaviour in soils a multi-element geochemical survey was carried out. The main objective of this study was to identify the main factors that control the spatial distribution of soil trace elements (TEs) in topsoil from the Mediterranean agricultural karst catchment by using multivariate statistics and geostatistics. Topsoil (0-30 cm) samples were taken from 75 locations within the 6830 ha of pastures and agricultural fields in the Vrana Lake Basin applying a stratified random sampling scheme, based on regular 500x500 m grid, with exceptions of inability to access certain sampling points. Core samples were bulked to obtain a composite sample from each location. The observation sites were spatially referenced using GPS and data were stored in different GIS layers. Main soil physico-chemical characteristics were determined along with the TEs contents (Ca, Cd, Co, Cr, Cu, Fe, Hg, Mg, Mn, Ni, P, Pb, Zn) after aqua regia extraction. Reactive pools of Cd, Cu, Pb and Zn were assessed as well using 0.43 M HNO3 extraction. The resulting data were examined through principal component analysis (PCA) that is the multivariate technique most extensively used for processing the results of TEs concentrations in soils. PCA enables the interpretation of the mutual influence or the presence some common or opposite characteristics, such as soil and landscape properties or source of the elements. The analysis of large number of variables (pH, organic matter, calcium carbonate, available K2O and P2O5, trace element contents, particle-size distribution: sand, silt and clay) proved that the main factors controlling TEs distribution are soil texture and variations in parent material and weathering. Land use and agricultural practices have been identified as secondary factors affecting topsoil TEs distribution relating mostly to pH and organic carbon variations. A geochemical baseline model is used to quantify metal enrichments (Cd, Cu, Pb and Zn) in the topsoil and separate them from natural metal concentrations in soils (using Fe2O3). Regression analyses showed very meaningful relationship in which topsoil samples are enriched, to varying extents, as compared to subsoil samples. Also, there is a linear relationship between total metal concentrations and enrichments. Geostatistical interpolation using co-kriging method was applied to map topsoil TEs distribution over the study area. This method included Gaussian anamorphosis modelling for TEs which did not have normal distribution and variogram fitting for selected TEs. Both the results of PCA and co-kriging method provided evidence of strong spatial structures attributed to different natural processes. The methodology used has proved to be a useful tool to separate geological and anthropogenic causes of variation in soil heavy metal content and to identify common pollution sources.

karst environment; geochemical survey; trace elements; principal component analysis; geochemical baseline model; geostatistical interpolation

nije evidentirano