Regression model for prediction availability of essential heavy metals in soils

摘要

The aim of paper was to compare relationshipsof total and plant available essential heavy metals in soils considering soil acidity and humus impacts. Regression models should provide additional data (EDTA-extractable fraction) from available data (heavy metal total content, soil pH). The 60 soil samples were analysed for pH, humus, phosphorus, potassium, and additionally for essential heavy metals (Fe, Mn, Zn, Cu, Ni) by extraction by aqua regia and by ethylenediaminetetraacetic acid (EDTA). The heavy metal with highest total content was iron, then manganese, zinc, nickel, and copper regardless of soil pH. The order of EDTA-extracted elements changed with copper jumping from last to third place, but the percentage of total content extracted by EDTAwas quite opposite: Cu, Ni, Zn, Mn, and Fe. Soil pH impacted on EDTA-extractable fraction and the regression models can predict the EDTA-extractable fraction of Zn, Cu and Ni with acceptable average errors, but not for Fe and Mn. The obligate input datafor models are total soil content of certain heavy metal and pH. Model errors decreased using humus content as additional input data. The lowest model errors were for copper in acid soils and nickel in calcareous soils.