Prediction of trace metal concentrations (Cd, Cu, Fe, Mn and Zn) in wheat grain from unpolluted agricultural soils
2013
Ivezić, Vladimir | Almås, Åsgeir Rossebø | Singh, Bal Ram | Lončarić, Zdenko
Regression modelling and statistical correlations are often developed to aid reasonable predictions of metal solubility and free metal ion in terrestrial ecosystems. In the present study, we test how regression model predictions of metal solubility and free metal ion correlate with metal concentrations in plants. We previously optimised a model for the prediction of metal solubility and free metal ion concentrations in soil solution from sites collected in the Danube basin of Eastern Croatia, containing total metal concentration (concentrated HNO₃), soil pH, soil organic matter (SOM) and dissolved organic carbon (DOC) as model parameters. In the present study, we report how our optimised models correlate with metal concentration in plants reported by Lončarić et al., which was based on a study from a smaller part of the same sampling area. Our default regression model and further optimised model predicted metal solubility and free metal ion concentrations. These predictions correlated well with Fe, Mn, Zn and Cd concentrations in wheat grain, but the different extraction methods and parameter optimisations affected the significance of correlations differently. Due to the low concentrations of total metal concentrations in soil and little variation in SOM, the soil pH was the only variable, in addition to the total metal controlling the metal solubility, free metal ion in soil water and also the concentrations in wheat grain. Results show that after optimisation for the few most important soil chemical variables regression models can predict metal concentrations in plants fairly well. For this particular study, optimisation for total metal, soil pH and perhaps also SOM and DOC are sufficient.
Показать больше [+] Меньше [-]Ключевые слова АГРОВОК
Библиографическая информация
Эту запись предоставил National Agricultural Library