Preliminary Stepwise Multiple Linear Regression Method to Predict Cadmium and Zinc Uptake in Soybean

Abstract

In cadmium (Cd)-contaminated soils, the ability to predict Cd and zinc (Zn) uptake in crops as determined by readily measurable ‘‘indicators’’ of Cd and Zn phytoavailability would be a useful management tool to rapidly evaluate and prevent food chain Cd contamination risk. The preliminary stepwise multiple linear regression results presented here indicate that for the range of Cd/Zn cocontaminated soils tested, the prediction of Cd and Zn uptake in soybean bean are best achieved by 0.005 M DTPA-extractable Cd and Zn alone, which account for over 80% of the variance in soybean seed Zn and Cd. Furthermore, the inclusion of pHw (1:5), clay content (%); and organic carbon (OC) to the stepwise multiple linear regression equation had no significant impact on this prediction. However, for the soils and soybean bean tested, the results indicate that readily exchangeable (0.01 M CaCl2-extractable) and sorbed + organically bound (0.005 M DTPA-extractable) forms of Zn and Cd are dependent on soil pH. This may have major implications for the management of phytoavailable Cd and Zn in contaminated soils.