Predicting Zn and Cu status in cereals – potential for a multiple regression model using soil parameters

Abstract

Prediction of micronutrient deficiency normally involves soil extraction tests and correlations between soil extractable micronutrients and their concentration in plant tissues. The present study reports that the inclusion of soil properties in addition to soil extraction data represents a powerful tool for improving models for a simple determination of phytoavailable Zn and Cu. Soil and plant samples were collected in spring 2000 from 19 cereal fields in Central and Southeast Norway at sites with known micronutrient deficiency. Based on the sample information, the model approach used stepwise regression, applying actual plant Zn and Cu concentration as the response, and soil extractable Zn/Cu and physico-chemical properties of the soil as model predictors. A model involving citric acid extractable Zn coupled with selected soil properties could predict the concentration of Zn in plant tissues of barley and oats grown under field conditions with a significantly high precision (R²=0·96, P<0·001). The soil parameters included in the model, and their importance for phytoavailability of Zn, were in the order: citric acid extractable Zn>clay content>total Zn>pH>total C>total N. Inclusion of DTPA extractable Zn in the model, rather than citric acid, resulted in a slightly lower precision in the prediction of phytoavailable Zn (R²=0·86, P<0·001). A corresponding model for phytoavailable Cu, determined by DTPA and soil parameters, gave less precise prediction (R²=0·59, P<0·01). DTPA was more suitable for predicting phytoavailable Cu than citric acid. Plant Mn concentration could not be correlated to any of the extracting agents.