Utilisation of agricultural databases for statistical evaluation of yields of barley and wheat in relation to soil variables and management practices

The objective of this study was to utilise existing databases for evaluating effects of soil and management factors on the yields of barley and wheat in Norway. In a multiple regression model, 17% and 19% of the variation of barley and wheat yield, respectively, was explained by 15 significant predictors. The highest-ranking predictors for barley yield were percent area of wheat (indicating more favourable climatic condition), irrigation, pH and silt. Other significant predictors were agricultural education, CV pH, man-hours on farm, (pH) 2 , loam, ln P-AL, ln area of farm, number of fields, (ln P-AL) 2 , (ln area) 2 and percent grass area. For wheat yield, the highest-ranking predictors were irrigation, percent winter wheat, pH and ln area of farm. Other significant predictors were CV pH, ln P-AL, (pH) 2 , loam, number of fields, (ln P-AL) 2 , man-hours per ha, percent grass area, agricultural education, other grain and (ln area) 2 .