Grain nutrient removal per bushel of yield appears to vary from year to year and region to region based on samples collected at a range of sites around Missouri from 2006 to 2008 for corn (PDF), soybean (PDF), and wheat (PDF).
The variability seen means that calculated grain removal values for N, P, and K will not always be accurate. Most fertilizer recommendations use these removal values as part of the basis for rate decisions. Current National Research Council feed grain values for N, which are currently the basis for University of Missouri nutrient removal calculations, are shown in golden brown.
Applying nutrients at rates to replace those removed by the crop is a sound approach, but putting too much faith in the ‘book values’ for removal could lead to problems. Soil testing should be used as another source of information about the nutrient status of soils.