Re‐evaluation reveals weaknesses of DRIS and sufficiency range diagnoses for wheat, corn, and alfalfa

Abstract

In prescient diagnostic analysis, conclusions about the need for fertilizer application based upon nutrient diagnostic methods are compared to independently‐determined correct diagnoses based upon yield responses to nutrient application. If a nutrient is diagnosed as being required (i.e., deficient or insufficient), the diagnosis is positive; no requirement is a negative diagnosis. Diagnoses are verified and classified as either true or false by their agreement with observed presence or absence of significant yield increases in response to application of the nutrient in question, A diagnostic method is considered acceptable only if: 1) at least 50% of all verifiable diagnoses are correct; 2) positive diagnoses are true more often than false; and 3) the net yield effect attributable to indicated nutrient treatments is positive. Using this approach, previously published Diagnosis and Recommendation Integrated System (DRIS) and published or derived concentration‐based diagnoses of wheat, corn and alfalfa were evaluated. DRIS and sufficiency range (SR) diagnoses for. wheat were acceptable based upon the above criteria for N, P and K, but only DRIS diagnosed S requirements satisfactorily. DRIS diagnoses of P and K in corn were acceptable, but N and S diagnoses exhibited marginal to unacceptable accuracy and yield responses; SR diagnoses were equal for P and K, and superior for N and S. For alfalfa, false positive N and K diagnoses at one location and false positive K and false negative P diagnoses at another location made accuracy and yield effects of DRIS unacceptable. These results demonstrate that neither SR nor DRIS diagnoses are consistently superior or even acceptable for wheat, com and alfalfa nutrient diagnoses.