Compositional Nutrient Diagnosis of the Greenhouse Tomato

Abstract

Compositional nutrient diagnosis (CND) norms were computed from a diagnosis and recommendation integrated system (DRIS) databank as means and sds of CND multinutrient ratios and CND factors characterizing the high-yielding subpopulation of greenhouse tomato (Lycopersicon esculentum Mill.) plants during the 3- to. 8-cluster developmental stages. A CND multinutrient ratio is a row-centered logratio corrected for the bounded-sum constraint to 100% of all components, including a filling value between known components and 100%. CND multinutrient ratios for N, P, K, Ca, and Mg were reduced to three dimensions by principal component analysis (PCA). Varimax-rotated factors–(N-K-Ca+), (Ca+Mg-), and (P–)—were designated as contrasts describing nutrient status and equilibria in diagnostic tissues. CND nutrient indexes for N, P, K, Ca, and Mg were highly correlated with their DRIS counterparts (r = 0.964 to 0.987), using 20 independent observations. Since CND is compatible with PCA, CND could expand DRIS to a multivariate diagnostic approach projecting structured information on nutrient data into a Euclidean space. A critical sphere specific to a developmental stage could delineate the high-yielding subpopulation for diagnostic purposes. The critical radius defined using a probabilistic approach to capture the high yielders should be validated with independent fertilizer experiments.