Forward Phenomics: Integrating Plant Ontogeny and Architecture to Predict Crop Yield.

Digitally repeated measurements of quantitative traits, and visual scores of categorical or qualitative traits on canopies of whole plots, single plants, leaves, reproductive modules and seed samples of  a large number of grain and oilseed crop genotypes in replicated field experiments were used to (1) identify plant traits that can integrate the effect of time (ontogeny, quantified by growing degree days for each crop species or genotype) and space (architecture, described by architectural and fractal dimensions of single plant samples at specific growth stages) on seed yield of grain crops, and on oil content and oil yield of oilseed crops; and (2) develop pyramiding phenotyping models based on field and laboratory of phenotypic and agronomic data while accounting for fixed and random sources of variation when interpreting components of phenotypic variance. A number of soil-based variables (soil series, physical and chemical soil analyses, apparent electrical conductivity, and pH) were measured throughout the growing season and were used as covariates to adjust for spatial variation; and together with weather variables were used in predicting crop yield and its spatial variability. The forward phenomic process identified and employed a minimum set of crop-specific phenotypic traits that can be digitally recorded at key growth stages of plant ontogeny with maximal discrimination power between crop genotypes within species and for reliable yield prediction.