Deploying high-throughput phenomics (HTP) into breeding programs requires more than the accelerated measurement of ‘existing’ traits. Spatial resolution and measurement cost are another two factors significantly affecting HTP’s application in practical breeding. Higher resolution measurements allow the use of ‘within-plot’ measurements of traits (i.e. crop cover, height etc.) to better account for spatial variability in trials, while low-cost methods may allow traits to be measured more frequently or across a wider range of environments. In parallel, the crop model capability to monitoring and modelling of environment conditions provides the opportunity to better interpret genotype-by-environment interactions in terms of responses to the environment. To better predict crop adaptation, new methods may be required to determine how to aggregate the value of these traits obtained from these available technologies, including into a complete modelled system.

The combination of HTP measurements across or within platforms (i.e. glasshouse, field etc) can be used to explain genotype responses to environment. These ‘derived traits’ extracted from HTP measurement may be used to parameterise crop simulation models in order to simulate environments and/or components of genotypic differences. Derived traits’ can be determined across multiple scales (spatial, temporal, multi-experiments), as for example the rate of leaf growth in response to temperature or traits affecting light interception. In addition, statistical models, such as splines, may be applied to temporal sequences of traits to determine the times when the greatest variance might be observed across treatments. When well-parameterised from phenotypic data, complete modelled systems of breeding populations and their underlying physiological and genetic drivers should provide a basis for assessing the value of phenomic measurements in accelerating yield advancement. This paper discusses some of our experiences in these areas and future opportunities in wheat.