“…Eco-physiological modelling has been widely used to resolve the complexity of grain yield under different environments (Soltani et al , 1999; Yin and Struik, 2010; Martre et al , 2011), by dissecting grain yield into its component traits or parameters. Most parameters in the model may be controlled genetically; therefore, eco-physiological models are believed to be able to quantify genotype–phenotype relationships for complex traits (Hammer et al , 2006; Bertin et al , 2010; Génard et al , 2016), using dynamic simulation on a daily or even shorter time-step basis. Unlike statistical approaches that require a large number of experiments (although on a single trait) to create a prediction model (Bustos-Korts et al , 2016), eco-physiological modelling can, in principle, rely on one or a few experiments for model parameterization because the prediction is made largely based on eco-physiological principles as captured by the models.…”