Modelling genotype × environment × management interactions to improve yield, water use efficiency and grain protein in wheat

Abstract

In breeding trials the complexity of interactions of genotypes (G) with their environment (E) requires extensive field experiments at many locations and seasons to test new traits. Management (M) treatments such as planting density and fertilizer supply are often neglected in such experiments. Moreover, several traits (e.g., increased specific leaf area and faster early root growth) are often changed simultaneously in new genotypes (e.g., early vigour lines) with the contribution of individual traits remaining uncertain. Using a well-tested simulation model can assist in exploring the impact of new genotypes and the contribution of individual physiological traits on yield by simulating genotypes across many locations, seasons and a range of management options. Recent advances in simulating grain protein concentrations in wheat also allow the analysis of the impact of yield-related physiological traits on grain protein percentage. Simulation analyses have shown that trait effects on yield are location- and season-specific. The contribution of traits to yield varies with growing-season rainfall and soil types. In Mediterraneantype environments, traits of increased specific leaf area and faster early root growth improve grain yields and water use efficiency on low water-holding-capacity soils, but can be detrimental on better waterholding- capacity soils. Increased specific leaf area is often only expressed if N management is adjusted while the positive effect of faster early root growth in sandy soils diminishes with increasing N supply. The sum of individual traits incorporated in early vigour (that is, increased specific leaf area, faster early root growth, earliness, reduced radiation use efficiency) and increased transpiration efficiency is often not the same as the effects when the traits are combined. Another trait for an increased capacity for storing water-soluble carbohydrates can be beneficial for yield and water use efficiency in terminal-drought environments, but is not as effective when growth is limited by water or nutrients during anthesis. The response of grain protein to yield-related trait modification (such as increased grain-filling rate) is usually negatively related to the response to grain yield. However, under combined water and N limitation, simulations indicated that this negative linear relationship can become non-linear, suggesting that a lower potential yield in such environments might improve grain protein concentrations without reducing the attainable yield