Increasing grain yield and water use of wheat in a rainfed Mediterranean type environment

Factorial experiments were conducted at eight sites in the central wheatbelt of Western Australia over two seasons. Time of sowing (mid-May, early June), cultivar (old tall, new semi-dwarf), nitrogen (N) fertilizer (- or +) and amount of seed sown (low and high) were combined as treatments, and grain yield, yield components, biomass, grain quality, water use, soil chemical and weather variables were measured. The aim was to increase grain yield by combining relevant agronomic inputs and increasing the seasonal water use or water use efficiency. Grain yields were increased by from 30 to over 100% by the combination of mid-May sowing, semi-dwarf cultivar, N fertilizer and increased seed level (high-inputs) compared to early June sowing, old tall cultivar, without N and lower seed level (low-inputs). The yield improvements mostly came from increased dry matter production at anthesis, largely due to increased applications of N and seed. Ear and kernel numbers were also increased by earlier sowing and N fertilizer and to a lesser extent by cultivar and increased weight of seed sown. Water use was increased at most sites, especially in the post-anthesis period and water use efficiency of grain production was increased at all sites. Soil evaporation was reduced by the high-input treatments and the low-input treatments did not use water supplies of > 250 mm efficiently in grain production. It was concluded that appropriate combinations of cultivar and agronomic practices can increase grain yields linearly up to about 5 t ha-1 at seasonal water use of about 400 mm, even in situations where considerable water stress occurs during grain filling. Grain protein concentration was generally increased and hectolitre weight and small grain sievings were not adversely affected by increasing agronomic inputs.