Early Sowing Systems Can Adapt Australian Wheat Production to Rainfall Decline.

Australia is an important region for global food security, annually producing ~25 Mt of rainfed wheat, most of which is exported. Currently, spring wheat cultivars are established on late autumn (April-May) rains, and flower at an optimal time in early spring defined by increasing incidence of drought and heat and decreasing frost risk. Australian wheat yield has plateaued since 1990 due to increasing spring temperatures and reduced April-May rainfall delaying establishment. If Australia is to increase wheat production in line with projected global population growth and demand, an improvement in yield is required. To this end, we compared a system of sowing winter genotypes in an unusually early sowing window (from early March, at which time rainfall has not declined), with the current practice of establishing spring cultivars in early- to mid-May. We grew near-isogenic lines of wheat which vary in alleles of major development genes in 24 experiments with multiple sowing dates, at 16 sites across the Australian wheat belt from 2012 to 2015. We found the yield advantage of the early sowing strategy over current practice was 0.7 t/ha (24%) in the Mediterranean low rainfall regions of SW and SE Australia, and 1.2 t/ha (30%) in the higher rainfall temperate regions of SE Australia. There was no advantage in the summer rainfall dominant regions of NE Australia. To evaluate the impact of potential practice change at the farm level and extend these results across multiple sites and seasons, we conducted simulation experiments using the APSIM crop production model. In SE and SW Australia, adopting the early sowing strategy increased predicted average farm wheat yield by 0.7 t/ha (range 0-2.5 t/ha) or 27%. Spatially interpolating the optimal strategy (i.e. retaining current practice where no benefit was demonstrated) showed a mean national benefit of 0.5 t/ha (s.d. = 0.4).