*fernanda.dreccer@csiro.au

Australia’s climate has warmed by 1.47 ± 0.24 °C since 1910, accompanied by a decline in cropping season rainfall in the southwest, particularly at sowing, stalling water-limited wheat yields (Hochman et al., 2017). Within the restrictions imposed by water availability, farmers try to minimise direct damage by frost and heat to yield, by choosing planting dates and cultivar combinations that lead to flowering within a thermally safe window. In this study we used historic weather records, knowledge of genotypic variation in phenology and a simulation model (APSIM NextGen) to investigate three topics.

Firstly, we asked if there had been historic changes in the thermally safe flowering window, in length and timing of occurrence. Using weather records in a 5km x 5km grid covering the cropping belt, we investigated the trends in the occurrence of the last day of frost (90th percentile, T<0°C) and first day of heat (20th percentile, T>32°C).

We show that a significant proportion of the wheatbelt is experiencing later frosts, advancing by ca. 1 day year-1 or more since 1970, extending a previous study (Zheng et al., 2015). At the same time, the first day of heat, is uniformly occurring earlier across the wheatbelt, by half a day or more per year. This has led to a significant shortening in the length of the thermally safe window in most regions. Secondly, we asked how historic changes in weather have impacted on the yield of cultivars of different phenology.

To this end, APSIM NextGen was used to simulate phenology and water limited yield of ca. 200 cultivars in sowing dates between 1st April and 29th July in the whole wheat belt. Across maturity groups and regions, the total cycle length is significantly shorter and yield potential has declined over time, with shorter cycles more impacted than long ones.

Finally, we characterized frost and heat patterns in terms of frequency and intensity during the critical period for yield formation (300°Cd before and 120°Cd after flowering) to analyze the likelihood of success of avoidance and sensitivity reduction targets.

Taking together all this evidence, we argue that as crop phenology is already a management tool to minimize losses (Hunt et al., 2019), future research should focus on addressing mechanisms to reduce the sensitivity to extreme temperatures during the critical period. Choosing frost as an example, we discuss to what degree sensitivity to frost would need to change during the reproductive period to impact significantly on yield loss and the boundaries of the thermally safe window for a given risk level.