Simulating the impact of extreme heat and frost events on wheat production: the first steps

Abstract

In Australia, wheat production occurs on over 13 million hectares, producing on average 19 million tonnes of wheat per year. Extreme weather events, such as frost and heat shock (short period of very high temperatures (>35°C)), can reduce wheat yields and represent a substantial management challenge. Damage due to frost and heat shock is greatest at ear emergence and around anthesis causing significant reductions in grain number and yield potential. Heat shock can also significantly reduce grain weight during the grain fill period when the risk of heat is greatest. Paddock-based crop models currently used to simulate crop production do not adequately account for the impact of extreme weather events such as frost and heat shock on yield components. While it is feasible to construct crop modules which capture this impact by extreme short term weather events, we felt it was important to quantify the frequency and spatial extent of the problem. This was important in determining whether the frequency and extent of potential grain losses from extreme weather events warranted the inclusion of added parameter input complexity within crop models. By taking into account the interactions between climate and crop phenology we were able to categorise areas as frequently affected by either frost or heat shock, those areas affected by both heat and frost and finally those areas which were rarely affected by either. Strategies to reduce the risks of extreme events will potentially be different for each of these regions. This paper investigates the spatial extent of where there is potential for improvements to the grains industry by having crop models which account for extreme heat and frost impacts linked to the key phenological crop stages. By incorporating phenological crop development, initiated by the autumn-break our analysis has established the actual frequency of overlap between extreme events and key phenological stages each year. This is important in determining the value of developing heat and frost modules to incorporate into crop models. To quantify the risk frequency of extreme heat and frost events across southern Australia's wheat growing regions the Catchment Analysis Tool (CAT, DEPI Victoria) was used. The study area (ca. 68 million hectares) incorporated agricultural land within the 200-1000 mm annual rainfall region and was significantly larger than the actual area sown to wheat annually. Two key periods were considered (a) a two week period centered on anthesis (50% of crop flowering) and (b) grain fill, for mid-season wheat variety using 50 years historical climate data. Based on our assumption of sowing at the autumn break, the occurrence of frost around anthesis and extreme heat during the grain fill period were important both in terms of frequency of occurrence and spatial area affected. Across the study region approximately 27% (ca. 18.5 million hectares) had a greater than 1:3 chance of both frost and extreme heat occurring at key crop phenological times, while 29% of the study area was generally affected by heat only during grain fill and 32% was generally affect by frost only. Additionally 12% (ca. 8 million hectares) had a less than 1:3 frequency of both frost and extreme heat occurring at key crop phenological times.