Impacts on winter wheat yield in Germany at 1º, 1.5º, 2º, 2.5º and 3º global warming

14:00 Wednesday 29 May

SS026 • OC152

Room S10


Michael Peichl (Germany) 1; Stephan Thober (Germany) 1; Andreas Marx (Germany) 1

1 – Helmholtz Centre for Environmental Research

After maize, wheat is the second most common cereal in the world and Germany is the ninth largest producer. In this way, wheat produced in Germany makes a significant contribution to the global security of food supplies. Here we estimate the effects of climate change on winter wheat in Germany for a 1, 1.5, 2, 2.5 and 3 K heating scenario.

A special challenge in the modelling of winter wheat is the long vegetation period in Germany. This usually leads to less predictability of statistical models than for crops with shorter growing seasons such as maize. Here, we integrate extreme weather conditions that influence winter wheat growth. The used extremes and their threshold values are derived and validated based on literature research and a co-production framework. In addition to temperature measurements such as extreme heat, which are used as standard in statistical approaches, we also consider the measures of extreme cold and water supply. This is particularly relevant for regions in the northern hemisphere for which water shortages are a limiting factor in contrast to high temperatures. Another focus is the modelling of the interaction between the individual growth phases. For this we use recursive partitioning methods, which are especially suited to model non-linearities and interactions between variables. In addition, these approaches allow a hierarchical selection of features. This is important for the design of adaptation measures, as knowledge of key factors is particularly important in this context. The models are estimated at district level for the period 1999–2017.

For climate projections a new data is used which has been generated within the EdgE project which consists of daily temperature and precipitation data as well as monthly soil moisture data with different soil layers on 5 km grid. Meteorological values, forced by three RCPs (RCP2.6, RCP6.0 and RCP8.5), are obtained for the period 1950-2099 obtained from five CMIP5 GCMs (HadGEM2-ES, IPSL-CM5A-LR, MIROC-ESM-CHEM, GFDL-ESM2M and NorESM1-M). Those data are then used as input to two hydrological models (mHM and PCR-GLOBWB) and two land-surface models (Noah-MP and VIC) to simulate soil moisture. The data are aggregated to administrative district level. In this study, the global warming values for 1, 1.5, 2, 2.5 and 3 K are determined using a time sampling method. Preindustrial warming between the periods 1881–1910 and 1971–2000 is subtracted from the warming levels for determining the 30-year periods for specific global warming.