Using Regional Climate Model output to force wheat simulations.

Regional Climate Model (RCM) output is increasingly being used to provide forcing data for crop simulation models in studies of the impacts of climate change on agriculture. This presentation describes investigations into the use of RCM output for this purpose. Specifically, the benefits of higher RCM resolution and of bias correction are addressed. The investigations focus on the wheat belt of New South Wales (NSW) in southeast Australia and involve running wheat simulations for 22 sites distributed across this region. Wheat simulations forced with climate model data and observational climate data are performed. An initial study focuses on simulated wheat yields for a recent climatological time period. “Errors” in simulated wheat yields for this period are defined as differences between wheat yields output by simulations forced with climate model output and wheat yields output by simulations forced with observational data. This definition allows wheat-­‐relevant errors in the simulation of the climate to be isolated from errors in simulated wheat yields arising from imperfections of the wheat model. The initial study considers the benefits of higher resolution climate model simulations by analysing output from a single set of nested climate model simulations – a Global Climate Model (GCM) simulation and WRF RCM simulations at resolutions of ~50km and ~10km. The study focuses on raw climate model output, as opposed to bias-­‐ corrected data. A subsequent study examines the ensemble of RCM simulations performed as part of the NSW and Australian Capital Territory Regional Climate Modelling project (NARCliM) to support work on climate impacts and adaptation in the region. The NARCliM ensemble comprises 12 sets of ~50km and ~10km nested WRF RCM simulations (four different GCMs providing forcing for three different WRF configurations) for each of the 1990-­‐2009, 2020-­‐2039 and 2060-­‐2079 periods. Increasing atmospheric greenhouse concentrations throughout the 21st century (consistent with the IPCC SRES A2 scenario) are assumed. The analysis extends to include both errors in wheat yields simulated for the recent period and simulated changes in wheat yields during the 21st century under the increasing greenhouse forcing of the climate. The effects of both higher RCM resolution and bias correction on wheat yield errors and simulated future changes in wheat yields are considered.