Effect of using different methods in the construction of climate change scenarios: Examples from Europe

To investigate the impacts of future climate change on agriculture in Europe and to aid the development of agricultural adaptation strategies, several scenarios of climate change were constructed and provided as input to crop-climate models at both broad and site-specific scales. These scenarios were based on results from a number of global climate model (GCM) experiments, including both equilibrium and transient climate change experiments. All GCMs provide 'internally consistent' scenarios of climate change, but at coarse horizontal spatial resolutions. Results from equilibrium experiments give no indication of the rate of climate change, whereas transient experiments simulate time-dependent climate change. Broad-scale climate change scenarios for Europe were constructed at 0.5 degrees latitude/longitude resolution using a Gaussian space-filtering routine. Site-specific scenarios at the daily timescale required by crop-growth simulation models were constructed using a stochastic weather generator in conjunction with GCM-derived information about possible future climate change. A regression technique to downscale broad-scale changes in temperature and precipitation means to specific sites was carried out at 2 European sites, Rothamsted, UK, and Seville, Spain. At other European sites, where crop-climate models were being run, the GCM-derived grid-box changes were used without downscaling. Use of a weather generator also enabled GCM-derived changes in climate variability to be incorporated into the scenarios by using the grid-box changes in the means and variances of the climate variables to perturb the parameters of the stochastic weather generator directly. The implications of using these different methods of climate change scenario construction for the resulting scenarios are discussed.