Comparing scenarios of canadian daily climate extremes derived using a weather generator

Stochastic weather generators are widely used for developing local climate scenarios from large-scale climate change scenarios simulated by GCMs. We used AAFC-WG, a stochastic weather generator developed at Agriculture and Agri-Food Canada (AAFC), to develop daily climate scenarios for the time period 2040-2069 for the agricultural regions of Canada, based on 4 climate change simulations conducted by global climate models CGCM3, HadCM3, ECHAM5/MPI-OM and CSIRO-Mk3.5. Annual and growing season (1 May to 30 September) extreme daily values of maximum temperature, minimum temperature and precipitation were estimated from these scenarios on fine grids (0.5° latitude × 0.5° longitude) and from direct daily global climate model (GCM) outputs on coarse grids (̃2.8° latitude × 2.8° longitude for CGCM3, 2.5° latitude × 3.75° longitude for HadCM3, and ̃1.865° latitude × 1.875° longitude for ECHAM5/MPI-OM and CSIRO-Mk3.5). Compared with the corresponding values from direct GCM outputs, the extremes from daily climate scenarios generated by the AAFC-WG provided more detail at the finer spatial scale. Although the spatial patterns of the changes in climate extremes were often similar, actual values of the extremes from localized scenarios may be more reliable than those from direct GCM outputs since climate extremes under the present climate were not often realistically reproduced by the GCMs studied here. It is understood that the actual values are important as they provide the basis for climate change impact studies and the development of adaptation strategies. © Inter-Research 2010.