Climate data and scenarios: synthesis of recent observation and modelling results, chapter 3.4.2

3.4.2 Statistically downscaled results from CMIP5 models

Statistical downscaling makes use of empirically-derived relationships between large and small scales, and allows for a range of relevant climate quantities to be estimated. An important underlying assumption is that the empirical relationships are unaltered by a changing climate. While this may be a limiting assumption, it is offset to some degree by the fact that these approaches reduce the effect of systematic biases that may be present in global and regional climate models. The reduction of systematic biases is essential for the projection of some extreme indicators that are based on threshold crossing, for example, heating or cooling degree days. Environment Canada has worked with the Pacific Climate Impacts Consortium (PCIC) to develop statistically downscaled climate scenarios based on the CMIP5 global climate projections and regional climate projections (NARCAPP and CorDEX^Footnote4). The projections for Canada are available via the PCIC Data Portal. Figure 14 shows the potential utility of statistical downscaling for projecting climate extremes. Projected changes in heating degree-days and cooling degree-days in Canada are shown for three future periods (see figure caption for further details).

Figure 14 - Illustration of potential utility of statistically downscaled projections of extremes. Projected changes in cooling (left panel) and heating (right panel) degree days (in degree-days) are shown for the 2016-2035 (top), 2046-2065 (middle), and 2081-2100 (bottom) periods. Projected changes are relative to the 1986-2005 mean estimated from the multi-model ensemble shown in Table 7 and downscaled using BCCAQ.