User Perspectives on the Application of Pattern-Scaled and Emulated Projections for Climate Change Impact Analysis

Abstract:

Pattern-scaled (PS) and emulated climate data are derived from climate model projections in order to offer a more comprehensive, though approximate, representation of uncertainties in future climate. These data are widely used by analysts studying climate change impacts, adaptation and mitigation, often in the form of "scenario generator" tools. Some of the advantages they provide include: rapid and flexible access to climate projections; representation of a range of uncertainties in projections, including multiple radiative forcings, time periods, variables and regional patterns; and (in some cases) access to probabilistic projections. They also present challenges to analysts, including: access to projections consistent with forcing levels not provided from climate models; selection of a manageable sample of projections; determination of a reference or baseline climate; issues around downscaling of projections; representation of future changes in climate variability at different scales; provision of variables other than temperature and precipitation; ensuring internal consistency between variables being projected; and consideration of climate projections alongside other uncertainties (such as socioeconomic futures and impact model projections).   In view of these advantages and challenges, and mindful of the potential needs of impact analysts affecting the design of forthcoming CMIP6 climate model simulation runs, this presentation offers some user perspectives on the kinds of data and information that may be required in the coming decades. The main messages have been developed out of discussions around this theme in a Workshop held at the National Center for Atmospheric Research, Boulder, Co in April 2014. In brief, some basic needs to be explored include: (1) evaluation of the PS/emulation techniques using formal statistical techniques, (2) compilation and critical evaluation of tools applying PS and emulation to scenario provision, (3) guidance on the use of PS/emulated outputs (and in the context of the new RCP/SSP global scenarios framework), and (4) examples of good practice application. Finally, some possible next steps for the research community will be suggested.