B1. Distinguishing regional climate changes from natural variability

The ability of GCMs participating in CMIP5 (Coupled Model Intercomparison Project Phase 5) to simulate climate extremes has recently been evaluated by Sillmann et al. (2012a, b). The CMIP5 models were found to simulate temperature extremes about as well as previous CMIP3 models, but demonstrate improvement for precipitation extremes. Further, the study of Murdock et al. (2012), using reanalysis-driven RCMs participating in NARCCAP, has revealed smaller biases in temperature extremes in the Columbia River Basin. An increasing body of work demonstrates that large-scale circulation features affect the frequency and intensity of temperature (e.g., Sillmann et al., 2011) and precipitation (e.g., Zhang et al., 2010) extremes. Understanding the role of circulation variations on temperature and precipitation extremes is important both from a scientific and risk management perspective. We will therefore first study the extent to which circulation variability simulated within the CRCM North American domain affects the occurrence and intensity of temperature and precipitation extremes. We will also examine the dependence of these effects on resolution, physics package, or nesting strategy, for several RCM simulations available for North America.

This project will also consider the potential for establishing and projecting changes in engineering design values, particularly design rainfall amounts that are used in infrastructure development. One key design rainfall product is the Probable Maximum Precipitation (PMP), defined as “the greatest depth of precipitation for a given duration meteorologically possible for a design watershed of a given storm area at a particular location at a particular time of year” (WMO, 2009). Estimation of PMP requires information on convergence and vertical motion and atmospheric water vapor. We will examine past and future changes in those quantities and estimated PMP at different durations, as well as intensity-duration-frequency (IDF) curves.

The scaling of precipitation change with temperature change (sometimes called the hydrologic sensitivity) is an important diagnostic for evaluating climate models and assessing the credibility of their projections, particularly with respect to extremes. The Clausius-Clapeyron (CC) relation predicts that the planetary sensitivity is about 6.5%/K (Boer et al., 1993; Allen and Ingram, 2002). GCMs simulate a substantially lower sensitivity in mean precipitation, but show sensitivities in extreme precipitation that correspond reasonably well with expectations (see Kharin et al, 2007, 2012). However, the sensitivity of GCM-simulated extremes over extra-tropical land areas, including Canada, tends to be less than the planetary sensitivity, and less than the CC prediction (Kharin et al., 2012). This contrasts with Westra et al. (2012) who show that the sensitivity of extreme precipitation from station data is about 7%/K on average over global land areas. We will evaluate the hydrologic sensitivity of the RCM family of models over North America at various points in the precipitation distribution, to establish how the sensitivity of extreme precipitation depends upon model resolution and physics package, and the dependence of sensitivity on the time-scale of temperature variation.