Extreme events can have large impacts on society. It is important to understand how circulation features are associated with extremes and to evaluate how skilfully this relationship is simulated by climate models in the current climate. The relationship between circulation features (such as atmospheric blocking) and cold extremes in high-resolution regional climate models (RCMs) is an interesting question because these circulation features can be located outside or near the edge of the typical RCM domain for North America.
‘Atmospheric blocking’ refers to an interruption of prevailing westerly winds and is often associated with a persistent region of high-pressure. Globally, blocking in the northern hemisphere mid-latitudes is most frequent in the eastern Atlantic and the eastern Pacific Oceans. We use an index of blocking frequency that counts the percentage of days per month when longitudes in the EP sector are classified as blocked. We define blocking based on geopotential height anomalies from ERA-Interim reanalysis (the same reanalysis that provides the lateral boundary conditions for the RCMs).
Block maxima (such as the inverted coldest day of the year) follow a generalised extreme value distribution (GEV). The GEV is described by three parameters: 1. the location parameter (μ), which is similar to the mean, 2. the scale parameter (σ), which controls the width of the distribution and 3. the shape parameter (ξ) that describes to which family of distributions the data belongs. The influence of atmospheric blocking on minimum temperature is quantified by including the index of blocking frequency as a covariate on the location and scale parameters.
In observations and reanalysis, blocking has a significant influence on minimum temperature over most of the continent. This pattern is well reproduced in the RCMs, indicating that the models are able to simulate the spatial extent of blocking influence.
A ‘20-year return value’ is a temperature that we expect to occur once every 20 years. We calculate the return values when blocking frequency is high and when it is low. The difference between these two numbers tell us the influence that blocking has on these “once-in-20-years” cold events. In the United States, a high-blocking regime is associated up to a 15°C decrease in the coldest day of the year, compared to low-blocking conditions. The opposite influence is found in Alaska, northern Canada and Mexico, where blocking is associated with warmer 20-year return values of minimum temperature (Figure).
We showed that atmospheric blocking in the eastern Pacific has a widespread and significant influence on the coldest days of the year. High-blocking is associated with substantially colder minimum temperatures in the United States with a smaller but significant warming effect in parts of Canada. The regional climate models used in this study were generally able to simulate the spatial pattern well but there were some discrepancies in the magnitude of the temperature response to blocking.
Read the paper for more details: http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-15-0493.1