Wintertime Cold Extremes: Mechanisms and Teleconnections with the Stratosphere

Abstract

Wintertime cold air outbreaks are predicted to decline in frequency and intensity as a result of climate change, but in spite of a robust warming trend over the last few decades it is unclear whether this decline has been observed. Some studies have found no trend or even a slight increase in North American cold air outbreaks, which is particularly remarkable when one considers the enhanced warming signal of 2-3 degrees C already observed in the wintertime Arctic, where most air masses resulting in mid-latitude cold air outbreaks originate. But with evidence from warmer climates in the far distant past, we know that cold air outbreaks should decline with global warming. Fossils of frost-intolerant species dating back to the Eocene warm climate period and found in the interior of North America indicate that the wintertime temperature never dropped below freezing while the average temperature was only 10 degrees C warmer than it is today, implying that cold extremes warmed by 2-3 times the average. This suggests that some mechanism may be acting to maintain cold air outbreaks in the modern climate in spite of the overall warming trend. Climate models have long-standing problems matching proxy records for winter temperatures at high latitudes during the Eocene, indicating that such a mechanism may be missing or improperly represented in models. Winter weather can also be mediated by teleconnections with geographically and dynamically distinct features. The stratospheric polar vortex has been hypothesized to exert a downward influence on surface weather, and more specifically to affect the frequency of cold air outbreaks, but the time scale and nature of this influence remains elusive. My dissertation is therefore presented in two parts: chapters one and two concern the mechanisms driving cold air outbreaks in near-modern and paleo climates, while chapter three considers the nature of teleconnections between the troposphere and the stratosphere in winter.

The first chapter looks at cold air outbreaks in a pre-industrial climate. Using output from the Community Earth System Model (CESM), I identify hundreds of cold air outbreaks over the deep interior of North America. To understand how those cold air outbreaks developed, I follow the air masses backward in time to see where they came from and how they evolved as they were swept from the Arctic into the mid-latitudes. I discover a significant role for diabatic cooling of the air masses on their way to produce cold air outbreaks, where negative surface sensible heat fluxes are mixed upwards by turbulence to cool near-surface air masses. I also identify an abrupt cutoff in the surface temperature distribution at the freezing temperature, which may indicate a significant role for the latent heat released by surface water as it freezes in maintaining above-freezing temperatures in warmer climates. In the second chapter, I expand this analysis to a comparison between a pre-industrial and a much warmer (Eocene-like) climate scenario. Here I find that increases in Arctic temperatures at the origin of cold air masses are responsible for the suppression of cold air outbreaks in much warmer climates. Surprisingly, the net diabatic temperature change along these air masses is the same in the two climate scenarios, where a balance between longwave cooling and warming from boundary layer mixing is maintained as both effects become stronger in the warmer climate.

The third chapter investigates teleconnections between the troposphere and the stratosphere in both the upward and downward directions. I use Maximum Covariance Analysis, which is based on singular value decomposition, on pairs of tropospheric and stratospheric fields over 60 years of reanalysis output to identify both the time scales and the spatial patterns of covariability. I find that the greatest covariance between the troposphere and the stratosphere occurs when the surface precedes the stratosphere by up to 9 days. Unlike previous studies, which focused on the time scale itself, this analysis method also enables me to identify the surface precursor, a wave-1 pattern in sea level pressure that is followed by changes in stratospheric potential vorticity, zonal wind, and EP flux. A second sea level pressure anomaly, similar to the first but rotated about the pole by about 60 degrees, was also found to precede stratospheric EP flux variations 2-3 days later. Counter to previous studies, which have suggested an important role for the stratosphere in surface cold air outbreaks, I find little evidence for a downward influence on minimum surface temperatures.