Patterns of Interannual Variability in the Northern Hemisphere Wintertime 850 mb Temperature Field

View More View Less
  • 1 Atmospheric and Environmental Research, Inc., Cambridge, Massachusetts
  • | 2 University of Lisbon, Lisbon, Portugal
© Get Permissions
Full access

Abstract

Interannual fluctuations of observed winter seasonal mean 850 mb temperatures over the Northern Hemisphere during 1958–73 are documented and compared with midtropospheric height variations. Interannual temperature variance maxima are found over the Eurasian and North American continents, in striking contrast to the height field which exhibits variance maxima over the midlatitude oceans. Patterns of interannual variability are defined objectively using eigenvector analysis. The first two spatial eigenvectors of temperature variability describe hemisphere-scale patterns. The gravest eigenvector contains elements of the Eurasian and Pacific–North American (PNA) height patterns defined in earlier studies. One-point correlation maps confirm the strong positive correlation between temperature fluctuations over Siberia and western Canada found in the first eigenvector but indicate that other elements of intercontinentality are not so strong as the eigenvectors suggest.

To isolate more regionalized patterns of variability, therefore, the leading temperature eigenvectors are subjected to varimax rotation. The leading rotated pattern contains North American centers coincident with the PNA height pattern and an additional Caribbean center, and its temporal fluctuations are highly correlated with a PNA index derived from 500 mb height anomalies. Over the Asian continent, two temperature patterns are found that incorporate north–south anomaly dipoles not clearly depicted in height patterns. One of these patterns also describes the tendency for positive correlation between temperatures over the Gulf of Alaska and Siberia. Another pattern contains two centers over Europe and a broad center across western and central Asia. A fifth rotated pattern describes temperature fluctuations associated with the North Atlantic Oscillation in the sea level pressure field.

Time series associated with patterns containing centers over Siberia and northwest North America are also correlated with interannual fluctuations of hemisphere-averaged temperature. In particular, temporal fluctuations of the PNA height pattern or the leading rotated temperature pattern are significantly correlated with hemisphere-mean seasonal temperature anomalies at 850 mb and at the surface.

The roles of anomalous temperature advection and transient eddy heat flux divergence in the heat balance of seasonal temperature anomalies are examined using correlation statistics. Over much of the hemisphere, seasonal anomalies of temperature advection appear to maintain seasonal temperature anomalies, while eddy heat flux divergence anomalies tend to dissipate them. However, advection anomalies are poorly correlated with temperature anomalies over the Asian continent west of about 90°E, so another term in the heat balance must play a dominant role. We speculate that radiative forcing due to snow cover anomalies may be important in this region.

Abstract

Interannual fluctuations of observed winter seasonal mean 850 mb temperatures over the Northern Hemisphere during 1958–73 are documented and compared with midtropospheric height variations. Interannual temperature variance maxima are found over the Eurasian and North American continents, in striking contrast to the height field which exhibits variance maxima over the midlatitude oceans. Patterns of interannual variability are defined objectively using eigenvector analysis. The first two spatial eigenvectors of temperature variability describe hemisphere-scale patterns. The gravest eigenvector contains elements of the Eurasian and Pacific–North American (PNA) height patterns defined in earlier studies. One-point correlation maps confirm the strong positive correlation between temperature fluctuations over Siberia and western Canada found in the first eigenvector but indicate that other elements of intercontinentality are not so strong as the eigenvectors suggest.

To isolate more regionalized patterns of variability, therefore, the leading temperature eigenvectors are subjected to varimax rotation. The leading rotated pattern contains North American centers coincident with the PNA height pattern and an additional Caribbean center, and its temporal fluctuations are highly correlated with a PNA index derived from 500 mb height anomalies. Over the Asian continent, two temperature patterns are found that incorporate north–south anomaly dipoles not clearly depicted in height patterns. One of these patterns also describes the tendency for positive correlation between temperatures over the Gulf of Alaska and Siberia. Another pattern contains two centers over Europe and a broad center across western and central Asia. A fifth rotated pattern describes temperature fluctuations associated with the North Atlantic Oscillation in the sea level pressure field.

Time series associated with patterns containing centers over Siberia and northwest North America are also correlated with interannual fluctuations of hemisphere-averaged temperature. In particular, temporal fluctuations of the PNA height pattern or the leading rotated temperature pattern are significantly correlated with hemisphere-mean seasonal temperature anomalies at 850 mb and at the surface.

The roles of anomalous temperature advection and transient eddy heat flux divergence in the heat balance of seasonal temperature anomalies are examined using correlation statistics. Over much of the hemisphere, seasonal anomalies of temperature advection appear to maintain seasonal temperature anomalies, while eddy heat flux divergence anomalies tend to dissipate them. However, advection anomalies are poorly correlated with temperature anomalies over the Asian continent west of about 90°E, so another term in the heat balance must play a dominant role. We speculate that radiative forcing due to snow cover anomalies may be important in this region.

Save