The Norwegian School, founded in the early 20th century by Wilhelm Bjerknes, is widely credited with marking the beginning of the modern study of meteorology. And yet, of the two cornerstones of the Norwegian School – air mass and frontal analysis - the second has been greatly modified from the original concept, while the first has been largely discarded. The Norwegian idea of a front as a near discontinuity extending through the whole depth of the troposphere and giving rise to cyclones and anticyclones, has been replaced by the concept of a baroclinic continuum whose instability produces baroclinic waves, whose deformation fields in turn concentrate temperature gradients – but only at the surface and around the tropopause. Without definite and extensive frontal boundaries, the concept of air masses becomes amorphous, and the pan-20th century shift in emphasis from diabatic influences on air mass evolution to adiabatic dynamics sealed its fate.

I this talk, I argue that we went too far. A glance at almost any sounding in the wintertime Canadian or Asian arctic shows the profound influence of radiative cooling in from the surface to as high as 500 mb, while a time series of such soundings reveals great spatial and temporal variability in radiative cooling. The existence of strong cooling through deep layers casts suspicion on purely adiabatic theories of blocking and suggests that seasonal prediction, if it is to have any success, will have to account for the subtleties of arctic air mass formation.

The strong radiative cooling that often dominates the wintertime lower troposphere over continental interiors produces near-surface air with potential vorticity (PV) more characteristic of the stratosphere. When such air surges equatorward, its leading edge tends to form PV fronts that are in many ways different from the idealized constant PV frontal zones described so well by the Hoskins-Bretherton results. In particular, near discontinuities in wind and temperature gradient can be present well above the surface and away from the tropopause. I will show some examples of PV fronts and present some simple theoretical ideas to describe them.