Dan Keyser: The Relentless Divider

The ubiquitous relationship between organized frontal zones and a large fraction of the cloud and precipitation distribution in extratropical cyclones, first introduced in the Norwegian Cyclone Model, is physically tied to the process of frontogenesis. This important dynamical idea was hinted at by Sutcliffe (1939), formalized by Petterssen (1940) and convincingly demonstrated as a dynamic principle by Sawyer (1955) and Eliassen (1962). In each of these studies changes in the magnitude of the potential temperature gradient were revealed to empower the production of the cloud and precipitation generating vertical motions. In the mid-1980s, Dan Keyser and Richard Reed (independently) began to consider the rate of change of the vector potential temperature gradient and thus introduced the notion of rotational frontogenesis to supplement the traditional scalar (or magnitude) frontogenesis. Application of this notion to the quasi-geostrophic Q-vector, led Keyser and colleagues to partition Q into its along- and across-isentrope components. The result was the first-ever glimpse at the vertical circulation associated with rotational quasi-geostrophic frontogenesis. This simple act of curiosity has led to the development of important new insights into the nature of frontogenesis, occluded cyclones, the process of upper frontogenesis and other advances. In this talk I will review some of the new insights made possible by the ingenious partitions imagined by Dan. I leave it to others to consider his innovative psi vector which allows separation of a three-dimensional vertical circulation into two two-dimensional components thus allowing objective assessment of the extent to which the overall vertical circulation is oriented in a preferred direction.