The Highly Curved Jet Model
If there is substantial curvature in the flow pattern (or height field) surrounding a jet streak, then we must account for the centripetal accelerations. Actually, one must consider the curvature of an air parcel trajectory, not the streamline curvature, because the jet streak is propagating, but that requires more analysis than is practical in a forecast environment.
A highly curved trough and attendant westerly jet stream (yellow region)
without a wind speed maximum (no along-stream variation of momentum).
Solid lines denote geopotential height contours. Thick black vectors
represent along-stream ageostrophic winds with magnitudes given by arrow
length, as determined by the centripetal acceleration. The flow is in
gradient wind balance.
This figure shows that the centripetal accelerations experienced by an air
parcel can be substantial enough that the ageostrophic wind is determined
primarily by those accelerations. According to the gradient wind equation,
the relationship between curvature (K) and the actual wind speed (V) is:
This equation shows that in the presence of cyclonic curvature (K>0), the
gradient wind speed is smaller than the geostrophic wind, or in other
words, the winds are "subgeostrophic". Now, since the ageostrophic wind is
the deviation of the gradient wind from the geostrophic wind:
Go to the Theoretical Principles and Conceptual Models Page.
then clearly the ageostrophic wind will be directed upstream at the base of
the cyclonically curved trough, as shown in the above figure. Conversely,
supergeostrophic flow at the anticyclonic ridge axis leads to
downstream-directed ageostrophic flow there. Synthesis of this information
leads to the pattern of divergence and convergence shown in the figure.
Thus, the end result of this reasoning is that sinking (rising) motions are
expected upstream (downstream) of the trough axis: a 2-cell pattern of
vertical motions. This is very different than the pattern predicted by the
simple straight jet streak model, yet the latter model is frequently used
(in fact, overabused) by the operational forecast community!