THE SPEED OF OBSERVED AND THEORETICAL LONG EXTRATROPICAL PLANETARY-WAVES

Planetary or Rossby waves are the predominant way in which the ocean a djusts on long (year to decade) timescales. The motion of long planeta ry waves is westward, at speeds greater than or equal to 1 cm s(-1). U ntil recently, very few experimental investigations of such waves were possible because of scarce data. The advent of satellite altimetry ha s changed the situation considerably. Curiously, the speeds of planeta ry waves observed by TOPEX/Poseidon are mainly faster than those given by standard linear theory. This paper examines why this should be. It is argued that the major changes to the unperturbed wave speed will b e caused by the presence of baroclinic east-west mean flows, which mod ify the potential vorticity gradient. Long linear perturbations to suc h flow satisfy a simple eigenvalue problem (related directly to standa rd quasigeostrophic theory). Solutions are mostly real, though a few a re complex. In simple situations approximate solutions can be obtained analytically. Using archive data, the global problem is treated. Phas e speeds similar to those observed are found in most areas, although i n the Southern Hemisphere an underestimate of speed by the theory rema ins. Thus, the presence of baroclinic mean flow is sufficient to accou nt for the majority of the observed speeds. It is,shown that phase spe ed changes are produced mainly by (vertical) mode-2 east-west velociti es; with mode-1 having little or no effect. Inclusion of the mean baro tropic flow from a global eddy-admitting model makes only a small modi fication to the fit with observations; whether the fit is improved is equivocal.