A Comparison of Observed and Model-Derived Structures of Caribbean Easterly Waves

Publisher Website
Google Scholar
Abstract
A linear primitive equation model has been used to test the hypothesis that the vertical structure of observed Caribbean easterly waves is determined by the interaction between convective heating and the environmental wind. The model determines the response to a propagating heat source in a specified basic state. The model allows for the inclusion of diffusion and cumulus momentum transports. The linear perturbations are assumed to have the form of a single Fourier component in the zonal direction. The frequency and zonal wavelength of the disturbance are taken from observations of the three-dimensional structure of a series of Caribbean easterly waves made by Shapiro. The structure of the basic state zonal wind, assumed to be a function of height, is based on observations near the latitude of largest observed wave amplitude. The maximum heating rate is 5 K day−1, centered at about 19°N. Very good agreement is found between the model-derived vertical structure of the waves and that observed by Shapiro (1986). In particular, the observed 90° westward phase shift between the 200 mb and near-surface troughs, and the westward tilt of the trough axis with height, are reproduced in the model solutions. Although linearization is not strictly valid for the observed wind amplitudes of ∼5 m s−1, the model's linear dynamical framework appears to represent the wave's structure well. The westward phase shift is found to depend on the downward flux of wave energy toward a near-critical layer near the ground. Experiments also suggest that the latitude of the disturbance may be as important a factor in the determination of the westward tilt of the trough axis as is the structure of the basic state zonal wind. An eastward tilt of the trough axis in the lower troposphere, such as that in the classical model of a Caribbean easterly wave, can occur at low latitudes, when the westward phase shift is in a narrow layer near the level of maximum heating. Cumulus momentum transports do not substantially change the structure of the forced wave disturbance. The model solutions are compared with similar experiments of Holton, and are related to results of Stevens, Lindzen and Shapiro.

This publication has 0 references indexed in Scilit: