| description abstract | Raymond's (1975) wave-CISK model is applied to several tropical convective storms observed in Venezuela, the eastern Atlantic and West Africa to predict their propagation velocity. Similar calculations are carried out with Moncrieff and Miller's (1976) analytical model for tropical cumulonimbus and squall lines. A comparison of the model predictions with the observed values is made. In some cases the models give good predictions, but not in others. In general, Raymond's model underestimates the propagation speed of the storms, while the Moncrieff-Miller model overestimates it. Raymond's model is poor when the cloud bases are very low. This result indicates that over tropical oceans wave-CISK models cannot give good results unless the mass flux due to the plumes, which is equated to the mass flux across cloud base, is treated in a more realistic way. The Moncrieff-Miller model gives better results if the mean wind component along the direction of motion is used rather than the mid-level wind. The wave-CISK model and steady-state models of storm motion are then considered in conditions of constant wind shear. In particular, their predictions are compared over a wide range of shear values, using realistic thermodynamic soundings. Despite the obvious differences between the models, it is found that, for Richardson number small (R<1) and very large, they give comparable predictions for the storm velocity. It appears that a very good approximation for the wave-CISK model over the entire R range is to put the storm speed proportional to the shear, plus a constant. An important conclusion is that the ability of storms to propagate relative to the environmental flow can be reproduced in the linear wave-CISK model and thus may not be a fundamentally nonlinear effect. It is therefore crucial to further examine forcing mechanisms of convective overturning and, in particular, to clarify the relationship between CISK and the implicit forcing involved in the steady model. | |
