Key points are not available for this paper at this time.
SMS-1 infrared brightness was estimated subjectively for each 1° square for the area from the equator to 20°N latitude and from 10°E to 30°W longitude for Phase III of GATE and the previous interphase period at 6 h intervals. Each grid square was assigned one of the following brightness categories: 0, no clouds or clouds with low tops; 1, clouds with tops at middle heights; or 2, clouds with coldest tops—presumably of convective origin. The percentage of area covered by category 2 clouds was then computed for squares 3° an a side. The data were filtered with respect to time to isolate 3–4 day period wave-related oscillations. In addition, over 160 individual cloud clusters, including a special type of "squall" cluster, were identified in the IR images and tracked on a 3 h basis. A progressive pattern of 3–4 day period fluctuations in convective activity was observed. A comparison of this pattern with time-filtered 700 mb wind data indicates a correspondence of convective activity to wave trough and ridge positions inferred from the wind data. Histograms of the wave phase location of maximum filtered convective cloud coverage indicate that convection was most enhanced at and ahead of the trough axis. However, convection was distributed to some extent throughout the wave. Similarly, convection was shown to be most suppressed at and ahead of the ridge axis. Greatest fluctuations of wave-related convective activity occurred along the path of the center of the 700 mb disturbance. Most large cloud clusters were found to he located ahead of the trough. These clusters tended to be the longest lived and to move along with the wave at slightly less than wave phase speed. Squall clusters were also most prevalent ahead of the trough. They were found to move at about twice wave phase speed, most moving forward through the wave to a position just behind the ridge before termination.
Payne et al. (Fri,) studied this question.