Tropical cyclone motion and recurvature in TCM-90

Abstract

Rawinsonde and satellite data collected during tie Tropical Cyclone Motion (TCM- 90) experiment, which was conducted during the summer of 1990 in the Western North Pacific, is used to examine tropical cyclone steering motion and recurvature. TCM-90 composite results a.re compared with those found in a previous composite study using twenty-one years (1957-77) of Western North Pacific rawinsonde data during the same August- September period and also for August-September months during this same 21- year period. Both data sets indicate that the composite deep-layer-mean (850-300 mb) winds 5-7° from the cyclone center provide an important component of the steering flow for tropical cyclones. However, despite the rawinsonde data enhancements of the TCM-90 experiment, data limitations prevented an accurate observation of steering flow conditions at individual time periods or for the average of only 5-10 time periods when composited together. Examination of environmental wind fields surrounding a recurving cyclone (Typhoon Flo, Sept. 1990), and those for non-recurving TCM-90 storms verify significant differences in the upper tropospheric zonal wind fields north and -northwest of the tropical cyclone one to two days prior to the beginning of the initial right turn of recurvature. Typhoon Flo actually began to recurve when 200 mb positive zonal winds had penetrated to within 6 degrees radius of the cyclones' center from the northwest. Tropical cyclones which did not recurve had negative zonal winds at this radius and azimuth. This special area to the north and northwest of the cyclone has been termed the "window of forecast opportunity". Basic statistical analyses of the typical spread of individual wind values at specific octants and 2 degree radial belts were ma.de for all TCM-90 rawinsonde and satellite wind data composites. The typical standard deviation about the mean of composited zonal and meridional winds in individual octants and radial belts was 5-6 m/s at lower levels and 6-7 m/s at upper levels. Zonal wind differences in excess of this threshold would be required for confidence in distinguishing between i1 dividual cases of recurvature and non-recurvature.