Abstract
A diagnostic model is proposed to use digitized satellite cloud brightness data to estimate objectively the large-scale flow patterns over data-void tropical regions. The model utilizes a linear barotropic vorticity equation with two primary assumptions: 1) that the area-averaged cloud brightness is positively correlated with large-scale divergence in the tropical upper troposphere; and 2) that the large-scale tropical flow is quasi-barotropic and quasi-non-divergent. It is designed to be used at any upper tropospheric level where divergence is important in determining the vorticity field. Three types of information are required: 1) boundary conditions determined from surrounding wind reports, 2) a mean zonal flow determined from climatology, and 3) an equivalent divergence forcing function constructed empirically from the brightness data.
The model is tested daily over a western North Pacific region for July-August 1971. Results for an 8-day representative period are presented and discussed. In general for 25% of the days tested, the model produces a flow field which accurately resembles the major features of the streamfunction field analyzed by the National Meteorological Center. In another 30% of the days it provides some valuable information about the flow patterns which would be difficult to obtain from boundary information alone. Experiments are also performed for two days in which the brightness data are enhanced by time-interpolated satellite infrared data. The resultant flow fields bear better resemblance to the NMC analysis. It is thus suggested that improved results may be expected when infrared and other types of advanced satellite data are available.