A METHOD FOR COMBINING PASSIVE MICROWAVE AND INFRARED RAINFALL OBSERVATIONS

Passive microwave observations of rainfall offer the ability to obtain very accurate instantaneous estimates of rainfall. Because passive mi crowave instruments are confined to polar-orbiting satellites, however , such estimates must interpolate across long time periods, during whi ch no measurements are available. In this paper the authors discuss a technique that allows one to partially overcome the sampling limitatio ns by using frequent infrared observations from geosynchronous platfor ms. To accomplish this, the technique compares all coincident microwav e and infrared observations. From each coincident pair, the infrared t emperature threshold is selected that corresponds to an area equal to the raining area observed in the microwave image. The mean conditional rainfall rate as determined from the microwave image is then assigned to pixels in the infrared image that are colder than the selected thr eshold. The calibration is also applied to a fixed threshold of 235 K for comparison with established infrared techniques. Once a calibratio n is determined, it is applied to all infrared images. Monthly accumul ations for both methods are then obtained by summing rainfall from all available infrared images. Two examples are used to evaluate the perf ormance of the technique. The first consists of a one-month period (Fe bruary 1988) over Darwin, Australia, where good validation data are av ailable from radar and rain gauges. For this case it was found that th e technique approximately doubled the rain inferred by the microwave m ethod alone and produced exceptional agreement with the validation dat a. The second example involved comparisons with atoll rain gauges in t he western Pacific for June 1989. Results here are overshadowed by the fact that the hourly infrared estimates from established techniques, by themselves, produced very good correlations with the rain gauges. T he calibration technique was not able to improve upon these results.