Global estimation of precipitation using opaque microwave bands

• Main Author: Chen, Frederick Wey-Min, 1975-
• Other Authors: David H. Staelin., Massachusetts Institute of Technology. Dept. of Electrical Engineering and Computer Science.
• Format: Thesis
• Language: English
• Published: Massachusetts Institute of Technology 2004
• Subjects: Electrical Engineering and Computer Science; Arctic
• Online Access: http://hdl.handle.net/1721.1/16696

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2004. Includes bibliographical references (p. 115-125). This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. This thesis describes the use of opaque microwave bands for global estimation of precipitation rate. An algorithm was developed for estimating instantaneous precipitation rate for the Advanced Microwave Sounding Unit (AMSU) on the NOAA-15, NOAA-16, and NOAA-17 satellites, and the Advanced Microwave Sounding Unit and Humidity Sounder for Brazil (AMSU/HSB) aboard the NASA Aqua satellite. The algorithm relies primarily on channels in the opaque 54-GHz oxygen and 183-GHz water vapor resonance bands. Many methods for estimating precipitation rate using surface-sensitive microwave window channels have been developed by others. The algorithm involves a set of signal processing components whose outputs are fed into a neural net to produce a rain rate estimate for each 15-km spot. The signal processing components utilize techniques such as principal component analysis for characterizing groups of channels, spatial filtering for cloud-clearing brightness temperature images, and data fusion for sharpening images in order to optimize sensing of small precipitation cells. An effort has been made to make the algorithm as blind to surface variations as possible. The algorithm was trained using data over the eastern U.S. from the NEXRAD ground-based radar network, and was validated through numerical comparisons with NEXRAD data and visual examination of the morphology of precipitation from over the eastern U.S. and around the world. It performed reasonably well over the eastern U.S. and showed potential for detecting and estimating falling snow. However, it tended to overestimate rain rate in summer Arctic climates. Adjustments to the algorithm were made by developing a neural-net-based estimator for estimating a ...