Extended Abstract (216K)JP1.17
An intelligent NN system for quality controlled retrieving the wind speed from SSM/I measurements
Vladimir M. Krasnopolsky, SAIC and NOAA/NWS/NCEP, Camp Springs, MD
A generic AI application presented in this paper illustrates a NN based intelligent integral approach in satellite retrievals when the entire retrieval system, including the quality control block, is build as a combination of several specialized NNs. Presented retrieval system retrieves the wind speed over the surface of the ocean, atmospheric moisture concentrations and sea surface temperature from satellite measurement performed by the Special Sensor Microwave Imager (SSM/I).
The system consists of a NN retrieval algorithm, which retrieves the surface wind speed, moisture concentrations, and sea surface temperature given SSM/I measurements and a NN forward model, which given the wind speed and other geophysical parameters simulates the SSM/I measurements. Combining together these two NNs result in creating an intelligent retrieval system, which not only produces accurate retrievals of the wind speed over the ocean surface, but also performs an analysis and quality control of these retrievals and environmental conditions, rejecting poor retrievals if they occur. This intelligent retrieval system successfully reduces both the wind speed RMS and maximum errors. This approach offers significant advantages for "real time" operational applications. The system is developed for implementation at the NCEP data assimilation system. Some constituents of the system have already been implemented there since 1998.
Joint Poster Session 1, Operational Applications and Artificial Intelligence (Joint between 12th Conference on Satellite Meteorology and Oceanography and Third Conference on Artificial Intelligence Applications to Environmental Science)
Monday, 10 February 2003, 2:30 PM-4:00 PM
Previous paper Next paper