SERVICE ORIENTED ATMOSPHERIC RADIANCES (SOAR): A Web Service Research Tool for the Gridding and Synthesis of Multi-Sensor Satellite Radiance Data for Weather and Climate Studies

摘要

Three decades of Earth remote sensing from NASA, NOAA and DOD satellites carrying successive generations of atmospheric instruments have resulted in petabytes of radiance data with successive increases in spatial and spectral resolutions stored at different data archives in various data formats. We describe here a web based Service Oriented Atmospheric Radiance (SOAR) prototype system built on the SOA architecture that will enable the science community to process these valuable climate data records according to their own gridding criteria. SOAR employs the standard XML based protocol suite of SOAP, WSDL and UDDI service descriptions for aggregating atmospheric instrument radiance data into user specified spatial grids. SOAR consists of three subsystems, a Client Server, a Directory Server, and a Process Server, connected to a high performance compute cluster and storage grid by a Service Bus. The process server employs optical communications to access data and invoke algorithms on the compute/storage cluster for on-demand spatial, temporal, and spectral subsetting Scientists can choose a variety of statistical averaging techniques for combining the footprints of satellite observed radiances from multiple instruments to form spatial-temporal grids for their respective studies. Animation services are also provided for viewing the results of the user specified service requests. Results are presented for subsetting and animating a multi-year high-resolution multi-instrument pre-gridded radiance field employing this initial version of SOAR.