A Content-Based Search Method and Its Application for EOS

Advances in Earth observing from space and associated Earth systems numerical simulations have resulted in rapidly growing data volumes for Earth science communities. The NASA’s Earth Observing System (EOS) program alone is producing massive data products with total rates of more than 1.5 Terabytes/day (King and Greenstone, 1999). To find interesting data, data users need an effective and efficient way to search through the data. Traditionally, metadata are provided in a database which support data searching by commonly used criteria such as spatial coverage, temporal coverage, spatial resolution, temporal resolution, etc. (Yang et al., 2002; (GCMD); NASA, 2003). However, metadata search itself may still result in large amounts of data which data users need to know about before determining the usefulness of the underlying data. Content-based data search, that is, searching data based on not only metadata but also actual data content will help data users to narrow down selected data. In fact, this may be the only way to find useful data in the future.