A variable-resolution data assimilation system employing the global stretched-grid approach has been developed and tested as an efficient alternative tool for producing regional analyses and diagnostics with enhanced resolution. The variable resolution stretched grid (SG) version of the Goddard Earth Observing System (GEOS) Data Assimilation System (DAS) incorporating the GEOS SG-GCM, is used for regional analysis and forecast as well as regional climate applications. The SG-design, with multiple regions of interest with enhanced horizontal resolution is used. These areas are: the U.S. (the major area of interest); El Nino/Brazil/Argentina; India-China; and Eastern Indian Ocean/Australia. The major monsoonal circulations are associated with these regions. Both the SG-DAS and reference (basic) uniform grid DAS, with the same amount of global grid-points, are compared in terms of regional products. The SG-DAS is capable of providing the efficient down-scaling to mesoscales for both instantaneous and time-averaged regional analysis and diagnostic fields. Moreover, the adaptive quality control procedure for the SG-DAS accepts more non-corrupted regional observations, especially near the surface. The SG-DAS regional analyses and diagnostics are used for validation of regional climate simulation experiments produced with the SG-GCM for the anomalous U.S. summer events of 1988, 1993, and 1998, the latter in the context of the 1997-99 ENSO cycle. Monthly mean precipitation for the events is closer to gauge precipitation observations for the SG-DAS than for the reference DAS.

The analyses and diagnostics are produced for the super-typhoon Paka development in the equatorial Pacific (December 1997). Both the depth of the center and associated winds are much more realistic for the SG-DAS than for the reference DAS.

The obtained results show that the SG-DAS is a viable candidate for producing high resolution assimilated products for regional studies and applications.