As part of the NCEP effort to unify the use of the advanced Noah Land Surface Model (LSM) in all NCEP regional and global models and their associated data assimilation systems, the Noah LSM has been implemented into the testbed for the NCEP Global Forecast System (GFS) in late 2002. The GFS coupled with the Noah LSM (referred to as GFS_Noah) has been used in various studies, such as seasonal forecasts, to evaluate the Noah LSM in the GFS and to assess its impact on GFS forecasts. Initial conditions for these studies are taken from the NCEP Global Data Assimilation System (GDAS) using the operational version of GFS (with old LSM), and thus, these studies could not reflect a fair evaluation of Noah LSM due to the inconsistency in initial land states and model land physics.

The GFS_Noah has been incorporated into the framework of the NCEP GDAS in early 2003. A retroactive data assimilation run starting from 1 Aug 2002 is in progress. The Noah cycled GDAS run employs reduced resolution (T62 L28) with respect to the operational run (T254 L64). A low-resolution GDAS run using the operational version of GFS (with old LSM) is run in parallel with the Noah cycled GDAS run. These data assimilation efforts aim to evaluate the impact of LSM upgrade on GFS forecasts as well as to provide continuously cycled Noah LSM land states in GDAS.

In this study, the GFS_Noah is used for warm-season seasonal hindcasts for 2003. Initial conditions are taken from: (1) low resolution GDAS using the operational version of GFS, and (2) low resolution GDAS using the GFS_Noah. The length of integration is 3 months, starting from late spring/early summer. The ocean surface boundary fields are from observed 1-deg NCEP Reyolds/Stokes SST. Results from two sets of GFS_Noah summer time integrations are analyzed and compared against observed near-surface temperature from CPC 344 Climate Division data and precipitation fields from CPC gauge date. The impact of using Noah cycled GDAS (which provide initial land states that are strictly self consistent with GFS_Noah land physics) on warm season predictions is assessed.