Land Surface Upgrades in the Next NCEP NEMS Implementation

The major causes to some long-standing biases related to the land surface processes in the NCEP GFS have been identified. By using the high resolution land surface characteristic data (IGBP vegetation type and STASGO soil type data) to replace the coarse data using in the current operational GFS, the blockiness patter in the surface fields has been gone. The surface cold biases over the cold regions have been reduced significantly after the new MODIS-based maximum snow albedo data and the improved ground heat flux forecast under the deep snow. The rapid temperature drop during sunset has been improved to a great extend after a decoupling prevention is applied to very stable situation. The whole land surface upgrade bundle has been extensively tested and is planned to be included in the next NCEP NEMS (hereinafter the replacement of GFS) implementation.