Recent GCIP-sponsored advancements in coupled land-surface modeling and data assimilation in the NCEP Eta mesoscale model

With support from the GEWEX Continental-Scale International Project (GCIP), the Environmental Modeling Center (EMC) of NCEP, NWS hydrologists in the Office of Hydrology (OH), and satellite land-surface remote sensing experts in NESDIS have worked together to improve the land-surface model (LSM) in the NCEP operational mesoscale Eta model, and verify Eta model performance following LSM changes. Currently the LSM coupled with the Eta model carries four soil layers with predictions of soil moisture using Richard's equation and soil temperature using the heat diffusion equation, as well as an explicit representation of vegetation, sub-grid runoff treatments, and satellite-derived (a) seasonal cycle of vegetation greenness and (b) snow cover/sea-ice (generated daily for operational use). In a major NCEP milestone on 03 June 1998, the LSM state variables of soil moisture, soil temperature, canopy water, and snow water became fully and continuously cycled state variables in the coupled Eta model's 4-D Data Assimilation System (EDAS). Herein, LSM performance must be good on many time scales (from hourly to annually). No nudging of the soil moisture is employed, neither to a soil moisture climatology or otherwise. A vivid example of EDAS soil moisture anomalies in the Texas and Oklahoma drought of summer 1998 will be shown, as well as long term model performance in terms of regional comparisons of two-meter temperature and relative humidity bias. Recent advances made to the LSM include frozen soil physics, snow cover patchiness, and improvements to the soil heat flux related to thermal conductivity and the presence of vegetation and snow cover. This latest LSM is coupled to the Eta model and is undergoing pre-implementation testing in different retrospective runs to assess model performance. Model verification also includes a new capability to validate LSM skin temperature against hourly 0.50-degree GOES-derived skin temperature. Future work will include realtime execution of the uncoupled LSM across the U.S. domain using observed atmospheric forcing. The resulting LSM output (land-surface forcing) will then drive the Eta model. The use of observed forcing reduces biases in the LSM state variables currently caused by using cycled Eta model atmospheric forcing (i.e, model precipitation and radiation) in the LSM. (See companion 'LDAS' abstract by Mitchell et al, this conference.) The LSM in the Eta model has been recently released to the public as the 'NOAH' LSM, acknowledging *N*CEP, *O*regon State University, the US *A*ir Force, and the NWS Office of *H*ydrology as the major contributors in the development of this LSM over the years. (The uncoupled NOAH LSM is available via anonymous ftp at 'ftp.ncep.noaa.gov/pub/gcp/ldas/noahlsm'.)