Recognizing the need to simulate more adequately land-surface processes in numerical weather prediction models, the National Centers for Environmental Prediction (NCEP) Environmental Modeling Center (EMC) implemented a modern land-surface parameterization scheme in the operational Eta model on 31 January 1996, with companion changes to the atmospheric surface layer parameterizations. Since that time, ongoing external studies and internal verification at EMC have resulted in refinements and improvements to the new schemes. The latest changes include an increase in the number of soil layers from two to four, and implementation of continuous self-cycling of soil moisture in the Eta Data Assimilation System (EDAS). These changes became effective operationally on 9 February 1998 and 3 June 1998, respectively. Before 3 June 1998, EDAS soil moisture was initialized from the NCEP Global Data Assimilation System (GDAS). The purpose of this study is to evaluate the warm-season performance of the improved schemes using Oklahoma Mesonet observations.
Prior to the 3 June 1998 implementation, initialization from GDAS resulted in a severe positive bias in top-layer soil moisture over Oklahoma in the operational Eta model. Verification against Oklahoma Mesonet observations during July 1997 also showed that ground heat flux appears to be underestimated in dry conditions, while net radiation at the surface is overestimated. Because the land-surface scheme applies the surface energy balance equation, the excess available energy afforded by the underestimated ground heat flux and overestimated net radiation must be realized in the turbulent fluxes of sensible and latent heat. Because of the erroneously large soil moisture, much of this excess was realized in latent heat flux, while sensible heat flux values were near the observed or somewhat overestimated. When the sensible heat flux is too large, diagnosed 2 m air temperatures are near observed. However, when sensible heat flux is near the observed, a cool bias develops. This scenario suggests a possible underestimate of warm air entrainment at the top of the planetary boundary layer (PBL). The net result of these factors during July 1997 over Oklahoma was a PBL that was often too cool, too moist, and too shallow.
In the current phase of this study, we have undertaken a comparison of July 1998 model performance to that of July 1997 over the Oklahoma Mesonet domain. The purpose of this comparison is to demonstrate whether the cited refinements in February and June 1998 have reduced the surface flux and PBL biases noted in 1997. The results of this comparison will be discussed in the preprint and presented at the conference