Proper partitioning of the surface heat fluxes that drive the evolution of the planetary boundary layer in numerical weather prediction models requires an accurate specification of the initial state of the land surface. The National Centers for Environmental Prediction (NCEP) operational Eta Model is used to produce land surface analyses by continuously cycling soil temperature and moisture fields. These fields previously evolved only in response to radiation budget constraints and modeled precipitation, but NCEP recently upgraded the self-cycling process so that soil fields respond instead to the radiation budget and observed precipitation. A comparison of 0000 and 1200 UTC Eta Model analyses of soil temperature and moisture at several soil depths with observations from the Oklahoma Mesonet during 2004 and 2005 shows that there are strong biases in soil temperature and a severe underestimation of soil moisture at all depths. After the change to a new assimilation scheme, there is notable improvement in the magnitude of the analyzed soil moisture fields, although a strong dry bias persists in the soil moisture field. A simple one-layer slab soil model quantifies the effect of such soil moisture errors on the diurnal cycle of soil temperature and reveals that these soil moisture errors alone may account for only 1.6°C increases in predicted maximum soil temperatures during the day and temperature reductions of the same magnitude at night. The much larger remaining soil temperature errors possibly stem from documented problems with the solar radiation and longwave parameterizations within the Eta Model.
All Science Journal Classification (ASJC) codes
- Atmospheric Science