The effect of land surface characteristics on simulated boundary layer meteorology

To predict atmospheric conditions in the boundary layer, it is necessary to have an accurate representation of the land surface processes and detailed input data such as land use data. The current USGS 25-category Land Use (LU) data and Land Cover (LC) data utilized in the MM5 modeling system is outdated (with the reference year 1990). The unrealistic land use map specified inside MM5 results in inaccurate predictions. Recently, Texas Forest Service (TFS) has generated LULC datasets for the Houston area and the surrounding 8 counties using LANDSAT satellite imagery for the reference year 2000. The TFS-LULC data is correctly reflecting the actual surface information for meteorological modeling.

The primary goal of this study is to derive a better understanding of how land surface processes affect simulated boundary layer meteorological characteristics. The secondary goal is to demonstrate the effects of LULC changes on the temperature, wind transport, PBL height structure and surface flux field. MM5 simulations are performed with NOAH land surface model (LSM) and MRF planetary boundary layer (PBL) scheme. Two sets of MM5 sensitivity tests are designed to utilize different LU datasets: USGS and TFS data. The change of the LULC data modifies the simulated surface heat flux structure and local wind transport. The result demonstrates the benefit of utilizing more precise and updated LU data.

The coupling of NOAH LSM and MRF PBL scheme inside MM5 could not accurately simulate daytime and nighttime wind speeds. The daytime wind speed issue is caused by the inappropriate convective velocity scale which enhances the surface momentum flux. To solve this problem, a more represented surface flux formula was implemented in the MRF PBL scheme. The high bias of nighttime wind speed is hypothesized to be caused by the inadequate parameterizations utilized for the strong stable atmospheric conditions or could be associated with the height of the MM5 lowest model layer. MM5 implicitly assumes that the surface layer parameterizations apply to the whole bottom layer. In reality, the PBL height at nighttime can be lower than 100 m AGL and the formulas used are expected to be valid within the surface layer (~about one tenth of the PBL height). Utilization of accurate formulas may overcome the current shortcomings in nighttime wind prediction. Future research will review the formulas suitable for stable condition, modify the PBL schemes in MM5, and compare the new simulations with available measurements in the Houston areas.