In this study, a LLJ system, that occurred on June 25, 2002 over the South Great Plain region and observed by Doppler lidar wind profiles obtained by the Goddard Lidar Observatory of Winds (GLOW) during the International H2O Program (IHOP_2002) is examined. The evolution of boundary layer wind structure associated LLJ system is revealed in great details by the GLOW wind profiles. The GLOW lidar wind profiles are then compared with an enhanced radiosonde at Homestead Station and the analysis of wind from NCEP Northern American Regional Reanalysis (NARR). It is apparent that the lidar wind speeds illustrate clearly a LLJ at night from 0100 UTC to 1100 UTC and convective activities with large wind speed over 7 ms-1 during the daytime from1200UTC to 1800 UTC. The large wind speeds correspond to daytime convection and extend higher than 1500 m at around 1300 UTC. The radiosonde speeds show little variation from 0300 to 0600 UTC due to a lower temporal frequency of the measurements. Diurnal variations from lidar and sonde show similar features, although the low-level jet measured from lidar is lower than that from the radiosonde and more detailed in its structure. In contrast, the NARR reanalysis data failed to capture the features of the LLJ: the low temporal and spatial resolution (horizontally about 30km, vertically about 250m in the lower troposphere, and temporally 3 hourly) makes the NARR data difficult to reveal the detail of structure and evolution of the LLJ. The intensity of the LLJ is also weaker in NARR reanalysis than in observations from lidar or sonde. The height of the LLJ core in reanalysis is approximately 200 m higher than the observed heights from either lidar or radiosonde measurements.
In order to examine the ability of high-resolution numerical simulations in representation of the LLJ and also to understanding the dynamic and physical processes associated with the evolution of the LLJ. A series of numerical experiments are conducted with the Weather Research and Forecasting (WRF) model to simulate the LLJ system. Results are validated with GLOW wind profiles. It is found that the numerical simulation of the LLJ is sensitive to the choice of physical parameterization schemes in WRF model and initial conditions. Larger sensitivities are found when varying land surface schemes, PBL parameterizations and initial conditions. Further evaluations are performed to understand the mesoscale convective processes and boundary layer characteristics related to the LLJ. Data assimilation experiment is also in progress to assimilate lidar profiles into the WRF model in order to improve overall analysis of the LLJ case. Detailed results will be reported during the conference.