Accuracy Assessment of Planetary Boundary Layer Height in the WRF Simulation using Temporally High Resolution Radiosonde Observations

Understanding the limitation in simulating Planetary Boundary Layer Height (PBLH) by mesoscale meteorological model is important for the accurate modeling of meteorological variables and air pollution. This study examined the accuracy for PBLHs using two of most generally used PBL schemes - MYJ and YSU - in the Weather Research and Forecasting (WRF) model. The simulated PBLHs were evaluated based on the radiosonde data observed at an interval of 3 hours. The radiosonde observations were conducted from August to December in 2014 near the Gumi weir over the Nakdong River in the Korean Peninsula. Observed PBLH were calculated using the parcel and Bulk Richardson number (Rib) methods. In order to reduce the simulation errors of PBLH, the impact of adjusting nudging coefficients (3.0×10-4 and 3.0×10-1) was assessed using the Four Dimensional Data Assimilation (FDDA) of the radiosonde observations. The assessment results showed that RMSE of PBLH with YSU scheme was smaller than that with MYJ scheme in the day and night. The YSU scheme showed smaller RMSE than the MYJ scheme on sunny, cloudy and rainy days, too. The WRF model with the YSU scheme produced lower PBLH than that with the MYJ scheme during night time. The simulation accuracy of PBLH was higher when structure of simulated wind speed profiles was similar to the observation. FDDA experiments showed positive impact on the structure of wind speed profiles and PBLH. Especially, the structure of wind speed profiles using the nudging coefficient of 3.0×10-1 (YSU_FDDA_2) became more similar to the observation than using the nudging coefficient of 3.0×10-4 (YSU_FDDA_1). The PBLH errors were reduced by as much as 40% by the incorporation of the YSU_FDDA_2: For example, observed PBLH was 1,610 m at 1500LST on 20 November 2014, whereas PBLH of YSU_FDDA_2 and YSU were 999 m and 414 m, respectively. The meteorological condition was effectively improved by the enhanced nudging coefficient of YSU_FDDA_2.