Sensitivity of WRF-simulated 10 m wind over the Persian Gulf to different boundary conditions and PBL parameterization schemes

Abstract

A sensitivity analysis of the Weather Research and Forecasting (WRF) mesoscale model with the Advanced Research WRF (ARW) dynamical solver for wind simulation at 10 m above ground level was conducted through different initial and boundary conditions datasets, and various planetary boundary layer (PBL) schemes throughout the year 2017 over the Persian Gulf region. Owing to the wide variety of approaches and development periods of analysis and reanalysis data (e.g. assimilation system) as well as different methods for PBL parameterization (closure formulations), this paper aims to obtain an efficient set up of the WRF model configuration in terms of the lowest error in simulating surface wind. Three datasets including ERA-Interim reanalysis, NCEP-R2 reanalysis, and NCEP-FNL analysis as initial and boundary conditions to the model and six PBL schemes including ACM2, BouLac, MYJ, MYNN, QNSE, and YSU accompanied by their relevant surface-layer schemes were used to accomplish this goal. Available observational wind data including 23 synoptic weather stations located in the region were used to compare the model wind simulation. Comparing WRF wind simulations with observations at synoptic weather stations indicates that; irrespective of the type of PBL scheme, ERA-Interim, and NCEP-FNL datasets exhibit better performance than the NCEP-R2. In addition, in the case of considering PBL schemes, results show that the configuration including the YSU scheme and ERA-Interim reanalysis data leads to the best estimation of wind speed and configuration including YSU and NCEP-FNL data produces the least error for wind direction. The performance of the model in summer shows higher errors compared to the winter season. It is due to the obvious differences between initial and boundary conditions data with observation data. The success of the YSU scheme in comparison with other schemes is rooted in the nonlocal closure of this scheme, which includes local gradients correction and entrainment sentences.