Evaluation of WRF regional climate model for simulating near-surface atmospheric characteristics

Abstract:

The atmospheric planetary boundary layer (PBL) affects the simulation of low-level winds, and simulation of near-surface winds is challenging because of the high interactions between local topography. Lower altitude atmospheric flow is a challenge for most of the Numerical weather prediction (NWP) models. We employed Weather Research and Forecasting (WRF) model to simulate near-surface atmospheric characteristics over Newfoundland and part of the North Atlantic Ocean, and assessed the skill of the model for the near-surface wind simulation. The WRF model employs several schemes for the PBL parameterization. The sensitivity of the wind estimation was evaluated by considering five widely used PBL parameterizations. The choice of PBL affects the choice of surface layer (SL) parameterization scheme because the PBL and SL parameterization actively interact with each other. We examined the Grenier-Bretherton-McCaa PBL scheme with the MM5 SL scheme; the Asymmetric Convective Model version 2 (ACM2) PBL scheme with the revised MM5 SL scheme; the Mellor Yamada (MYJ) PBL scheme with the ETA SL similarity scheme; the Quasi-Normal Scale Elimination (QNSE) PBL scheme with the QNSE SL scheme; the Yonsei University (YSU) PBL scheme combined with the MM5 SL similarity scheme. A two-way nested domain with a 3:1 grid spacing ratio between the 15 km parent domain and the 5 km nested domain was considered. The NCEP climate forecasting system version 2 reanalysis data (CFSV2) was used to initialize the model and to provide the lateral and surface boundary conditions. Build-up of errors over time were reduced using forecast simulations executed for 36 hours with 12 hours of spin-up time. Observational data recorded by Environment Canada were used for the verification process. Preliminary results indicated that the PBL schemes tend to show better performance during the summer. Moreover, the simulations tend to underestimate the wind speed in both the summer and winter time. Among the different PBLs, MYJ with the local closure approach provides better results during the cold season, whereas the non-local YSU scheme prevails in summer.