Impact of WRF model PBL schemes on air quality simulations over Catalonia, Spain

Here we analyze the impact of four planetary boundary-layer (PBL) parametrization schemes from the Weather Research and Forecasting (WRF) numerical weather prediction model on simulations of meteorological variables and predicted pollutant concentrations from an air quality forecast system (AQFS). The current setup of the Spanish operational AQFS, CALIOPE, is composed of the WRF-ARW V3.5.1 meteorological model tied to the Yonsei University (YSU) PBL scheme, HERMES v2 emissions model, CMAQ V5.0.2 chemical transport model, and dust outputs from BSC-DREAM8bv2. We test the performance of the YSU scheme against the Assymetric Convective Model Version 2 (ACM2), Mellor-Yamada-Janjic (MYJ), and Bougeault-Lacarrère (BouLac) schemes. The one-day diagnostic case study is selected to represent the most frequent synoptic condition in the northeast Iberian Peninsula during spring 2015; regional recirculations. It is shown that the ACM2 PBL scheme performs well with daytime PBL height, as validated against estimates retrieved using a micro-pulse lidar system (mean bias=−0.11km). In turn, the BouLac scheme showed WRF-simulated air and dew point temperature closer to METAR surface meteorological observations. Results are more ambiguous when simulated pollutant concentrations from CMAQ are validated against network urban, suburban, and rural background stations. The ACM2 scheme showed the lowest mean bias (−0.96μgm−3) with respect to surface ozone at urban stations, while the YSU scheme performed best with simulated nitrogen dioxide (−6.48μgm−3). The poorest results were with simulated particulate matter, with similar results found with all schemes tested. [Display omitted] •Planetary boundary-layer and land-surface interactions can invoke implications to air quality forecast system simulations.•Three WRF model PBL schemes are compared to the current scheme used in the Spanish air quality forecasting system (AQFS).•WRF PBL scheme choice show large differences in CMAQ-simulated surface pollutant concentrations.•The Barcelona micropulse lidar (MPL), newly introduced to the NASA MPLNET, is used to validate model-simulated PBL height.•It is found that the nonlocal ACM2 scheme may be a better choice than the current operational scheme (YSU) for complex areas.