Sensitivity Study of WRF Parameterization Schemes for the Spring Sea Fog in the Yellow Sea

Sea fog is a water vapor condensation phenomenon, which happens in marine atmospheric boundary layer (MABL). Low atmospheric visibility caused by sea fog brings huge threat to maritime transportation, fishery and oil-drilling operations. Therefore, it is becoming increasingly important and being paid more and more attention. In recent years, meso-scale atmospheric numerical modeling has become a dominant way for the mechanism study and numerical modeling of sea fog.Previous studies on sea fog indicate that sea fog modeling is very sensitive to initial conditions, especially realistic representation of temperature and humidity profile in MABL. Besides initial conditions, turbulence process and cloud generating process are the other important aspects for sea fog modeling. In a meso-scale atmospheric numerical model, the turbulence process is described by planetary boundary layer (PBL) scheme, and the cloud generating process is determined by microphysics (MP) scheme. Due to the uncertainties of the modeling result and the complexities of turbulence and cloud microphysics processes, many options of PBL and MP schemes are available for choice focusing on different modeling purposes.Based on the Weather Research and Forecasting (WRF) model and cycling three-dimensional variational method, sensitivity study of WRF PBL and MP schemes for the Yellow Sea fog is conducted, focusing on 10 typical widely-spread sea fog cases. The result indicates that simulated sea fog area mostly depends on PBL scheme but little on MP scheme; density and depth of simulated sea fog are affected by MP scheme with cloud droplet number being predicted and how it is prescribed. The best combination of PBL and MP schemes is YSU and Lin, while the worst is Mellor-Yamada and WSM5. The Mellor-Yamada and QNSE scheme brings about much stronger turbulence simulation, resulting in much higher boundary layer, and therefore it's not favorable to the development and maintenance of sea fog, while turbulence intensity and boundary layer height produced by MYNN and YSU schemes benefit sea fog developing. MYNN scheme can match YSU scheme in general, however, the latter performs better in most cases while the former is better in certain ones. In depth investigation is needed to tell whether MYNN or YSU PBL scheme is better for a given sea fog case. These information can provide hints to choose and improve PBL and MP schemes of WRF for the Yellow Sea fog numerical prediction system in the near future.