Development and evaluation of a regional ocean-atmosphere coupled model with focus on the western North Pacific summer monsoon simulation: Impacts of different atmospheric components

A regional ocean atmosphere coupled model (ROAM) is developed through coupler OASIS3, and is composed of regional climate model RegCM3 and CREM (Climate version of Regional Eta Model) as its atmospheric component and of a revised Princeton ocean model (POM2000) as its oceanic component. The performance of the ROAM over the western North Pacific summer monsoon region is assessed by the case simulation of warm season in 1998. Impacts of different atmospheric model components on the performance of ROAM are investigated. Compared with stand-alone simulation, CREM (RegCM3) produces more (or less) rainfall over ocean area with inclusion of the air-sea coupling. Different biases of rainfall are caused by the different biases of SST derived from the coupled simulation. Warm (or cold) SST bias simulated by CREM_CPL (RegCM3_CPL) increases (or decreases) the evaporation at sea surface, then increases (or decreases) the rainfall over ocean. The analyses suggest that the biases of vertical profile of temperature and specific humidity in stand-alone simulations may be responsible for the SST biases in regional coupled simulations. Compared with reanalysis data, the warmer (or colder) and moister (or dryer) lower troposphere simulated in CREM (RegCM3) produces less (or more) sea surface latent heat flux. Meanwhile, the more unstable (or stable) lower troposphere produces less (or more) cloudiness at low-level, which increases (or decreases) the solar radiation reaching on the sea surface. CREM (RegCM3) forced by observed SST overestimates (or underestimates) the sea surface net heat flux, implying a potential warm (or cold) heat source. After coupling with POM2000, the warm (or cold) heat source would further increase (or decrease) the SST. The biases of vertical profile of temperature and specific humidity may be ascribed to the different representation of cumulus convection in atmospheric models.