A discussion about the role of shortwave schemes on real WRF-ARW simulations. Two case studies: cloudless and cloudy sky

A wide range of approaches for radiative transfer computations leads to several parameterizations. Differences in these approximations bring about distinct results for the radiative fluxes, even under the same atmospheric conditions. Since the transfer of solar and terrestrial radiation represents the primordial physical process that shapes the atmospheric circulation, these deviations must have an impact on the numerical weather prediction (NWP) model performance. In this paper, an analysis of the role of shortwave schemes on the Weather Research and Forecasting (WRF-ARW) model is presented. The study compares the effect of four parameterizations (Dudhia, New Goddard, CAM and RRTMG) in two cases: i) cloudless and ii) cloudy sky situations for a domain defined over Catalonia (northeast of the Iberian Peninsula). We analyze the direct and the indirect feedback between the dynamical aspects and the physical parameterizations driven by changes in the radiative transfer equation computation. The cumulative effect of these variations are studied through three simulation windows: current day (0-23 h), day-ahead (24-47 h) and two days ahead (48-71 h). These analyses are focused on several NWP model fields. From the most directly related to shortwave schemes such as global horizontal irradiance or the heating rate profile, to apparently secondary outcomes such as wind speed or cloud composition among others. The differences observed between model runs using different solar parameterizations increase with the simulation horizon, being more important in the cloudy scenario than in the cloudless sky.