Comparison between several parameterization schemes in WRF for solar forecasting in coastal zones

The numerical weather prediction (NWP) models have widely demonstrated their usefulness in solar energy forecasting. Nevertheless, the high resolution and accuracy needed today for the operational management of solar resources necessitates a constant development of such models and a continuous effort in the accurately representation of the physical processes involved. In coastal California, the principal source of NWP model inaccuracy is the presence of marine stratocumulus clouds. These clouds have a strong influence on the radiative balance and solar irradiance at the surface. In this work, the effect that several combinations of parameterization schemes have on simulating marine stratocumulus clouds is examined. Specifically, the cloud microphysics, cumulus, radiation and planetary boundary layer parameterization schemes implemented in the Weather Research and Forecasting Model (WRF) are analyzed. Additionally, the influence of the initial and boundary conditions in the model have been tested. In order to quantify the overall model performance, WRF- simulated cloud water path was compared to MODIS cloud data, WRF'-simulated GHI was compared to Solar Anywhere data, and WRF'-simulated vertical temperature/mixing ratio profiles were compared against atmospheric radiosondes. In general, the results revealed that the choice of parameterization in addition to the initialization dataset strongly influence the accuracy of the simulated cloud fields, the thermodynamic properties of the atmosphere, and the solar radiation reaching the surface.