Solar radiation modeling under 3D reconstructed cloud fields

The distribution of diffuse sky radiance over the sky hemisphere is frequently required in many fields and applications, like predicting the solar irradiance on sloped planes or titled solar collectors, quantifying the spectral angular effects on photovoltaics (PV) device performance, etc. Atmospheric clouds constitute the most dominant factor for solar radiation attenuation in the atmosphere, and their accurate representation in radiative transfer models is more than necessary. However, cloud information in radiative transfer models remains a challenging process characterized by uncertainties due to the incomplete knowledge of cloud optical properties, vertical structure, and their high spatiotemporal variability, reflecting high uncertainties in solar resource estimations. These uncertainties frequently stem from the use of two-dimensional (2D) information about clouds, like pixel-level image information from geostationary satellites. Radiative transfer calculations cannot neglect the inherent three-dimensional (3D) cloud structure in order to alleviate the uncertainties in the estimation of solar resource. In this study, we investigate the potential to derive the 3D morphology of clouds with a network of all-sky imagers. Their spatial displacement and synchronized exposure times provide the necessary information for a 3D reconstruction within the area around the cameras. The main objective is to combine, under different cloud conditions, the 3D reconstructed cloud fields and radiance measurements from all-sky imagers in radiative transfer simulations. In particular, the 3D reconstructed cloud fields will be used as input to 3D radiative transfer simulations (Monte Carlo code for the physically correct Tracing of photons In Cloudy atmospheres (MYSTIC) algorithm, included in the libRadtran software package) in order to calculate solar irradiance maps.