Improved Heliosat4 cloud parallax and shadow corrections using gridded cloud top height data

Kurzfassung

The CAMS radiation model utilizes images from geostationary meteorological satellites to assess cloud presence and cloud properties to estimate solar irradiance at the Earth’s surface. Previous studies have investigated the effects of satellite viewing parallax and shadow effects of the cloud position, but a correction for such effects has not yet been implemented on the current CAMS model. The parallax and shadow displacements are based on cloud top height, solar and satellite viewing angles. This study focuses of an improved geometrical correction for a future implementation of these effects on the current model to obtain an improved estimation of Global Horizontal Irradiance (GHI) using the Heliosat-4 Method. These adjustments were estimated by: (i) retrieving pixel wise satellite estimated cloud top temperature data to calculate cloud top height (CTH) using atmospheric temperature profiles. (ii) Using CTH on geolocated cloudy pixels and viewing angles of each pixel to estimate the ground displacement from the parallax effect. (iii) The geolocation of the pixels for each of the displacements are found using haversine formulae and solar angles of those locations are determined to calculate the shadow displacement. (iv) Cloud Optical Thickness (COT), Cloud Coverage, and Cloud Type of the displaced pixels are shifted to new geolocated points and a bilinear interpolation is performed to estimate the values of those parameters on the original gridded map. (v) Global horizontal irradiance (GHI) of pixels where ground meteorological stations are available is estimated using the Heliosat-4 method. The corrected GHI values found were compared against solar surface irradiance (SSI) ground observation data and CAMS uncorrected data. Results showed and overall 0,71% MBE, 1,61% MAE and 2,55% RMSE reduction, when validated against the current model. The parallax/shadow adjusted irradiation estimates enable better solar resource retrieval. This demonstrates the potential of the implementation of the correction method with the Heliosat method to better estimate SSI and advance solar modeling capabilities. Further refinements to the parallax correction approach and model parametrization are discussed as avenues for future research