Publicações

The downward global horizontal irradiance (GHI) modulates various physical processes on Earth and is essential in climate, ecology, and renewable resources studies. In recent years, the use of geostationary satellites for estimating downward global solar irradiance (GHI) has gained increasing interest due to their regional coverage, lower maintenance, and calibration requirements. The GHI estimates using satellite-based models rely on calculating the cloud cover index (Ceff) that depends on the assumed reflectance for totally overcast condition (ρcloud). This work investigates how the spatial resolution of satellite imagery and ρcloud variability affect Ceff evaluation and the skill metrics of satellite-based models to estimate GHI over a tropical region. The results show the robustness of the cloud index parameterization against pixel resolution and solar geometry. A small detachment was observed when the resolution was reduced to 1 km. Zenith and satellite phase angles induced no bias change in GHI estimates. On low cloud cover conditions, models overestimated GHI while underestimating otherwise, suggesting an internal compensation to minimize the error on all cloud cover conditions. Ranging ρcloud from 0.3 to 1.0 led to a bias deviation from 200 Wm−2 (−40%) to 40 Wm−2 (8%), recalling the importance of adequately evaluating ρcloud from satellite imagery to calculate Ceff. The clear sky model and the overcast-representative cloud reflectance, ρcloud, were the most sensitive parameters affecting model performance. In summary, when calibrating the cloud index model for a specific region, it is essential to consider both the clear sky model and the maximum cloud reflectance to achieve optimal results.