Comparing ground and satellite imagery for cloud identification and their impact on solar radiation

Recent advances in solar energy have driven the need for a better understanding of interactions between atmospheric phenomena and solar irradiance variability. Different aspects of cloud physics can explain the significant part of these interactions. Latest cloud observation techniques cover a broad spectrum of cloud identification, classification, flow direction, speed and height estimation on different temporal and spatial scales. Satellite imagery based techniques can provide information on macro-level for intra-day applications while ground based imaging techniques covers micro-level phenomenon of cloud development and are more suitable for intra-hour applications. Our work includes a comprehensive review of recently proposed techniques for both ground and satellite observation techniques. Available methods have been concurrently applied on satellite and ground based images to identify and track the clouds. The results of cloud identification at micro and macro scale are compared to identify 'blind-spots' in cloud formation and flow at various altitudes in the atmosphere. To validate our findings, the cloud index based on ground and satellite derived Global Horizontal Irradiance (GHI) data are compared to ground measurements. The evaluation shows the potential of ground and satellite images to model the solar radiation variability with an assumption that sudden ramps in solar irradiance are caused by the formation and movement of the clouds. Various image features are computed for both, satellite and ground images and are correlated with ground solar variability measures. Features with highest correlation are selected and their application is shown for cloud detection and solar irradiance forecasting.