Cloud nowcasting: motion analysis of all-sky images using velocity fields

July 18th and 19th 2013 Barcelona 61 Cloud monitoring and prediction is crucial to air traffic and weather forecast. Clouds play a key role in solar radiation balance. These issues have prompted Sieltec Canarias S.L. and AEMET to develop an automatic system of cloud observation (SONA). This system provides cloud cover percentage using all-sky images and data from a multichannel photometer, which are processed by neural networks. Once the system has detected and recorded cloud images, a cloud motion technique makes possible to obtain several “velocity fields”, corresponding to different layers of clouds. The evolution in time of these velocity fields is a powerful tool to provide very short term forecasting (less than 1 hour) of cloudiness (known as cloud nowcasting). In the future, this information, combined with Radiative Transfer Models, may be used to predict solar radiation (Global Horizontal Irradiance –GHIand Direct Normal Irradiance –DNI-) (A) Original image. (B) Image analysed by the neural network with cloud pixels marked as green. (C) Image with two clustered velocity fields marked as green and yellow, respectively. Cloud Nowcasting: Motion Analysis of All-sky images using velocity Fields Yezer Gonzalez148, Cesar Lopez148, Emilio Cuevas149 17VS053