ABSTRACT
Solar forecasting must have a physical aspect, in that, sky camera imagery, satellite imagery, and numerical weather prediction ought to be involved when making solar forecasts. This chapter outlines the base forecasting methods, which has hitherto been used by expert solar forecasters. The first section of the chapter focuses on numerical weather prediction, which is a two-step procedure, one diagnostic and the other prognostic in nature. Various concepts related to numerical weather prediction, such as the primitive equations, physical parameterization, or data assimilation, are outlined carefully to guide those who are not familiar with these atmospheric science topics. Following that, the satellite-based forecasting is introduced in terms of how irradiance can be retrieved from remote-sensing data and how the retrieved field can be advect forward in time to generate forecasts. The last section of the chapter deals with statistical forecasting with ground-based data. Particularly emphasis is placed on the spatio-temporal statistical methods, which have clear advantage over univariate methods owing to their ability to capture evolution of irradiance fields.
