Probabilistic Semi-Dynamical Nowcasts of Solar Radiation Data Globally via Solcast Freemium Web API Distribution.

At higher solar penetration levels, grid integration becomes challenging due to inherent solar production forecast uncertainty. Minimising the risk and cost associated with reserve generation, demand response and energy storage requires not only accurate forecasts, but sharp and reliable probability information to inform conservative decision making under uncertainty. Between 15 minutes and 4 hours ahead, probabilistic cloud "nowcasts" can be made using the current observed cloud state from satellites, resulting in a narrowing envelope of solar radiation uncertainty at shorter lead times, which can inform utility decision making and help drive automated systems. We present and demonstrate a global, probabilistic, rapid-update nowcasting system for solar radiation data and PV power which can immediately provide live (real-time) and nowcast data globally through a web API (https://solcast.com). Two types of real-world case studies are demonstrated. The first is improved production forecasting for a utility-scale solar farm, the second case is improved electricity demand forecasts accounting for the impacts of collective behind-the-meter PV production. [ABSTRACT FROM AUTHOR]