16A.2 Solar Energy Nowcasting Case Studies Near Sacramento

Friday, 3 July 2015: 10:45 AM
Salon A-2 (Hilton Chicago)
Jared A. Lee, NCAR, Boulder, CO; and S. E. Haupt, P. A. Jimenez, T. C. McCandless, M. A. Rogers, and S. D. Miller

With increasing penetration of solar energy, the need for reliable solar power forecasts is also increasing. The intermittency and variability of solar power generation provides challenges for electrical grid operators. In order to perform load balancing properly, accurate forecasts of both the irradiance and its variability at solar farms are required on the nowcasting time scale (minutes to hours ahead) with high temporal frequency.

To meet these challenges, the SunCast solar energy forecasting system was developed. SunCast includes several components that can be blended together to form a single 15-minute average irradiance (and power) forecast in intervals of 15 minutes out to 6 hours. These components include site-based statistical learning methods (StatCast), satellite cloud advection (CIRACast), and a modified version of the Weather Research and Forecasting (WRF) numerical weather prediction model (WRF-Solar).

In this study we compare the performance of StatCast, CIRACast, and WRF-Solar irradiance forecasts over four case days near Sacramento, California, at pyranometers operated by the Sacramento Municipal Utility District. Each of the four case days chosen is from a different sky cover regime: clear, mixed clouds and sun, morning fog or marine stratus, and overcast. Each of these sky cover regimes presents unique challenges to the SunCast component systems. This comparison will reveal the strengths and shortcomings of the various techniques, and point to how we can further improve these models.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner