3.2 The Impact of Early Morning Infrared Satellite-Based Cloud Advection Forecasts on Solar Forecast Error in Hawaii

Monday, 7 January 2019: 10:45 AM
North 129A (Phoenix Convention Center - West and North Buildings)
Steven H. Young, UL, Albany, NY; and D. Kirk-Davidoff

Accurate short-term forecasts of solar generation are of great value to electrical system operators, allowing for more efficient deployment of reserve generation capacity to balance electrical supply and demand. Solar irradiance estimates based on geostationary satellite observations provide a high-resolution map of estimated irradiance in the absence of a dense network of surface based sensors. These maps can be used to estimate cloud motions and predictions of irradiance from minutes to a few hours into the future. When the solar elevation angle is greater than about 15°, visible imagery alone is sufficient to provide significant forecast value. However, the quality of visible satellite-based cloud estimates degrades for solar elevations less than about 15°. When the solar elevation is below about 7°, visible satellite based estimates have little to no value. Here we report on 1) the use of infrared channels to estimate cloud fields in the early morning from several hours before sunrise to shortly after sunrise when the solar elevation angle is low, and 2) the use of infrared-based cloud estimates to generate short-term cloud-motion based forecasts for the first few hours after sunrise.
In this project, we generate predawn and early morning forecasts for the state of Hawaii from infrared satellite data for several months in 2018. We evaluate the benefit to our operational solar forecast for 386 sites on the islands of Hawaii, Maui, Molokai, Lanai and Oahu when this forecast method is added to the other methods available to the optimized ensemble forecast generator. This method will be extended to use the additional infrared channels and higher resolution imagery as soon as sufficient GOES-17 data is available.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner