A predictive tool was tailored by the U.S. National Center for Atmospheric Research and the Electric Power Research Institute to help New York state manage its expanding solar power infrastructure. NYSolarCast generates 15 minute resolution predictions of global horizontal irradiance (GHI) on a 3 km grid covering the state. The GHI data, which provide an indicator of potential electricity production from solar plants, are updated every 15 minutes at select solar farms and on an hourly basis for each zip code to estimate solar power generation. Forecasts of irradiance across New York state are produced every 15 minutes by NYSolarCast. Source: U.S. National Center for Atmospheric Research

NYSolarCast combines numerical weather prediction with machine learning techniques, constrained by ground-based observations. An analysis of the accuracy of the forecasting tool detailed in the journal Solar Energy reveals that certain key pieces of data, such as weather conditions or solar irradiance levels, are only relevant for a short period, around 45 minutes before conditions change.

Forecasts made by NYSolarCast over a one-year period were observed to be superior to those derived from tools situated precisely at weather stations. This is likely because the system could analyze real-time data straight from active solar farms, unlike weather stations, which had limited access to real-time data.

The source code for NYSolarCast is available for download on GitHub.