kdayday/solarbenchmarks: Intra-Hour and Intra-Day Probabilistic Solar Forecast Benchmarks

Implements common benchmark probabilistic solar forecast methods for intra-hour (very short-term) and intra-day (short-term) forecasting. Benchmark methods are intended as baselines against which forecast improvements are assessed. Six benchmark classes are available: climatology, complete history persistence ensembles, persistence ensembles, numerical weather prediction ensembles, Gaussian error distributions, and Markov- chain mixture distributions. Open-source SURFRAD (irradiance observations) and CAMS McClear (clear-sky irradiance estimates) data are packaged for easy comparison across the 7 SURFRAD locations in the U.S.

install.packages("remotes")
remotes::install_github("kdayday/solarbenchmarks")