Solar forecast error can be represented in electric utility production
cost modeling to assess impacts of uncertainty in different scenarios.
We use reforecasts (a.k.a., historical forecasts, hindcasts) that were
developed to be concurrent with synthetic solar generation profiles
(matching each hour in the same year), along with asynchronous forecasts
(i.e., with errors remapped from different hours and/or years), to
explore impacts on model results. We demonstrate that production cost
increases non-linearly with forecast error, and that our method for
remapping asynchronous forecast errors to a solar generation profile
produces results that are similar to the concurrent forecasts, but with
about 20\% lower cost impacts due to forecast error.