A Solar Forecasting Model Based on The National Digital Forecast Database

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Sunday, 4 January 2015
Andrew Polasky, Carleton College, Northfield, MN; and R. Ellingson

We report on the development of a medium-term solar-irradiance forecast model, based off of data from the National Digital Forecast Database (NDFD). Using the predicted values of sky cover, dew point, and temperature as inputs we created empirical and semi-empirical models for predicting irradiance up to 51 hours in advance. Validation against ground measured irradiance values has shown moderately good correlations for both of our models. However, while well correlated, the variability in the measured irradiance values is poorly captured in the models, and leads to limited forecast viability.