Maximum Likelihood Estimation (MLE) Approach for Determining Most Representative Solar Resource Data Set for United States

Solar projects in the US benefit from having many irradiation data sources available for comparative analysis. Not many countries in the world have a database extending back to the 1960s as well as multiple providers of satellite-derived irradiance estimates. The National Solar Resource Database ("NSRDB") program envisioned to support the growth of the solar market by laying a strong foundation for irradiation measurement. The "site-specific" NSRDB modeling approach transitioned in 2014 to use a satellite-derived Physical Solar Model ("PSM"). Currently, several additional satellite-derived data sets are commercially available which provide more recent estimates of irradiation data. The key question is: which one of these data sets is likely to be most representative of the 'long term' resource for a site? This paper explores using the concept of Maximum Likelihood Estimation (MLE) to be applied to several equally relevant data sources to provide feedback on long-term weather selection.