Verification of the short-term forecast of the wind speed for the Gibara II wind farm according to the prevailing TSS

Abstract: In Cuba, short-term predictions have been developed for wind speed in the Gibara wind farms. These predictions present an absolute mean error (EMA) that sometimes exceeds 3 m/s. This study has the aim of verify the wind forecast generated by SisPI using the Synoptic Situation Types Catalog (TSS), a wind speed observation data provided by the anemometers installed in the wind turbine. The study period spanned from May 2020 to April 2021. For the evaluation were used the metrics: root mean square error (RMSE) and EMA, and the analysis was made in the rainy and dry seasons, through the methodology developed by (Patiño, 2022). Results indicate that the subtype 3 (Extended undisturbed anticyclonic flow) was the one with the highest frequency of cases between very good and good in both seasonal periods. Subtype 19 (migratory anticyclone in an advanced state of transformation) was the system that produced the worst results in the dry season, with the largest number of cases of bad wind speed forecasts. The results of the statisticians: bias (BIAS) and Pearson's Correlation Coefficient (r), were very favorable.