Atmospheric mesoscale modeling to simulate annual and seasonal wind speeds for wind energy production in Mexico

Numerical models have been used widely to reproduce wind resources around the globe. Mexico's vast territory has a wide range of geographical characteristics with abundant wind potential. This work explores WRF simulations applied to reproduce the wind speed and capacity factor (CF) of 22 wind masts, organized into seven regions delimited by geographic conditions and consisting of 33 years of data. Biases, correlations, dispersion indexes and terrain gradient are selected to study the model and experimental data annually and seasonally. Results indicate that WRF simulations show a persistent positive bias in all regions, leading to overestimating CF. In a seasonal analysis, 86% of the CF data falls between the -0.1 and 0.1 bias range. Bias is not related to a physical seasonal phenomenon; instead, it appears to be related to geographic conditions. The findings indicate that different combinations of settings should be chosen to better reflect the geographical conditions and physical phenomena that affect the intricate Mexican landscape for wind energy production. This research identify regions with best reproducibility and suggests potential areas for future research on wind energy forecasting.