Date of Award

4-2025

Degree Name

Master of Science

Department

Geography

First Advisor

Laiyin Zhu, Ph.D.

Second Advisor

Kathleen Baker, Ph.D.

Third Advisor

Lei Meng, Ph.D.

Keywords

Cumulative distribution function, ERA5-land, teleconnection, Weibull, wind, XGBoost

Access Setting

Masters Thesis-Open Access

Abstract

This study investigates Michigan wind speeds by combining climatology with wind energy, identifying wind drivers and applying them to wind farm infrastructure. Using eXtreme Gradient BOOSTing (XGBoost), this thesis develops a highly accurate predictive model for wind speeds. Coupled with Shapley Additive exPlanations (SHAP) analysis, it is found that proximity to the nearest Great Lake is the ultimate influencer of wind speeds, followed by local environmental variables and finally remote climate phenomena. Other analyses find wind speeds are greatest near the Great Lakes and during fall and winter, while inland summer wind speeds were lowest. Wind energy infrastructure is assessed through a climatological and economic lens. While wind infrastructure considers many criteria when choosing a site, it is found that many wind farms are not sited in areas with favorable wind speeds. Despite having 50 wind farms with hundreds of turbines, Michigan is found to be generally unsuitable for wind energy.

These findings illuminate important considerations in wind and energy forecasting, which ultimately enhance energy security and strategic siting of Michigan renewable energy infrastructure. By marrying climatology and renewable energy, this thesis emphasizes the importance of the integration of climate science with renewable energy siting to enhance sustainability and efficiency.

Download

Share

COinS