Parametric and Machine Learning Approach for describing Hurricane Offshore Winds over the U.S Northeast Outer Continental Shelf

Assessment of hurricane offshore wind structures are of high importance to site characterization for risk and safety assessments over the U.S Northeast Outer Continental Shelf. The parametric approach for modeling hurricane winds is very efficient for catastrophe risk analysis in coastal areas and easy to implement with few TC parameters but has not been assessed for use in characterizing hurricane wind fields for offshore wind energy over the U.S Northeast Outer Continental Shelf. This study assesses the adequacy of existing parametric hurricane models and investigates the need of utilizing a machine learning approach for describing winds over the full profile of the turbine. We assess the cases of historical hurricanes that impacted the U.S Northeast Outer Continental Shelf from variants of parametric wind field models (e.g., Holland; Emanuel) to determine how they perform across hurricane characteristics and locations for offshore wind farm applications. Implications for risk and safety assessments that use such parametric models will also be discussed.