BIO

I am a sophomore meteorology student from Tallahassee, Florida. My research interests include improvement of intensity forecasts for tropical cyclones and severe weather events. More broadly, I am passionate about programming and its applications to meteorology. My hobbies are gaming, exercising, and keeping up with college sports.

Machine Learning for Loop Current Detection

Abstract

The Loop Current (LC) is a warm ocean current located in the Gulf of Mexico. It provides a robust heat source allowing tropical cyclones to rapidly intensify. A recent example of this phenomenon is Category 4 Hurricane Ida in 2021. While manual detection of the LC is trivial, automated detection can be challenging due to the complex dynamics of the LC and its eddies. A variable known in the literature as Sea Surface Height (SSH) was derived, and a 17 cm threshold for SSH was applied to an oceanic dataset to find the LC. Once the LC threshold was obtained, computer vision (CV) techniques were applied to make the LC continuous and isolated. A contour line was drawn around the boundary of the LC, and its perimeter was computed from the dataset. The evolution of its perimeter over time can be used to describe some dynamics of the LC, such as stretching and shedding events. In addition, extreme events, such as hurricane interactions with the LC, may affect its perimeter.

Keywords: machine learning loop current oceanography