Secondary Eyewall Detection in Passive Microwave Imagery

Tropical cyclone secondary eyewall formation complicates forecasting as they often result in a larger area of hazardous conditions and fluctuations in the tropical cyclone intensity. However, the factors behind the formation and evolution of secondary eyewalls are largely unknown. Due to limitations in ground-based observing platforms, the formation and evolution of secondary eyewalls are rarely observed in situ. Instead, to study secondary eyewalls, forecasters and researchers rely on remote sensing platforms like satellite passive microwave sensors, which allow an observer to look underneath cloud tops and at the tropical cyclone convective structure like secondary eyewalls.

The Tropical Cyclone Precipitation, Infrared, Microwave, and Environmental Dataset (TC PRIMED) is a dataset centered around satellite passive microwave observations of tropical cyclones. TC PRIMED’s extensive collection of tropical cyclone passive microwave observations, location and intensity, and environmental characteristics make it an excellent data source to study secondary eyewalls. However, with over 197,000 observations of tropical cyclones, manually identifying secondary eyewalls in TC PRIMED requires extensive labor.

Here, we present preliminary work to develop a machine learning model to detect secondary eyewalls in satellite passive microwave observations. To reduce biases from the different characteristics of the different sensors in TC PRIMED, we first reanalyze a subset of TC PRIMED passive microwave observations onto a cylindrical grid using a radial and azimuthal filter. Then, we manually identify secondary eyewalls in this observation subset using an expert-based consensus approach. Using this labeled dataset, we train our machine learning model and investigate the model performance. By developing this machine learning model, we will be able to identify all secondary eyewall observations in TC PRIMED, which allows for a more holistic study of tropical cyclone secondary eyewalls.

Disclaimer: The scientific results and conclusions, as well as any views or opinions expressed herein, are those of the author(s) and do not necessarily reflect those of NOAA or the Department of Commerce.