Improved Eyewall Replacement Cycle Forecasting Using a Modified Microwave-Based Algorithm

The use of microwave imagery for Eyewall Replacement Cycle (ERC) forecasting has been mostly subjective up until now, and also rather limited because of the lack of quantitative tools to handle the information of eyewall development. Here we apply the ARCHER TC center-fixing algorithm to quantify the presence of multiple eyewalls (or, even more importantly, emerging eyewalls) and use this information in a statistical model to predict the onset of an ERC. The model, ‘M-PERC,’ is calibrated with 1577 images from 65 North Atlantic TCs using Principal Component Analysis of the ARCHER output, and achieves a Brier Skill Score of 22% in independent testing. We will demonstrate the performance of this model on the 2017 North Atlantic season, and discuss this method’s ability to reduce TC intensity error in an operational forecasting environment.