Skip to content
 

Deep-learning Structure Analysis for Tropical Cyclones and its Application for Studies of Concentric Eyewalls and Climate Change

Presented by: Buo-Fu Chen - Center for Weather Climate and Disaster Research, National Taiwan University, Taipei, Taiwan
Date: September 17, 2024 12:00 am
Location: In-person in the CIRA Commons (Coffee and light refreshments provided just prior to the presentation)

Deep learning (DL) is useful in various regression tasks for tropical cyclone (TC) analysis and forecasting, including regression to the current TC radial wind profile and regression to future TC structure parameters (i.e., statistical forecasts of intensity or size). The first part of the presentation showcases the usefulness of DL for reconstructing homogenized and trustworthy global TC wind profile datasets since 1981, thus facilitating an examination of climate trends of TC structure/energy extremes. By training with uniquely labeled data integrating best tracks and numerical model analysis, our model converts multichannel satellite imagery to a 0-750-km wind profile of axisymmetric surface winds. The model performance is verified to be sufficient for climate studies by comparing it to independent satellite-radar surface winds. Moreover, the integrated kinetic energy (IKE) calculated based on the AI winds has an R2 = 0.99 against aircraft observation.

Understanding past TC trends and variability is critical for projecting future TC impacts on human society, considering the changing climate. Based on the new homogenized dataset, the major TC proportion has increased by ~13% in the past four decades. Moreover, the proportion of extremely high-energy (IKE) TCs has increased by ~25%, along with an increasing trend (> one standard deviation of the 40-y variability) of the mean total energy of high-energy TCs. Although the warming ocean favors TC intensification, the TC track migration to higher latitudes and altered environments further affect TC structure and energy.

On the other hand, long-lived and short-lived concentric eyewalls (CEs) are accompanied by diverse structural parameters and can affect TCs’ intensity. Therefore, we used the DL TC wind profiles, along with a CE dataset, to examine how CEs affect TC development pathways revealed by IKE-intensity (K-V) diagrams. Results show that short-lived CEs (duration < 20 h) tend to maintain TC intensity and IKE, while long-lived CEs (25% of all CEs) even favor IKE growth, contributing to TCs with extremely large circulation.

This study showcases that DL-generated data may help accelerate classical research and enhance our scientific understanding of TCs.