Abstract:Anthropogenic influences have been linked to tropical cyclone (TC) poleward migration, TC extreme precipitation, and an increased proportion of major hurricanes [1, 2, 3, 4]. Understanding past TC trends and variability is critical for projecting future TC impacts on human society considering the changing climate [5]. However, past trends of TC structure/energy remain uncertain due to limited observations; subjective-analyzed and spatiotemporal-heterogeneous "best-track" datasets lead to reduced confidence in the assessed TC repose to climate change [6, 7]. Here, we use deep learning to reconstruct past "observations" and yield an objective global TC wind profile dataset during 1981 to 2020, facilitating a comprehensive examination of TC structure/energy. By training with uniquely labeled data integrating best tracks and numerical model analysis of 2004 to 2018 TCs, 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. 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 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/energy. This new deep learning method/dataset reveals novel trends regarding TC structure extremes and may help verify simulations/studies regarding TCs in the changing climate.

Reduced Tropical Cyclone Genesis in the Future As Predicted by a Machine Learning Model

Basin Dependence of the Tropical Cyclone Genesis Environment and the Future Changes Revealed by Machine Learning Methods

Prediction of Tropical Cyclone Genesis from Mesoscale Convective Systems using Machine Learning

A Multiscale‐Model‐Based Near‐Term Prediction of Tropical Cyclone Genesis Frequency in the Northern Hemisphere

Tropical Cyclone Genesis Prediction Based on Support Vector Machine Considering Effects of Multiple Meteorological Parameters

Projected future changes of tropical cyclone genesis frequency in the Northern Hemisphere based on a multi-timescale regression model

Future Projection of Tropical Cyclone Genesis in the Western North Pacific Using High-resolution GCMs and Genesis Potential Indices

TC‐GEN: Data‐Driven Tropical Cyclone Downscaling Using Machine Learning‐Based High‐Resolution Weather Model

Tropical Cyclone Genesis over the Western North Pacific Simulated by Coupled Model Intercomparison Project Phase 6 Models

Probabilistic Forecasting of Tropical Cyclones Intensity Using Machine Learning Model

Patterns and frequency of projected future tropical cyclone genesis are governed by dynamic effects

Future Changes of Tropical Cyclone Activity over the West Pacific under the 1.5°C and 2°C Limited Warming Scenarios Using a Detecting and Tracking Algorithm

A Neural Network Framework for Fine-Grained Tropical Cyclone Intensity Prediction

A Toy Model for the Global Annual Number of Tropical Cyclones

Future Projections of Multiple Tropical Cyclone Events in the Northern Hemisphere in the CMIP6‐HighResMIP Models

Climate Trends of Tropical Cyclone Intensity and Energy Extremes Revealed by Deep Learning

A Deep Learning-Based Global Tropical Cyclogenesis Prediction Model and Its Interpretability Analysis

A Study on Tropical Cyclone Genesis Characteristics in the Northwest Pacific Based on Multidimensional Analysis Using EWM and TOPSIS Methods

How Does Tropical Cyclone Genesis Frequency Respond to a Changing Climate?

A Machine Learning-Based Parameterized Tropical Cyclone Precipitation Model

A Statistical Intra-Seasonal Prediction Model of Extended Boreal Summer Western North Pacific Tropical Cyclone Genesis