Machine learning-based correlation analysis of decadal cyclone intensity with sea surface temperature: data and tutorial

TL;DR

This paper analyzes the correlation between sea surface temperature and cyclone intensity over 40 years using machine learning, providing datasets, visualization, and insights into cyclone genesis.

Contribution

It introduces a curated dataset and open-source tools for analyzing the relationship between SST and cyclone intensity with machine learning methods.

Findings

Strong positive correlation between SST and cyclone wind speed

Linear regression effectively models cyclone intensity

Open dataset enables further research on cyclone genesis

Abstract

The rising number of extreme climate events in the past decades has motivated the need for a thorough consideration of tropical cyclone genesis and intensity, given the sea-surface temperature (SST). In this paper, we present an analysis of the relationship between the increasing global SST with cyclone genesis using linear regression machine learning models. We extract and curate a dataset of tropical cyclones across selected ocean basins with their associated SST over the past 40 years. We provide correlation analysis using linear regression and visualisation strategies. Our preliminary results show a strong positive correlation between SST and high wind speed across selected ocean basins via linear regression and machine learning models. Our dataset and available open-source code offer a novel perspective for the investigation of the genesis and intensity of tropical cyclones. Alongside the time and position of each cyclone, we also provide the related Saffir-Simpson category, season, wind speed, and SST for 15 days before and after the tropical cyclone genesis.