Global Tropical Cyclone Intensity Forecasting with Multi-modal Multi-scale Causal Autoregressive Model

TL;DR

This paper introduces MSCAR, a novel causal autoregressive model that integrates multi-modal data for improved global tropical cyclone intensity forecasting, supported by a comprehensive new dataset.

Contribution

The paper presents MSCAR, the first model combining causal relationships with large-scale multi-modal data for TC intensity prediction, and introduces the SETCD dataset.

Findings

MSCAR outperforms state-of-the-art methods in global and regional forecast accuracy.

Achieves maximum error reductions of 9.52% globally and 6.74% regionally.

Provides a publicly available dataset for TC research.

Abstract

Accurate forecasting of Tropical cyclone (TC) intensity is crucial for formulating disaster risk reduction strategies. Current methods predominantly rely on limited spatiotemporal information from ERA5 data and neglect the causal relationships between these physical variables, failing to fully capture the spatial and temporal patterns required for intensity forecasting. To address this issue, we propose a Multi-modal multi-Scale Causal AutoRegressive model (MSCAR), which is the first model that combines causal relationships with large-scale multi-modal data for global TC intensity autoregressive forecasting. Furthermore, given the current absence of a TC dataset that offers a wide range of spatial variables, we present the Satellite and ERA5-based Tropical Cyclone Dataset (SETCD), which stands as the longest and most comprehensive global dataset related to TCs. Experiments on the dataset show that MSCAR outperforms the state-of-the-art methods, achieving maximum reductions in global and regional forecast errors of 9.52% and 6.74%, respectively. The code and dataset are publicly available at https://anonymous.4open.science/r/MSCAR.