Data-Based Models for Hurricane Evolution Prediction: A Deep Learning Approach

TL;DR

This paper introduces a deep learning approach using RNNs to predict hurricane trajectories accurately and quickly, leveraging historical data to improve short-term forecasting and reduce potential damages.

Contribution

The study develops and compares two RNN-based models for hurricane trajectory prediction, demonstrating their efficiency and accuracy over traditional ensemble methods for short-term forecasts.

Findings

Many-To-One models are less accurate than Many-To-Many models due to error accumulation.

Both models perform similarly for 6-hour predictions.

RNN models are faster than ensemble models while maintaining comparable accuracy.

Abstract

Fast and accurate prediction of hurricane evolution from genesis onwards is needed to reduce loss of life and enhance community resilience. In this work, a novel model development methodology for predicting storm trajectory is proposed based on two classes of Recurrent Neural Networks (RNNs). The RNN models are trained on input features available in or derived from the HURDAT2 North Atlantic hurricane database maintained by the National Hurricane Center (NHC). The models use probabilities of storms passing through any location, computed from historical data. A detailed analysis of model forecasting error shows that Many-To-One prediction models are less accurate than Many-To-Many models owing to compounded error accumulation, with the exception of $6-hr$ predictions, for which the two types of model perform comparably. Application to 75 or more test storms in the North Atlantic basin showed that, for short-term forecasting up to 12 hours, the Many-to-Many RNN storm trajectory prediction models presented herein are significantly faster than ensemble models used by the NHC, while leading to errors of comparable magnitude.