Image credit: Pixabay/CC0 Public Domain
Time is of the essence when forecasting tropical cyclones: the more advance warning a community has, the better prepared its members will be when a storm makes landfall. Currently, the path and type of tropical cyclones can only be predicted up to five days in advance.
But new research published in Journal of geophysical research investigated a new method for predicting these storms that could extend the lead time to two weeks.
The researchers created a hybrid model for longer-range tropical cyclone forecasting that combines the power and high resolution of the physics-based Weather Research and Forecasting (WRF) model with the large-scale circulation and storm path modeling capabilities of a machine learning model called Pangu-Weather (Pangu).
WRF divides the Earth’s surface into a grid of squares measuring just 2 kilometers on each side, simulating the processes involved in the development of a tropical cyclone. However, physics-based models such as WRF have some limitations in predicting storm intensity because they do not always capture changing environmental factors such as sea surface temperature or interactions with other storms.
Machine learning-based forecast models are better at predicting the tracks of tropical cyclones, but Pangu’s 25-kilometer resolution means that smaller-scale variations within a developing storm cannot be detected.
To reduce these limitations, the researchers combined the approaches of the WRF and Pangu models. They conducted six experiments over the course of two weeks to test their model design. After adjusting their models, they tested their approach as a case study on Freddy in 2023 – the longest-lasting tropical cyclone on record.
They found that the hybrid approach significantly improved tracking and intensity predictions compared to using just one modeling method. Their approach also enabled accurate predictions from five to seven days, accurately matching Freddy’s path and intensity over the entire two weeks.
The authors point out that by testing more tropical cyclones, their modeling approach could extend warning times to more than two weeks, helping vulnerable communities better prepare for severe storms.
Further information:
Liu et al., A hybrid machine learning-physics-based modeling framework for 2-week extended tropical cyclone forecasting. Journal of Geophysical Research: Machine Learning and Computation (2024). DOI: 10.1029/2024JH000207. agupubs.onlinelibrary.wiley.co…10.1029/2024JH000207
This story is republished with permission from Eos, hosted by the American Geophysical Union. Read the original story here.
Quote: Physics meets machine learning for better cyclone predictions (21 August 2024), accessed 21 August 2024 from https://phys.org/news/2024-08-physics-machine-cyclone.html
This document is subject to copyright. Except for the purposes of private study or research, no part of it may be reproduced without written permission. The contents are for information purposes only.
