Sensitivity of simulated storm surge to tropical cyclone wind forcing around islands: insights from Hurricane Irma
Abstract
In September 2017, Category-5,Hurricane Irma made landfall on the small islands of Barbuda and Saint Martin, causing exceptional and devastating storm surges. Understanding such extreme surges is essential for improving their forecast and designing effective risk management strategies. One of the main challenge for their accurate modelling lies in the strong uncertainties in the characterisation and representation of tropical cyclone (TC) atmospheric conditions. The present study investigates the sensitivity of simulated surge to TC position, wind intensity, structure, and surface stress formulation, through experiments based on scenarios exploring their range of uncertainty. Simulations with the Coastal and Regional Ocean COmmunity model reveal that the surge peak (over 2 m) at two tide gauge locations is primarily driven by extreme winds in the cyclone eyewall, and is correctly captured only when specific cyclone track and wind stress intensity are met, while being completely missed when relying on global atmospheric reanalysis winds. Relative to a reference simulation, changes in the wind stress formulation or the TC position reveal variations in the peak surge amplitude of up to 52% and 44%, respectively. Meanwhile, the peak surge duration is highly sensitive to the wind structure. The impact of structure and track is also shown to be strongly site-dependent. These results highlight the need for realistic cyclone intensities, structures, and tracks for an accurate surge modelling, especially when intense cyclones pass over, or at a few kilometres, from volcanic islands. They also provide guidance on the TC uncertainties to be accounted for when simulating surge, since deriving cyclone characteristics, even from state-of-the-art observations, remains challenging.
Related
- Skillful Seasonal Tropical Cyclone Prediction with AI Weather--Climate Models Forced by Dynamically Predicted SSTs
- Tibetan Plateau uplift increased western North Pacific tropical cyclone frequency
- When anticipation fails: A forensic analysis of Tropical Cyclone Ditwah reveals whole-system performance limits in a small-island flood-risk system
- Hurricane Laura-induced loads on wind turbines -- Part I: High-fidelity modeling and sensitivity to turbulent inflow conditions
- Tropical Cyclone-Following Dataset for Limited-Area Machine Learning Applications
BibTeX
@article{zribi2026sensitivity,
title = {{Sensitivity of simulated storm surge to tropical cyclone wind forcing around islands: insights from Hurricane Irma}},
author = {A. Zribi and S. Jullien and N. Herry and X. Bertin and G. Dodet},
journal = {Ocean Modelling},
year = {2026},
month = jun,
doi = {10.1016/j.ocemod.2026.102780},
url = {https://hal.science/hal-05676260v1},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Sensitivity of simulated storm surge to tropical cyclone wind forcing around islands: insights from Hurricane Irma
AU - A. Zribi
AU - S. Jullien
AU - N. Herry
AU - X. Bertin
AU - G. Dodet
PY - 2026
DA - 2026/06/15
JO - Ocean Modelling
DO - 10.1016/j.ocemod.2026.102780
UR - https://hal.science/hal-05676260v1
AB - In September 2017, Category-5,Hurricane Irma made landfall on the small islands of Barbuda and Saint Martin, causing exceptional and devastating storm surges. Understanding such extreme surges is essential for improving their forecast and designing effective risk management strategies. One of the main challenge for their accurate modelling lies in the strong uncertainties in the characterisation and representation of tropical cyclone (TC) atmospheric conditions. The present study investigates the sensitivity of simulated surge to TC position, wind intensity, structure, and surface stress formulation, through experiments based on scenarios exploring their range of uncertainty. Simulations with the Coastal and Regional Ocean COmmunity model reveal that the surge peak (over 2 m) at two tide gauge locations is primarily driven by extreme winds in the cyclone eyewall, and is correctly captured only when specific cyclone track and wind stress intensity are met, while being completely missed when relying on global atmospheric reanalysis winds. Relative to a reference simulation, changes in the wind stress formulation or the TC position reveal variations in the peak surge amplitude of up to 52% and 44%, respectively. Meanwhile, the peak surge duration is highly sensitive to the wind structure. The impact of structure and track is also shown to be strongly site-dependent. These results highlight the need for realistic cyclone intensities, structures, and tracks for an accurate surge modelling, especially when intense cyclones pass over, or at a few kilometres, from volcanic islands. They also provide guidance on the TC uncertainties to be accounted for when simulating surge, since deriving cyclone characteristics, even from state-of-the-art observations, remains challenging.
SN - 1463-5003
SN - 1463-5011
ER -