Estimating Extreme Gusts via Optical Flow Tracking from Defoliated Tree Motion
Abstract
Peak 3-second gusts (s₃) during severe storms are critical drivers of wind damage, yet their fine-scale dynamics remain difficult to measure. Visual Anemometry (VA) unlocks the possibility of spatially dense wind speed estimates, yet its performance under extreme winds remains poorly understood. Here, we combine physics-based simulations of tree motion with Optical Flow Tracking Velocimetry (OFTV), allowing a first assessment of s₃ estimation for winds of 25-144 km/h across camera angles of 0-180°. Modelled tree velocities increased approximately linearly with input wind speed and were captured well by OFTV (R² ≈ 0.9, p < 0.05), mechanistically demonstrating the validity of VA for extreme winds and structural loading. Against expectations, both predictive performance and the relationship between video motions and estimated s₃ are not significantly affected by camera angle (0°-180°), importantly indicating that multi-branch tree motion produces an angularly isotropic response. Moreover, directional effects reduce, becoming negligible, under strong winds. Demonstrating the robustness of tree-based VA for extreme wind gusts (e.g. no direction-dependent calibration needed) is a foundational step towards enhanced gust detection, hazard mapping, and risk assessment for planning and insurance.
Related
- Experimental validation of an open-source low-cost single-camera 6-DOF tracking of floating-body motion in wave tanks
- Estimating the Completeness Magnitude m c and the b ‐Values in a Snap
- Approaches and challenges in estimating economic loss from tsunami damage models
- Machine learning and spatial regression approaches to estimating willingness to pay for ecosystem services in Jakarta and Taoyuan
- Estimating Regional Snow Water Equivalent Using a Physics‐Informed Graph Convolution Network Model
BibTeX
@misc{kulkarni2026estimating,
title = {{Estimating Extreme Gusts via Optical Flow Tracking from Defoliated Tree Motion}},
author = {Sai Kulkarni and John Hillier and Tim Marjoribanks and Sarah Bugby and Daniel Bannister and Jonny Higham and Rufus Dickinson and Matthew Baddock},
howpublished = {EarthArXiv},
year = {2026},
month = sep,
doi = {10.31223/x5780b},
url = {https://eartharxiv.org/repository/view/15172/},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Estimating Extreme Gusts via Optical Flow Tracking from Defoliated Tree Motion
AU - Sai Kulkarni
AU - John Hillier
AU - Tim Marjoribanks
AU - Sarah Bugby
AU - Daniel Bannister
AU - Jonny Higham
AU - Rufus Dickinson
AU - Matthew Baddock
PY - 2026
DA - 2026/09/23
PB - EarthArXiv
DO - 10.31223/x5780b
UR - https://eartharxiv.org/repository/view/15172/
AB - Peak 3-second gusts (s₃) during severe storms are critical drivers of wind damage, yet their fine-scale dynamics remain difficult to measure. Visual Anemometry (VA) unlocks the possibility of spatially dense wind speed estimates, yet its performance under extreme winds remains poorly understood. Here, we combine physics-based simulations of tree motion with Optical Flow Tracking Velocimetry (OFTV), allowing a first assessment of s₃ estimation for winds of 25-144 km/h across camera angles of 0-180°. Modelled tree velocities increased approximately linearly with input wind speed and were captured well by OFTV (R² ≈ 0.9, p < 0.05), mechanistically demonstrating the validity of VA for extreme winds and structural loading. Against expectations, both predictive performance and the relationship between video motions and estimated s₃ are not significantly affected by camera angle (0°-180°), importantly indicating that multi-branch tree motion produces an angularly isotropic response. Moreover, directional effects reduce, becoming negligible, under strong winds. Demonstrating the robustness of tree-based VA for extreme wind gusts (e.g. no direction-dependent calibration needed) is a foundational step towards enhanced gust detection, hazard mapping, and risk assessment for planning and insurance.
ER -