Physics-Based Framework for Wind Speed Estimation from Wind-Borne Debris Flight Imagery

Non-synoptic wind events, such as tornadoes and downbursts, account for significant weather-related fatalities and billions of dollars in property damage annually in the United States. Yet understanding of near-surface wind characteristics remains limited, in large part due to measurement challenges. Debris flight videos captured by smartphones, drones, and cameras offer widespread documentation of extreme wind events, but applying physics-based equations of motion to estimate wind speeds requires tracking six-degree-of-freedom (6-DOF) motion, which single-view camera technology cannot reliably capture. This study presents a framework for estimating wind speeds from debris flight imagery by simplifying debris motion to two degrees of freedom (2-DOF) representing translational kinematics, while retaining physics-based equations of motion that account for debris mass and aerodynamic properties. The framework integrates a 6-DOF debris flight model with computer graphics techniques to generate a virtual testbed that projects synthetic debris trajectories onto 2-D image planes, enabling systematic evaluation under various uncertainties. Comprehensive Monte Carlo validation demonstrates that the 2-DOF simplification yields reliable wind speed estimates when key parameters are properly characterized, achieving high accuracy for compact debris and demonstrating practical applicability across non-compact debris geometries including plates and rods. Global sensitivity analysis quantifies the relative influence of uncertainty sources, establishing which parameters are critical for practical implementation. This framework provides a practical foundation to leverage widespread video documentation from tornadoes and other extreme wind events for addressing persistent gaps in near-surface wind speed measurements.