Natural Hazards Research Summit 2024: Estimating Near Surface Wind Characteristics from Wind-borne Debris Flight Media

Non-synoptic wind events, including thunderstorms, downbursts, and tornadoes, account for most weather-related fatalities in the United States and cost billions of dollars in property damage each year. Despite advancements in incorporating non-synoptic winds into wind design codes and standards, there remains a significant gap in understanding the near-surface wind characteristics of these events. However, with the advent of smartphones, video monitoring systems, and remote sensing technologies, a wealth of media now exists that capture the real-time dynamics of non-synoptic storms, including the motion of wind-borne debris, which could be a useful source of data. This study aims to establish a framework to infer near-surface wind characteristics from debris speed and other debris attributes in debris flight videos and other media. The proposed framework involves three key steps: (1) accurate extraction of debris trajectories from videos in 2D or 3D space as a function of time using computer vision techniques, (2) estimation of debris attributes relevant to debris-flight motion (e.g., mass, area, aerodynamic coefficients) from the video, and (3) application of equations of motion to (1) and (2) to infer background wind flow characteristics. This study focuses on applying the framework to unstructured debris-flight media (publicly shared videos) and structured debris flight media from controlled experiment at the NHERI Wall of Wind shared-use facility at Florida International University. Overall, this study shows the possibility to enhance our understanding of near-surface wind and debris characteristics using wind-borne debris trajectories extracted from debris flight media.