Files in this item



application/pdfRHEE-DISSERTATION-2021.pdf (15MB)Restricted to U of Illinois
(no description provided)PDF


Title:Near-surface tornado wind field estimation using damage patterns
Author(s):Rhee, Daniel Minkee
Director of Research:Lombardo, Franklin T
Doctoral Committee Chair(s):Lombardo, Franklin T
Doctoral Committee Member(s):LaFave, James M; Trapp, Robert J; Cha, Eun Jeong
Department / Program:Civil & Environmental Eng
Discipline:Civil Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Near-surface Wind Speed
Damage Pattern
Abstract:Tornadoes are considered the most violent and dangerous wind hazards and are emerging as a major public concern, causing a significant amount of property loss and casualties in the past decades. In response, the engineering community in the United States is implementing tornado-based design in codes and standards. To ensure robust tornado-based design, an accurate estimate of the near-surface (10 m above ground level) wind field is essential. However, the near-surface wind field of a tornado is not well understood due to the lack of in-situ measurements and limitations of the wind speed estimation from structural damage. As a result, a method of estimating the near-surface wind field of a tornado using tree damage, known as tree-fall analysis, is often used as an independent wind speed estimation method. Tree-fall analysis is exceptionally useful in forested areas where structural damage is limited. In addition, field observations showed similar fall patterns from other damage indicators (e.g., crops, traffic signs). In this dissertation, the foundation and applications of tree-fall analysis using different damage indicators are presented with an overarching goal of improving and accurately estimating the near-surface wind field of tornadoes using damage patterns. This dissertation covers the following topics: 1) documenting tornado damage, 2) improving tree-fall analysis, 3) analyzing failure wind speed of cantilever-like damage indicators and tornado wind field models, 4) apply tree-fall analysis to different tornadoes. Post-storm damage from a total of nine tornadoes was documented in detail both on the ground and in the air. To rapidly collect the tree-fall directions of a large-scale tornado from aerial photographs, a new method that can automatically detect tree damages and obtain tree-fall patterns is developed using an image-processing technique. The method shows a 95% accuracy of detecting downed trees and 74% of the downed trees have less than 45 degrees difference in median fall-direction from the traditional method (manually “tagging” trees). Tree-fall analysis analyzes the characteristics of a tornado by examining the tree-fall pattern generated by simulating a translating idealized Rankine vortex. Different types of tree-fall patterns are examined and a method to compare the simulated pattern to the observed pattern and calibrate the vortex parameters is introduced. Herein, the critical wind speed of a cantilever-like damage indicator is used as an input parameter to simulate fall patterns. Thus, the different influential factors and methods to estimate the critical wind speed are also examined. In addition, the validity of different idealized vortex models and the feasibility of incorporating them into tree-fall analysis are investigated. Possible techniques to modify the idealized vortex models to accommodate external effects (e.g., RFD surge, topographic effects) are presented. The parameters of real tornadoes are then estimated and the near-surface wind field is recreated using the estimated parameters. The estimated wind speeds are compared with independent wind speed estimates using other methods and other damage indicators. In particular, the near-surface wind field of the Naplate, IL tornado was estimated using different damage indicators (residential buildings, trees, signs). The general agreement in wind speed estimation supports the application of any subset of the methods. The near-surface wind field of Sidney, IL, and Bondurant, IA tornadoes were estimated using tree-fall analysis applied to crops. Both cases are possible evidence of the EF Scale method underrating the tornado intensity due to no structures present in the vicinity. The tree-fall patterns of Tuscaloosa, AL, and Alonsa, MB tornado show a general pattern of increase in tornado intensity, size, and tangential flow over time. Finally, empirical fragility curves of structures and trees are also established using the estimated wind fields, which quantify the vulnerability of structures and trees to tornadoes.
Issue Date:2021-04-21
Rights Information:Copyright 2021 Daniel Rhee
Date Available in IDEALS:2021-09-17
Date Deposited:2021-05

This item appears in the following Collection(s)

Item Statistics