Damaging tornadoes are possible in all U.S. regions beyond the traditionally recognized "tornado alleys" in the southern plains and mid-south. Hence, this research examines how study participants respond to rare tornado events. Specifically, how disaster experiences affect risk perceptions and protective actions when facing a rare tornado event. This study used the protective action decision model (PADM) as the theoretical lens to investigate this issue. The PADM shows that personal characteristics, including personal disaster experience, could affect risk perceptions and how protective action decisions are made. This study applies the theory by recruiting 136 human subjects from the state of Washington to participate in a social experiment. The participants take part in a study that includes hypothetical tornado scenarios, ranging from thunderstorm warning to a tornado emergency then answering questions regarding the presented scenarios. Findings suggest that disaster experience and demographics influence how people perceive and respond to tornado disasters. Overall, people with direct and indirect tornado experiences showed significantly lower risk perceptions in the early stages of the threat when compared with those without any tornado experiences. Their tornado risk perceptions later increased when the threats were higher, and they tended to be more complacent regarding protective actions. This study applies the PADM as a theoretical basis for decision making and could assist emergency managers in creating better plans for risk information dissemination. The study could also potentially aid in gauging preparedness in different areas of a city or population, which can help with appropriate tornado protective actions.

No description available.

No description available.

Strong tornadoes have impacted the central Appalachian Mountains multiple times in recent years. The topography of this region leads to unique spatial patterns of tornado damage as the tornado vortices pass over ridges in forested areas, and this damage can be detected with vegetation indices derived from remotely sensed imagery. The objectives of this study were to 1) Classify forest damage from the April 19, 2019 EF-3 tornado in Franklin County, VA using remotely-sensed images, 2) Quantify the spatial patterns of forest damage intensity across the path using derived vegetation indices and terrain variables (primarily slope, aspect, elevation, and exposure), and 3) Use regression models to determine if relationships exist among terrain variables along the and forest damage patterns. I generated EVI and NDII vegetation indices from Sentinel-2 imagery and compared the derived damage to the underlying terrain variables. Results revealed that the two vegetation indices were effective for classifying tornado damage, and discrete damage classes aligned well with NWS EF-scale tornado intensity estimations. ANOVA testing suggested that EF-3 equivalent damage was more likely to occur on downslope topography, leeward of the tornado's direction of travel. OLS and geographically weighted regression (GWR) modeling performed poorly, suggesting that an alternative method may be more suitable for modeling, the scale of assessment was inadequate, or that important predictor variables were not captured. Overall, the intensity of the tornado was clearly modified by terrain interactions, and the remote sensing methodology used was effective for reliably identifying and rating damage in forested areas. / Master of Science / Strong tornadoes have impacted the central Appalachian Mountains multiple times in recent years. The topography of this region leads to unique spatial patterns of tornado damage as the tornado vortices pass over ridges in forested areas, and this damage can be detected with vegetation indices derived from remotely sensed imagery. The objectives of this study were to 1) Classify forest damage from the April 19, 2019 EF-3 tornado in Franklin County, VA using remotely-sensed images, 2) Quantify the spatial patterns of forest damage intensity across the path using derived vegetation indices and terrain variables (primarily slope, aspect, elevation, and exposure), and 3) Use regression models to determine if relationships exist among terrain variables along the and forest damage patterns. I generated EVI and NDII vegetation indices from Sentinel-2 imagery and compared the derived damage to the underlying terrain variables. Results revealed that the two vegetation indices were effective for classifying tornado damage, and discrete damage classes aligned well with NWS EF-scale tornado intensity estimations. ANOVA testing suggested that EF-3 equivalent damage was more likely to occur on downslope topography, leeward of the tornado's direction of travel. OLS and geographically weighted regression (GWR) modeling performed poorly, suggesting that an alternative method may be more suitable for modeling, the scale of assessment was inadequate, or that important predictor variables were not captured. Overall, the intensity of the tornado was clearly modified by terrain interactions, and the remote sensing methodology used was effective for reliably identifying and rating damage in forested areas.

Tornado warnings are life-threatening situations, and since the public uses television as the main source for tornado warning information, it is important to know how different visuals and messages are influencing the opinions and intentions of people. This research found that participants were more apt to say they would shelter if they were shown a tornado video than a radar video displaying storm-relative velocity or correlation coefficient, while seeing the tornado brought out a heightened level of anxiety and fear. Participants were most influenced by a call-to-action statement spoken by the meteorologist. Participants expressed negative feelings after being asked to both shelter and send in tornado footage at the same time. The standard call-to-action video and one with the addition of viewer-submitted tornado footage were statistically equivalent. Participants more correctly identified the location of a tornado on the radar product explained to them than one that was not.

The 2019 Dallas Tornado struck a densely populated area, was the costliest tornado in Texas history, and had minimal warning lead time, yet there were no serious injuries or fatalities. To understand why, this study examines individuals' decision-making processes during this tornado using the protective action decision model (PADM). Specifically, it investigates the factors affecting threat belief and evaluation, the facilitators and impediments to protective action, and the effects on future risk perception and hazard adjustment measures. Semi-structured telephone interviews were conducted with 23 survivors to explore their experiences and decision-making processes during this tornado. Interviews were analyzed through inductive coding and a constant comparative approach. Key findings of this study suggest that clear and direct warning messages, coupled with rapid, heuristic-driven reactions, can overcome the impediment of a short-fuse warning time and motivate those at risk to take protective action. Additionally, this study identifies condominium owners as a housing population with unique needs and impediments in the tornado recovery process. Furthermore, results illustrate how the hazard scenario and contemporary technological culture nuance protective action decision-making and future hazard adjustment measures.

[pt] A consideração dos efeitos da incidência de tornados sobre estruturas é prática corrente em vários países, principalmente no que diz respeito a instalações industriais sensíveis, nas quais as conseqüências de uma eventual ocorrência de tal fenômeno são de maior risco. No Brasil, a despeito de múltiplos registros em algumas regiões ao longo dos últimos anos, tais estudos são ainda incipientes. Objetiva-se, em particular, a avaliação de efeitos mecânicos de tornados sobre sistemas estruturais. Para isso, utiliza-se o modelo de campo de vento proposto por Wen (1975), baseado na solução qualitativa de Kuo (1971), e três modelos de pórticos tridimensionais, de alturas 20, 60 e 100 m. Através de ensaios numéricos, apresentam-se as respostas estática, cinemática e dinâmica dessas estruturas quando atingidas pelo tornado modelado. Análise comparativa entre os efeitos de inércia e arrasto é feita e é proposto um espectro de resposta cinemática a tornados. Os efeitos globais são também avaliados, a fim de se inferirem informações gerais sobre os efeitos mecânicos do fenômeno sobre sistemas estruturais. Comparações com a norma brasileira de vento são apresentadas e a simplificação de se considerar a estrutura como ponto material perante o sinistro é avaliada. / [en] The consideration of the effects of the incidence of tornadoes on structures is common in many countries, mainly in sensible industrial installations, in which the consequences of an eventual occurrence of such phenomenon are of large risk. In Brazil, in spite of multiple registers in some regions during the last couple of years, such studies are still incipient. A methodology for the evaluation of tornado mechanical effects on structural systems is presented. For this, the wind field model proposed by Wen (1975), based on the qualitative solution of Kuo (1971) is used, and three models of 3-D framed structures, 20, 60 and 100 m high, are analyzed. Through numerical tests, the static and dynamic response of these structures to a sampling tornado are evaluated and analyzed. Comparative analysis of the inertia and drag effects is made and a tornadic response spectrum is presented. Global effects are also evaluated, in order to infer general information on the mechanical actions of the phenomenon on structural systems. Comparisons with the Brazilian wind code provisions are presented and the simplification of considering the structure as a material point during the analysis is criticized.

Correlation tests were run on yearly snowfall and tornado activity data collected between water-years 1965/66 and 2010/11. Snowfall activity was evaluated using two separate measurements, which included snowfall days and daily snow depth. Tornado activity was measured through tornado days and total yearly tornado occurrences. Tornado days were defined as a 24-h period (0600–0600 UTC) during which either one or more tornadoes occurred within a chosen southern region boundary. Correlation tests revealed an absence of any mutual relationship between the snowfall and tornado activity. Three prominent teleconnections (ENSO, NAO, & AO), 6-month (Oct-Dec and Feb-Apr) and 3-month means (Oct-May) were also analyzed to reveal possible correlations with the tornado and snowfall activity. Significant negative correlations were found between ONI × tornado days; ONI × tornado totals; NAOI × snowfall days; NAOI × snow depth; AOI × snowfall days; and AOI × snow depth

abstract: Severe weather affects many regions of the United States, and has potential to greatly impact many facets of society. This study provides a climatological spatial analysis by county of severe weather warnings issued by the National Weather Service (NWS) between January 1st, 1986 to December 31st, 2017 for the contiguous United States. The severe weather warnings were issued for county-based flash flood, severe thunderstorm, and tornado phenomena issued through the study period and region. Post 2002 severe weather warnings issued by storm warning area were included in this study in the form of county-based warnings simultaneously issued for each affected county. Past studies have researched severe weather warnings issued by the NWS, however these studies are limited in geographic representation, study period, and focused on population bias. A spatial analysis of severe weather warning occurrences by county identify that (a) highest occurrences of flash flood warnings are located in the desert Southwest and Texas, (b) severe thunderstorm warning occurrence is more frequent in Arizona, portions of the Midwest, the South, and the Mid and South Atlantic states, (c) the tornado activity regions of Tornado Alley and Dixie Alley (i.e. Colorado, Kansas, Oklahoma, Arkansas, Texas, Louisiana, Mississippi, Alabama, Tennessee, and Illinois) contained the highest occurrences of tornado warnings, and (d) the highest instances of aggregate warning occurrences are found in the desert Southwest, the Midwest, and the Southern regions of the United States. Generally, severe weather warning “hot spots” tend to be located in those same regions, with greater coverage. This study concludes with a comparison of local maxima and general hot spot regions to expected regions for each phenomenon. Implications of this study are far reaching, including emergency management, and has potential to reduce risk of life. / Dissertation/Thesis / Masters Thesis Geography 2019

Thesis (M.S.)--University of Tennessee, Knoxville, 2009. / Title from title page screen (viewed on Oct. 23, 2009). Thesis advisor: Wayne Clatterbuck. Vita. Includes bibliographical references.