Recommendation of Research Needs

The workshop participants convened into smaller breakout groups that coincided with their expertise in one of the six workshop topics. These breakout groups then discussed the challenges in their selected topic and what would be required to advance design for tornado effects from the current state of the art to the long-term vision. Each group discussed the research needs and then prioritized them in their breakout session (Table 4-1).

The full group of workshop participants voted on these research needs to identify the top priority research needs for tornado design, which are summarized in Section 5.2.

The Tornado Climatology and Near-Surface Wind Characteristics breakout session (Figure 4-1) comprised the following members:

Moderator:

Frank Lombardo

Scribe:

Zach Wienhoff

Reporter:

Anthony Lyza

Participants:

Guangzhao Chen

Fred Haan

Jana Houser

Joe Kanney

Karen Kosiba

Larry A. Twisdale, Jr.

Joshua Wurman

Grace Yan

Michael Zimmerman

The breakout session panel, which was selected to cover the wide array of relevant research topics related to the group, was asked to provide a short presentation on the relevant facets of their work and their vision for the future of tornado climatology and near-surface wind characteristics given the current state of the art. The group discussed and debated many present and future research ideas to benefit tornado science geared toward informing tornado-based structural design. Following the discussion, the scribe and reporter collected the ideas of the breakout group and summarized the discussion into two overarching objectives:

From these overarching objectives, the group agreed upon six specific goals, which are highlighted and discussed briefly as follows.

Continued investigation of tornadoes using a suite of in situ and remote sensing techniques are desired to further expand the available tornado observations database at all scales of motion (i.e., storm scales to turbulence scales). As described, the collection of useful observations of tornadoes for tornado-based design poses many logistical difficulties. Though other methods, including laboratory and numerical simulations, alleviate many of the logistical issues associated with tornado observations, known deficiencies with these techniques leave numerous important questions unanswered. As such, observations remain desired to expand the understanding of tornado flows and to assist in the validation and improvement of these alternate techniques. As observing technology and forecasting techniques continue to improve, assuming that the success rate at observing tornadoes will increase is reasonable. However, the many logistical difficulties, particularly those related to life safety, will likely remain a barrier to more efficient data collection. This goal will likely remain at the forefront of tornado science research in perpetuity as the scientific community always seeks to characterize the near-surface flow field more completely in tornadoes.

Building on the previous goal, using direct observations in tandem with techniques like numerical and experimental simulations will allow estimation of numerous additional load-related quantities and scenarios, particularly for situations in which direct observations are not feasible or are impossible to collect. An example would be using high spatial and temporal resolution radar data to inform a tornado-scale numerical model via data assimilation. Such data could improve a model's ability to handle realistic tornado flows, including further quantification of the three-dimensional flow in tornadoes of various structures and the turbulence content associated with the tornado and surrounding environmental wind field. In addition, as simulations become increasingly realistic, additional complexities can be introduced in a model framework such as the addition of artificial terrain, civil infrastructure, and debris, all of which will dynamically impact the tornado and will in turn likely impact the loading characteristics for any case. Integration of these techniques will require integrating disciplines and the people working within those disciplines to succeed.

As the ASCE/SEI Wind Speed Estimation in Tornadoes Committee continues standard development, the need for additional studies that integrate and compare all available wind speed estimations is emphasized. Using all methods will expand the understanding of each method's strengths and weaknesses, will bolster confidence in the accuracy and error ranges associated with these estimates, and will allow for modifications and improvements to the individual methods with time (e.g., modification of the EF-scale wind speed bounds motivated by wind speed estimates derived from other methods). As the standard continues to be developed, the committee needs to serve as a “home” for these additional studies that integrate and compare all available wind speed estimations.

With the recent mainstream emergence of artificial intelligence, these techniques can likely be utilized in ways that strengthen data analysis and simplify wind speed estimation techniques that currently require significant person hours to complete. Examples of areas where artificial intelligence may be useful are the analysis of structural damage, tree and crop fall patterns, and debris patterns, and other analyses that may involve large numbers (tens-to-hundreds of thousands) of data points. At present, techniques to automate these processes exist but, in many cases, still require human intervention, or in some cases manual labor entirely. Utilizing automated methods could streamline these processes allowing these analyses to be readily used to inform initial tornado intensity estimates for public consumption and for various other research-related purposes.

As tornado damage surveys act as a key data source for documenting parameters important to tornado climatology, a more rigorous, engineering-driven procedure for post-tornado survey conduction must be established. Tornado damage surveys are vital to, at a minimum, confirm a tornado and document some of its basic characteristics including the path length, width, and intensity. In some cases, damage surveys are conducted with exceptional detail, especially for strong tornadoes or those that impact populous areas of the country. By establishing uniform data collection, documentation, and data storage procedures, the needs of all associated parties can be met. Additionally, establishing such a rigorous data collection procedure can ensure that the data will be available not only for current analyses but also for future reanalysis as techniques evolve with time. Given the many varied goals of both meteorological and engineering research groups, convergence on a set of universal goals and methods is necessary (see research goal F, Table 4-1).

F. Fostering stronger relationships among social scientists, meteorologists, and engineers along with National Weather Service partners

For all these purposes, continuing to grow relationships among social science, meteorology, and engineering research partners and with the National Weather Service to meet current and future goals remains a vital need. The development of unified research objectives and procedures will require a seat at the table for all relevant and interested parties, and working to understand where the needs of each group lie within the overarching goals remains key for the success of all tornado-based design research.

The Tornado-Structure Interaction breakout session (Figure 4-2) comprised the following members:

Moderator:

Peter Vickery

Scribe:

Lauren Mudd

Reporter:

Chris Letchford

Participants:

Sudhan Banik

Girma Bitsuamiak

Lakshmana Doddipatla

Emily Kim

Joy Pauschke

Maryam Refan

Dan Rhee

Partha Sarkar

Delong Zuo

During the tornado-structure interaction breakout session, five research needs surfaced. The highest-priority research needs to better characterize tornado-structure interaction centered around tornado simulators and the ability to accurately measure loads.

The first research priority is to develop a standard for tornado simulators, similar to ASCE/SEI 49-21 (ASCE 2021), to define the minimum requirements for conducting and interpreting tornado simulator tests to determine tornadic wind loads on buildings and other structures. The standard should address the scaling of tornado simulators; develop common definitions of design tornado characteristics, reference velocity, and pressure; and provide guidance on how to quantify uncertainty through many realizations of non-Gaussian tornado events.

Following the development of a standard for testing in a tornado simulator, another research need is for a series of benchmarking studies to be conducted across a range of facilities with varying sizes, layouts, equipment, and capabilities. For the benchmark experiments, a standard building model should be defined to be tested in each tornado simulator facility, similar to previous benchmarking studies performed across multiple atmospheric boundary layer wind tunnels (e.g., the CAARC high-rise or Aylesbury low-rise buildings). Such a benchmarking study would allow researchers to better translate or compare experimental results among different tornado simulators.

The development of standard tornado simulator guidance and benchmarking studies to translate results among facilities will enable researchers to improve measuring techniques within tornado simulators for characterizing the three-dimensional near-surface flow. Current measurement techniques do not allow for reliable measurements inside of the tornado core, and capturing vertical wind speeds properly is often difficult. Better measurement techniques will ideally provide greater insight regarding near-ground vertical profiles, modeling of atmospheric pressure change and internal pressures, differences in single- and multi-vortex tornadoes, quantifying of the effect of roughness, and loading due to wall suction and roof corner vortices.

Even with significant standardization of testing and improvements to measuring capabilities at tornado simulators, gaps will likely remain in physical testing and simulation capabilities. Computational fluid dynamics (CFD) should be used to fill these testing gaps. The results of a benchmark study could be used to validate the CFD models and ensure results are consistent with previous loading conditions. CFD could be particularly useful in developing tornado-specific windborne debris requirements, as the current requirements in ASCE/SEI 7-22 are largely based on observational data from past hurricane events.

Finally, comparisons should be made between atmospheric boundary layer (ABL) loading and tornado loading for a given wind speed. If large differences occur, this could have significant effects on the design philosophy of tornado-induced wind loading within ASCE/SEI 7-22, which currently follows a similar approach to ABL loading. Cost implications of any code changes should also be quantified as part of this comparison study.

The Design of Residential Structures breakout session (Figure 4-3) comprised the following members:

Moderator:

John van de Lindt

Scribe:

Blythe Johnston

Reporter:

David Roueche

Participants:

Omar Amini

Anne Cope

Gary Ehrlich

Jennifer Goupil

Douglas Rammer

Tim Reinhold

Jotsana Sunder

Although the research needs in this section will vastly improve the understanding of residential building behavior in a tornado event, these learnings will not translate to improved building performance unless improved design and construction is required by standards and codes and interdisciplinary, community-engaged research is conducted to bridge adoption gaps. This more social science–centered research will need to investigate various incentives to determine what compels home builders and homeowners the most. Social scientists will also need to conduct social cost-benefit analyses to account for not only direct damage costs but also the cascading effects of disaster. Furthermore, what communities deem acceptable levels of building performance for various tornado intensities needs to be investigated. As with all modern social science work, these investigations should strive to be non-extractive and build in ways to use their public engagement as an opportunity to disseminate information that will promote life safety in a tornado event such as retrofitting opportunities for homes and tornado safety protocols. That researchers continue to advance knowledge in residential structural behavior under tornadic loading is undoubtedly important, but practice must come along with it. This means not only distilling this knowledge into design standards and building codes, but also ensuring that home builders, contractors, and the public knows both the box that needs to be checked to meet a requirement (i.e., hurricane straps) and the logic behind a requirement. In other words, they need to know not only the what but also the why of a given retrofit or construction best practice.

Some other exciting areas with research potential that surfaced during the breakout session include the following. First, open-source modeling software such as OpenSEES for seismic loading has had years of success. Hence a growing possibility exists that a similar program could be developed for tornadic loading. This would require investigation and development of design tornadoes for the software to simulate loadings. That these simulated loadings are accurate and reliable will be of utmost importance for such software to gain traction among design professionals and researchers. Thus, the development of these design tornadoes and their resultant loading are a significant research task in and of itself.

Second, the breakout group expressed excitement around leveraging technology to reimagine design, construction, and inspection of residential buildings. This included discussion of using innovative materials that could bolster environmental commitments, lower workforce requirements, reduce construction times, and of course allow for more tornado-resistant design. Many of these materials do not yet have an extensive presence in residential design, but with advancements in cost efficiency this potential could grow. As mentioned elsewhere, improving inspection processes was also included in the discussion of researching how to leverage new technologies to improve real-world residential design. Other concepts were briefly discussed but not extensively enough to expound on here. In reviewing these recommendations and the breakout group discussion, a grounding element throughout these conversations was the need for future work to center on the broader impact and the potential for this research to increase survivability and community-level resilience against tornado hazards.

The design using ASCE/SEI 7-22 breakout session (Figure 4-4) comprised the following members:

Moderator:

Cherylyn Henry

Scribe:

Korah Parackal

Reporter:

John O'Brien

Participants:

Jubayer Chowdhury

Dustin Cole

TiAwna Moffat Daniels

Dines Jong

John Kilpatrick

Mehedy Mashnad

Murray Morrison

Justin B. Nevill

Tom Smith

Katie Stueckle

The Design Using ASCE/SEI 7 breakout session prioritized research needs and specified six topics. Additional topics were discussed and are included at the end of this section.

The highest research priority for design is reexamining the target reliabilities and MRIs for tornado winds versus straight-line winds. A reliability study was performed during the last revision of ASCE/SEI 7 to determine the suitability of using the straight-line wind MRIs for tornadoes. The results of the study were that the tornado load criteria in Chapter 32 gave reasonable consistency with the reliability in the Chapter 26 and 27 straight-line wind criteria. However, structural engineers who have used the Chapter 32 provisions have found that unless a Risk Category III building has a large effective plan area (greater than 400,000 ft² (37,161 m²), tornado loads almost never govern. Is this a correct result of the provisions? And should the load factors for tornado winds be the same as for straight-line winds? Another consideration is incorporating Risk Category II structures into the tornado provisions. Currently, the tornado wind loads would not govern over the straight-line winds. Adjusted MRIs may capture these structures.

The positive internal pressure coefficient of +0.55 for tornado design correlates well with the partially enclosed design provisions for straight-line winds and in the case of tornadoes is meant to capture the contribution of the atmospheric pressure change. However, is +0.55 the right number?

Another research need is the impacts of tornadoes on high-rise buildings. In recent use of the provisions, structural engineers have found that straight-line winds, and not tornado winds, govern taller buildings. Researchers and practitioners need to understand how a building's environment, such as a suburban location versus a downtown setting, might impact a taller building. Another consideration is the role of tornado-borne debris. These topics point to the current understanding of a straight-line wind profile versus a tornado wind profile. Is it probable that straight-line wind governs but glazing must be impact resistant? Further research of taller buildings could answer this question.

Load cases were discussed, and straight-line wind and tornado wind load combinations should be the same. Further discussion included the case where a much smaller tornado impacts a large structure. A localized impact could, and likely will, have different failures than a similar sized structure as compared to the tornado size, that is squarely in the tornado path. Should different load cases be considered depending on the size of the building to address potential localized failures? And is there a difference between a direct hit versus a glancing blow to a structure?

Wind (tornado)-borne debris is a major consideration in tornado design, and it is a greater hazard than wind (hurricane)-borne debris. The potential exists for larger and faster-moving debris, and this debris can puncture walls and windows, creating an opening in the building envelope. This permits wind-driven rain to penetrate the building, often rendering the building uninhabitable until repairs are made. To address tornado-borne debris, what is a reasonable standard tornado missile?

The final priority research need for design is the impacts of topography and exposure (ground surface roughness) on the tornado structure and in turn on the building the tornado strikes. These effects are difficult to measure near the surface of the tornado, and wind tunnel studies have been used to better understand what is happening closer to the ground. These studies suggest that the characteristics of the tornado are modified by exposure and topography, but to what extent is not known. The challenge lies in simulating the numerous exposure and topography variables and obtaining consistent results.

Other research topics discussed included modifications to existing buildings, building envelope projects, and PV panels on existing roofs. These are not in the scope of ASCE/SEI 7-22, though guidance in these situations would be helpful to structural engineers.

The Design of Storm Shelters and Safe Rooms breakout session (Figure 4-5) comprised the following members:

Moderator:

E. Scott Tezak

Scribe:

Shane Crawford

Reporter:

Glenn Overcash

Participants:

Benchmark Harris

Matthew Holland

John Hutton

Brian Orr

Donald Scott

Pataya Scott

The Design of Tornado Shelters and Safe Rooms breakout group identified 13 topics areas where research is needed, which are ordered by importance in Table 4-1. The research needs include improved science to understand the loading and resistance criteria for the design of protective structures, human needs for sheltering, and improved social science to ensure that shelter requirements are adopted and implemented nationwide, and that the public understands its risk from tornadoes given its members’ sheltering decisions. The top research need, which is cuts across several breakout sessions, is the need for improved understanding of design tornadic wind speeds. Requirements for storm shelters and safe rooms differ from other structures because current design maps are deterministic, whereas all other straight-line wind and tornado design wind maps are probabilistic. To align protective structure design and better enable risk-based design thinking, probabilistic storm shelter and safe room wind speed maps are needed.

The second research need is a broad catch-all for human needs criteria, which includes sheltering duration followed by duration-based needs including square footage and water/sewage requirements. The current design standards assume a 2-h duration in shelters and safe rooms. While this is likely to be satisfactory in most situations, longer occupancy times may be required at times. In these situations, each person would likely require more square footage for comfort and safety, and water and wastewater systems would become more critical.

Tornado-borne debris impact testing and approvals for nonproprietary sections and assemblies are another research need that could prove very impactful for the design community, which faces challenges in finding low-cost options for protective structure design. This would require compliance with current codes and standards and the creation of a publicly accessible database of assemblies. To achieve this, professional associations should be engaged to help fund and conduct testing.

Research is needed to better account for the wind pressure protection of soil covering shelters, including amount and types of soil required to adequately resist the combined forces of uplift, scour, and debris impacts. Research is also needed to develop analytical methods for evaluation of debris impact resistance, in addition to testing.

Another significant research need for storm shelters and safe rooms is the reconciliation of wind zones and pressures of discipline-based standards such as American Society of Heating, Refrigerating, and Air-Conditioning Engineers (ASHRAE) standards, ASCE/SEI 7-22, and ICC 500. Currently, ASHRAE standards require straight-line wind design, which is likely not appropriate for tornadoes. This can cause issues between structural and energy performance and the design practitioners. Similarly, research need E from Table 4-1 deals with loads and pressure test methods for wind on mechanical equipment in protective structures. Laminar air flow is assumed in HVAC equipment, which may not be correct when experiencing tornadic winds. If incorrect, internal pressure could exceed allowable limits and cause failures of equipment such as dampers that could then become hazards for shelter and safe room occupants. Additionally, the internal pressurization could harm interior furnishings and equipment, which could similarly become hazards and could harm occupants.

A known research need that has received inadequate attention is the need for standardized assessment criteria for identification of the Best Available Refuge Areas in existing structures that lack shelters and/or safe rooms. Most structures nationwide do not include protective structures; therefore, it is critically important for life safety to identify the safest locations in those structures.

Operational issues for protective structures are another area where more research is needed. FEMA criteria require a 5-min max travel time for occupants to arrive at safe rooms, and the IBC requires a 1,000 ft (300 m) door-to-door travel distance. Social science research is needed to understand the feasibility and acceptability of these provisions in real-world applications.

The next set of research needs are based on structural design criteria for various aspects of protective structure design. Research into criteria for using soil cover in the design of protective structures is needed to understand how much and what types of soil can offer adequate protection to the structure. More research is needed to determine whether the current impact load factor for laydown and falling debris hazards is adequate in practice and in the creation of similar factors for walls. Also needed is the development of test methods and standards for components and cladding in protective structures. Full door testing is needed to determine whether current design criteria for doors is adequate. This testing is needed because previous tests did not adequately account for the bottom of the door in strike/catch scenarios of tornado-borne debris impact. An analytical design method for tornado-borne debris loads is required to supplement physical testing, which would allow for more economical design creations. Last on the list of research needs but within the topic area of structural design criteria needs is the development of test methods and standards for components and cladding in protective structures. Achievements in these topic areas would allow for more economical and easily implementable designs for protective structures, which would allow for more of these structures to be built nationwide to reduce risks to life safety.

Little research has been conducted on fire safety requirements in storm shelters and safe rooms, which creates a gap in the understanding of risk in this area. Fire resistance ratings and egress requirements are needed for situations where host structures fail and cause fires. In these situations, occupants of protective structures must choose between two hazards, which is not ideal in life safety protection structures.

The final research need identified in the breakout session is development of a strategy to optimize the size and placement of protective structures to provide the highest level of protection to the greatest number of people. This research should include consideration of social vulnerability as a trade-off, understanding that socially vulnerable populations are more likely to live and work in less-resistant structures than other population groups.

The Tornado Effects on Non-building Structures Beyond ASCE/SEI 7-22 breakout session (Figure 4-6) comprised the following members:

Moderator:

Alex Griffin

Scribe:

John Haney

Reporter:

Tom Mara

Participants:

Bianca Augustin

Matthew Browne

Bryan Lanier

Andrew Sarawit

Ting Shi

Greg Soules

DongHun Yeo

The Non-building Structures and Beyond ASCE/SEI 7-22 breakout session identified nine research focus areas, which this section discusses further. One of the primary questions from the session was whether the tornado design equations and approach outlined in ASCE/SEI 7-22 accurately represent the true effect of tornadoes on non-building structures, specifically whether critical load cases are captured. The ASCE/SEI 7-22 Chapter 32 provisions focus primarily on building design, including the effective plan area and the tornadic internal pressure coefficients, while tornado wind loads for non-building structures are treated very similarly to straight-line wind loads. For tanks, atmospheric pressure change is addressed in ASCE/SEI 7-22, but further research is needed to verify the magnitude of this phenomenon. Also, research is needed to verify if the current code equations accurately capture the correct tornadic loading or if additional criteria, such as eccentric or torsional load cases, should be considered for non-building structures. For example, for a lattice transmission or telecommunication tower, omitting potential torsional loading from tornadoes could prove detrimental to the overall stability of the structure.

In addition, ASCE/SEI 7-22 and other industry standards are relatively silent (except for the nuclear and LNG industries) on the application of tornado-borne debris loading to non-building structures. Currently, ASCE/SEI 7-22's tornado provisions do provide some guidance on “clinging” debris loading for lattice towers; however, no guidance on impact missile loading is provided. For non-building structures, missile loading could prove significant in design given that most non-building structures are low-profile with exposed structural members that could be easily damaged by tornado-borne debris. The risk and consequence of failure for non-building structures varies by industry, so additional research and guidance are needed for the applicability and magnitude of tornado-borne debris loading based on a structure's risk category.

The breakout group also discussed the public's knowledge gap in understanding building performance expectations for tornado design, specifically how the tornado loads are included (i.e., designing for tornado risk vs. designing for a specific EF tornado event). Many non-building structure industries are resistant or slow to adopt new changes in design, particularly if those changes may impact cost or schedule. Additionally, some of these structures could fail but have redundancy in the overall system, leaving the public unaffected, and the structure could be replaced relatively quickly. The significance of this will vary by, and within, each industry. By researching the most beneficial ways to establish mutual understanding of the community resilience benefits and long-term benefit-cost ratios of improved tornado design, the effects of tornado resilience can be expedited, and unnecessary pushback and confusion can be avoided.

Another research topic was the lack of guidance on the area of influence of a tornado in the event of a direct hit. Much of the current tornado provisions treat the wind effects of a tornado as a global loading over the entire structure; however, for some larger non-building structures, this may be inaccurate and conservative, particularly for resistance to smaller EF0–EF2 tornadoes. A better understanding of a tornado's area of influence can enable engineers to more economically design structures for its potential effects.

Another key research topic concerns the effective plan area for non-building structures, including those in a spatially distributed system. Unlike buildings, which will likely have a concentrated effective plan area, a facility with non-building structures of multiple risk categories (such as LNG facilities) spread out across hundreds of acres, is not as intuitive when it comes to the effective plan area used for design. For spatially distributed systems such as transmission lines, designing the entire system for a tornado load occurring simultaneously along the systems’ length would not be realistic. Thus, additional research on the size effect and angle of attack is warranted. Further research and guidance on how these structures should be evaluated would be beneficial to avoid future confusion and inconsistencies in design while avoiding design inefficiency and excessive cost.

Due to the critical and high-risk facilities considered in the breakout session, a key point discussed was the need for a procedure to perform site-specific tornado analysis, particularly in relation to topographic features. Currently, the tornado provisions in ASCE/SEI 7-22 do not account for topographic effects. Further research into this topic could help determine if topographic factors need to be included for tornado loading, which could impact the siting of nuclear and LNG facilities.

Lastly, solar PV structures are developing rapidly across the nation. However, as mentioned in previous sections, information on how to properly design these structures against tornadoes is currently lacking. Therefore, the breakout session recommended the development of tornado response coefficients for solar arrays to understand how to design these structures for both ground-mounted and roof-mounted configurations. As the solar industry continues to grow and the US energy grid relies more on its prevalence, this information could be crucial to improving community resilience.