Prioritization and Benefits of Recommended Research Needs

Following the breakout sessions, the workshop participants reconvened into a single group and reviewed the combined recommended research needs from the breakout sessions. Table 4-1 summarizes these research needs by breakout session.

After discussion of the breakout sessions’ recommended research needs, the WSC combined similar research needs and identified an overall list of needs. Then, the workshop participants identified the priority research needs and determined the order of urgency (Table 5-1). The table lists the priority research needs in order of priority and their estimated costs and times. Section 5.3, Summaries of Research Priority Needs, describes these needs in greater detail. These summaries include a description, estimated cost, estimated time, measurement science challenges and potential solutions, stakeholders and roles, and impacts on standardization and application in practice. Sections 5.4, 5.5, and 5.6 describe the comprehensive budget and schedule, interrelationships among research activities, and their benefits.

The WSC provided the cost estimates, based upon its members’ knowledge of the costs of similar research efforts. Estimated costs for each research topic are provided using one of the following ranges: less than $1,000,000 (low cost); $1,000,000–$3,000,000 (moderate cost); and more than $3,000,000 (high cost).

Similarly, the WSC estimated the time requirements to properly address each research topic, based on member experience with comparable research efforts. Estimates are provided using the following time-period ranges: 1–2 years (short time period), 2–5 years (moderate time period), and 5–10 years (long time period).

Some of the research needs identified in the individual breakout sessions were similar in scope. For that reason, the WSC combined similar research needs into those listed in Table 5-1. These research needs were then prioritized based on the combined votes received from the workshop participants. The following summarizes how these research needs were combined. The breakout session and research need from these breakout groups are noted, using the letter, for example (A), indicated in Table 4-1.

The WSC developed the following in-depth summaries of the priority research needs identified in Section 5.2. They include a description, estimated cost, estimated time, measurement science challenges and potential solutions, stakeholders and roles, and impacts on standardization and application in practice.

Description: The vast majority of buildings in U.S. communities are residential wood frame often exceeding 90%. Community resilience is defined as the ability to plan, prepare for, and adapt to hazards and conditions to rapidly recover from natural hazard events (Presidential Policy Directive 21, 2014). Rapid return to functionality of these wood frame and other types of residential buildings is a condition necessary for a community to be resilient to tornadoes, however, currently codes or standards are lacking to achieve the necessary performance of these types of structures. The objective of this research priority is to quantify the relationship between individual building performance and community-level resilience goals for tornado hazard and use that information to identify the levels of performance residential buildings need to be designed for, to achieve those community resilience goals. These performance levels will be used directly in the development of resulting codes and standards, and construction and retrofitting guidance documents.

The community models needed to support the process described above are complex and have interdependencies between all physical and socio-economic systems in a community. Thus, this approach should consider damage and loss of functionality to key community components, including residential buildings, businesses, schools, hospitals, and the supporting infrastructure (roads; bridges; and the power, water, sewer, and communications networks), and the effect on households and the ability of social institutions to perform their functions. Also, a means by which to translate this damage to the built environment into the impact on the local population and economy is needed. Additionally, because tornado loading is now considered in the building code, a specific, targeted approach to improving residential housing resilience is needed that includes the inclusion of residential buildings in current codes and standards.

NIST Special Publication 1190, Community Resilience Planning Guide for Building and Infrastructure Systems: Volume 1 (NIST 2014) and NIST Special Publication 1190GB-16, Community Resilience Planning Guide for Buildings and Infrastructure Systems: A Playbook (NIST 2020) (referred to collectively as NIST CRPG) should be treated as the guiding documents for the hazard intensity levels, and this research need should not focus directly on design code levels but rather on the community resilience levels a community agrees to during its planning process. The specific tasks envisioned as part of this robust effort to connect community goals to individual building performance include (1) reviewing of the NIST CRPG and related community engagement workshops to better understand community goals; (2) using existing models including but not limited to INCOR or HAZUS, develop a suite of community level models that can be used as archetypical to systematically quantify the building performance to community relationship described earlier; (3) perform validation of models from existing tornado field studies; and (4) develop design and construction standards, guides, and retrofitting guidance documents. While Tasks (1), (2), and (3) are dependent and linear, Task (4) could be started immediately and refined as the other Tasks are completed. These tasks are expanded upon in the Appendix.

Estimated Cost: $4,000,000 to $7,000,000

(1) $300,000 to $500,000

(2) $2,000,000 to $3,000,000

(3) $500,000 to $1,000,000

(4) $1,000,000 to $2,000,000

Estimated Time: 4 to 5 years total

(1) 1 year

(2) 2 years

(3) 3 years

(4) 3–6 years

Description: Debris lofted and transported in tornadoes has been widely documented to be a significant contributor to tornado-induced damage. The types of debris range in size from gravel to automobiles and whole roof structures. The debris includes both restrained (e.g., building components from damaged buildings) and non-restrained types (e.g., roof gravel, automobiles). Quantifying the debris hazard requires:

The Electric Power Research Industry funded the development of a tornado missile hazard model (TORMIS) in the 1970s, which the nuclear power industry in North America still uses. The TORMIS model does not include a structure fragility model but simply injects building components into the tornado wind field at a prescribed wind speed.

Models for these five components have been published in the literature, and hurricane-windborne debris models have been developed, but an all-encompassing probabilistic based tornado-borne debris model has not been developed. Such a model could be developed by either private industry or universities.

Estimated Cost: $1,500,000 to $3,000,000

Estimated Time: 2 to 3 years

Description: In situ observations remain desired to expand understanding and characterization of tornado flows. Characterization of flow fields and links with the associated response of buildings and other structures is likely fundamental to many other research topics listed in this section. Specifically, four-dimensional characterization (x, y, z, t) in high resolution (< 1 s, core vortex diameter) is highly desirable. The Tornado Climatology and Near-Surface Wind Characteristics breakout group suggested collection of “footprints” of tornado observations at all scales of motion (i.e., storm scales to turbulence scales) to capture potentially relevant flow field characteristics. Improved characterization of flow fields is also important in experimental facilities as tornado characteristics vary in time and space and these detailed measurements are needed to accurately measure loads. The call for improved characterization includes (1) an increase in measurement campaigns and (2) the study, development, and advancement of innovative technologies to characterize tornadic wind flow.

As observing technology and forecasting techniques continue to improve, assuming an increasing success rate at observing tornadoes is reasonable. However, the many logistical difficulties, particularly those related to life safety, cost, and the transience of tornadoes, 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 engineering community seeks to characterize the near-surface flow field more completely in tornadoes.

Estimated Cost: > $3,000,000

Estimated Time: 5 to 10 years

Collection and subsequent characterization of tornado wind fields are likely to have wide-ranging impacts for standardization and application in practice. Wind engineering has traditionally relied upon data in the field to validate rigorous experimental simulation (i.e., wind tunnels). Experiments are conducted and used in both standardization and subsequent practice.

Description: The current tornado load provisions are based on the same target reliabilities as for straight-line winds, which based on the reliability analyses performed by Li et al. (2024). These target reliabilities are from the first row of Table 1.3-1 in Chapter 1 of ASCE/SEI 7-22, nominally for ductile failure. However, the workshop showed that practicing engineers often do not know this rationale for target reliability. Additionally, the commentary of ASCE/SEI 7-22 currently does not provide this information regarding the derivation of tornado loads. Perhaps the “sudden failure” or the fourth row of Table 1.3-1 might be appropriate.

Due to the small size of tornadoes as compared with weather phenomena causing straight-line winds, the probability of a building being hit directly by a tornado is already very low. Therefore, matching reliability levels of tornado loads to straight-line winds result in the design for low-level tornado events.

Designers and owners of buildings classified as essential facilities may be required to design for a direct hit of a particular tornado severity, which probabilistically would require a much higher reliability level than required for straight-line winds. As such, current provisions in Chapter 32 of ASCE/SEI 7-22 are not intended for this type of design.

This research priority involves study of the target reliabilities required to enable designers to design for a direct hit of a tornado of a particular intensity. Approaches may include requiring a target reliability for a non-ductile failure mode, according to Table 1.3-1 in ASCE/SEI 7-22. Alternatively, a thorough reexamination of the paradigm used for target reliabilities of Risk Category III and IV buildings when considering tornado loads can be undertaken. For example, a community-level approach may be taken for the design of essential facilities to ensure that a community can continue to function even with the extremely low likelihood of direct tornado impact on such buildings.

Furthermore, revisions to the commentary of ASCE/SEI 7-22 should be undertaken to provide designers with the rationale for the tornado load provisions in Chapter 32.

Estimated Cost: $1,000,000 to $3,000,000

Estimated Time: 2 to 6 years

Description: Given the difficulty in collecting measurements of the tornado wind field, the need for research to improve existing and develop new methods and techniques of wind speed estimation, largely from damage to the built and natural environment, is emphasized. Wind speed estimation from damage will be the only way tornado wind speeds are derived in nearly all tornadoes for research purposes.

Methods and techniques include (1) remote sensing advances, including satellite, Uncrewed Aerial Vehicles, and other technologies; (2) improved forensic analysis, including experimental testing and numerical and probabilistic modeling to assess response of the built and natural environment to tornadic wind loading; (3) studies that integrate and compare all available tornado wind speed estimations; and (4) artificial intelligence and machine learning techniques, especially for very large, already existing data sets, which are plentiful. Utilization of methods will expand understanding of each method's strengths and weaknesses, will bolster confidence in the accuracy and error ranges and potential bias associated with these estimates, and will allow for modifications and improvements to the individual methods with time. Examples of areas where improved methods may be useful are in the analysis of tree and crop fall patterns, debris patterns, and other analyses that may involve very large numbers of data points. Methods that could streamline these processes such that these analyses could be readily used to inform initial tornado intensity estimates for public consumption, and for various other research-related purposes, are needed.

Estimated Cost: > $3,000,000

Estimated Time: 5 to 10 years

A standard committee has already been formed for this purpose. Impacts, such as the use of new and improved methods, should be easily incorporated and used by the practicing engineers in forensic analyses and other related areas.

Description: To ensure that measurements of various straight-line wind-induced loads and pressures made in boundary layer wind tunnels are acceptable for use in design, a wind tunnel facility and the software used in the facility must meet minimum standards as given in ASCE/SEI 49-21. As more tornado simulators come online and more experiments are performed in these simulators, for use in the design of structures and in informing loading standards, these facilities must meet minimum standards. Developing a standard for tornado simulation facilities is more challenging than for boundary layer wind tunnels owing to the additional complexity associated with measurements of wind speeds in the nonstationary translating tornado and the wandering of a translating tornado and with measuring the atmospheric pressure within the tornado. The short duration of the tornado event coupled with the wandering of the tornado requires that each experiment be repeated many times to obtain statistically valid results, with the number of simulations used varying from facility to facility.

The problem is further complicated by the lack of full-scale data upon which to benchmark the model tornado characteristics and by scaling issues, which are currently not well-defined.

Estimated Cost: $1,000,000 to $3,000,000

Estimated Time: 3 to 5 years

Ensures tornado loads developed by tornado simulators meet minimum requirements and can be used with confidence.

Description: Non-tornadic winds create internal pressure inside of buildings because of the airflow to and from the building through openings in the building envelope. The APC affects the magnitude of this internal pressure. The atmospheric pressure is significantly lower in the center of the tornado than it is farther away from the center. While the internal pressure is still a function of the external pressure, the relative size of the tornado with respect to the building can be important. The tornado size effect also plays a significant role in estimating additional loads acting on the building due to the APC within the tornado and complicates the computation of internal/external pressures (see Research Need No. 9). The effect of the APC within a tornado is currently handled through the internal pressure coefficient alone in ASCE/SEI 7-22 and does not account for building size relative to tornado size.

A similar situation exists for “sealed” structures such as tanks and requires further investigation and quantification of the appropriate internal pressure coefficient.

The myriad combinations of tornado sizes, intensities, structure sizes, building aspect ratios, porosities, and orientations essentially require the use of a simulation model that can incorporate these parameters and subsequently frame the problem from a structural reliability point of view.

Estimated Cost: $1,000,000 to $3,000,000

Estimated Time: 3 to 5 years

Provides appropriate tornadic internal pressure coefficients for various building envelopes.

Description: Most buildings at risk of damage and loss of functionality from tornado loading will not be new construction but existing buildings. The tornado provisions added to the ASCE/SEI 7-22 design standard for Risk Category III and IV structures are likely not satisfied by most existing buildings in the United States, nor do people realize that they can design for a higher wind load if their community desires better resilience against wind hazards such as tornadoes. In addition, most jurisdictions may not have a comprehensive understanding of and/or availability to check capacities of existing structures given the complexities associated with as-built construction, material deterioration, and older noncompliant structures from past editions of codes and standards and, in some cases, before standards were adopted.

The research topic proposed here would be a tornado version of the current ASCE/SEI 41. This is proposed to be a two-phase project, with the first phase involving research of evaluation methodologies and the second being the development of retrofit guidelines for wind.

Phase 1: The project to build out a robust method for evaluation of existing structures will require research into assessing building stock in situ and non-destructively. This will focus on evaluation methods to determine if retrofit is needed and the optimal strategies.

Phase 2: The retrofitting guideline project will focus on three key topics and may need to be separated by types of structure, for example, Risk Category II, including single family residential, versus Risk Categories III and IV, and so on. The first topics will be providing common approaches and details for a continuous vertical load path. The second concerns a calculation methodology for all building types to make sure that the failure plane is not altered through some type of systematic check or overstrength rule. The third involves the management of a detailed database with appropriate review.

Estimated Cost: Phase 1: $1,500,000 to $2,000,000

Phase 2: $3,000,000 to $4,000,000

Estimated Time: Phase 1: 3 years

Phase 2: 5 years (can overlap by 1 to 1.5 years)

Description: Components and cladding and main wind force resisting system loads in tornadoes result from both the direct action of the wind speeds and the effects of the APC within the tornado. The effective loads are strongly influenced by the porosity of the building, with more porous buildings reducing the effects of the APC. Pressures measured on model buildings in tornado simulators inherently include the combined effects of the wind speed–induced loads and those associated with the APC. Significant issues are associated with separating the APC loads from the wind-induced loads because the translating tornadoes in simulators tend to wander, thus each experiment is slightly different. Due to this, determining where the center of the tornado was during the experiment is very difficult, and the pressure distribution within the tornado is not known precisely. Furthermore, the wandering effect results in rapid local variations in the APC-induced loads due to the variation about the “mean” track.

Some tornado simulator experimentalists believe that separating the two load effects should not be attempted and simulation results should be used to define the tornado loads in ASCE/SEI 7-22. Other experimentalists believe that such an approach is not necessary, and the problem can be addressed using other methods. While likely more accurate, the countless combinations of tornado size, intensity, translation speeds, and building sizes and orientations present a perhaps insurmountable problem.

The current approach of separating the two load effects using relatively simple models is much more amenable to incorporation into a load-and-resistance model for determining structural reliability/annual failure probabilities matching those specified in Chapter 1 of ASCE/SEI 7-22. Ideally, some type of hybrid analytical-experimental models could be developed to address the differing approaches.

Estimated Cost: $5,000,00 to $10,000,000

Estimated Time: 5 to 10 years

Improved loading requirements for tornado-resistant design.

Description: Several issues are related to the design of mechanical components for tornado storm shelters that are not well understood and in need of research. In the meantime, the ASHRAE standards for straight-line winds are used even though they are likely not appropriate for tornadoes. Specific needs include research on tornado loads and development of pressure test methods for tornadic wind on louvers, dampers, ducts, and air exiting ducts or openings.

Estimated Cost: $500,000 to $1,000,000

Estimated Time: 1 to 3 years

Guidance for storm shelter designers where none exist today for design of expensive mechanical components.

Based on the priority research summaries provided in Section 5.3, Table 5-2 summarizes the proposed program budget and schedule for the first 10 years. The WSC made efforts to identify where, and which research efforts depend on or need subsequent efforts. Section 5.5 explains these relationships in more detail.

Tables 5-1 and 5-2 list the top 10 identified research needs from the workshop. Each of these research needs seeks to improve the built environment through development of standards and techniques that will allow the practicing structural engineer to better determine the tornadic wind loading and the effects caused by tornado wind events, which are required for the structural design of their projects. Consequently, completion of certain research needs will depend on the status, development, and perhaps completion of other research needs. The WSC offers the following commentary regarding the likely interrelationships of the research needs:

The benefits of the recommended research program include the following:

For the nation, implementation of the proposed research program will yield the following major benefits:

Development of further tornado wind speed estimation, tornado pressure coefficients, and guidance for considering community resilience for tornadic events will enable the development of more resilient communities and help prevent the economic loss and loss of life in tornadic events.