Seismic Retrofit Ordinances Part 3: Concrete Tilt-up & Wood Soft-Story

To view the figures and tables associated with this article, please refer to the flipbook above.

Concrete tilt-up buildings were originally observed as requiring attention based on their poor performance in the 1971 San Fernando earthquake. For many buildings, either the exterior walls fell away from roofs and/or their roofs pulled apart internally leading to roof collapse, allowing the exterior walls to then fall or lean outward precariously. This was the result of roof to wall lateral force connections relying on cross-grain bending of the wood ledgers and the lack of internal cross-ties in the roof diaphragm.

It may come as a surprise to many engineers, but prior to 1971, there were no specific provisions for the design of tilt-up buildings in the building code since the structures did not fit neatly into any of the listed building systems. Tilt-ups weren’t concrete buildings, and they weren’t wood buildings either. And dynamically they don’t perform like any other building, then or now, but this was not recognized for many years. In some ways, the original concrete tilt-ups were like the original brick bearing wall buildings, with modern concrete instead of archaic brick.
For those unfamiliar with original concrete tilt-ups, they were typically constructed with:

Following the 1971 San Fernando earthquake, the design provisions for tilt-ups in the Uniform Building Code were changed on several occasions during the following decades, in a manner that some engineers have characterized as “trial and error.” The changes to the code mainly focused on the prescriptive out-of-plane wall anchorage forces. The required force coefficients were incrementally increased for 0.2g to 0.3g in the 1979 UBC, to 0.45g in the 1991 UBC (but only in the center of the diaphragm span) and to 0.63g in the 1997 UBC (now over the full diaphragm width) over a period of around 25 years. See Figure 1 for a graphical representation of the changes.

This incremental “trial and error” approach could have been avoided, in my opinion, had engineers spent more time assessing how concrete tilt-ups and their cousins—concrete masonry wall buildings—actually respond dynamically in earthquakes rather than attempting to develop prescriptive solutions that addressed observed performance in the last earthquake rather than the cause. Each time, these code changes were thought to have solved the problem until the next large earthquake occurred, and the code was found lacking again. Then the process was repeated. The performance in the 1994 Northridge earthquake showed that the code changes made over the previous two decades had not gotten to the root of the problem, as shown in Figure 2. The building was a modern structure with out-of-plane connection hardware (using holdowns) (Fig. 3).

Many cities in California have enacted their own retrofit ordinances, but somewhat surprisingly, given the large number of such structures throughout California, there is no statewide retrofit ordinance. The retrofit ordinances tended to focus on the early versions of these buildings which posed the greatest hazards, those designed using the 1976 UBC or earlier editions. The City of Fremont was an earlier adopter and enacted a retrofit ordinance titled Earthquake Hazard Reduction in Existing Tilt-up Concrete and Reinforced Masonry (TRM) Buildings. The ordinance required retrofit in accordance with 1997 Uniform Code for Building Conservation (UCBC), used for existing structures, which specifies lateral forces equal to 75% of the 1997 UBC provisions for new buildings. As the name of the ordinance suggests, the intent is hazard reduction.
The retrofit of tilt-ups is a quite straightforward affair, often done in two phases. The first phase involves adding or increasing the strength of the out-of-plane wall anchors and creating diaphragm cross-ties using off-the-shelf holdown anchors. Two carpenters on a scissor lift with a drill can accomplish the task. The second phase involves improvement to the diaphragm nailing. This work is usually put off until the next re-roofing project when the cost of additional nailing is small as a percentage of the entire project.

In a July/August 2003 STRUCTURE article by Dave McCormick of Simpson Gumpertz & Heger (SGH), Seismic Retrofit – Strengthening Tilt-up Structures, the authors state, “Along with unreinforced masonry buildings, older wood frame buildings with parking below, and older concrete frame buildings, older tilt-ups have proven to be among the worst performing building types in an earthquake.”

The Structural Engineers Association of California (SEAOC) developed two documents that can be used by engineers to reduce the vulnerabilities in existing tilt-ups: Guidelines for Seismic Evaluation and Rehabilitation of Tilt-up Buildings (Guidelines), and Chapter 2 of Guidelines for Seismic Retrofit of Existing Buildings (GSREB). The GSREB is an appendix to the 2003 International Existing Building Code. Both the GSREB and the Guidelines are based on extensive research, by SEAOC members, of tilt-ups in past earthquakes including the 1994 Northridge earthquake.

It can be argued (some would say with considerable hindsight) that the key to understanding the seismic performance of tilt-ups is to focus almost entirely on the roof diaphragm and the strength of the connections between the roof and the walls and within the roof itself since these components constitute the “system” that responds dynamically to an earthquake. Even today there is no designated structural system for tilt-ups with an appropriate R value. Tilt-ups are still characterized as concrete or masonry bearing wall buildings with the diaphragms designed for the same R and out-of-plane anchorage forces specified in a separate section. Some advancements have been included in ASCE 7-22, Chapter 12, Section 12.10.4 which contains provisions for roof diaphragm design. They are still too prescriptive to me.

What is still overlooked is that the wall panels, the foundations, and the connections of the wall panels to the foundations are essentially rigid and non-yielding, and regardless of the actual wall design details, are almost never the cause of poor seismic performance.

Dynamically, the wood roof diaphragm and the tributary out-of-plane mass from the concrete walls perpendicular to the earthquake load form a single degree of freedom system: essentially a weighted, simple span beam spanning between two reaction points (the walls parallel to the assumed EQ direction). The reaction points (the parallel walls) have far more strength than the force that the roof diaphragm can generate, so the roof diaphragm is where any yielding will occur. The roof diaphragm is the “yielding” element and the connections between the roof and the walls and the walls themselves are the “non-yielding” elements.

Thinking back to structural dynamics, it only takes about 10 cycles of input for “things to blow up” when the frequency of input matches the natural period/frequency (resonance) of the structure even if the input demands are not that great. This is what happens with tilt-ups. The key to proper seismic performance is determining the actual strength of the roof, usually controlled by the diaphragm shear strength based on the provided plywood nailing and then designing the wall to roof connections for that load or greater with an appropriate R-factor, rather than a code-based out-of-plane load. Using more nails in the roof diaphragm than required, either by design or unintentionally by the contractor (isn’t more, better than less?) is a bad thing.

The required out-of-plane force is:

w (force per foot) = shear strength of diaphragm * length of diaphragm in direction of the load *2 / width of the diaphragm perpendicular to the load

This is an average force assuming a uniform acceleration across the length of the diaphragm. One could also assume that the acceleration shape is more of a bell curve, with higher accelerations in the center region of the diaphragm.

Once in this performance-based mindset, the designer will quickly realize that the out-of-plane connection loads will be highest in the longitudinal direction of a building. This is counterintuitive to many engineers. Minimum shear nail spacing of six inches on center in the longitudinal direction for a long building typically creates a diaphragm that is much, much stronger than it needs to be based on prescriptive code design provisions. Since something will fail, the out-of-plane connections becomes the weak link.

Those particularly interested in this topic should also review FEMA P-1026, Seismic Design of Rigid Wall-Flexible Diaphragm Buildings – An Alternative Procedure (Second Edition).

The 1989 Loma Prieta and 1994 Northridge earthquakes showed that “vintage” buildings with large open parking areas on the ground level or open areas in crawl spaces have significant seismic vulnerabilities and are prone to collapse in strong earthquake shaking. Many were designed before the advent of building codes, but many are more modern buildings with tuck-under parking, code-compliant at the time of their design. Figures 4 and 5 show damaged buildings in the Marina District of San Francsico after the 1989 Loma Prieta earthquake.

Common design flaws include:

In addition, these buildings are mostly older structures that were designed to lower lateral forces levels than today, sometimes with straight sheathed walls or stucco clad stud walls without underlying sheathing (straight, diagonal or plywood).

In 2013, nearly 25 years after the 1989 Loma Prieta earthquake, the City of San Francisco embarked on an ambitious and groundbreaking endeavor: the mandatory seismic retrofit of its wood framed soft-story apartment buildings. The Loma Prieta earthquake caused considerable damage to such buildings in the Marina District and exposed the vulnerability of buildings with soft and weak first stories. Similar performance occurred in the 1994 Northridge earthquake.
Yes, even wood framed buildings, thought by most engineers to be the most naturally earthquake resistant type of structure due to their lightweight nature and reserve strength, can collapse under the right (or perhaps wrong) circumstances. According to a 2016 report by the Association of Bay Area Governments, San Francisco had 6,700 soft-story buildings, far more than the rest of the region combined. The goal of the San Francisco Ordinance was two-fold: 1) address an obvious structural hazard with a threat to life-safety and 2) protect the city’s housing stock by improving resilience. Approximately 10% to 15% of the San Francisco population live in roughly 5,000 buildings covered by the ordinance. Since the enactment of the San Francisco Ordinance, Los Angeles, and Oakland have enacted similar ordinances.

San Francisco, despite its downtown of steel and concrete high-rises, is really the land of wood structures—long, narrow, multi-story buildings with zero side setbacks. Whether in single family homes or multi-tenant apartment buildings, parking on the ground level is almost universally nose-to-tail. Most of the oldest buildings, built before the automobile era, have been modified to allow for parking too. Linear parking configurations make transverse shear walls impossible to use and the result was thousands of weak and soft story buildings.

The Mandatory Seismic Retrofit Program addresses wood framed buildings three-stories or taller, or two-story buildings over a basement or crawl space, with five or more dwelling units, constructed under a permit dated before January 1, 1978, and with no prior seismic strengthening.

Buildings were grouped into four Tiers, with the largest and most vulnerable in Tier 1 (special, institutional, and educational), then Tier 2 (15 or more units), then Tier 3 (5 to 14 units) and then Tier 4 (buildings with ground floor commercial spaces). Commercial buildings were placed last because it was judged (as it turned out rightly) that building owners would need additional time to plan and work with tenants. It turned out that businesses often had to close for several months or moved to other buildings to stay in business.

The Ordinance requires retrofit work in the weak/soft or “target” story only. The target story is considered weak/soft if the number of walls and the wall layout are significantly different from the typical stories above. San Francisco’s residential buildings commonly have identical or nearly identical plan layouts in the upper stories with a large number of interior walls around small rooms, and open ground levels consisting of undeveloped crawl spaces or developed ground levels with large unobstructed areas used for parking or storage. To be subject to the Ordinance, the lateral force resisting system in the target story must be wood framed elements.

Most retrofits involve the addition of plywood sheathing to the existing wood stud walls. When this is not feasible (due to parking or access requirements) steel moment-resting frames are often added in the parking stalls either in front of or behind the existing columns (Fig. 6). The size of frame elements can impact parking (this is the prime driver of any design) so having more frames means smaller columns and less demand on the existing structure (more attachment points). The frames can be of a traditional configuration or inverted, with the beams in the ground and the columns cantilevering upward to the floor above (Figs. 7 and 8). Whenever possible, plywood shear walls or high capacity “strong walls” manufactured by Simpson and Mi-Tek are often used instead of steel moment frames to save money.

Over time, as retrofits were designed and constructed, it became evident that the technical requirements were not as clear as originally hoped, which led to a problem with the consistency in the application of the program’s technical provisions. San Francisco looked to the Existing Building Committee of the Structural Engineers Association of Northern California (SEAONC) for guidance, which improved the situation but did not eliminate all inconsistencies.

How Strong?—California Existing Building Code (CEBC), Appendix 4, allows retrofits to be designed to 75% of the design base shear coefficient used for new buildings which is based on the 475-year earthquake (i.e. repeat of the 1906 Earthquake). Engineers found that the strength of the stories above the target story, due to the considerable length of interior wood lath and plaster finishes and exterior wood siding, sometimes also with stucco, in older buildings, often had lateral strengths far in excess of the base shear required as a result of their weight alone. As a result, designing to the code design base shear improved the expected performance but is considered by many engineers to be insufficient to eliminate the soft- or weak-story condition.

Multiple R-Values—Retrofits often involved multiple systems with different R-values. Plywood shear walls are almost always used (R=6.5) but due to the parking issues, ordinary steel moment frames (R=3.5) and cantilevered column systems (R=2.5) are often used. After much debate, R values could be considered on a line-by-line basis, largely because of the assumption of flexible diaphragms.

Foundations—Original continuous footings below the wood bearing walls are commonly of unreinforced concrete. In the oldest buildings, some footings are of brick, predating concrete. Although there was no consensus, standard practice was to not replace or strengthen the foundations unless there was a large overturning moment that the existing foundations could not resist. Providing new foundations to resist overturning moments or to address ACI anchor bolt provisions would increase construction costs astronomically and were judged to not significantly improve the expected performance relative to the extra cost.

As one can see, developing code provisions for a class of buildings that are relatively similar is difficult. It is far better to define the problem, present information on how such buildings have performed in past earthquakes and why, and then let the structural engineer decide on the appropriate design. There is still an important role for plan review by the building department to confirm that a proper design has been developed and constructed. ■

Part 2 of this series appeared in the March 2025 issue of STRUCTURE.

About the Author

John A. Dal Pino is a Principal with Claremont Engineers, Inc. in Oakland, California. He serves as the Chair of the STRUCTURE Editorial Board (jdalpino@claremontengineers.com).