Potential Acceleration Values During a Wasatch Fault Earthquake
By Brent Maxfield, SE and Eric Hoffman, SE, Reviewed by the SEAU Technical Committee
The purpose of this article is to review the code methodology for setting seismic response acceleration values (S_{S} and S_{1}). We focus on how these code values compare to the response accelerations that could potentially be experienced during a Wasatch fault earthquake. The Wasatch fault is thought to have the potential to generate acceleration levels that exceed code prescribed values of S_{S} and S_{1} by more than a factor of two. The building code methodology, which is used uniformly across the United States to set ground motions, may not provide the level of protection expected by communities along the Wasatch Front. This article provides background information and describes what the Technical Committee is doing to better understand the effects that a Wasatch fault earthquake could have on our code-based designs. Prior to writing this article the Technical Committee has discussed these concepts with members of the USGS, UGS, U of U Seismograph Stations, ASCE 7 Seismic Code Committee Members, and other leading experts in the field.
Nearly every structural engineer has obtained the spectral response values needed for design, S_{S} and S_{1}, from a hazard website using the latitude and longitude of a building site. The hazard websites get their values from the USGS. However, many structural engineers may not understand the additional information provided from these hazard websites. We use the www.hazards.atcouncil.org website and a location in downtown Salt Lake City with coordinates of 40.769, and -111.891 as an example to discuss some of this additional information as shown in Figure 1.
Figure 1 - ATC Seismic Hazards
The S_{S} and S_{1} values we obtained for the example location are S_{S}=1.484g, and S_{1}=0.541g. These two values define the response spectrum or MCE_{R} that is used in most of our designs. Although these are the only values required for design, note that there is an important additional section called, oddly enough, “Additional Information.” We will discuss four of those values (SsRT, S1RT, SsD, and S1D).
SsRT and S1RT are the “Probabilistic risk-targeted ground motions.” They are derived by taking SsUH*CR_{S} and S1UH*CR_{1} to obtain a response acceleration that is expected to achieve a 1% probability of collapse within a 50-year period (See ASCE 7-16 Section 21.2.1 and Commentary Sections C21.2.1, C21.2.1.1, and C21.2.1.2). For this site SsRT=1.484g and S1RT=0.541g. Note they are the same values as S_{S} and S_{1} (see more on this below).
The other two values of particular interest for the Wasatch Front, are SsD and S1D. For this site SsD=2.694g, and S1D=1.16g. These values are the “factored deterministic acceleration values”-the spectral acceleration values that have about an 84% probability of not being exceeded during a Wasatch fault earthquake (See ASCE 7-16 Section 21.2.2). (Think of the word deterministic as accelerations from a scenario earthquake assuming the earthquake has occurred.) They are not the largest potential acceleration values that a Wasatch fault earthquake might generate, but they represent the 84^{th} percentile (or one standard deviation above the median) acceleration values. The deterministic values are 1.81 times higher than SsRT and 2.14 times higher than S1RT.
The code methodology takes the lesser of the probabilistic and the deterministic values and does not require engineers to consider the higher deterministic acceleration levels (See ASCE 7-16, Section 21.2.3). At our example site S_{S} is the lower of SsRT and SsD, and S_{1} is the lower of S1RT and S1D.
There are several factors that determine the probabilistic values of SsRT and S1RT along the Wasatch Front, but two primary factors are: 1) the level of acceleration that the Wasatch fault can generate, and 2) the return period of a Wasatch fault rupture. Other factors include other nearby faults, the magnitude of earthquakes those faults can generate, and the return period of these faults. Additionally, some gridded background (random) seismicity is also included to account for potential earthquakes not directly associated with any particular, known fault.
The potential acceleration levels for the Wasatch fault are very high, but its return period is very long. The long return period tends to lower the values of S_{S} and S_{1}. If the return period for the Wasatch fault were 500 years instead of roughly 1250 years, then the S_{S} and S_{1} acceleration values would be considerably higher, not because the acceleration potential has increased, but because the return period has decreased.
The code methodology also treats the earthquake return period as a random (Poisson) event, meaning that it does not consider the time since the last earthquake occurred in calculating the fault’s risk of rupture. If the code used the probability of a Wasatch fault rupture in the next 50 years rather than any given 50 years, the S_{s} and S_{1} values would be higher along large portions of the Wasatch Front.
To add perspective to these values we have compared the Wasatch fault’s deterministic values of SsD and S1D with other high seismic cities along the west coast (see Table 1). The deterministic values show that although the Wasatch fault moves far less frequently than some of the more infamous West Coast faults, when it does rupture, the peak accelerations have the potential to be just as or even more violent.
City |
Lat, Long |
SsD |
S1D |
Salt Lake City, UT |
40.769, -111.891 |
2.69 |
1.16 |
Provo, UT |
40.233, -111.662 |
3.03 |
1.30 |
Ogden, UT |
41.225, -111.973 |
2.99 |
1.24 |
San Francisco, CA |
37.770, -122.443 |
1.50 |
0.60 |
Los Angeles, CA |
34.052, -118.235 |
2.48 |
0.78 |
Portland, OR |
45.504, -122.674 |
1.50 |
0.60 |
Seattle, WA |
47.605, -122.332 |
2.79 |
1.16 |
Table 1 - Comparison of Various Western Cities Deterministic Accelerations
We are working to understand the impact that a Wasatch fault earthquake could have on the Wasatch Front if areas of the community experience acceleration levels about double the S_{S} and S_{1} (MCE_{R}) values.
According to the collapse fragility curve defined in ASCE 7-16, Section 21.2.1.2, the probability of collapse is 4.5 times higher if a Risk Category II building experiences an acceleration level of twice the S_{s} or S_{1} values. (See ASCE 7-16 Commentary Figure C16.4-1 (Complete Collapse Fragility), which shows that if the spectral acceleration Sa is twice the MCE_{R}, the collapse probability is 0.45 vs. 0.10 at MCE_{R}.) The code expected performance for a Risk Category II building in an MCE_{R} event is Collapse Prevention and Life Safety at 2/3 MCE_{R} (See ASCE 7-16 Commentary Figure C11.5-1).
The above data shows that designing buildings per the current code methodology along the Wasatch Front results in design forces well below what a Wasatch fault earthquake can potentially generate. The Technical Committee is working to answer the “How big of a deal is this?” question. To accomplish this, the Committee proposes two studies.
The first study will compare potential losses given different scenarios. Potential losses have been studied previously using median predicted accelerations (50^{th} percentile deterministic) for a Wasatch fault earthquake. However past earthquake records have shown that median accelerations are not felt everywhere in a large earthquake, but pockets of higher and lower accelerations occur throughout the affected area. This is part of why the code uses 84^{th} percentile rather than 50^{th} percentile to calculate deterministic accelerations.
For this study we will use Hazus in conjunction with the SP3 Risk Model to study the losses (fatalities, damage costs, and downtime) using a series of ShakeMap inputs that will be populated with areas of various acceleration levels above and below the median predicted acceleration levels for a Wasatch fault earthquake. We will then rerun the same ShakeMap scenarios but change the Hazus building database for all buildings constructed after 2002, as if they were engineered for an S_{s} and S_{1} equal to the higher 84^{th} percentile deterministic SsD and S1D acceleration levels. This will make the buildings in the database built since 2002 stronger, potentially resulting in lower losses. We will then have loss estimates from: 1) a scenario with median predicted acceleration levels everywhere, 2) scenarios with varying acceleration levels, with some areas higher and some areas lower than median, and 3) the same scenarios as in 2, but with stronger buildings in the Hazus database. Comparing these loss estimates will help us better understand potential losses from a Wasatch fault earthquake and what effect strengthening building stock would have on those losses.
The second study will look at the construction cost increases if a building were engineered using the higher deterministic SsD and S1D acceleration levels. We will be engaging local engineering firms to help with these analyses.
The results from these two studies will allow us to have informed discussions about the potential effects of continuing the status quo and the potential costs and benefits of increasing design forces to potentially address a scenario Wasatch fault rupture. We are currently working on funding these two studies.
We invite SEAU members to be a part of the SEAU Technical Committee which covers the various technical topics affecting structural engineers including seismic, wind and snow loads, code questions and amendments, existing buildings, resiliency and emergency response, residential and commercial construction, construction quality and any other technical topics that need to be addressed by SEAU. Please contact the Committee Chair, Eric Hoffman to get invited to our meetings. [email protected].