St. Louis area earthquake hazards mapping project; seismic and liquefaction hazard maps

Chris H. Cramer; Robert A. Bauer; Jae-won Chung; David Rogers; Larry Pierce; Vicki Voigt; Brad Mitchell; David Gaunt; Robert Williams; David Hoffman; Gregory L. Hempen; Phyllis Steckel; Oliver S. Boyd; Connor M. Watkins; Kathleen Tucker; Natasha McCallister

doi:10.1785/0220160028

USGS · 70189628

St. Louis area earthquake hazards mapping project; seismic and liquefaction hazard maps

Chris H. Cramer·Robert A. Bauer·Jae-won Chung·David Rogers·Larry Pierce·Vicki Voigt·Brad Mitchell·David Gaunt·Robert Williams·David Hoffman·Gregory L. Hempen·Phyllis Steckel·Oliver S. Boyd·Connor M. Watkins·Kathleen Tucker·Natasha McCallister

View original source ↗Export metadata ↓

Abstract

We present probabilistic and deterministic seismic and liquefaction hazard maps for the densely populated St. Louis metropolitan area that account for the expected effects of surficial geology on earthquake ground shaking. Hazard calculations were based on a map grid of 0.005°, or about every 500 m, and are thus higher in resolution than any earlier studies. To estimate ground motions at the surface of the model (e.g., site amplification), we used a new detailed near‐surface shear‐wave velocity model in a 1D equivalent‐linear response analysis. When compared with the 2014 U.S. Geological Survey (USGS) National Seismic Hazard Model, which uses a uniform firm‐rock‐site condition, the new probabilistic seismic‐hazard estimates document much more variability. Hazard levels for upland sites (consisting of bedrock and weathered bedrock overlain by loess‐covered till and drift deposits), show up to twice the ground‐motion values for peak ground acceleration (PGA), and similar ground‐motion values for 1.0 s spectral acceleration (SA). Probabilistic ground‐motion levels for lowland alluvial floodplain sites (generally the 20–40‐m‐thick modern Mississippi and Missouri River floodplain deposits overlying bedrock) exhibit up to twice the ground‐motion levels for PGA, and up to three times the ground‐motion levels for 1.0 s SA. Liquefaction probability curves were developed from available standard penetration test data assuming typical lowland and upland water table levels. A simplified liquefaction hazard map was created from the 5%‐in‐50‐year probabilistic ground‐shaking model. The liquefaction hazard ranges from low (<40% of area expected to liquefy) in the uplands to severe (>60% of area expected to liquefy) in the lowlands. Because many transportation routes, power and gas transmission lines, and population centers exist in or on the highly susceptible lowland alluvium, these areas in the St. Louis region are at significant potential risk from seismically induced liquefaction and associated ground deformation

Explore related subjects

Seismological Research Letters Article

Source-reported bounding extent: 34.23451236236987° to 41.86956082699455° latitude; -96.1962890625° to -81.49658203125° longitude. This indicates report coverage, not an exact sampling location. View area on OpenStreetMap.

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Chris H. Cramer, Robert A. Bauer, Jae-won Chung, David Rogers, Larry Pierce, Vicki Voigt, Brad Mitchell, David Gaunt, Robert Williams, David Hoffman, Gregory L. Hempen, Phyllis Steckel, Oliver S. Boyd, Connor M. Watkins, Kathleen Tucker, Natasha McCallister. 2016-11-09. St. Louis area earthquake hazards mapping project; seismic and liquefaction hazard maps. https://doi.org/10.1785/0220160028

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related USGS reports

USGS2026

Seismic evidence of the transition from debris flow to hyperconcentrated flow at Mount Rainier, USA

Debris flows are a seasonal hazard in Tahoma Creek in Mount Rainier National Park, Washington, United States. We present seismic recordings spaced every 100–200 m along the path of a debris flow triggered by a glacial outburst on 15 August 2023. These recordings capture changes as the debris flow deposits ∼24,000 cubic meters of material along an ∼800 m stretch of channel, approximately 2.5 km from the glacial source. As the flow deposits material, the seismic signal onset shape associated with the debris flow snout recedes, resulting in an elongation of the delay between the signal onset and the peak amplitude from 100 to 300 s. In parallel, the peak frequency increases by more than 20 Hz, and energy below 20 Hz fades, indicating deposition of larger particles and changes in sediment concentration within the flow. We interpret these changes to correspond to the transition from debris flow to hyperconcentrated flow, a distinction based on sediment concentration by volume and on rheological criteria such as yield strength. We find that the frequency content of the seismic signal depends on grain size and sediment concentrations within the flow, among other factors, complicating the starting assumptions of many physical models of debris flows.

Washington↗

USGS2026

Time for an updated macroseismic scale

Macroseismology plays a crucial role in earthquake hazard and risk analyses, linking past earthquake occurrences and impacts with those of the present and future. The popularity of internet-based community contributions and the usefulness of intensities for communicating seismic hazards, early warning, and impacts to the public have all bolstered the use of macroseismology beyond its traditional utility in characterizing sources and shaking from historical and modern earthquakes (Wald et al., 2023). Nevertheless, many modern approaches for collecting building damage data and assigning macroseismic intensities suffer limitations. Many countries do not or no longer routinely collect such data, and even after significant and even catastrophic earthquakes, building damage data are often not collected or shared systematically, nor are field-based macroseismic intensities professionally assigned. Additionally, the United States (U.S.), Canada, and other nations still use variations of the Modified Mercalli Intensity (MMI) scale (Wood and Neumann, 1931) for assigning macroseismic intensities. At nearly 100 years old, the MMI scale predates most modern building types. As a result, the scale is silent about and blind to building damage that now matters to most domestic occupants. It says nothing about single- or multi-unit wood-frame buildings, the predominant type of residential structure in the US and Canada, whose damage is a particularly important indicator of shaking intensity, nor about the collapse or safety of modern building types. Importantly, the dependence on the U.S. Geological Survey’s “Did You Feel It?” (DYFI) for domestic intensity data collection limits their accuracy type at higher, damaging intensity levels, where attention to building type is requisite. The MMI and some of its widely used variants are often poorly documented; there is no official version, nor is it even clear which version of the scale was employed in many studies (Musson, 2002). The MMI scale is consistent with—but is more limited than—the 1998 European Macroseismic Scale (EMS-98, Grünthal, 1998). The new International Macroseismic Scale of 2025 (IMS-25; Grünthal et al., 2025) builds on EMS-98—a well-documented, modern scale already widely used in Europe and well beyond. IMS-25 further expands EMS-98 internationally by enhancing the EMS-98 vulnerability table and damage grade collections to include more worldwide building types and their evolving design codes. Because the IMS requires well-defined vulnerability classes and damage grades specific to various structural types, its worldwide adoption will help unify post-event data collection on building damage. Similarly, the IMS requires statistical analyses of damage levels for specific structural types, ensuring that building damage sampling is sufficient for other important engineering analyses, such as the derivation of fragility functions, calibration of loss models, and the assessment of losses and damages after significant events

Seismological Research Letters↗

USGS2026

Intensity prediction equations for intraslab earthquakes in the Cascadia Subduction Zone

In the U.S. Pacific Northwest (PNW), intraslab earthquakes located within the subducting Juan de Fuca slab are responsible for most of the region’s moderate‐magnitude seismicity. Historically, these earthquakes have caused significant damage, such as the 2001 M 6.8 Nisqually, 1965 M 6.5 Seattle, and 1949 M 6.8 Olympia events. Intensity prediction equations (IPEs) aim to relate source‐to‐site distance and earthquake magnitude to the resultant felt shaking intensity at a particular location, and help to estimate hazard from earthquakes. We find that a commonly used IPE developed by Atkinson et al. (2014) does not accurately describe shaking intensity from most intraslab events in the PNW, likely because it was developed using primarily crustal fault earthquakes. In this study, we develop two new IPEs that provide an improved fit to observed intensities from Cascadia subduction zone intraslab earthquakes. To develop our IPEs, we compute linear mixed‐effects regressions using community decimal intensity data from the U.S. Geological Survey’s ‘Did You Feel It?’ program from 36 intraslab events that occurred in British Columbia, Canada, Washington, Oregon, and northern California from 2000 to 2025. We also investigate the relationship between the new intraslab IPEs and select parameters related to near‐surface conditions and deeper geologic structure that may be related to ground‐motion intensities, but find no correlations. These new IPEs provide a better fit to observations of intraslab earthquakes in the PNW than Atkinson et al. (2014) , particularly within 300 km of the epicenter. They also provide a better fit to the 2018 M 7.1 Anchorage, Alaska, intraslab earthquake, suggesting that they may have broader applicability to global subduction zones.

California, Oregon, Washington↗