Missouri University of Science and Technology Scholars' Mine Geosciences and Geological and Petroleum Geosciences and Geological and Petroleum Engineering Faculty Research & Creative Works Engineering 01 Jun 2012 Seismic Site Classifications for the St. Louis Urban Area Jaewon Chung J. David Rogers Missouri University of Science and Technology, [email protected] Follow this and additional works at: https://scholarsmine.mst.edu/geosci_geo_peteng_facwork Part of the Geology Commons Recommended Citation J. Chung and J. D. Rogers, "Seismic Site Classifications for the St. Louis Urban Area," Bulletin of the Seismological Society of America, vol. 102, no. 3, pp. 980-990, Seismological Society of America, Jun 2012. The definitive version is available at https://doi.org/10.1785/0120110275 This Article - Journal is brought to you for free and open access by Scholars' Mine. It has been accepted for inclusion in Geosciences and Geological and Petroleum Engineering Faculty Research & Creative Works by an authorized administrator of Scholars' Mine. This work is protected by U. S. Copyright Law. Unauthorized use including reproduction for redistribution requires the permission of the copyright holder. For more information, please contact [email protected]. Bulletin of the Seismological Society of America, Vol. 102, No. 3, pp. 980–990, June 2012, doi: 10.1785/0120110275 Seismic Site Classifications for the St. Louis Urban Area by Jae-won Chung and J. David Rogers Abstract Regional National Earthquake Hazards Reduction Program (NEHRP) soil class maps have become important input parameters for seismic site character- ization and hazard studies. The broad range of shallow shear-wave velocity (VS30, the average shear-wave velocity in the upper 30 m) measurements in the St. Louis area results in significant uncertainties between the actual spot values and the averaged values used to assign NEHRP soil classes for regional seismic hazard studies. In the preparation of an NEHRP site classification map of the St. Louis urban area, we anal- yzed 92 shear-wave velocity (VS) measurements, supplemented by 1400+ standard penetration test (SPT) profiles in areas bereft of VS measurements. SPT blow counts correlated to VS values based on the published correlations. The data were then com- piled for respective surficial geologic units and bedrock type. These data suggest that the reciprocal of VS30 exhibits a fairly linear relationship with depth to bedrock, likely because VS30 is a function of the thickness of surficial materials exhibiting relatively low VS values. The VS30 values were interpolated by summing the regressed VS30 on the depth to bedrock and kriged values of the regression residuals. The resulting NEHRP site classification maps predict that upland areas of the St. Louis area are spatially classified as soil site classes SB to SD, while the low-lying floodplains are consistently classified as SD to SF. Introduction The evaluation of seismic site conditions is commonly and a method of interpolating these measurements generally made by comparing the thickness of unconsolidated materials requires the employment of some assumptions. One of the and shear-wave velocity (VS) of these materials with that of fundamental assumptions is that VS values depend on the the underlying bedrock, generally referred to as the impe- physical properties of materials. This arose from early dance contrast. The intensity of seismic shaking tends to in- studies that sought to correlate VS data with surficial geology crease where sediments of low density and low VS exist and/or late Quaternary stratigraphy and then interpolate those (Fumal and Tinsley, 1985). Site classifications for new con- (Tinsley and Fumal, 1985; Park and Elrick, 1998; Wills et al., struction are generally best resolved with site VS testing of 2000; Stewart et al., 2003). Most of these studies were recognized surficial geologic units. Previous studies have performed in California and were consistent in showing a V V demonstrated that the average S in the upper 30 m ( S30) significant correlation between geologic age and VS, because is inversely correlated with the average horizontal spectral of increased cementation with age. amplification of earthquake ground motion (Borcherdt and Holzer, Bennett, et al. (2005) produced an NEHRP site Gibbs, 1976; Borcherdt et al., 1991). To assess the suscept- condition map of the central San Francisco Bay (140 km2) ibility to ground amplification, in 1994 the National Earth- using 210 VS30 values, which were derived from the average quake Hazards Reduction Program (NEHRP) defined six soil VS values for surficial geologic units and their estimated S S profile types ( A to F) following the study by Borcherdt thicknesses. It is noted, however, that this approach would (1994), which suggested a consistent relationship between be difficult to apply for larger areas (Wills et al., 2006), site response and VS30 (Building Seismic Safety Council because it might require a much larger pool of VS measure- [BSSC], 2003; R. Borcherdt, personal comm., 2008). The six ments, as well as the accurate thickness of each stratigraphic V site classes, defined in terms of S30, are described in Table 1. unit. These variables are seldom known with any significant reliability over broad areas, especially if the stratigraphic Site-Conditions Mapping Method horizons are more than 20 m deep (because far fewer wells The fundamental complication in estimating seismic site or borings penetrate the deeper horizons). This approach also response in urban environments of considerable land area tends to ignore variations of VS with depth. The regional site 2 (>1000 km ) is that VS values are measured at discrete conditions have generally been delineated according to the points (due to cost, permission, and space constraints), arithmetic mean of the VS30 values for the surficial geologic 980 Seismic Site Classifications for the St. Louis Urban Area 981 Table 1 NEHRP Site Classification (BSSC, 2003) Site Class VS30 (m/s) N30 General Description SA > 1500 N.A. Hard rock SB 760–1500 N.A. Rock with moderate weathering SC 360–760 < 50 Very dense soil and soft rock SD 180–360 15–50 Stiff soil SE < 180 < 15 Soft clay soil SF N.A. N.A. Soils requiring site-specific evaluations unit and their corresponding NEHRP site classes (Borcherdt et al., 1991; Borcherdt, 1994; Wills et al., 2000; Wills and Clahan, 2006). As a consequence, this method implies that areas lying upon the same surficial geologic units would be designated as the same NEHRP site class. Uncertainty of NEHRP Site Classes Assignments Our previous compilation of VS30 site-conditions (Chung, 2007) for the St. Louis metro area suggested con- V siderable variance and an uneven distribution of S30 data, Figure 1. Histograms of VS30 values within the surficial geo- including logic units blanketing the St. Louis area showing (a) the wide dis- tribution of VS30 values and outliers, (b) instances where the mean • V The wide distribution of S30 values within mapped units VS30 values fell on the border between NEHRP site class boundaries, exists between the NEHRP site class categories. Outlier and (c) an inadequate number of VS tests to adequately assess the values of VS30 were occasionally identified, which tend alluvial fan unit. to affect the overall average of VS30 values, shifting a unit’s assignment to the next higher soil site class (Fig. 1a). In the St. Louis area study, we employed the following • V The average value of S30 assigned to a specific unit is procedures: often within 20 m=s of the designated site class bound- aries (Fig. 1b, c). 1. Analyze VS profiles, as well as standard penetration test • Very few VS tests were carried out on alluvial fan facies, and (SPT) profiles. The SPT profiles are less accurate but al- V these were classified as zones SE (VS30 < 180 m=s) and SD low us to fill in areas bereft of reliable S measurements. (VS30 180–360 m=s; Fig. 1c). If more tests had been taken in They are also helpful in identifying anomalous outliers, these spatially-restricted materials, we could expect the which may be due to buried bedrock knobs or measure- average values to favor either one site class or the other. ment errors. 2. Identify observable relationships between VS30 values and the respective depths to bedrock, in order to charac- The variability of VS30 within the surficial geologic units terize the variance of VS30 with bedrock depth. is likely attributed to (1) natural variations in stratigraphy, the 3. Prepare detailed regional NEHRP site classification maps, depth to rock horizons, and the range of sedimentation using VS30 estimates. For this step, we used 92 VS30 mea- (Bauer et al., 2001; Romero and Rix, 2001; Gomberg et al., surements and more than 1400 SPT profiles (taken from 2003; Holzer, Padovani et al., 2005); (2) misclassification of engineering borings), surficial and bedrock geologic these same units by using increasingly inaccurate geologic maps, and the depth-to-bedrock maps. maps of smaller scale (Park and Elrick, 1998; Wills and Clahan, 2006); or (3) the result of human interpretation and/or instrumental errors (Scott et al., 2006; Bauer, 2007). Study Area These factors often result in significant uncertainties between Geologic Setting and Earthquake Hazards the actual VS30 values and the averaged values used to assign NEHRP soil classes in regional seismic hazard studies, which The study area encompasses 12 U.S. Geological Survey can lead to erroneous site-condition maps. Wills et al. (2000) (USGS) 7.5 minute quadrangles (a land area of 1800 km2)in found that 25% of VS sites within California were misclas- the greater St. Louis area, Missouri, and Illinois (Fig. 2). This sified. These problems imply that the mapped units can be area is located on Quaternary deposits, which consist of assigned with multiple NEHRP site classes. (1) Holocene alluvial deposits in major channels (Mississippi, 982 J. Chung and J.