Unique Site Conditions and Response Analysis Challenges in the Central and Eastern U.S. James R. Martin, C. Guney Olgun, & Morgan Eddy Civil and Environmental Engineering World Institute for Disaster Risk Management at Virginia Tech ()
Earthquake Engineering Issues in the CEUS Unique geologic conditions fall line (Columbia) 160 km coast line (Charleston) Soft rock sediments B-C Classification ~1 km Hard Rock
Site Response Issues in the CEUS Unique geologic conditions in the CEUS: High shear wave velocity rock (>2,500 m/s) close to surface, overlain by sediments (~250 m/s), resulting in very high impedance contrasts (i.e., Columbia) Deep sediment stacks (NEHRP B-C material) resulting in amplification of long period motions (i.e., Charleston) NEHRP maps developed for generic geologic conditions NEHRP seismic design procedures biased toward western US sites (i.e., soil amplification factors) Lack of recorded time histories. Challenges in finding candidate motions for site response analyses.
1. Hard rock near surface Columbia, SC Main culprit Actual Vs profile and transitions Assumed Vs profile USGS (generic)
Columbia Effect of High Impedance Contrast 0.4 0.2 0.0 0 Generic USGS 0 Columbia, SC 0.4 0.2 0.0-0.2-0.4 0 10 20 30 40 Soil -0.2-0.4 0 10 20 30 40 50 500 Deconvolute Depth, D (m) 1000 7500 Depth, D (m) 100 150 Transition Site response Assumed profile for CEUS 1996 Nationsl Hazard Maps 200 Hard Rock 0.4 0.2 0.0-0.2 8000-0.4 0 10 20 30 40 0 1000 2000 3000 4000 Shear Wave Velocity, V s (m/s) 250 0 1000 2000 3000 Shear Wave Velocity, V s (m/s) 0.4 0.2 0.0-0.2-0.4 0 10 20 30 40
Computed Site Amplification Factors Response Spectrum Ratio 4.0 3.0 2.0 1.0 pga 0.1g UHS (0.23g) pga 0.5g short long C-D Site F a NEHRP Site Amplification Factors 3.0 2.5 2.0 1.5 E D 1.0 C Short Period 0.5 100 200 300 400 500 4 0.0 0.01 0.1 1 10 Period, T (seconds) Spectral Amplification Ratio Ratio of the response spectra Ground surface to base rock (normalized to Site Class B) F v 3 E 2 D C 1 Long Period 0 100 200 300 400 500 V s,30 (m/s)
Computed Site Amplification Factors 4.0 pga 0.1g UHS (0.23g) pga 0.5g 3.0 2.5 NEHRP Site Amplification Factors Response Spectrum Ratio 3.0 2.0 1.0 short long F a 2.0 1.5 E D 1.0 C Short Period 0.5 100 200 300 400 500 4 0.0 0.01 0.1 1 10 Period, T (seconds) Spectral Amplification Ratio Ratio of the response spectra Ground surface to base rock (normalized to Site Class B) F v 3 E 2 D C 1 Long Period 0 100 200 300 400 500 V s,30 (m/s)
Comparison with NEHRP Simplified Spectra 1.5 "C" Sites 1.5 "D" Sites Spectral Acceleration (g) 1.0 0.5 Spectral Acceleration (g) 1.0 0.5 IBC Spectrum (MCE) Site Class C 0.0 0.01 0.1 1 10 Period, T (seconds) IBC Spectrum (MCE) Site Class D 0.0 0.01 0.1 1 10 Period, T (seconds) Significant amplification, beyond the code, especially at site-class C. The reason for higher amplification at C sites has to do with the match between frequency content of the rock motion and the site period at these sites.
Lessons Learned from Columbia, SC Very hard rock close to surface creates abrupt impedance contrast and large amplifications at short periods NEHRP simplified procedure based on VS-30 misleading at such sites; we predict much higher Fa values, somewhat lower Fv values Need to recognize such sites as special conditions that require site-specific analysis Need to model top 20-40 m accurately Transition of Vs from hard weathered rock to very hard crystalline rock difficult to determine, but has big influence on motions Input ground motions have big influence on site response results Site conditions like this common throughout CEUS Sites near fall line have unique wedge effect - enough soil to amplify short periods, but not enough to damp them (1886 MMI s?)
1886 Charleston Earthquake MMI contours Fall line wedge effect?
2.Deep Sediment Stack Charleston (0.8-km deep) Depth, D (m) 0 500 1000 7500 Top of marl Charleston Profile profile USGS USGS Assumption Spectral Acceleration (g) 2.0 1.5 1.0 0.5 IBC Spectrum (MCE) Site Class D Potential concern here 8000 0 1000 2000 3000 4000 Shear Wave Velocity, V s (m/s) Typically up to 20 meters of soft/loose soils at the top (Vs< 200 m/s). Underlain by Tertiary deposits (Vs ~ 700 m/s) about 800 meters thick 0.0 0.01 0.1 1 10 Period, T (seconds) Code highly conservative at T < 0.4 sec. Potential concern at longer periods, esp. T > 3 sec. (T fund sediment stack ~ 4-5 secs)
Comparison with NEHRP Amplification Factors 3 2 F a per IBC: S s = 1.28g Hard rock pga 0.5g Hard Rock pga = 0.75g 4 3 Period Band [0.4 to 10.0 sec.] Hard rock pga 0.5g Hard Rock pga = 0.75g Fa Fv 2 1 1 F v per IBC: S 1 = 0.41g Period Band [0.4 to 2.0 sec.] 0 100 200 300 400 500 V s30 (m/s) 0 100 200 300 400 500 V s30 (m/s) Short-period amplification factor is average of RRS between 0.1 and 0.5 seconds. Mid-period amplification factor is average of RRS between 0.4 and 2.0 seconds (for Charleston, this misses amplification at 4 to 5 seconds)
Lessons Learned - Charleston (deep stack) Coastal Plain sediments softer than USGS generic Vs assumption Much uncertainty in damping characteristics (D,Q) of sediment stack Analyses show low-period motions strongly de-amplified, high-period motions slightly amplified relative to NEHRP; suggests NEHRP overconservative at low periods, potential problem at higher periods Psuedo-nonlinear analyses (i.e., SHAKE) tends to overdamp highfrequency motions for deep stack; non-linear analyses often needed Site response controlled mainly by depth to marl and depth to hard rock; results not sensitive to frequency content of strong ground motion input Should model hard rock base as halfspace (esp. for long-period structures); using soft rock (marl) as base may be unconservative if input motions do not contain appropriate long-period energy Perform parametric variation of factors that most influence problem Recognize that such site conditions require site-specific analysis
Example: using marl as base instead of hard rock
Interactive EQ Hazard Maps Developed for SCDOT VT developed seismic hazard maps (2%/50-yr., 5%/50-yr., & 10%/50-yr) that account for unique geological conditions in SC Maps officially adopted by SCDOT for design VT developed interactive GIS-based analysis system for maps that automatically generates synthetic time histories and performs site response analysis based on input of longitude and latitude and soil profile data; similar to USGS web site Dr. Martin Chapman was lead VT investigator
Geological Conditions Programmed into System
SCDOT Seismic Hazard Maps (2%/50 yr.) Determine S s and S 1 from the maps: Ss (0.2 sec) map S1 (1.0 sec) map
Major Cities Used as Main Grid Points
Other Common CEUS issues: Where to model half space? Lots of soil profiles with increasing Vs with depth (i.e., Savannah River Site, Charleston marl) Deaggregation of seismic hazard to determine scenario earthquakes, requires care. For instance, PGA not as meaningful as for liquefaction evaluation, etc.; key on 0.5 or 1 Hz SA (all pga s controlled by close-in small-magnitude events) USGS map values for soft rock (B-C), not hard rock uuugh!! (makes necessary deconvolution of B-C motions using USGS transfer function to establish hard rock motion) Lack of real records, lack of well-established procedures relative to western US Lack of adequate communication between seismologist, geotechnical, and structural engineer More inconsistency of seismic design product relative to western US
Interpretation of Results at this Class D site? Pseudo-spectral acceleration (g) 1.4 Profile 1 - Seed 1 Profile 1 - Seed 2 SEE 2,500 Event year MCE Profile 2 - Seed 1 Profile 2 - Seed 2 1.2 Profile 3 - Seed 1 Profile 3 - Seed 2 Profile 4 - Seed 1 1.0 Profile 4 - Seed 2 Average Spectrum NEHRP SCDOT Simplified Spectra 0.8 0.6 0.4 0.2 0.0 0.01 0.1 1 10 Lacking the benefit of multiple recordings to cover the natural variability between different events, regional conditions and to develop systematic procedures Limitation of using synthetic motions or scaled records from other regions South SC Coastal Santee River Plain Bridge Class D Site Period (sec) Site Class E Site Class D
Typical Questions & Decisions How many EQ input motions to use? What type of motions? Synthetics? Real motions? What scaling rules to use? Can you find good candidate motions recorded on really hard rock? Sensitivity of results to inputs? Is the answer controlled by input motions? site response? soil behavior? structural characteristics? How should final design spectrum be established from sitespecific analysis results? What confidence level should be associated with the design spectrum? Median? +1 Std. Dev.?; that is, how do we systematically judge the results and provide consistent products? such questions tougher to answer in CEUS.
Summary and Conclusions: Unique geological conditions not captured by current simplified NEHRP procedures Must recognize conditions that require site-specific analysis High impedance contrasts with hard rock near surface can cause unusual amplification of ground motions. C sites may possibly amplify as E sites. Site classification based on top 30 meters (V s-30 ) can be misleading in these areas. Deep sediment stacks act as filters and deamplify the short period motions. Code may be unconservative at long periods Lack of real EQ data and less guidance, leads to inconsistency in CEUS seismic design products
Effect of SC Coastal Plain on Ground Motions Soft Rock Hard Rock Q=30, K=0.05
USGS Map Values for B-C (soft rock) not hard rock Hard rock motion USGS B-C to rock transfer function B-C motion
USGS s Hard Rock to B-C Amplification Curve
Attenuating Effect of Coastal Plain sediments Columbia Charleston 5 Hz 7 Hz 10 Hz