A Comparative Analysis of 2-D MASW Shear Wave Velocity Profiling Technique

 

Thanop Thitimakorn

Department of Geological Sciences and Engineering,
University of Missouri-Rolla, Rolla, MO 65409, USA

Neil Anderson

Department of Geological Sciences and Engineering,
University of Missouri-Rolla, Rolla, MO 65409, USA

David Hoffman

National Hazards Mitigation Research Institute,
277 Butler Carlton Hall, 1870 Miner Circle, Rolla, MO 65409, USA

Ahmed Ismail

Department of Geological Sciences and Engineering,
University of Missouri-Rolla, Rolla, MO 65409, USA

ABSTRACT

Multi-channel surface wave data were acquired along near-linear traverses at four separate test sites in southeast Missouri. Each of the acquired surface wave data sets were processed using multichannel analysis of surface wave (MASW) software and transformed into 2-D MASW shear-wave velocity profiles. Shear-wave velocity control extends from the surface to sub-bedrock depths.

The interpreted 2-D MASW shear wave velocity profiles were compared to available crosshole seismic and seismic cone penetrometer shear wave velocity data, and to borehole (depth to bedrock) and cone penetrometer control provided by the Missouri Department of Transportation. The 2-D MASW profiles correlate very well with available geotechnical and subsurface velocity control, supporting the conclusion that the MASW technique can be used to generate reliable and interpretable 2-D shear-wave velocity profiles of the shallow subsurface.

Keywords: Shear wave velocity, Surface wave, MASW, SCPT

INTRODUCTION

The Missouri Department of Transportation (MoDOT) routinely acquires multiple seismic cone penetrometer (SCPT) shear-wave velocity data sets as part of the investigation and NEHRP (National Earthquake Hazards Reduction Program) classification of soils at highway structures and other geotechnical sites in the Mississippi Embayment area of southeast Missouri. The shear-wave velocity of soils is a critically important design criterion because it can be used to determine how highways and highway structures will respond to an earthquake.

In an effort to ensure their geotechnical investigations are as effective and efficient as possible, MoDOT funded a study of the utility of the 2-D MASW shear wave velocity technique. For comparison purposes, 2-D MASW shear wave velocity profiles and SCPT shear wave velocity control were acquired at 4 separate test sites in southeast Missouri. Test site #1 was along a segment of interstate I-70 in downtown St. Louis; test sites #2, #3 and #4 were in the Poplar Bluff area, southeast Missouri (Figures 2 and 3). The MASW data were acquired, processed and interpreted by the University of Missouri-Rolla; the SCPT data were acquired and processed by MoDOT. The shear wave velocity profiles were analyzed comparatively, and the two alternate methods were ranked in terms of accuracy, functionality, cost, other considerations and overall utility to MoDOT.

Test Site #1 MASW profile data (2-D) were acquired along a 1.95 km paved segment of interstate I-70 in downtown St. Louis, Missouri (Figure 2). Multiple SCPT tests were conducted and multiple test boreholes were drilled by MoDOT in the right-of-way (ROW) immediately adjacent to the MASW traverse. The subsurface in the I-70 study area consists of mixtures of clay, silt and sand of variable density. Depth to acoustic bedrock along the 2-D MASW profile varies between 6 and 13.4 m. Bedrock is described in MoDOT driller’s reports as weathered to intact limestone, which in places contains clay seams and/or clay-filled vugs.

MASW shear wave velocity profile data were also acquired at Poplar Bluff test sites #2, #3 and #4 along 152.4 m long traverses (Figure 3). Five SCPT tests were conducted at 38.1 m intervals along each MASW traverse. Crosshole (CH) seismic data were acquired at test site #2 for control purposes. The Poplar Bluff test sites #2, #3 and #4 are characterized by stream-deposited alluvial soils composed mostly of sand with some silt, clay and gravel. The alluvial soils are typically about 30 m thick, except where they thin immediately adjacent to the uplands. Bedrock is relatively flat except in proximity to the uplands.

 


Figure 1. Map of SE Missouri showing St. Louis and Poplar Bluff test sites.

 

ACQUISITION AND PROCESSING

Crosshole (CH) Data

The acquisition of the CH shear wave velocity data at Test Site #2 was relatively straightforward and was conducted according to ASTM standard D 4428. A high frequency shear wave source was lowered to the base of one of two twinned (4.5 m separation), PVC-cased, air-filled boreholes. A shear wave geophone (receiver) was lowered to the same depth in the adjacent borehole. The source and receiver were locked in-place. The acoustic borehole source was discharged (in an upward direction); the receiver recorded the arrival time and amplitude of the acoustic shear wave energy that traveled directly from the source to the receiver (cross-hole seismic field record). The source was then discharged in a downward direction thereby generating an opposite polarity field record.

 


Figure 2. Map of I-70 showing locations of MASW traverse, boreholes and SCPT.

 


Figure 3. Map of Poplar Bluff study area showing locations of MASW traverses.

The source and receiver were raised to the surface at 1.5 m increments (test intervals). At each test depth, they were temporarily coupled to the casing and the source was discharged twice (in opposite directions). Two reverse polarity shear wave seismic field records were thereby generated and recorded for each test depth. The reverse polarity seismic field records for each test depth were plotted on the same graph. The transit time of the shear wave, from source to receiver, was determined for each test depth on the basis of the cross-over time of the reverse polarity seismic records.

The separation between the twinned boreholes at each test depth was determined using a borehole deviation tool. The transit time and borehole separation data were then used to determine the in-situ shear wave velocity of the soil at each depth interval tested.

Good quality CH shear wave velocity data, acquired in relatively uniform soil, are generally accepted as more reliable than SCPT or MASW data. Consequently, CH shear wave velocity data were acquired at Test Site #2 with the expectation it could be used as a “yard stick” for evaluating the accuracy of the acquired SCPT and MASW shear-wave velocity data.

CH shear wave velocity data are generally considered to be more accurate than SCPT data for several reasons. First, the CH source signal is higher frequency than the SCPT source signal; hence the arrival time of the CH shear wave pulse can be determined with greater precision. Second, SCPT field data are generally noisier than CH data; hence the arrival time of the shear wave pulse on CH field data can be determined with greater accuracy. Third, CH velocities are typically measured using source/receiver separations on the order of 4.5 m; SCPT velocities are determined using travel distances on the order of 1 m. Therefore the calculation of CH velocities is less affected by small errors in the determination of the travel distances and/or travel times. Fourth, the direct compressional wave and direct shear wave are clearly visually separated on CH field data; this is not the case for SCPT field records acquired at shallow depths.

CH shear wave velocity data are generally considered to be more accurate than MASW shear wave velocity data for several reasons. First, in the CH technique, shear wave travel times and travel distances are measured directly and used to compute shear wave velocities. The MASW technique, in contrast, measures phase-dependent surface wave velocities, and uses these to estimate shear wave velocities as a function of depth. Second, MASW shear wave velocities represent “average” velocities over lateral distances typically on the order of 30 m. Third, MASW shear wave velocities represent “average” velocities over depth intervals that increase with increasing depth of burial. CH velocity data acquired in fairly uniform soil is not subject to significant lateral and/or vertical averaging.

Seismic Cone Penetrometer (SCPT) Data

The SCPT shear wave velocity data were acquired by MoDOT’s CPT crew using a Hogentogler cone penetrometer unit mounted on a dedicated rig. The acquisition process was relatively straightforward (Campanella et al., 1986). A horizontally-polarized geophone, connected to the tip of the seismic cone, was pressed into the subsurface to a depth of 1 m. A shear wave source (hammer and block located almost directly above the tip of the cone) was discharged twice at the surface with opposite directional impacts thereby generating two opposite polarity shear wave field records. These field records were recorded digitally. This process was repeated as the cone was pressed into the subsurface and halted momentarily at depth intervals of 1 m. The cone was pressed into the subsurface until refusal.

The reverse polarity seismic field records for each test depth were plotted on the same graph. The transit time of the shear wave, from source to receiver, was determined for each test depth on the basis of the cross-over time of the reverse polarity seismic records. The transit time and source/receiver separation data were then used to determine the in-situ shear wave velocity for each soil interval tested. The entire SCVPT data set was processed by MoDOT.

Standard CPT data were acquired simultaneously with the SCPT data. The CPT and SCPT logs (collectively) show the following channels: tip resistance, sleeve friction, friction ratio, pore water pressure, and shear wave velocity.

Multichannel Analysis of Surface Wave (MASW) Data

The generation of the 2-D MASW profiles was relatively straightforward. At test site #1, 161 surface wave data sets, spaced at 12 m intervals, were acquired using twenty-four 4.5 Hz geophones spaced at 1.5 m intervals and a sledge hammer source discharged at an offset of 9 m. The acquired surface wave data were processed using the Kansas Geologic Survey (KGS) software package SURFSEIS (Park et al., 2000). Each of the 161 surface wave data sets was transformed from the time domain into the frequency domain using Fast Fourier Transform techniques. These field-based data were used to generate 161 site-specific dispersion curves, each of which was transformed into a vertical 1-D shear-wave velocity curve. Each of these 161 shear-wave velocity curves constitutes a single trace on the 6400 ft test site #1 2-D MASW shear wave velocity profile.

At test sites #2, #3 and #4, forty-nine surface wave data sets, spaced at 3 m intervals, were acquired using twenty-four 4.5 Hz geophones spaced at 1.5 m intervals and a sledge hammer source discharged at offsets of 9 m. These field-based data were used to generate 49 shear-wave velocity curves, each of which constitutes a single trace on the corresponding 146.3 m long 2-D MASW shear wave velocity profile.

I-70 TEST SITE #1 SHEAR WAVE VELOCITY DATA

The interpreted version of the test site #1 MASW shear wave velocity profile is presented as Figure 4. The “horizon” correlated across the MASW profile (depths of 6-13.4 m) is interpreted as limestone bedrock. The bedrock horizon closely follows (with minimal smoothing) the 304.8 m/s shear wave velocity contour. This interpretation was based on 3 borehole logs provided by MoDOT at the onset of the study. Thirteen additional borehole logs were provided by MoDOT after the fact to assess the reasonableness of the MASW interpretation.

 


Figure 4. Interpreted I-70 MASW Profile.

 


Figure 5. B-39 borehole log and MASW #84.

A comparison of borehole depths to bedrock and MASW estimated depths to bedrock at one test locations along the MASW traverse indicates the MASW interpretation compares favorably to ground truth (Figure 5). On average, MASW estimated depths to bedrock at this test location exceed borehole depths to bedrock by ~0.2 m. This average depth differential is remarkably small, given the variable depth to bedrock in the study area and the fact that the boreholes were not situated exactly on the MASW traverse.

 


Figure 6. I-70 SCPT profile (4 traces) with corresponding MASW curves superposed

In Figure 6, SCPT shear wave velocity data from four test locations along the MASW traverse is presented.The closest corresponding 1-D MASW data sets (extracted from 2-D profile) have been superposed for comparison purposes.

POPLAR BLUFF TEST SITE #2 SHEAR WAVE VELOCITY DATA

The interpreted version of the test site #2 MASW shear wave velocity profile is presented as Figure 7. The “horizon” correlated across the MASW profile is interpreted as limestone bedrock. The bedrock horizon closely follows (with minimal smoothing) the 380 m/s shear wave velocity contour. This interpretation was based on the twinned borehole logs, provided by MoDOT, used to acquire CH seismic data (Figure 8).

 


Figure 7. MASW profile for test site #2.

 


Figure 8. Map of test site #2.

 


Figure 9. CH, MASW, and SCPT data for test site #2.

 


Figure 10. 2-D SCPT profile for test site #2 with superpose MASW traces.

 


Figure 11. SCPT data for test site #2.

The twinned boreholes encountered bedrock at a depth of approximately 34.4 m. However, because of obstructions or the shortness of the PVC casing, CH data were obtained only to a depth of 33.5 m. MASW control extends to a depth of approximately 36.5 m. SCPT data were acquired only to depths of 13 m because of soil resistance.

The CH shear-wave data and corresponding SCPT curve and 1-D MASW trace are presented in Figure 9. The visual inspection of the CH data indicates that the shear wave velocity of soil, with minor fluctuations (on the order of + 30.8 m/s), increases gradually with depth of burial (from a low of about 182.9 m/s to a high of about 304.8 m/s). The soil at test site #2 consists almost exclusively of sand, silt and clay, and the observed minor fluctuations in the CH shear wave velocity are attributed to minor changes in lithology and grain size.

The 1-D MASW shear wave velocity values are similar to the CH shear wave velocity values at depths less than 21.3 m. Within this depth interval (0-21.3 m), the CH velocities range from about 182.8 m/s to about 243.8 m/s. The corresponding MASW velocities range from about 182.8 m/s to about 236.2 m/s. The MASW velocity values at depths greater than 21.3 m are consistently10-15% lower than the corresponding CH velocity values. These differences may be due to the fact that MASW velocities are laterally and vertically averaged or they could be due to the fundamental inaccuracy of the MASW method.

The SCPT shear wave velocity values are comparable to both the CH and MASW values except that both the shallowest layer and the deepest layer on the SCPT velocity curve exhibit anomalously high shear wave velocities (“spikes”). SCPT data acquired at four other locations at test site #2 also correlate fairly well with corresponding MASW traces except for the presence of anomalously low- and high-velocity “spikes” (Figure 10). Such “spikes” are not uncommon on SCPT profiles particularly at shallow depths and are not believed to be real rather they are assumed to result from minor inaccuracies in either the placement of the SCPT receiver or the timing of transit times. Indeed the high-velocity “spikes” observed on SCPT profile # 5 do not correlate with zones of high tip resistance on the CPT data set (Figure 11). Nor do any of the “spikes” on the other SCPT shear wave velocity curves correlate with zones of anomalously low or high tip resistance.

POPLAR BLUFF TEST SITES #3 AND #4 SHEAR WAVE VELOCITY DATA

The interpreted versions of the 2-D MASW shear wave velocity profile for sites #3 and #4 are presented as Figures 12 and 13, respectively. The “horizon” correlated across the MASW profiles is interpreted to be limestone bedrock. The” bedrock” horizon closely follows (with minimal smoothing) the 380 m/s shear wave velocity contour. This contour interval was selected because it was consistent with the borehole and velocity control available for test site #2.

Interpreted bedrock on the site #3 and site #4 MASW profiles exhibits relatively little relief (< 1 m) compared with the test site #2 MASW profile. This is not unexpected as these three sites are situated within the Mississippi Embayment, at least a couple of kilometers from the edge of the uplands, whereas site TS3 is situated immediately adjacent to the uplands at the mouth of a major ancestral stream (Figure 3).

In Figure 14 and 15, suite of five SCPT shear-wave velocity profiles (and corresponding 1-D MASW traces) for sites #3 and #4 are presented, respectively. We believe that the MASW shear-wave velocity curves are more reasonable for several reasons. First, the shallowest (or next to shallowest) layer on three of the five SCPT profiles of site #3 has been assigned an anomalously high velocity irrespective of its depth. Inasmuch as the soil is fairly uniform at site #3 and inasmuch as this high velocity layer is not present on SCPT profiles #4 and #5, we conclude these “spikes” are artifacts of processing rather than an indication of near surface conditions. This conclusion is supported by the tip resistance curves on the corresponding CPT data. Analyses of these CPT logs show that the layers that are assigned anomalously high shear-wave velocities are not characterized by unusually high tip resistance.

In Figure 15, the suite of five SCPT shear-wave velocity profiles (and corresponding 1-D MASW profiles) from Test Site #4 are presented. The MASW shear-wave velocity curves are believed to be more reasonable because the anomalously high velocity layer (shallowest or next to shallowest unit) on SCPT profiles (#3, #4, and #5) has been assigned an anomalously high velocity irrespective of its depth. Inasmuch as the soil is fairly uniform at Site #4, we conclude these “spikes” are artifacts of processing rather than an indication of near surface conditions similar to site #3.

 


Figure 12. MASW profile for test site #3.

 


Figure 13. : MASW profile for test site #4.

 


Figure 14. 2-D SCPT profile for test site #3, with superposed MASW traces.

 


Figure 15. 2-D SCPT profile for test site #4, with superposed MASW traces.

COMPARATIVE ANALYSES OF SHEAR WAVE METHODS

The MASW and SCPT tools were evaluated in terms of accuracy, functionality, cost-effectiveness and overall utility. These parameters were evaluated for the purposes of determining the shear-wave velocity of soils and from the perspective of the end user, in this case, the Missouri Department of Transportation.

ACCURACY

The relative accuracy of the MASW and SCPT shear wave velocity curves was estimated using two different approaches: average standard deviation and calculated USGS velocity. These determinations were based on the data acquired at test site 3 only, because CH data was available for this site only.

In the first approach, the CH data were assumed to be the most accurate. Average deviations of the MASW and SCPT data (relative to the CH data) were determined based on velocity differences at 1.5 m intervals over the depth ranges tested to a maximum of 30 m. Using this approach, the MASW data set is the second most accurate with average deviations of 15 m/s. The SCPT data set is least accurate with a standard deviation of 165 m/s. The high deviation associated with the SCPT tool is attributable to the presence of the velocity “spike”.

In the second approach, a NEHRP velocity was calculated for each profile using shear wave velocity values over the depth range tested to a maximum of 30 m using the following formula (NEHRP, 1997):

 

The CH NEHRP velocity of 220 m/s is assumed to be accurate. The MASW data set are the second most accurate with a NEHRP velocity of 205 m/s; the SCPT data are the least accurate with a NEHRP velocity of 635 m/s.

Functionality (Data Acquisition)

Data acquisition functionality was based on each tool’s capability to provide shear wave velocity control at any potential test site in the Mississippi Embayment to a depth of 30 m, as per NEHRP guidelines.

The MASW tool was ranked first in terms of data acquisition functionality. This tool can be deployed anywhere a 36 m geophone array can be laid out and a 9 km sledge hammer source employed. The geophones do not need to be physically coupled to the ground so the tool can be used on any solid surface (paved or asphalt roadway, exposed rock, rocky soils, sandy soils, etc.). Data can also be acquired on sloped surfaces, beneath bridges, etc.

The SCPT method was ranked second because it can be used only where a dedicated rig can be securely anchored. The tool cannot penetrate pavement, asphalt, rock, or rocky soil or even some stiff soils. Depth penetration at Poplar Bluff test site 3 (Figure 9) was limited to about 13 m because stiff soils were encountered. Also, great care must be taken to ensure the strike block and geophone is properly placed every time the source is discharged. The presence of “spikes” on the MoDOT SCPT data set suggest it can be difficult for a production crew to exercise the care necessary to ensure placement errors do not occur.

Functionality (Data Processing)

Data processing functionality was based on the necessity for qualitative input and the potential for resultant human error.

The MASW tool was ranked first in terms of data processing functionality because all of the MASW dispersion data acquired in the Poplar Bluff area were readily “picked” during data processing, and because the results were reproducible to the extent that different processors consistently “picked” the same dispersion curve and generated essentially the same shear wave velocity profiles.

The SCPT tool was ranked second because it was difficult to “pick” shear wave arrival times on records acquired at shallow depth. Additionally, perhaps because of noise, or human acquisition or processing errors, low and high velocity spikes are present on many of the SCPT shear wave velocity curves provided by MoDOT.

Cost

The MASW tool was ranked first in terms of cost. MASW data can be acquired at a single test site by a 3-person field crew in less than one hour (start to finish after arriving on-site). The equipment is portable; normally the crew and equipment are transported in a single vehicle. The data can be processed in the field on a laptop.

The SCPT tool was ranked second because of the high capital cost of the tool and dedicated rig and the cost of transporting the SCPT tool and rig to the field sites. Additionally, the acquisition of SCPT data at each field site took a couple of hours.

Other Considerations

The SCPT and MASW tools were ranked tied for first because both methods can be of significant utility to geotechnical engineers. The CPT data that are acquired simultaneously with the SCPT data are very useful to MoDOT. The MASW tool is also very useful because it can be used to estimate depth to bedrock.

Overall Utility to MoDOT

The MASW tool was ranked first in terms of overall utility. The MASW was ranked tied with the SCPT tool in terms of other considerations, ranked slightly ahead in terms of accuracy and processing functionality, and ranked significantly ahead in terms of acquisition functionality and cost.

DISCUSSION

Theoretically, SCPT shear-wave velocity profiles should be more accurate than the CH shear-wave velocities. This is because the SCPT velocities are measured over shorter source/receiver separations (typically 1 m) and hence are subject to less “averaging”, and because the SCPT tool measures interval velocities over vertical travel paths. The CH tool in contrast, measures interval velocities along horizontal travel paths, and uses these velocities to generate a vertical shear-wave velocity profile.

In practice however, SCPT shear-wave velocities are generally considered to be less reliable than CH shear-wave seismic velocities. There are several reasons. First, the CH source signal is higher frequency than the SCPT source signal, hence the arrival time of the CH shear-wave pulse can be determined with greater precision. Second, SCPT field data are generally noisier than CH data, hence the arrival time of the shear-wave pulse on CH field data can be determined with greater accuracy. Third, CH velocities are measured using source/receiver separations on the order of 4.6 m. SCPT velocities are determined using travel distances on the order of 0.9 m. As a result, the calculation of CH velocities is less affected by small errors in the determination of the travel distances and/or travel times. Fourth, the direct compressional wave and direct shear wave are clearly visually separated on CH field data. This is not the case for SCPT field records acquired at shallow depths.

In our opinion, the SCPT shear-wave velocity profiles are also not as accurate as MASW shear-wave velocity profiles. We believe this is because the SCPT method is particularly sensitive to slight errors in the estimations of the depth of the geophone and slight timing errors. These errors are manifested as high and/or low velocity “spikes” on the SCPT shear-wave velocity profiles. In general, such “spikes” should not be misinterpreted as indicative of subsurface geologic conditions.

On the upside, the SCPT shear-wave velocity profiles (“spikes” excepted) are very similar to both the CH and MASW shear-wave velocity profiles. For geotechnical engineering purposes, the SCPT profiles would yield essentially the same result as the CH or MASW profiles, excepting for the depth limitations associated with the SCPT method. The advantage of the SCPT in comparison with other methods is the fact that the simultaneously acquired CPT data provide information about static soil properties such as point bearing (qc), sleeve frictional resistance (fs) and stratigraphy, as well as ground proofing the site.

CONCLUSIONS

CH shear wave velocity data are much more reliable than either SCPT data or MASW data. However, in our opinion, the cost of acquiring CH data generally does not justify the expense associated with drilling and casing twinned (or tripled) boreholes down to the base of the zone of interest. We do not recommend the acquisition of CH shear wave velocity data as part of routine geotechnical site investigation work.

SCPT shear wave data are less reliable than either CH or MASW data. The SCPT tool also suffers from significant operational limitations. For examples, SCPT data cannot normally be acquired in areas inaccessible to drill rigs, on paved roadway, in bedrock, or in dense or rocky soil. Indeed, most of the SCPT shear wave velocity profiles acquired in the Poplar Bluff study area were terminated (because of penetration limitations) at depths much shallower than 30 m (base depth of interest). On the upside, the CPT data (acquired simultaneously with SCPT data) may have significant benefit to MoDOT. We recommend that MoDOT acquire SCPT data only when/where CPT control is required.

MASW shear wave velocity data are more reliable than SCPT data and only slightly less reliable than CH data. The MASW tool has significant advantages over both the CH and SCPT tools. MASW data are much less expensive than CH data and can normally be acquired in areas inaccessible to drill rigs. MASW data are less expensive than SCPT data and can normally be acquired in areas inaccessible to SCPT rigs, on paved roadway, in bedrock, or in dense or rocky soil. Indeed, the MASW shear wave velocity data acquired in the Poplar Bluff study area routinely extended to depths below 30 m. One other real advantage the MASW tool has over both the CH and SCPT tools is that it can be used to map variable depth to bedrock.

On the basis of the comparative analyses of the shear wave velocity data acquired in the Poplar Bluff study area, we conclude that the MASW method is by far the most cost-effective tool for determining the shear wave velocity of soils for geotechnical site investigation purposes.

REFERENCES

  1. ASTM: American Sociaty for testing and Materials (2003) D4428: Standard test methods for crosshole seismic testing, Annual Book of ASTM Standards Vol. 04.08 Soil and Rock (I).
  2. Campanella, R.G., P.K. Robertson, and D. Gillespie (1986) Seismic Cone Penetration Tests, Use of In-Situ Tests in Geotechnical Engineering, ASCE GSP 6, 166-130.
  3. NEHRP (1997) NEHRP Recommended Provisions for Seismic Regulations for New Buildings and Other Structures, Part 1: Provisions, Building Seismic Safety Council, Washington, D. C.
  4. Park, C.B., R.D. Miller, J. Xia, and J. Ivanov (2000) Multichannel seismic surface-wave methods for geotechnical applications: Geophysics 2000, FHWA and MoDOT Special Publication, 4:7.1-4:7.11.

 

© 2006 ejge