Hammer-Impact SH-Wave Seismic Reflection Methods in Neotectonic Investigations: General Observations and Case Histories from the Mississippi Embayment, U.S.A.

INTRODUCTION

The relationship between contemporary seismicity and shallow geologic structure is one of the most important topics of research related to seismic hazard evaluation in the Mississippi Embayment region of the central United States. Seismic activity in the Mississippi Embayment is dominated by earthquakes that occur in the New Madrid seismic zone, the most active seismic zone in the central and eastern United States (Fig. 1) and home to the great New Madrid earthquake sequence of 1811–1812 (Johnston and Schweig, 1996). Determining the association between seismicity and surface deformation in the Mississippi Embayment is hindered by the presence of thick, water-saturated, unconsolidated sediments. The upward continuation of basement faults into the nearsurface (Tertiary and Quaternary) sands, gravels, silts, and clays is often masked by the inability of the soft sediments to propagate large fractures, making identification and age determination of near-surface faults a difficult problem because deformation is often distributed over a complex zone of disturbance.

Figure  1.  Map showing the seismicity (open circles representing earthquakes of M=1 to M=6) of part of the Mississippi Embayment and locations (squares) of S-wave seismic reflection profiles discussed in this article [earthquake data (1974-2002) from CERI catalog, University of Memphis]. Squares: M. Monticello; BC. Big Creek; PG. Porter Gap; NM. New Madrid; T. Tamms. Stars: Pad. Paducah; Mem. Memphis. MO. Missouri; IL. Illinois; KY. Kentucky; TN. Tennessee; AR. Arkansas; MS. Mississippi.

DownLoad: Full-Size Img PowerPoint

Although P-wave (compressional wave) reflection methods have been used for 30 years for imaging faults in the Paleozoic through Tertiary deposits of the Mississippi Embayment (e.g., Stephenson et al., 1999; Williams et al., 1995; Schweig et al., 1992; Sexton and Jones, 1986; Zoback, 1979), only in the last two decades has detailed seismic investigation of deformed Quaternary sediments, primarily using S-wave (shear wave) methods, been undertaken (e.g., Cox et al., 2006, 2001, 2000; Harris and Sorrells, 2006; Baldwin et al., 2005, 2002; Woolery et al., 1999, 1996, 1993; Harris et al., 1998). One of the principal justifications for using S-waves is the potential for increased seismic resolution (compared with conventional Pwave methods) in the water-saturated, shallow sediments of the Mississippi Embayment—an important consideration in light of the modest surface deformation in the region.

Through direct comparison with shallow P-wave reflection data, this article shows the influence of near-surface geology (sediment composition and water saturation) on seismic data quality, and the increased resolving power of S-wave seismic reflection methods in unconsolidated, water-saturated sediment sequences. Finally, five brief case histories are presented in which low-cost, hammer-impact S-wave seismic reflection profiling was directly integrated into neotectonic investigations (supporting surficial mapping, shallow drilling, and paleoseismologic trenching) of surface deformation in the Mississippi Embayment.

REGIONAL GEOLOGY

The structural configuration of the Mississippi Embayment is a south-plunging syncline containing a thick sequence of unconsolidated to poorly consolidated gravels, sands, silts, and clays, and weakly lithified sedimentary rocks. The sediments and sedimentary rocks filling the Paleozoic bedrock trough range in age from Late Cretaceous to Quaternary and thicken southwestward from ~100 m at Paducah, Kentucky, near the northern boundary of the Mississippi Embayment, to >1 km in the vicinity of Memphis, Tennessee.

The first suggestion for the origin of the Mississippi Embayment, by Burke and Dewey (1973), was that its formation was due to a failed rift arm. Ervin and McGinnis (1975) expanded this idea and presented the Mississippi Embayment as a Late Precambrian or Early Cambrian rift zone. During the Late Mesozoic, the southeastern United States moved across the Bermuda Hotspot, causing reactivation of the rift complex followed by subsidence and deposition of marine sands and clays from the southwest and fluvial clastics from the north (Cox and van Arsdale, 2002, 1997). Instability of the crust in the vicinity of the rift complex is believed to be responsible for earthquakes in the upper Mississippi Embayment (Herrmann and Canas, 1978). Reactivation of favorably oriented zones of weakness in response to the regional stress field has been presented as the driving mechanism for the seismicity.

BASIS FOR USE OF SH-WAVE METHODS

The choice of the SH-wave (horizontally polarized) as the preferred shear wave phase is based on the theoretical determination that SH-waves are easier to identify because pure SH-wave energy reflects and refracts only as an SH-wave and, unlike P- and SVwaves (shear waves polarized in a vertical plane), do not experience mode conversion. The consistent character of shallow SH-wave reflections should, therefore, make their identification and interpretation less complicated.

SH-wave seismic reflection methods have been used to map shallow geologic features in unconsolidated, water-saturated sediments (Young and Hoyos, 2001; Harris et al., 2000; Goforth and Hayward, 1992; Hasbrouck, 1991; Suyama et al., 1987). In the Mississippi Embayment, SH-wave methods have been tested at numerous sites, mostly in support of earthquake hazard investigations (Cox et al., 2006, 2001, 2000; Harris and Sorrells, 2006; Baldwin et al., 2005, 2002; Woolery et al., 1999, 1996, 1993; Harris et al., 1998). Figure 2 shows representative field data (from Memphis, Tennessee and Paducah, Kentucky) exhibiting the range of depths (< 10 m to >100 m) imaged by hammer-impact, SH-wave reflection methods in the Mississippi Embayment.

Figure  2.  Example S-wave reflection field records from (a) Memphis, Tennessee, and (b) Paducah, Kentucky. The Memphis shot gathers show numerous shallow reflections from the Quaternary section (between 100 and 300 ms) corresponding to depths of < 10 to ~25 m. The target for the Paducah data (~500 ms) is Mississippian limestone (>100 m deep).

DownLoad: Full-Size Img PowerPoint

GENERAL OBSERVATIONS OF SH-WAVE REFLECTIONS IN MISSISSIPPI EMBAYMENT

Influence of Near-Surface Geology on Data Quality

Amplitude and frequency characteristics of Pwave reflection data are highly influenced by the depth of water saturation in near-surface materials. Because S-waves travel with the velocity of the sediment framework, they are not greatly affected by the degree of saturation, and often lead to more consistent, high-quality data in unconsolidated, watersaturated sediment sequences. For example, Fig. 3 shows a comparison between SH- and P-wave data collected at two sites (PW-11 and H-3) in the unconsolidated alluvial sediments of the lower Ohio River Valley near Paducah, Kentucky (see Fig. 1 for location). The SH-wave sections (Figs. 3a and 3c) are similar in that both show a direct arrival, a refracted arrival from a shallow horizon, and a reflection from the top of Mississippian-age limestone bedrock (~650 ms at site PW-11 and ~550 ms at site H-3). The quality of the P-wave data (Figs. 3b and 3d), however, varies quite dramatically. The site PW-11 section shows a strong refraction from the top of a shallow water table (calculated to be ~2 m deep) and an easily identifiable bedrock reflection (~160 ms), while the data from site H-3 exhibit severe high-frequency attenuation and provide little, if any, useable information. Based on a borehole log at site H-3, water was first encountered at a depth of 15 m and the near-surface, unsaturated material consisted of a heterogeneous deposit of silt, clay, and pebbly sand. These shallow conditions (thick, attenuative unsaturated zone) impede the collection of highquality P-wave data in that area.

Figure  3.  Example from the Ohio River Valley near Paducah, Kentucky, showing the effects of near-surface saturation conditions on the quality of S- and P-wave data. Notice the decrease in P-wave data quality (as opposed to the relative consistency in S-wave data quality) between the sites PW-11 (water table ~2 m deep) and H-3 (water table ~15 m deep).

DownLoad: Full-Size Img PowerPoint

In near-surface SH-wave reflection profiling, the presence of well-developed surface wave arrivals (Love waves) has commonly been viewed as an obstacle to collecting high-quality data and it has been shown that they can easily be misinterpreted (Miller et al., 2001). Pullan et al. (1990) observed strong Love wave arrivals interfering with shallow S-wave reflections in the presence of large S-wave velocity gradients in the near surface, and Eslick et al. (2008) demonstrated the influence that the thickness of a low-velocity surface layer (waveguide) has on the development of dispersive Love waves. Figure 4 presents four composite (walkaway) gathers collected in the Ohio River Valley near Paducah, Kentucky, showing the varying influence that the Love wave has on viewing the bedrock reflection (~100 m deep). The increasing development of the low-frequency Love wave from northeast to southwest (away from the Ohio River) is evident. The bedrock reflection can clearly be seen at sites J-3 and H-3 between 500 and 600 ms on the longer offset traces. At site J-4, the bedrock reflection is subtle, but can be viewed on the longer offset traces at about 450 ms. The bedrock reflection is masked by the surface wave at site H-1. The velocity of the surface layer (Quaternary loess and alluvium) remains fairly constant along the profile (175–220 m/s). The surface wave is well developed where the underlying unit, with a velocity of >450 m/s, is within 5–6 m of the surface.

Figure  4.  SH-wave walkaway profiles collected in the Ohio River Valley near Paducah, Kentucky. Comparison of the profiles clearly shows the development of the dispersive Love wave from NE to SW (away from the Ohio River Valley) where the ~5-m-thick low-velocity surface deposits (loess) are underlain by higher-velocity upland deposits (sand, gravels, and clays).

DownLoad: Full-Size Img PowerPoint

An interesting observation regarding data quality and source-receiver offset for near-surface S-wave reflection data is related to the influence of the earth's surface. At incident angles greater than critical [sin^-1(V_S/V_P)], amplitude, phase, and frequency changes, caused by interactions with the free surface, can distort the arriving waveform, while arrivals within the "shear-wave window" are generally undisturbed (Evans, 1984). High V_P/V_S ratios, commonly found in the shallow sediments of the Mississippi Embayment, reduce the width of the shear-wave window and can cause distortions on shallow reflection data recorded at offsets with incidence angles beyond critical. Furthermore, this narrowing of the shear-wave window often forces the offset range for unaltered shallow reflections to be coincident with the Love wave noise cone, and special attention is needed to insure that Love wave arrivals aren't stacked as shear-wave reflections (Miller et al., 2001). Figure 5 shows an example shot gather from a seismic profile in western Kentucky (Harris, 1996), on which the shear-wave window has been identified. Analysis of first arrivals (both S-wave and P-wave) suggests a shear-wave window ranging from 17º to 28º. Based on the V_P/V_S ratios and source-receiver offsets (3 to 36 m) used to image the top of a shallow gravel layer (approximately 35 m deep), incidence angles of 2º to 27º were calculated, suggesting that some of the longer offset reflection data were likely to have been modified by recording outside the shear-wave window.

Figure  5.  Example shot gather from western Kentucky showing distorted reflection arrivals recorded outside the shear-wave window (from Harris, 1996).

DownLoad: Full-Size Img PowerPoint

Vertical Resolution

Due to the small-scale deformation encountered in many shallow reflection surveys supporting neotectonic investigations, seismic resolution is frequently the most important consideration when choosing a survey method and making data acquisition decisions. Vertical resolution depends on wavelength (velocity divided by frequency), so for seismic energy of the same frequency, because Swaves travel with lower velocities than P-waves, Swavelengths are shorter and vertical resolution is higher. Even though S-waves are rarely observed in the same frequency range as P-waves (in the nearsurface deposits of the Mississippi Embayment they are commonly 0.5 to 0.25 those of P-waves), higher vertical resolution is most evident in water-saturated, alluvial material where the P-wave velocity is regularly 5 to 10 times higher than the S-wave velocity. Figure 6 compares the resolving capability of S-wave and P-wave reflections from sites with similar geology in northeastern Arkansas. The target reflection is from the Quaternary–Tertiary boundary (35–45 m deep), and an estimate of vertical resolution using the 1/4 wavelength criteria (Widess, 1973) suggests that resolving power is nearly two-thirds greater for the S-wave compared with the P-wave (1.3 m for the S-wave and 2.0 m for the P-wave). In the Mississippi Embayment, the upward continuation of basement (seismogenic?) faults into the soft sediments of the near surface (Tertiary and Quaternary) leads to a subdued, tectonically produced surface topography and makes identification and age determination of surface/near-surface deformation a difficult problem. The most common methods used to investigate shallow subsurface deformation in the Mississippi Embayment include detailed mapping of surficial geology, paleoseismologic trenching, and shallow coring, all supported by seismic reflection profiling.

Figure  6.  S- and P-wave field records from northeastern Arkansas. A comparison of vertical resolution for the reflection from the Quaternary– Tertiary boundary suggests that S-waves have the potential for considerably higher resolving power than P-waves in unconsolidated, water-saturated sediments (P-wave data from Williams et al., 1995, Fig. 9).

DownLoad: Full-Size Img PowerPoint

CASE HISTORIES

Data Acquisition and Processing

The last section of this article contains example case histories showing the use of SH-wave seismic reflection methods in imaging near-surface deformation associated with neotectonic processes in the Mississippi Embayment. At all five sites (see Fig. 1 for locations), the active spread consisted of 12 receivers and the seismic energy was generated by horizontal impacts (oriented perpendicular to the spread) of a sledgehammer on a steel I-beam (Fig. 7). At one site, Tamms, Illinois, the seismic data were collected using a landstreamer, while at the other four sites the individual receivers were coupled to the ground using 6-cm spikes. The landstreamer, with geophones mounted on metal sleds tethered together by nylon webbing, was pulled behind a vehicle that stopped at regular intervals to collect the data. A major advantage of landstreamer acquisition is time efficiency (Pugin et al., 2004), although data quality sometimes suffers. However, the packed-gravel and asphalt road surface, with little traffic, offered excellent conditions for landstreamer acquisition at the Tamms site and, based on observations in the field, the data quality appeared equal to that of spike plant geophones. Table 1 summarizes the seismic data acquisition parameters for the five sites.

Figure  7.  Hammer-impact SH-wave data acquisition. (a) 4.5-kg sledgehammer and 10-kg steel I-beam (Monticello, Arkansas); (b) 1.8-kg sledgehammer and 5-kg steel I-beam (Tamms, Illinois); also shown is the 12-channel landstreamer.

DownLoad: Full-Size Img PowerPoint

Table  1.  SH-wave seismic data acquisition

 | Show Table

DownLoad: CSV

Processing (on a laptop or desktop PC) followed a standard sequence for shallow common-mid-point (CMP) reflection data (e.g., Baker, 1999; Steeples and Miller, 1990) and included trace editing, bandpass filtering, and automatic gain control. The seismic data were stacked as 6-fold sections. Because there was significant topography on several of the profiles, special care was taken with elevation statics and velocity analysis.

Because low-fold, hammer-impact seismic reflection data commonly may have low signal-to-noise ratios, only basic processing steps were employed in order to minimize processing artifacts (and maximize interpretation confidence) produced by "overprocessing" noisy data (Steeples and Miller, 1998). Furthermore, the stacked sections were compared with the field records to confirm that interpreted reflections were visible on shot gathers (see Fig. 12 for example from Tamms, Illinois).

Monticello

A 210-m SH-wave reflection profile (Fig. 8) was acquired across a zone of surface faulting near Monticello, Arkansas (see Fig. 1 for location). As mapped at the surface in a paleoseismologic trench (Cox et al., 2000), the faults displace Tertiary clay and gravel and Holocene silt. The seismic profile shows strong reflection energy to depths in excess of 150 m. The most coherent reflection is from a depth of 110 m and, based on correlation with local well data, is interpreted to represent the top of the Eocene Cockfield (sand) Formation. The reflections are disrupted by two south-dipping normal faults that correlate updip with faults mapped in the paleoseismologic trench. The style of folding and fracturing identified in the trench suggests that the faults were reactivated as a left-slip system in a Quaternary east-west compressional stress field.

Figure  8.  Interpreted SH-wave reflection profile near Monticello, Arkansas, showing the relationship with a coincident paleoseismologic trench (Cox et al., 2000).

DownLoad: Full-Size Img PowerPoint

Big Creek

A 135-m SH-wave reflection profile was shot across a section of the northeast-southwest trending Big Creek escarpment in eastern Arkansas (see Fig. 1 for location). The scarp was first recognized to be of tectonic origin based on geomorphic evidence (Fisk, 1944). Krinitzsky (1950) and Smith and Saucier (1971) produced borehole profiles that show down-to-thesoutheast normal faulting of Tertiary and Quaternary sediments to be coincident with the scarp. The interpreted seismic reflection profile across the Big Creek escarpment (Fig. 9) shows that the scarp is underlain by a high-angle fault that extends upward into Quaternary sediments (shallower that 45 m). Warping of shallow reflections indicates compression and suggests that the principal fault has been active as a reverse or transpressional fault in the contemporary, east-west compressional stress field.

Figure  9.  Interpreted SH-wave reflection profile across the Big Creek escarpment (eastern Arkansas) and correlation with an adjacent borehole profile (modified from Smith and Saucier, 1971).

DownLoad: Full-Size Img PowerPoint

Porter Gap

A 330-m SH-wave reflection profile was collected at Porter Gap, Tennessee (see Fig. 1 for location), across the margin of the Chickasaw bluffs of the Mississippi River (Cox et al., 2001). The Porter Gap profile (Fig. 10) exhibits strong, continuous reflections at depths of 50–75 m and 90–120 m, both in the Eocene section. Both reflections are disrupted by a nearvertical fault complex coincident with the west-facing bluff margin. Separation on the fault is up-to-the-west, although structural relief along the line is higher on the east (consistent with topography). Therefore, the fault is interpreted to have been structurally inverted during Late Tertiary or Quaternary. This is similar to the deformational history suggested by the reflection profile across the Big Creek escarpment, over 200 km to the southwest.

Figure  10.  Interpreted SH-wave reflection profile across the Chickasaw bluffs at Porter Gap, Tennessee (Cox et al., 2001).

DownLoad: Full-Size Img PowerPoint

New Madrid

The 445-m SH-wave reflection profile at New Madrid, Missouri (see Fig. 1 for location), was shot across a topographic scarp that has been recognized as the northwest extension of the Reelfoot scarp (van Arsdale et al., 1995), last active during the M=8, February 7, 1812 New Madrid earthquake (Kelson et al., 1996). Based on shallow coring and paleoseismologic trenching, the Reelfoot scarp has been interpreted as a fault propagation fold above the west-dipping Reel foot reverse fault. The interpreted profile (Fig. 11) across the New Madrid scarp shows a west-dipping reverse fault (located east of the scarp) that propagates upward from Tertiary (>40 m deep) to Quaternary sediments. This structural setting is almost identical to that imaged on the Reelfoot fault in west Tennessee (Sexton and Jones, 1986) where scarp topography and near-surface folding/extensional faulting (Kelson et al., 1996; van Arsdale et al., 1995) mimics folding and extensional faulting in the hanging-wall of the reverse fault. This is the first seismic image that clearly shows the continuation of the Reelfoot fault into Quaternary sediments, displaying the relationship between scarp topography and near-surface folding in the hangingwall of the Reelfoot fault.

Figure  11.  Interpreted SH-wave reflection profile from New Madrid, Missouri. The shallow borehole information superimposed on the topographic profile is from van Arsdale et al. (1995).

DownLoad: Full-Size Img PowerPoint

Tamms

The northeast-trending Commerce Geophysical Lineament (CGL), a series of gravity and magnetic anomalies extending from northeast Arkansas to central Indiana, is associated with Quaternary deformation northwest of the New Madrid seismic zone. In southern Illinois, a previous study imaged Quaternary sediments deformed along the Penitentiary fault (PF), a northeast-striking fault coincident with the CGL (Odum et al., 2002). The PF is coincident with a linear northeast-trending, east-facing bluff line that separates Late Quaternary Cache River Valley (CRV) deposits on the east from Paleozoic upland rocks on the west.A nearly 1-km SH-wave reflection profile, intersecting the PF and the CGL, was acquired near Tamms, Illinois (see Fig. 1 for location). The seismic profile (Fig. 12) displays coherent reflectors at both the bedrock-CRV unconformable contact and within the Quaternary deposits. Depth to bedrock was interpreted as approximately 30–50 m deep in the CRV and approximately 10 m deep west of the bluff line. Coincident with the bluff line and the PF is a complex zone of high-angle faults that may project into overlying Quaternary alluvium. In the eastern part of the section, transpressive (?) faulting is interpreted as projecting upsection across several reflectors within the Quaternary CRV sediments. These eastern faults may be associated with a previously unidentified fault zone east of the PF.

Figure  12.  Interpreted SH-wave reflection profile from Tamms, Illinois. The bedrock reflection is shaded. Also shown (above the profile) are two shot gathers from the eastern part of the profile displaying shallow reflection events identified in the field and targeted for the survey.

DownLoad: Full-Size Img PowerPoint

CONCLUSIONS

This article illustrates the effective use of shallow (10–100 m deep) hammer-impact shear-wave seismic reflection data applied to neotectonic studies in unconsolidated sediments of the Mississippi Embayment. The examples show that shallow SHwave reflection data are most useful in situations where high resolution is required and supplemental geologic information (paleoseismologic trenching/ coring and borehole data) is either already available or field work is being planned. Data acquisition is lowcost, minimum-impact (environmentally), and rapid (due to typically small study areas). Processing is relatively straightforward (desktop/laptop PC and inexpensive software) and a preliminary section can be produced in near real time to aid in selecting paleoseismologic trench locations, choosing drilling sites, and interpreting geologic features in the field.