Review of MASW and Refraction Surveys

by Nicholas MacIntyre
Bachelors Geology, Humboldt State University
OSOP Staff Geophysicist

Abstract

The purpose of this paper is to explain the theory and methods behind refraction and MASW surveys
and to accompany the OSOP's 'Quick guide to MASW and P wave refraction field survey and post
processing using SARA's DoReMi and Geopsy Software.' In this paper we will follow a survey
performed at a nearby baseball field in Volcan, Panama. The focus of this paper is on understanding
methods and touch on the physics of MASW and refraction survey techniques.

Introduction

Near-surface geophysical surveys such as MASW and refraction are gaining popularity throughout the
world due to the fact that they are non evasive, easy to perform and accurate. These surveys produce
knowledge of sub surface structure, thickness of layers, velocities of body waves, and soil
amplification parameters like Vs30; all of which are important in earthquake engineering (Moura,
2012).

This survey was performed at a local baseball field in the city of Volcan, Panama. Volcan sits in the
shadow of Volcan Baru, a volcano responsible for the hummocky topography and local geology. The
last few major volcanic events covered the city of Volcan in Pleistocene avalanche-debris deposits
(figure 1). Avalanche-debris are a result of slope failure due to the unsuitability caused by a volcanic
events. These large volume deposits are said to cover an area of 600 square kilometers on the west side
of Baru and range in thickness form 10 to over 100 meters (Sherrod et al., 2008).

Fig. 1 Geologic map showing local geology of Chiriqui Provence, Panama. Black arrow points to area of field study.
(Modified form figure 1 Sherrod et al., 2008).
The upper geology of the baseball field consists of a firm organic soil with many embedded pebble to
coble sized pyroclastic clasts. Initial estimates of surface velocities, based on geologic interpretation,
predicted S wave velocities to be between 200 m/s to 500 m/s and P wave velocities to be between 500
m/s to 900 m/s (Tables 1 and 2).

Field survey

This field survey consisted of 12 vertical geophones positioned linearly and spaced 4 meters apart
(figure 4). Our seismic source for both the refraction and MASW surveys was a 10 pound hammer
struck on a metal shock plate. For the refraction test, only P waves were recorded. Seismic waves or
shots were initiated from both the north and south end of the line, allowing for a more detailed view of
the sub surface structure. For the MASW survey, shots were initiated from just the south end of the line
and the vertical component of Rayleigh waves were recorded. Shots were initiated at 2 and 4 meters
from the nearest geophone and later combined using SARA software to form a 24 channel system with
geophones spaced 2 meters apart. Array techniques described here for MASW are based off the F-K
method and dispersion curves are inverted using GEOPSY software.

fig. 4 Photo showing the linear array of vertical geophones at the baseball field survey sight.

Performing the tests in conjecturer allows for a more detailed and accurate interpretation of geologic
structure and soil amplification estimates. For the same amount of seismic energy, refraction can not
see to as great a depth that MASW can (Wathelet, 2005). However refraction tests are easily interpreted
giving unique results of layers such as velocities of body waves, thickness and tilt. Although MASW is
unable to interpret structure other than a 1 dimensional stratigraphic column, it is insensitive to
inversion of layers which is a limiting factor in refraction. By using results from both refraction and
MASW surveys will further strengthen velocity vs. depth models and the soil amplification Vs30
accuracy.
Refraction

Refraction surveys use the refractive properties of body waves in order to interpret wave velocity and
geologic structure. Refraction is the the phenomenon in which a wave changes its direction due to a
change in media (Lee et al., 2002). Waves refract following Snell's Law which states that the ratio
between the sine of the angle of incidence and the angle of refraction is equal to the ratio of the phase
velocities in the two media or:

If the underlying layer is denser than the previous layer, than at a certain angle the wave will be
refracted along the boundary between the two layers and will propagate with velocity V2. This angle is
known as the critical angle. Waves that travel along the boundary are again refracted back into the
upper layer and are know as head waves. The principle in refraction is that the critically refracted wave
has a faster velocity than the direct wave due to the physical properties of the second layer. At some
point the head waves will reach the geophones before the direct wave (Figure 2).

Fig. 2 Diagram showing ray paths of direct wave, refracted wave and the head waves in a two layer model. Due to the
velocity of the refracted wave, at a certain distance form the source, the head waves will arrive at the geophones before the
direct wave. V2 > V1.

In refraction, wave trains are recorded and first arrival times are used to determine velocity of the P or
S waves. This is done by measuring the time it takes for the wave to propagate across a known distance
and can be calculated by fitting a line through first arrival points on a travel time graph
(http://www.iris.edu/hq/resource/how_shallow_earth_structure_is_determined). The slope of the line
being the slowness or the inverse of velocity (figure 3). Since the velocity in a homogeneous half space
is constant, we can easily see multiple layers based on the presence of consecutive linear lines with
greater slopes. However, refraction can only detect successive layers of increasing velocity.

Taking shots starting at different ends and in the middle of the survey line allows us to see a more
detailed image of the structure including tilt of layers and possible faults. Odd and confusing results in
refraction are an indicator of a complex geologic structure. Refraction tests are useful because solutions
are obtained without the use of inversion, and therefore yield a single and unique solution. In contrast,
the analysis of the dispersion of surface waves yields many models and it is the job of the geophysicist
to pick a model that best represents their field area.

Fig. 3 Multichannel shot record of a refraction survey showing first arrival picks and the line connecting them. Two layers
are clearly shown by the presence of two distinct velocities. Thicknesses of the first layer can be calculated using the two
velocities and the distance at which the two velocity lines intercept.
(Figure modified from http://www.iris.edu/hq/resource/how_shallow_earth_structure_is_determined)

Multichannel Analysis of Surface Waves (MASW)

MASW tests use the dispersive properties of surface waves. Dispersion is the phenomenon in which the
phase and group velocities of a surface wave depends on frequency. In a layered media, the frequency
of a surface wave is related to the elastic and physical properties of the material (Lee et al., 2002).
Geophones used in MASW surveys record a wave train formed by waves of different frequencies
arriving at different times. Amplitudes recorded are a function of all frequencies generated within the
layered media. Since surface waves are generated as a result of P and Vs (vertical shear-wave) waves
interfering, once the appropriate dispersion curve is picked from the multichannel shot record of the
propagating surface waves, it is then inverted to obtain the P and Vs wave velocities as a function of
depth (Wathelet, 2005). The dispersion curve is an interpretation of the different modes or harmonics of
the surface wave as it propagates through a given media. The Vs30 parameter is calculated by taking a
weighted average of the Vs wave velocities calculated for inversion within the first 30 meters. Vs30 is
given by the equation:
Where hi is the depth and Vi is the velocity at that respected depth.

Refraction data processing

To process Refraction curve data, use Geopsy software to easily pick first arrivals of P waves (Figure
5). Four shots were initiated to obtain four refraction travel time records, two from both the north and
south ends of the line. For all records, first arrivals were picked. The records are plotted on a distance
vs. time graph in the Geopsy plug-in and the slope between adjacent picks yields the velocity of the P
wave between each pair ofgeophones. Wave velocities in a given layer should remain relatively
constant. First arrivals form a constant layer follow the slope of the linear velocity line and are
therefore easily predictable. Faster layers are shown by first arrivals that appear earlier than the
previous line predicts. Velocity of consecutive layers are given by plotting a new line though the new
first arrivals. Slower layers are never seen because they will never appear as first arrivals. The
thicknesses of a given layer is determined from its own velocity, the velocity of the underlying layer
and the distance at which the two waves cross each other.

Fig. 5 One multichannel shot record, as shown in the Geopsy plug-in, displaying first arrival picks for P wave refraction.
Channels are shown on the y axis and time on the x axis. Red numbers represent slope of line between adjacent geophones
based on one unit separation between channels. To obtain velocity, multiply the slope by distance between geophones. In
this case, 4 meters. The green line shows average slope.
In figure 5, the first arrival picks from the P wave refraction survey shows one layer with a consistent P
wave velocity of 566 m/s. No second layer was found and it is postulated, due to the length of the line
and weak energy source, that this survey could see no more than a depth of 10 meters however there is
no precise way of knowing. With the absence of a second layer, the thickness of the initial layer cannot
be calculated. Results from the refraction test, in this case velocity of the initial layer, are used as input
for the MASW survey to drive inversion to more realistic results.

MASW data processing

For the MASW survey multiple shot records were recorded. Shots were initiated at 2m and 4m from
the first geophone and latter combined to give a 24 channel record. A higher density of geophones is
recommended to increase the signal-to-noise ratio (Taipodia et al., 2012). The multichannel shot record
is uploaded into the Geopsy plug-in and a spectral graph is produced (Figure 6) based on the arrival
times of the different frequencies, their energy or amplitudes, and their apparent phase (Wathelet,
2005). The spectral graph shows the fundamental and higher modes or harmonics of the Rayleigh
waves in the layered medium. Dispersion curves picked from the spectral graphs are interpretations of
the modes. Generally, a dispersion curve representing the fundamental mode is enough for the
inversion process.

Unfortunately, the most problematic issue with MASW processing is in picking the dispersion curve
(Figure 6). Their is no clear or straight forward criteria to distinguish between the fundamental and
higher modes. Within the spectral graph, the fundamental mode is often the most dominant one
(Taipodia et al., 2012), however this is not always the case. Bad picks and the mis-identification of
modes will lead to unreasonable results. Because of this, several attempts at inversion are often
necessary to obtain a decent model.

Fig. 6 Spectral graph obtained from MASW test. The black line is the dispersion curve based on the fundamental mode.

Since Rayliegh waves are the product of the interference between P and Vs waves, the inversion of the
dispersion curve returns P and Vs wave velocities as a function of depth. Through inversion, hundreds
of models for possible solutions are produced. Inputing additional information into the inversion
process limits models to more realistic results and it is up to the surveyor to find the best fit model. In
Geopsy software, to steer a model to more appropriate results, one can input parameters such as the
Poisson ratio, geologic structure, wave velocities and soil density.

To reach an accurate model input into Geopsy software the velocity of the P wave as measured in the
refraction survey. Realistic models should yield an Vs wave velocity that is roughly half that of the P
wave. Assuming, from the refraction survey, that the first layer is a few meters thick, appropriate
models should match.

The depth of investigation for Rayliegh wave surveys is related to the longest wavelength generated.
The longest wavelength generated depends on the impact power of the source and physical properties
of the substraight. The penetration depth of Rayleigh waves is about 0.4 times the longest wavelength
(Schuler, 2008). Therefore, the depth of investigation can be estimated by using the dispersion curve.
Since wavelength is equal to velocity divided by frequency we can estimate the depth of penetration
using the equations:

D= 0.4*lambda
lambda= 0.4*Vr / f

Where D is depth of penetration, lambda is wavelength, Vr is Rayleigh wave velocity and f is

frequency. Combining the equations we get:
D= 0.4*Vr/f

By selecting the lowest frequency on our dispersion curve (figure 6) we can estimate that this survey is
accurate to about 15 meters depth. D= 0.4 * (460m/s)/(12Hz) = 15.3m.

Fig. 7After inversion of the dispersion curve, we obtain graphs of P wave velocities as a function of depth. Based on the
refraction survey, the best models is one that shows a constant layer for at least 6 meters depth.
 Fig. 8 Vs wave velocities obtained from MASW inversion. The vertical blue line shows weighted average for the Vs30
parameter.

Discussion

Refraction data shows that a single homogeneous layer is present with a velocity of 534m/s. Based on
the weight of the hammer, the surveys depth of penetration is postulated to be no more than 10 meters
and without a second layer it is impossible to know estimate the thicknesses of the upper layer. MASW
data of P wave velocities (Figure 7) shows that a second layer is present around 8 meters depth with a
velocity of around 2000m/s. The absence of this new layer on the refraction data suggests that the
surveys depth of penetration is less than 8 meters. When referencing the Vs wave velocities (Figure 8),
we can see that it agrees with a constant first layer up to around 8 meters depth. Possible explanations
of this increase at 8 meters is the presence of the water table or a denser layer. The Vs30 parameter is
calculated to be 534m/s which by NEHRP standers designate it as site class C; dense soil and soft rock
(Table 2).
 Table 1 Chart showing tested materials and their respected S wave velocities.
Designation of NEHRP (right)) Site class or EC8 (left) Ground type is based on Vs30 calculations. (Moura et al. 2012)

Table 2 Chart showing tested materials and their respected P wave velocities. Chart taken form ASTM.org.
References

ASTM:"ASTMD577700(2011)e1StandardGuideforUsingtheSeismicRefractionMethod
forSubsurfaceInvestigation."Astm.org.N.p.,n.d.Web.24Oct.2013

IRIS:"HowShallowEarthStructureIsDetermined."
http://www.iris.edu/hq/resource/how_shallow_earth_structure_is_determined.NationalScience
Foundation,n.d.Web.23Oct.2013.'

Lee,WilliamHungKan.InternationalHandbookofEarthquakeandEngineeringSeismology.
Amsterdam:Academic,2002.Print.

Moura,R.M.,Noronha,F.,(2012)Vsmeasurmentsthroughdispersiveavemethodsintheurban
environmentofPorto(NorthPortuga):CGUP,UniversiyofPorto,FacultyofSciences,Portugal

Schuler,Juerg.(2008)JointinversionofsurfacewavesandrefractedPandSwave.Masters
ScienceThesis,EidgenossischeTechnischeHochschuleZurich,SwissFederalInstituteof
TechnologyZurich.

Sherrod,D.R.,Vallance,J.W.,TapiaEspinosa,A.,andMcGeehin,J.P.,2008,VolcanBaru
eruptivehistoryandvolcanohazardsassessment:U.S.GeologicalSurveyOpenFileReport2007
1401,33p.,1plate,scale1:100,000

Taipodia,Jumrik.,Dey,Arinada.,(2012)AReviewofActiveandPassiveMASWTechniques:
EGCEG,CBRI,Roorkee

WatheletM(2005)Arrayrecordingsofambientvibrations:surfacewaveinversion.PhDthesis,
UniversitdeLige,Belgium