Park’s 1999 MASW Algorithm: A major change in Dispersion Analysis

The 1999 publication of "Multichannel analysis of surface waves" by Choon B. Park, Richard D. Miller, and Jianghai Xia in the journalGeophysicsMarked a significant departure from traditional seismic site characterization techniques. This seminal work introduced the Multichannel Analysis of Surface Waves (MASW) method, a seismic survey technique designed to estimate the shear-wave velocity (Vs) profile of the shallow subsurface. By utilizing the dispersive properties of Rayleigh waves—where different frequencies propagate at different velocities depending on the material properties of the layers they traverse—the authors provided a strong framework for non-invasive geophysical investigation.

Before the development of the MASW algorithm, engineers and geophysicists primarily relied on the Spectral Analysis of Surface Waves (SASW) method, developed in the 1980s. While SASW provided a foundation for using surface wave dispersion to infer subsurface stiffness, it was limited by its reliance on only two receivers. The 1999 paper addressed these limitations by proposing a multichannel approach that utilized an array of geophones, typically 12, 24, or more, to record the wavefield simultaneously. This transition allowed for a more detailed capture of the seismic signal and provided the necessary data density to employ sophisticated signal-processing techniques that were previously unavailable for near-surface applications.

What changed

The introduction of the MASW algorithm fundamentally altered the workflow and accuracy of seismic site investigations. Several key technical shifts occurred as a result of the 1999 methodology:

Receiver Configuration:The shift from a two-receiver setup (SASW) to a multichannel array (MASW) allowed for the simultaneous recording of the entire wavefield across a specified distance, reducing the time required for data acquisition.
Wavefield Transformation:The algorithm introduced a wavefield transformation method—specifically the phase-shift method—that converts data from the time-space (t-x) domain to the frequency-phase velocity (f-c) domain. This transformation creates a visual "dispersion image" or "power spectrum."
Mode Separation:Unlike previous methods, MASW enabled the visual and mathematical separation of the fundamental mode of the Rayleigh wave from higher modes and other types of seismic energy, such as body waves and source-generated noise.
Noise Rejection:By utilizing multiple receivers, the algorithm applied spatial redundancy to filter out incoherent noise. The summation process inherent in the multichannel transform naturally attenuated signals that did not match the expected propagation characteristics of surface waves.
Redundancy and Reliability:The use of multiple geophones provided a more representative sample of the subsurface over the length of the array, making the resulting shear-wave velocity profiles more stable and less sensitive to localized soil heterogeneities.

Background

Surface wave analysis is rooted in the physical principle of dispersion. In a homogeneous half-space, Rayleigh waves travel at a constant velocity regardless of frequency. However, the Earth’s near-surface is rarely homogeneous; it is typically composed of layers with varying densities and elastic moduli. In such stratified media, Rayleigh waves of different wavelengths (and thus different frequencies) penetrate to different depths. Shorter wavelengths are influenced only by the properties of the shallowest layers, while longer wavelengths sample deeper materials. This frequency-dependent velocity is known as dispersion.

In the decades preceding the 1999 paper, the geotechnical community sought ways to use this phenomenon to measure shear-wave velocity without the high cost and invasiveness of drilling boreholes for downhole or crosshole seismic tests. The SASW method, pioneered by Nazarian and Stokoe, used the phase difference between two receivers to calculate a dispersion curve. However, SASW required highly skilled operators to interpret complex phase-wrapping diagrams and was notoriously susceptible to noise. The presence of reflected waves or traffic noise often rendered SASW data difficult to interpret, as the two-receiver setup could not distinguish between the primary surface wave and secondary interference.

Technical Impact of the MASW Algorithm

The core innovation of the Park, Miller, and Xia (1999) algorithm is the phase-shift method. This method treats each multichannel record (a "shot record") as a single entity. The process begins with a Fourier transform of the time-domain data to the frequency domain. The algorithm then applies a series of phase shifts to the data, corresponding to a range of trial phase velocities. By summing the amplitudes across all receivers for each trial velocity, the algorithm identifies the velocity that produces the highest constructive interference. This peak energy corresponds to the phase velocity of the Rayleigh wave at that specific frequency.

This approach effectively "images" the dispersion curve. In a plot of phase velocity versus frequency, the dispersion curve appears as a high-amplitude trend. This visual representation allows geophysicists to trace the curve with high confidence, even in the presence of noise that would have obscured the results in an SASW analysis. Furthermore, the 1999 paper demonstrated that by observing the entire wavefield, researchers could identify "higher modes" of propagation—waves that travel at higher velocities than the fundamental mode. Accounting for these higher modes is critical in environments where a stiff layer overlies a soft layer, a scenario that often leads to erroneous results in simpler analysis models.

Comparison of SASW and MASW Regarding Noise Rejection

The robustness of MASW in noisy environments is perhaps its most significant practical advantage. In SASW, any noise—whether from a passing vehicle, wind, or industrial machinery—recorded at one of the two geophones directly affects the phase calculation. Because there is no spatial redundancy, the algorithm cannot determine if a phase shift is due to the soil properties or a localized noise source. In contrast, the MASW algorithm uses the entire array to "beamform" the signal. Noise that does not propagate coherently across the geophone array at a consistent velocity is naturally attenuated during the summation process in the wavefield transformation.

This spatial filtering capability allowed seismic surveys to move out of quiet, rural research sites and into active construction zones and urban centers. The MASW algorithm proved that meaningful subsurface data could be extracted from a wavefield even when the signal-to-noise ratio was relatively low, provided that a sufficient number of receivers were used to sample the wavefield effectively.

Case Studies in Complex Urban Environments

Since 1999, the MASW algorithm has been applied to numerous case studies involving shear-wave velocity mapping in difficult urban terrains. One significant application is the mapping of Vs30—the average shear-wave velocity in the top 30 meters of the subsurface—which is a standard metric used in building codes for seismic site classification. In densely populated cities like Los Angeles, Tokyo, and Istanbul, MASW has been used to create high-resolution maps of soil stiffness, allowing engineers to design buildings that can better withstand ground shaking during an earthquake.

In many of these urban case studies, the presence of buried utilities, pavements, and foundations creates a highly heterogeneous medium that scatters seismic energy. Documented applications have shown that MASW can effectively ignore the "scatter" from localized objects like pipes or small voids while capturing the bulk properties of the surrounding soil matrix. For instance, in projects involving the construction of new subway tunnels, MASW has been used to detect variations in bedrock depth through thick layers of urban fill. The algorithm's ability to handle lateral variations in soil properties by using a