Standardizing Ambient Noise Data: A Review of the SESAME Project Protocols

The Site Effects Assessment Using Ambient Excitations (SESAME) project, a research initiative funded by the European Commission between 2001 and 2004, represents the most significant effort to standardize the application of ambient noise techniques in seismic site characterization. Specifically, the project focused on the Horizontal-to-Vertical Spectral Ratio (HVSR) method, often referred to as the Nakamura technique. This method measures the ratio of the horizontal components of ground motion to the vertical component to estimate the fundamental resonance frequency of soil deposits, a critical factor in earthquake engineering and geotechnical site investigations.

Before the culmination of the SESAME project, the use of microtremor measurements was widely adopted due to its cost-effectiveness and non-invasive nature, yet it lacked a rigorous mathematical and procedural framework. The 2004 final report and the subsequent user guidelines established a set of twelve specific criteria designed to ensure the reliability of the H/V spectral ratio curve and the resulting peak frequency estimates. These protocols are now the international benchmark for researchers and engineers investigating acoustic wave propagation within heterogeneous solid-state media and complex geological stratigraphies.

At a glance

• Official Title:Site Effects Assessment Using Ambient Excitations (SESAME).
• Operational Window:2001 to 2004 (funded under the European 5th Framework Programme).
• Core Objective:To provide scientific validation and standardized protocols for the H/V spectral ratio technique using ambient vibrations.
• Primary Output:A set of 3 criteria for the reliability of the H/V curve and 9 criteria for the reliability of the H/V peak.
• Methodological Focus:Passive seismic source surveys (microtremors), Rayleigh wave ellipticity, and spectral analysis.
• Technological impact:Influenced the calibration standards for three-component geophones and the development of inversion algorithms for subsurface imaging.

Background

The investigation of seismic surface waves, particularly Rayleigh and Love waves, is essential for understanding the dynamic properties of the shallow subsurface. In urban environments or regions where active seismic sources (such as explosions or hammers) are prohibited or impractical, ambient noise provides a continuous source of excitation. This noise consists of microtremors caused by natural phenomena (wind, ocean waves) and anthropogenic activities (traffic, machinery). By analyzing the spectral characteristics of these waves, geophysicists can infer the lithological characterization and elastic moduli of the soil.

In the late 20th century, the Nakamura technique gained popularity for its ability to identify the resonance frequency (f0) of soft sedimentary layers overlying a stiff basement. However, different research groups often produced disparate results from the same sites due to variations in sensor placement, recording duration, and signal processing workflows. The SESAME project was initiated to address these inconsistencies. It sought to differentiate between peaks caused by actual subsurface interfaces and those caused by transient localized noise or instrumental artifacts. Through extensive numerical simulations and experimental testing across various European geological settings, the project participants established the "SESAME guidelines," which define the boundary between reliable scientific data and noise-induced error.

The 12 Criteria for H/V Reliability

The SESAME project categorized its standards into two main groups: those ensuring the statistical reliability of the H/V curve itself and those ensuring the clarity and significance of the identified peak. Compliance with these criteria is mandatory for any survey claiming to follow international good methods in passive seismic imaging.

Reliability of the H/V Curve

To ensure that the H/V curve is representative of the site and not a product of random fluctuations, the following three conditions must be met:

1. Window Length and Frequency:The length of the individual time windows used for spectral analysis (lw) must be sufficient to capture at least 10 cycles of the fundamental frequency. Formally, f0 > 10 / lw.
2. Number of Significant Cycles:The total number of significant cycles (nc) must be greater than 200. This is calculated as the product of the number of windows and the window length, multiplied by the frequency. This ensures statistical stability in the averaging process.
3. Standard Deviation Threshold:The standard deviation of the H/V ratio amplitude must be within narrow bounds. Specifically, the standard deviation should be lower than a factor of 2 for frequencies above 0.5 Hz, or lower than a factor of 3 for lower frequencies.

Reliability of the H/V Peak

Once a curve is deemed reliable, the peak frequency (f0) must be validated using nine sub-criteria (often consolidated into six primary metrics in simplified reports) to confirm its geological origin:

• Peak Magnitude:The amplitude of the H/V peak must be greater than 2. A ratio lower than 2 suggests a lack of significant impedance contrast between the soil and the bedrock.
• Frequency Stability:The standard deviation of the peak frequency must be small (typically < 5% or < 10% depending on the frequency range).
• Peak Prominence:There should be a clear, distinct peak. SESAME recommends checking that the H/V amplitude at frequencies slightly away from the peak (e.g., 0.25f0 and 4f0) is significantly lower than the peak amplitude.
• Azimuthal Consistency:The peak frequency should remain stable regardless of the horizontal direction (azimuth) of the sensor components. If the peak shifts significantly when rotating the horizontal axes, it may indicate a 2D or 3D geological structure or localized industrial noise rather than a 1D soil layer resonance.
• Vertical Component Behavior:The vertical spectrum should not show a trough at the same frequency as the H/V peak, as this could artificially inflate the ratio without representing a true resonance of the soil column.

Myth vs. Record: Transient Noise and Wind Impacts

One of the most persistent debates in seismic surface wave studies involves the impact of external noise sources. The SESAME project meticulously documented the effects of wind and traffic, debunking several common myths while reinforcing the need for controlled field conditions.

The Wind Impact

Myth:Strong wind simply adds high-frequency noise that can be filtered out during processing.
Record:SESAME research demonstrated that wind is particularly detrimental to low-frequency seismic signatures. Wind pressure on the sensor or nearby vegetation (trees) causes infinitesimal tilting of the geophone. In three-component sensors, this tilt is projected as a large signal on the horizontal components relative to the vertical one, creating a false H/V peak at low frequencies. To mitigate this, the SESAME protocols mandate that sensors be buried or shielded with heavy covers and that measurements be avoided during high-wind conditions.

The Traffic Noise Paradox

Myth:Traffic noise is an unwanted pollutant that invalidates passive source surveys.
Record:While nearby transient traffic (a single truck passing within 10 meters) can saturate sensors and introduce non-stationary artifacts, distant traffic is often the primary source of the high-frequency ambient vibrations (above 1 Hz) required for shallow subsurface imaging. The project found that "diffuse" traffic noise is beneficial, provided that the processing algorithm uses anti-triggering windows to remove the localized, impulsive signals from the data stream. By analyzing the dispersion curves of these microtremors, researchers can effectively perform non-destructive testing of foundations and detect buried utilities.

Practical Implementation and Data Quality Checklist

For practitioners at Surface Wave Hub and similar institutions, verifying data quality requires a systematic checklist derived from the SESAME final report. This ensures that inversion algorithms for inferring density and porosity are based on high-fidelity ground-motion signatures.

The legacy of the SESAME project remains the cornerstone of modern seismic site effect studies. By providing a rigorous framework for HVSR measurements, it transformed a qualitative observation tool into a quantitative scientific instrument. These protocols allow for the precise calibration of ground-motion sensors and provide the necessary data quality for the development of sophisticated inversion algorithms, ultimately leading to more resilient infrastructure and a deeper understanding of the earth's shallow subsurface characteristics.