Extensive surface geophysical prospecting for seismic microzonation

Abstract

This paper presents an overview of the geophysical activities for the seismic microzonation of 138 municipalities belonging to four Italian regions (Abruzzo, Lazio, Marche and Umbria) that were severely damaged by the seismic sequence of Central Italy (August 2016–January 2017). This study is the result of a collaborative effort between research Institutions and professional geologists with the support of local Administrations and the Italian Civil Protection Department and sets an unprecedented large-scale example of geophysical investigations supporting detailed seismic microzonation studies. This manuscript presents the methodological approach adopted for the geophysical activities, including the technical protocols and procedures, the best practices, the final products and the results supporting a detailed microzonation study of III level. The first step of the study was the collection and critical review of all available geophysical and geological information for planning the new geophysical surveys (specifically their type and location), in order to assess the subsoil geometry and the seismic characterization of the areas under investigation. Integration with the newly acquired geophysical data allowed the identification of zones with homogeneous local seismic hazard as well as the reference seismo-stratigraphy for each area, defining for each geological unit the ranges of the relevant properties in seismic amplification studies: layering and thicknesses, density, P-wave and S-wave seismic velocity. We also present a few representative case studies illustrating the geophysical investigation for different geomorphological situations. These examples, together with the findings of the entire project, are discussed to point out the strength points and the criticalities, as well as the necessary requirements in the application of geophysical methods to detailed microzonation studies.

Appendices

Appendix: Synthesis of the protocol for acquisition and processing of geophysical data

In the following we present a synthesis of the protocol for acquisition and processing of geophysical data distributed by CMS in the extended form (in Italian) to the professionals involved in Microzonation studies in Central Italy. These indications integrate the “Italian Guidelines for Seismic Microzonation” (ICMS 2008).

1.1 HVSR

HVSR protocol follows and integrates the recommendations of the SESAME project (2004).

1.1.1 Acquisition

• Three-component velocimetric sensor with eigen-frequency 1–2 Hz, 0.70 damped.

Before in-field operation, it is recommended the calibration of the three-components by a clustering procedure (comparison of two or more homologue components in the same site). It is also suggested to calibrate the response curves of each component (N, E, Z) of each sensor in order to evaluate the real frequency range in which the HVSR estimation is reliable.

• Data-logger with a dynamic range ≥ 20 bit.

• Recording length ≥ 30 min.

• Sampling frequency ≥ 100 Hz.

• Horizontal sensor orientation towards North.

Note 1: In general, the recording should guarantee an analysis of about 10 min noise without transients. The recording length must, however, be extended to have almost 20 noise windows of 25–30 s length without transients. It is recommended to verify in field the acquisition of the minimum number of useful windows after the application of the STA/LTA filter, with different ratio values, for the elimination of transients. A raw field processing it also recommended to verify the presence of anomalous peaks and polarized noise.

Note 2: Each area must be sampled with at least 5 HVSR measurements, placed on a grid (regular or irregular) with distance between the nodes not exceeding 200–300 m. When anomalous peaks are observed it is recommended the repetition of the acquisition in an area of about 50 m around the site.

Note 3: The total number of measurements and their location will be decided on the basis of: the extension of the area and the geo-lithological and morphological characteristics identified in the MOPS of the I and II level microzonation studies.

Each MOPS, with extension at least 300 m, should be sampled with a measurement and, if possible, sample the bedrock.

The previous available geophysical measurements (surface and/or borehole) and seismo-stratigraphies should be controlled with new measurements. These could be used as calibration sites to extend the seismo-stratigraphy in the surrounding area and to integrate the measurement points.

1.1.2 Processing

• Selection of the time windows for spectral analysis and calculation of the HVSR. The selection may be performed manually or using a STA/LTA transient filter (suggested: short window 1 s; long window 25 s; STA/LTA = 2.5). At least 20 windows of 25–30 s.

• No filtering.

• Tapering cosine 5% without overlap of time windows.

• The quadratic mean of the horizontal components N (F) and E (F) is suggested. Other mean types are indicated if the difference observed between the components is significant, for example the geometric one.

• Frequency range 1–20 Hz with at least 200 values ​​spaced on a logarithmic scale.

• Smoothing with Konno–Ohmachi function (Konno and Ohmachi 1998) (suggested smoothing parameter 40; any suitable smoothing operator could be used; we choose this to homogenize the processing between different users).

• Analysis of HVSR directionality to evaluate the polarization of the noise. Clearly polarized HVSR must be discarded.

• Evaluation of the quality of the HVSR according to the SESAME criteria.

• Picking and classification of the HVSR peak frequency in 1–20 Hz range (0.05–1 s).

• Interpretation of fundamental resonance frequency: F_o = Vs/4 h.

• Inversion of ellipticity curves to obtain 1D Vs models. Constrained with the Vp and Vs seismic-stratigraphies derived from P and S refraction, MASW, seismic arrays and/or down-hole measurements.

1.1.3 Products

• HVSR field sheet for each measurement in text format

• Original time series in digital format according to geophysical-seismological or ASCII standard formats.

• HVSR in ASCII format giving for each frequency: average value, minimum and maximum values and/or standard deviation value.

• ASCII file of the F_o frequencies (and other clearly detectable peak frequencies i) relating to all HVSR measurements containing: site-code, coordinates and elevation,  F₀ values ​​and corresponding HVSR peak values ​​(notes on shape: Offset, Narrow Peak, Wide Peak).

• Map of the F_o (or T_o).

  The results of the HVSR measurements allow to construct the A(F_o), (or A(T_o)), map of the investigated area. The maps graphically summarize the results of the HVSR analyses in order to highlight:

  • The presence of expected resonance phenomena: classification of the peak frequency (period) of the HVSR within defined frequency classes (colour),

  • The expected impedance contrast: classification of the HVSR peak within defined amplitude classes (radius).

• Processing and interpretation report. The report must contain the description of the main steps of acquisition, processing and interpretation and instruments and software used.

Refraction in P and S wave and MASW

2.1 Acquisition

The acquisition parameters and the acquisition geometry depend on the sampling of the signal in time and space (wavefield), the maximum and minimum depth of investigation, and survey geometry.

Wavefield sampling

Wavefield sampling mainly depends on the spectrum of the source signal (its max frequency, F_max) and phase velocity of the terrain (its minimum velocity, V_min). Signals must be sampled, in the time domain, with a frequency F_s, greater than 2F_max and spatially with geophone spacing less than V_min/(2F_max). Over-sampling in time and space is always recommended to reduce error in the picking of arrival times and to better reconstruct the velocity spectrum in the F-K or F-slowness domains.

Max and min investigation depths

Refraction: maximum depth about 1/4 of the maximum length of the profile, minimum depth of the layers about ½ of the geophonic distance;

MASW: approximated maximum depth of investigation Z_max ≤ V_R/(2F_min), approximated minimum depth of the layers Z_min ≥  V_R/(2F_max); where F_min and F_max are the minimum and maximum frequency values of the signal spectrum and V_R is the average of Rayleigh phase velocity.

We suggest acquiring P wave refraction and MASW data using the same instrumentation and survey geometry.

2.2 Geometry of P wave refraction / MASW and S wave refraction surveys

• Seismic baseline with minimum 24 geophones.

• Geophone spacing not exceeding 3 m, for long refraction survey 5 m spacing should be used.

• Minimum number of energizations in P and S wave and MASW:

  • 3 internal to the base (centre, 1/4 and 3/4 of the base),

  • 2 at the ends of the base,

  • 2 external to the base with minimum offset of 1/4 of the base (also for MASW)

  • 2 external to the base with minimum offset of 1/2 of the base (also for MASW).

• 1 HVSR acquisition along the profile (suggested in the central part).

Note 1: Recording surface should not be characterized by a high roughness (elevation difference <  = ¼ of geophone spacing; local slope not exceeding 15°).

Note 2: In order to acquire redundant data for the MASW analysis, the acquisition can be integrated with external energizationsat 4–6 m from the ends of the baseline. Further, to reduce spurious effects (e.g. surface waves scattering near the source), the minimum offset should be comparable with the cross-over distance detected with P refraction survey. To avoid under and overestimates of P, S and Rayleigh phase velocities (apparent velocities) the sources should be located ensuring a sampling of the subsoil in different directions. It is suggested, for reciprocal sources, a field inspection of the velocity of the surface-wave cone and of the apparent velocity of the refracted arrivals to evaluate the presence of dipping layers.

2.3 Acquisition parameters P wave refraction and MASW

• Digital seismograph ≥ 24 channels (dynamic range ≥ 20 Bit).

• Vertical geophones with a eigen-frequency of 4.5 Hz. If it is necessary to enhance high-frequency contents of the first arrival, the P-wave refraction acquisition could be also carried out using geophones with a eigen-frequency ≥ 10–14 Hz.

• Recording length ≥ 2 s.

• Sampling frequency ≥ 1000 Hz.

• Energization P with sledge-hammer (mass ≥ 5 kg, 10 kg recommended). To improve the source-terrain coupling we suggest using a thick aluminium plate partially buried in the terrain.

• Trigger system with error ≤ 1 ms.

• Data must be recorded without filter application (filter should be applied only for field control).

Note: The data should be recorded after each energization without automatic stacking to avoid the sum of the spurious signals.

2.4 Acquisition parameters S wave refraction

All the parameters previously described with the following differences:

• Horizontal geophones with eigen-frequency ≥ 4.5 Hz (or ≥ 10 Hz for high frequency enhancement) all oriented in the perpendicular direction to the profile.

• S wave energization with a hammer impacting horizontally on a beam or plinth coupled with the ground. Energization sequence: P wave source with vertical impact; S⁺ and S⁻ wave sources with opposite horizontal impacts in the direction perpendicular to the profile.

• Registration window ≥ 2 s.

2.5 P and S wave refraction processing and interpretation

• Geometry entry and trace edit and kill.

• Spectral analysis.

• Pre-processing: filtering and picking of the first arrivals. It is recommended to perform the data picking on the unfiltered data and controlled on filtered data. Reciprocal times for reciprocal source configurations must be verified.

• 1D interpretation with intercept times method using the first arrivals of the reciprocal sources.

• 2D interpretation with delay time method or GRM using the first arrivals of the reciprocal sources.

Error/uncertainty in depth and in velocity of layer/interface must be given.

• Tomographic interpretation using all sources picking. Model hit-counts and corresponding errors/uncertainties should be calculated.

Note 1: The 1D and 2D interpretation allows the detection of genuine layer interfaces, contrarily the tomography allows the detection of lateral and vertical variations of the velocity field. The two interpretations should be integrated.

2.6 P and S wave refraction processing products

• Refraction / MASW acquisition field sheets in text format, one for each survey.

• Seismic data files stored in SEG2, SEGY or ASCII format.

• ASCII file of the first arrivals picking for each profile.

• 1D model ASCII file reporting the depth of the top of the layer, the corresponding P (and/or S) velocity intervals and the assignment of the seismo-stratigraphic unit.

• 2D model file reporting for each layer the top of the interface identified by the metric coordinate (x, z) along the profile, the P (and/or S) velocity layer intervals and the assignment of the seismo-stratigraphic unit.

• Tomographic model file with the metric coordinates (x, z) values of the velocity pixels and the corresponding velocity value. P (and/or S) velocity intervals and their assignment to the seismo-stratigraphic unit.

• Processing and interpretation report. The report must contain the description of the main steps of acquisition, processing and interpretation and the used instruments and software.

2.7 MASW processing and interpretation

• Geometry entry, trace editing and bad trace kill.

• Filters should not be applied.

• Selection of data related to off-end sources and/or recordings with offsets greater than or equal to 6–10 m. However, the recordings selection must be as that the maximum difference of offsets is greater than the maximum depth to be investigated.

• Calculation of Rayleigh velocity/slowness spectra (frequency range 5–50 Hz).

• Stacking of the velocity spectra. The coherence of the velocity spectra calculated for different sources, including the reciprocal, must be evaluated before the stacking.

• Picking of the dispersion curve relative to the fundamental mode starting from single and stacked Rayleigh velocity (or slowness) spectra. It is suggested to pick the minimum and maximum dispersion curves to assign the picking errors in the inversion.

• 1D Vs inversion of the dispersion curves: (1) free without constraints, (2) with constraints obtained from P-wave and/or S-wave analysis (or with constraints from available geo-stratigraphy).

• 1D Vs joint inversion of dispersion curve and HVSR ellipticity curve. Always carry out with constraints from the P and/or S refraction and/or MASW and or available geo-stratigraphy.

2.8 MASW processing products

• Refraction / MASW field sheet for each profile in text format.

• Seismic data in SEG2, SEGY or ASCII format.

• Phase velocity (slowness) spectra.

• Picking values of dispersion curves, and, if evaluated, the related error in ASCII format.

• 1D model file reporting for each layer the depth of the top and the corresponding S velocity intervals and seismic-stratigraphic unit attribution. This file integrates the P model file.

• Processing and interpretation report. The report must contain the description of the main steps of acquisition, processing and interpretation and the instruments and software used.

2.9 Result uncertainty

For a complete assessment of the results of the inversions and interpretation, it should appropriate to carry out an estimate of the uncertainty of the seismo-stratigraphy (Vp, Vs models) obtained for each homogeneous microzone. This uncertainty is linked both to non-uniqueness of the subsoil velocity model and to the possible local lateral variations in the microzone itself. In order to assess this uncertainty, it is suggested to perform several constrained and unconstrained inversions by collecting the models that produce similar misfits (differences between observed and calculated data). The set of these models, obtained from different geophysical measurements carried out in the same microzone, must be used to define a reference seismic-stratigraphy of the itself microzone.