Passive noise datasets at regolith sites

The data presented in this article contain datasets of passive noise measurements at regolith sites in Adelaide, South Australia. The data were acquired using three component (3C) LE-3Dlite Lennartz seismometers with an eigenfrequency of 1 Hz. The data were acquired at eight sites across Adelaide׳s regolith in a hexagonal array layout. Four tests, each with a duration of 30 min, were conducted at different times. The ambient noise data can be used for both horizontal to vertical spectral ratio (HVSR) analysis and array analyses, which are essential to obtain the site fundamental frequency and the ellipticity of the fundamental mode Rayleigh waves at the measured site. The array analyses are useful to obtain the dispersion curves, which are needed to estimate the shear wave velocity profile.


component (3C) LE-3Dlite Lennartz seismometers with an eigenfrequency of 1 Hz Data format
Raw, filtered, analyzed Experimental factors The passive noise measurements were conducted for at least 2 h at a sampling frequency of 100 Hz Experimental features The 3C LE-3Dlite Lennartz seismometers were equipped with an analogto-digital recorder & a global positioning system (GPS) Data source location Adelaide City, South Australia Data accessibility Data are included with this article

Value of the data
The ambient noise data can be used in the development of further experiments at other regolith sites The data can be compared to other measurements for provide greater insight The HVSR curves serve as a benchmark for other researchers The data are important to evaluate the reliability of ambient vibration data analysis and for comparison of other sites with similar or divergent geophysical characteristics.

Data
The data in this article contain a series of measurements of ambient noise (microtremor) at regolith sites in Adelaide, South Australia (Fig. 1). Eight sites across Adelaide's regolith were measured in a hexagonal array layout. The acquired data contains four continuous ambient noise tests, each with a duration of 30 min. The ambient noise data, which were acquired in arrays consisting of three instruments in a triangular arrangement, are important for horizontal to vertical spectral ratio (HVSR) analyses. The HVSR method was introduced by [1] based on the work of [2]. This HVSR method was popularized by [3]. Due to its simplicity this HVSR method has been used extensively since 1989. The data from all seismometers in the hexagonal array can be used to evaluate the dispersion curves in array analyses, such as spatial autocorrelation (SPAC) analysis [4]. HVSR analyses are carried out to Fig. 1. Locality of the collected data sites [5]. obtain the site fundamental frequency and ellipticity of the fundamental mode Rayleigh waves at each of the measured sites. The use of the array analyses to evaluate the dispersion curve is motivated by the objective of analyzing Raleigh waves and excluding Love waves.

Experimental design
A hexagonal array with a radius of 50 m was used in the measurement process, as shown in Fig. 2(a). Eight ambient noise field measurements were conducted in the parklands that surround the Adelaide city area (Fig. 1) using seven sets of 3-component seismometers. These seismometers record the three orthogonal components of vibration: two horizontals (i.e. east-west and north-south) and one vertical. The data were recorded and saved to the internal memory storage within each instrument.

Equipment
All the seismometers used for the noise data acquisition are three component (3C) LE-3Dlite Lennartz seismometers with an eigenfrequency of 1 Hz, as shown in Fig. 2(b). These 3C seismometers were connected to an analog-to-digital recorder [Kelunji digital data recorder, as shown in Fig. 2(b)], a global positioning system (GPS), antenna and a battery. The digital data recorder was equipped with a memory disk to store the acquired data. A laptop computer was used for initial setting up and checking the status of the recording process.
In order to be provide a stable measurement platform at each site, the seismometers were placed onto of a 20 mm thick circular concrete slab over a generally firm to a stiff ground surface, which was previously cleared of vegetation and any pebble-sized rubble. The seismometers were leveled by adjusting the legs so as to minimize any instability during recording. For consistency with the measured horizontal vibration, all the seismometers were also oriented to magnetic North. Furthermore, to maintain the stability of the seismometer during the data acquisition, all the seismometers were protected from wind-induced vibrations by means of a plastic container and stabilized with a paving block on top, as shown in Fig. 2(c).

Equipment repeatability
Prior to data acquisition, all of the equipment sets were examined to assess their repeatability. All equipment sets were run simultaneously at the same location, with a separation distance of approximately 0.5 m from one another for at least two days. The results of this huddle test are presented in Table 1 (where f 0 is the fundamental frequency; n is the number of windows selected for the average HVSR curve; A 0 is the HVSR peak amplitude at frequency f 0 ; and i, ii, and iii are the criteria suggested by [6] to assess the reliability of the HVSR curves. The HVSR curves from the huddle tests that passed the SESAME reliability criteria, were plotted with respect to HVSR amplitude and frequency. The results are presented in Fig. 3, which demonstrates very similar curves, particularly between 0.8 and 10.0 Hz. A high dispersion is shown from a frequency of 0.8 Hz, downward to 0.25 Hz, which suggests a low noise source intensity over this frequency interval. Further investigation of this huddle test is available in [7].

Detection of industrial origin
The Geopsy [8] damping toolbox was used to detect the presence of any vibration originating from an industrial source. Data from an industrial origin is concluded if the damping is much lower than 1% and the frequency is sustained. This identification is important in the HVSR analysis. Example results are shown in Fig. 4. Further details of all such analyses are included in Appendix A of the Supplementary Data associated with this paper.

Field data
At each survey location, noise measurements were conducted for at least 2 h, at a sampling frequency of 100 Hz. Field data were filtered using a cutoff frequency of 50 Hz. Only recordings on three instruments (in a triangular arrangement) were used for the horizontal to vertical spectral ratio (HVSR) analysis and the vertical seismometer readings from all instruments were used to analyze the spatial autocorrelation (SPAC). At each of the 8 sites, four tests, with a duration of 30 min, were conducted at different times using all instruments. An example of the field data sheet is shown in Fig. 5. All the field data sheets are included in [9].

HVSR analysis
Analysis of the spectral ratio, between the Fourier amplitude spectrum of the horizontal (H) and vertical (V) components of the ambient noise, is fundamental to the horizontal to vertical spectral ratio (HVSR) method.
The selection of the windows with the most stationary wave forms is a crucial initial step in computing HVSR spectral ratios. This selection is used to exclude transient vibrations [6]. Subsequently, in each selected window, the Fourier spectra of each HVSR component are smoothed and merged by adopting the geometric mean.
The computation of HVSR is performed by taking the root mean square of the horizontal components of the Fourier amplitude spectra (F NS and F EW ) divided by the vertical component frequency spectrum (F UD ) [10], as shown below.
In the process, several parameters (window length, threshold of the short-time average/long-time average (STA/LTA) and the lengths of STA/LTA) were employed. Each selected window was then smoothed using a smoothing constant of 40 [11]. HVSR analysis was carried out using the method proposed by [11] to obtain the HVSR ellipticity curves, from which the site fundamental frequency is obtained. The results of this HVSR analysis, at all measured sites, are included in Appendix B of the Supplementary Data associated with the present paper. Prior to the HVSR analysis, as presented in this paper, several tentative window lengths (i.e. 25, 30, 35 and 40 s) were trialed to obtain as many reliable HVSR curves as possible, as suggested by [8], and the 40 s window length was found to be the optimal. Selected HVSR curves from the study areas are shown in Figs. 6 and 7.

Shear wave velocity profile
Shear wave velocity profiles were obtained by inverting the HVSR ellipticity curves using the process recommended by [8] and 20 best shear wave velocity models were extracted from the results of the inversion. The profiles are presented in Figs. [8][9][10]. From the 20 best models, arithmetic mean and median values of the shear wave velocity were calculated.