The Infrasound station I48TN in Tunisia is one of the most important Infrasound stations in the International Monitoring System (IMS) network with its location in the heart of the IMS infrasound network (Figure 1(a)). The objective of the IMS infrasound network in the verification regime is to detect nuclear explosion in the atmosphere and the surface of the earth where all other sources are considered as noise. The ambient noise, includes both incoherent wind noise and real coherent infrasonic waves [1] can affect the infrasound analyst work, which can be much easier if he is aware of the ambient coherent infrasound sources, in order to distinguish between the coherent noise and the sources of interest. For this reason, this study focused on one of the 60 IMS infrasound station (I48TN) to define and characterize the mines and quarries sources in the region recorded by the Tunisian station in order to contribute to the ongoing effort of infrasound experts to define and characterize the existing coherent noise which is around 85% to 90% of the detections [2] for all Infrasound IMS network stations, as most of the existing studies focused on some separate events and not for long term observations to define and characterize the coherent Infrasound sources.

I48TN is located in Latitude 35˚.80523 and Longitude 9˚.32302 at 800 m above the sea level in the north west of Tunisia. It is the only infrasound station in North Africa (Figure 1(a)) and very close to Europe. Data from I48TN is used for the production of the European Infrasound Bulletin [3].

I48TN is an infrasound array that is composed of seven sites (Figure 1(b)) and one meteorological station within the site 1. Each site is equipped with a Microbarometer MB2005 (Figure 1(c)) from 2006 to 2017 and MB3a from 2017 up to now. Both sensors were developed by the CEA, France [4], and evaluated by SANDIA laboratory [5].

The software used for data analysis in this study are provided by the International Data Centre of the Comprehensive Nuclear Test Ban Treaty Organization (CTBTO) to the National Data Centres of the state signatories within the

NDC-in-A-Box package. The DTK_GPMCC version 6.3.0 to analyse the infrasound data, DIVA version 3.4.3 [6] [7] (Copyright © 2016, Commissariat à l’energie atomique et aux énergies alternatives (CEA)) to visualize long-term detections and Geotool version 2.3.91 (Copyright © 2015, CTBTO) to analyse the seismic data. Geotool uses Generic Mapping Tools (GMT) to generate maps [8]. The NDC-in-A-Box package software is used by the National Data Centres (NDCs) and for the NDC training cycle in the International Data Centre [9]. The 1/3 octave configuration is used for the Infrasound data processing [10]. The National Data Centre in Madagascar provided the Raytracer software used to perform the raytracing. Eleven years of collected Infrasound data from the I48TN station and some specific periods of time from the seismic station KEST were used for this study. Data are available for authorized users from the state parties of the Comprehensive Nuclear Test Ban Treaty Organization. Also, the data can be available for academic purpose via a virtual Data Exploration Center after signing an agreement with the CTBTO.

The speed of sound for an ideal gas is given by:

C i d e a l = γ g R c T a (1)

where γ_g gives the ratio of the specific heats, R_c is the gas constant and T_a represents the absolute temperature. This equation applies only when the sound wave is a small perturbation with respect to the ambient pressure. For air at room temperature, the multiplication of γ_g with R_c equals 402.8 m²∙s⁻²∙K⁻¹ and so the speed of sound in air is given by:

C a i r = 402.8 T a (2)

The atmosphere influences infrasound propagation, with the composition, temperature and wind as the important controlling variables. Using equation 1 for C_ideal and taking into account this wind effect as well, [11] describe the so-called effective sound:

C e f f = γ g R c T a + n ^ x y ⋅ u ¯ (3)

The inner product between the wind speed vector u ¯ and the unit vector n ^ x y gives the contribution of the wind for infrasound waves traveling from source location x ¯ to receiver location y ¯ . This means that C_eff is positively affected by wind vectors having a component in the direction of propagation and negatively affected by winds with a component opposing the propagation direction of the infrasound waves [12]. Refraction of infrasound follows from Snell’s law and the bending of infrasound waves towards the earth’s surface depends on the combined effect of temperature and the wind. As the temperature and wind vector change when traveling upwards through the atmosphere, so does the effective sound speed.

C_eff relates to the square root of the temperature. The combined effect of this temperature effect and the contribution of the wind vector controls the refraction of infrasound waves. In general, due to the temperature increase with altitude in the stratosphere and thermosphere, these two layers have a high potential of acting as a duct for upward traveling infrasound waves [13] [14]. Attenuation of infrasound is frequency dependent. Both, the classical amplitude absorption coefficient and the rotational amplitude absorption co-efficient relate to the square of the frequency [15]. As the total amplitude coefficient is the sum of these two coefficients, the attenuation of infrasound is related to the square of the frequency.

The long-term period detection from 2006 to 2012 visualized by DIVA software (Figure 2), shows the most coherent sources azimuths in the station I48TN. Most of those sources are repetitive and from different azimuths and will be defined starting from this observation.

The observation was continued for the last four years (2016-2019) to confirm the azimuths and all parameters related to the repetitive infrasound sources in the station I48TN. The long-term observation (Figure 2 and Figure 3) shows some

repetitive infrasound sources and the influence of the zonal wind on the detections at the station. In Summertime most of the detections are Westward and in wintertime most of the detections are Eastward (Figure 3).

As shown in Figure 4, some known infrasound sources were ignored (black color), as the microbaroms from the Ocean; Mediterranean Sea and the two Volcanos (Etna and Stromboli), in order to focus on the other coherent sources. Most of the sources in red are quarries and mines observed by the IMS station IS48 in the last few years. Most of the coherent infrasound sources in Tunisia are in the South-West and some of them in the neighboring countries like Algeria (Figure 5).

The detected infrasound sources (mines and quarries) will be presented one by one; based on the Long-term observation of infrasound detections at I48TN and the Pg and Lg arrivals from the IMS seismic station (KEST) in Tunisia, situated at the Latitude: 35˚.7317 and Longitude: 9˚.346. The IMS station “KEST” is a broadband 3-component primary seismic station, equipped with a seismometer CMG-3TB [16]. The sensor is installed in a 100 m deep borehole.

Table 22 shows that the results of analysing infrasound and seismic data for the defined coherent sources are very close to the results in the Infrasound Reference Event Database source of the International Data Centre [17]. Only for the source in Mechtat ouled Saad where the origin time in the International Data Centre results doesn’t much with the IDC Pg arrival time. In our study, Pg arrival is 12:57:51 and in the IDC results is 12:57:52 with 1s difference which is acceptable but the correct time origin in our results is 12:57:31 which is acceptable as the difference between Pg arrival and the origin time is 20 seconds, but for the IDC results is 12:09:18 which cannot be correct as the difference between Pg arrival and the origin time is more than 48 minutes.

In Figures 41-43, some examples of the ray tracing of three mine explosions in Bou Kadhra, Djebel Onk and Mechtat Ouled Saad shows the low atmosphere propagation under 1km from the sources to the I48TN station.

As shown in Figure 44, there is a high speed of wind from 10 m/s to 30 m/s between 5 km to 20 km. This low altitude high speed wind and the ground is the wave guide that allows the low propagation possible. Attenuation at low altitude is less than 0.001 dB/km for 1 Hz. Thus, possibility of propagation [18] [19].

The results show the most important local coherent infrasound sources in the region (mines and Quarries) as recorded by the IMS station I48TN, which are considered as coherent noise to the Tunisian infrasound station (Table 23 and Figure 45). The Progressive Multi-Channel Correlation (PMCC) results include information on the Back azimuth (BackAz), speed and frequency of the coherent signals which typically exhibit systematic seasonal variations [20]. The influence of the zonal wind is very clear on the I48TN station detectability as the dominant detections are coming from West side of the station in Winter and the sources in the East side are dominant in summertime of the year as shown in the long term observations presented in this study.

The origin times of the events were based on the seismic arrivals at KEST station and the Back Azimuths from I48TN were very important to distinguish the right source from different sources in the same region as for the three mines in Mdhilla, Metlaoui and Redeyef.

The comparison of the study results to the International Data Centre results of the analysed events and which are included into the Infrasound Reference Event Database source of the International Data Centre (IRED) shows an arrival time error for the event detected in Mechtat ouled Saad. In our study Pg arrival is 12:57:51 and in the IDC results is 12:57:52 with 1s difference which is acceptable,

but the time origin in our results is 12:57:31 which is acceptable as the difference between Pg arrival and the origin time is 20 s while for the IDC results is 12:09:18 which cannot be correct as the difference between Pg arrival and the origin time is more than 48 minutes.

The low altitude high speed wind from 10 m/s to 30 m/s between 5 km to 20 km and the ground is the wave guide that allows the low propagation possible as shown for the three mines explosions in Bou Kadhra, Djebel Onk and Mechtat ouled Saad. Six Infrasound coherent sources defined in the South-West of the I48TN station are considered as the main ambient noise to the station from that side.

Several conclusions can be drawn from this study. This will help the infrasound analysts to distinguish between the coherent noise and the events of interest. The synergy between infrasound and seismic data is very important to distinguish the right mine or quarry source from a couple of sources in the same region, especially when there is a low seismic network coverage. However, there is a software limitation in data fusion between Infrasound and Seismic data and this needs to be addressed by software engineering and waveform experts to work more on the existing tools to address this limitation.

Six Infrasound coherent sources defined in the South-West of the I48TN station will help the Infrasound data analyst to better monitor the station and distinguish between the coherent sources and any other event of interest from the South-West region for I48TN.

The Infrasound Reference Event Database source in the International Data Centre (IRED) needs to be updated according to this study regarding the origin time of the event in Mechtat Ouled Saad. The historical Infrasound Reference Event Database needs to be revisited [21].

More studies on the other observed coherent sources in the region as microbaroms from the oceans and the Mediterranean Sea, Etna and Stromboli volcanos and airports, will help the infrasound analyst in focusing more on the events of interest and can distinguish them from coherent noise to the station, as the main goal for the International Monitoring System is to detect nuclear explosions.

The result of this study can be transformed into a script to allow the infrasound analyst to automatically categorize the predefined coherent sources to the I48TN station. Cataloging the coherent infrasound sources for each Infrasound station will improve the monitoring of the earth for any nuclear explosions in the atmosphere and the surface of the earth.

The authors are grateful to the Comprehensive Nuclear Test Ban Treaty Organization for the data and software used for this paper.

CTBTO IMS data, IDC products and software package used for this article are available to CTBTO member states. For academic purposes, Infrasound data can be requested at the CTBTO International Data Center (IDC) via the virtual Data Exploration Center.

The authors declare no conflicts of interest regarding the publication of this paper.

Agrebi, A., Ramanantsoa, A.H., Rambolamanana, G. and Rasolomanana, E.H. (2021) Characterisation of the Coherent Infrasound Sources Recorded by the Infrasound International Monitoring System Station I48TN in Tunisia (Mines & Quarries). Atmospheric and Climate Sciences, 11, 214-243. https://doi.org/10.4236/acs.2021.111014