Project Topics
Get Complete Project Material File(s) Now! »
Observation network and methods
This section describes the observation system and survey methods. Sections relating to the observation system detail not only geometrical network parameters but arrays configuration (both seismic and infrasound) and error estimation for backazimuth and apparent velocity. Such estimations are critical for understanding the reasons for the mismatch between the observed and expected backazimuths. Amplitude and phase-frequency characteristics are included for all the sensors used. This data is very important for the seismic portion of the investigation. The point is that unlike microbaroms, seismic sensors have non-flat amplitude-frequency characteristics within the target frequency band. Noise levels have been compared for each array locations for the summer and winter months. These data are sufficient to testify that within a frequency band of 0.1 – 0.4 Hz, noise levels in the summer months are lower. It is evident on infrasound records where the microbarom peak is significantly weaker. The similarity of the noise energy spectrum on collocated seismic and infrasound records also indicate that both microseisms and microbaroms registered by these arrays most likely have common sources. In addition to the observation system and tools used, this section also reviews available observation data and processing tools. The introduction section 1 details microbarom survey findings using IMS infrasound subnetwork. Similar surveys were conducted for Kazakh arrays as well (Smirnov et al., 2010). It shall be noted however that a simplified source simulation approach was used in 2010 and a shorter one-year data fragment was studied. Source parameters were simulated for one winter month only, atmospheric temperature and wind profiles influence on infrasound distribution and bathymetry effect on microseism source intensity were not considered. The goal of the 2010 survey was aimed at qualitative analysis and validation of the nature of low-frequency signals registered by the Kazakhstan network. The current survey is conducted to characterize the sources and uses more precise methods for source parameters simulation. The geography for the survey was extended from the North Atlantic to global coverage. The method section details all the applied techniques including PMCC detector, source simulation, a method for consideration of atmospheric influence on infrasound propagation, and bathymetry effect on microseism source intensity. In the 2010 survey it was shown that the expected and observed microseisms backazimuths differ significantly (Smirnov et al., 2010). An attempt was made to explain these differences with heterogeneities in the Earth crust along the microseism propagation path. Source-Specific Static Corrections of the surface waves were assessed based on strong North Atlantic earthquake detection data. In this survey, this method was expanded. Findings of the whole processing tract are presented in the section conclusions (section 5), including signals detection, source simulation, and comparison of the observation data (Smirnov et al., 2018). The tract was tested with the limited data set, and validity of the selected configuration was demonstrated. The section ends with a description of the technique for quantitative assessment of the matching expected and observed signal azimuths and amplitudes.
Observation system
The Kazakh seismo-acoustic network (KNDC, 2019) operated by IGR is unique for microbarom and microseism study, as it contains a five seismic and three infrasound arrays Figure 4 (BVAR seismic array is not shown as its data were not used in this study).
Stations in the network are part of other global networks such as the IMS (CTBTO), IRIS consortium, etc. KNDC closely cooperates with the institutions responsible for these networks and leading seismic and infrasound centers such as the International Data Center (IDC, Austria) of the CTBTO, Air Force Technical Applications Center (AFTAC, USA) and Commissariat à l’énergie atomique et aux énergies alternatives (CEA, France).
Figure 4. IGR monitoring network. Yellow and red stars are seismic and infrasound arrays, respectively. Seismic and infrasound arrays are collocated at two sites. IS31 infrasound and ABKAR seismic arrays are located ~200 km apart KURIS and MKIAR have been operating since 2010 and 2016, respectively. Microbarometers MB2000 and MB2005 are used at IS31 and KURIS, and Chaparral Physics M25 microbarometers are installed at MKIAR. Figure 6 (a) and (b) shows the frequency and phase responses of the MB2000/MB2005 and Chaparral M25.
The frequency responses of the sensors are flat from 0.01 to 5.0 Hz. Analyzed together infrasound observables recorded by this network allows discriminating regional natural and anthropogenic sources (Smirnov, 2015; Smirnov et al., 2011, 2018).
The seismic network consists of Kurchatov Cross array and MKAR part of the IMS network, ABKAR and KKAR part of the (AFTAC, USA) network (Figure 7 and Table 1). Kurchatov cross array differs from the others with 20 elements arranged as a cross with an aperture of 22 km (Figure 8).
Figure 5. Infrasound arrays of the IGR monitoring network: IS31 (2 km aperture), KURIS and MKIAR (1 km aperture) Kurchatov Cross consists of CMG-3V sensors. Although the 0.1-0.3 Hz frequency band is at the edge of the sensors frequency response, they can record microseisms. The configuration of ABKAR, BVAR, KKAR and MKAR are similar with nine elements and an aperture of ~5 km. ABKAR array configuration is shown in Figure 7.
These arrays are equipped with GS-21 short period vertical sensors with flat response for frequencies above 1 Hz. Figure 6 (c) and (d) shows the frequency and phase response of GS-21 and CMG-3V.
Figure 7. Configuration of ABKAR seismic array, which includes a central point, inner and outer circles of 3 and 5 elements, respectively waves are also registered on GS-21 short period sensors. Although, in the frequency band of interest (0.1-0.3 Hz), the signal attenuation is about 30 dB, all stations detect microseisms effectively due to their large amplitude above the background noise.
A peculiarity of the network is that infrasound and seismic arrays are collocated at two sites (KURIS and Kurchatov Cross; MKIAR and MKAR) or installed relatively close to each other (IS31 and ABKAR are 220 km apart (Figure 4). With such setting, this network can be used to develop synergetic approaches to better constrain microbarom sources and evaluate propagation effects.
Array configuration and errors
It is important to take into account uncertainties in azimuth and apparent velocity estimations identified in microbarom studies. The uncertainties of the estimated wave parameters of microseisms can be large due to the relatively small aperture of the seismic arrays. Uncertainties in wave parameter estimates are calculated considering the array geometry of the above mentioned infrasound and seismic arrays following (Szuberla and Olson, 2004) (Table 1).
For the infrasound arrays, the horizontal velocity is set to 340 m/s. For the seismic arrays, the value of 3000 m/s is chosen corresponding to the average speed of the Rayleigh wave. The uncertainties for the seismic arrays are significantly higher for the body waves due to higher velocities.
Power spectral density of the noise at seismic and infrasound arrays
The aim of the noise spectral content analysis is the comparison of conditions in summer and in winter. One-day long waveform segments are selected for the analysis. The selection criteria are low wind noise well expressed in the infrasound data. The power Spectral Density (PSD) is calculated in adjacent 1-hour windows with 30 % overlap. Figure 9 shows the processing results for the seismic and infrasound arrays. The situations in winter and summer differ from each other in a similar manner at infrasound and seismic arrays. Microbarometric and microseismic peaks are 0.2 Hz and have larger amplitudes in October. At infrasound arrays, the summer peak is detected at IS31 only. The difference in spectra in both seismic and infrasound stations is expressed more distinctly in October and December than in July.
Observation network and methods
Figure 9. One-day long PSD calculated on 1-hour windows for the infrasound arrays of the Kazakh monitoring network. Low-noise one-day long record intervals studied on December, 2017 (a), and July, 2017 (b); for the seismic arrays on October, 2017 (c), and July, 2017 (d). Comparison of noise spectra at collocated KURIS and Kurchatov Cross arrays on October, 2016 (e), and July, 2017 (f)
Review of existing databases (continuous recordings, detection bulletins, seismicity catalogs)
State of the art observations of microbaroms and microseism in Kazakhstan
Observations of microbaroms and microseism were carried out in Kazakhstan (Smirnov et al., 2018). After the first investigation by Smirnov et al. (2011), the Kazakh infrasound monitoring network increased by two infrasound arrays KURIS and MKIAR.
Observation network and methods
Existing bulletins of detections
launched for continuous recordings of data from the 8-element I31KZ infrasound array near the city of Aktobe (Aktyubinsk) in Kazakhstan (Smirnov et al., 2011). The detection algorithm is based on the Progressive Multi-Channel Correlation technique (PMCC) (Cansi, 1995).
PMCC detects coherent propagating signals across several sensors of the array, and that are delayed by time-shifts consistent with an acoustic planar wavefront. The coherent wavefront can be either impulsive, or made of transient signals or continual signals of longer duration. PMCC detects and classifies both types of signal efficiently and generates a bulletin of infrasonic signals detected at the array, with the characteristics of each detection described by parameters such as the arrival time, backazimuth, apparent velocity, frequency, and amplitude. Figure 10 shows azimuthal distribution of the signals detected by I31KZ from January 1 to January 31, 2008. Figure 11 shows the azimuthal distribution versus frequency of the detected signals.
From Figure 10, it is clear that there are several preferred directions of arrivals, indicating repeating or continual infrasound sources. Two most prominent concentrations of detections are between 180˚ and 195˚, and between 290˚ and 330˚. Figure 11 indicates that at 180˚-195˚ back-azimuth range are dominated by signals with high frequencies (0.5 to 4 Hz) whereas the signals from 290°-30° are of lower frequency (below 1 Hz). A temporary infrasound array installed at Akbulak seismic array, together with satellite images, revealed the source of most of the infrasound detections from the south: gas flares from Zhanazhol gas and oil field (Smirnov, 2007). The back-azimuth estimates associated with lower amplitude signals from northwest direction are consistent with microbaroms generated in the North Atlantic Ocean.
The following concept was used for source modeling at that stage. Microbaroms were first observed by Benioff and Gutenberg (1939) who noted similarity between the low-frequency signals on an electromagnetic microbarograph and the microseisms typically observed on seismographs. They suggested that the origin of the signals were low pressure systems in the North Pacific Ocean. In 1950, Longuet-Higgins (1950) formulated the basis of modern notions about the generation mechanism for microseisms. He demonstrated that microseisms could be generated by standing waves which resulted from wave groups of approximately the same frequency travelling in opposite directions. (Kedar et al., 2008) first used this theory for modelling seismic microseisms using ocean wave numerical models.
In the area of standing water waves (SWW), pressure changes are generated on the ocean floor which do not attenuate with depth (Longuet-Higgins, 1950; Tabulevich et al., 2001). These pressure changes are manifested in low-frequency seismic signals recorded at distances up to thousands of kilometers and are referred to as storm microseisms. SWW field at the rear side of cyclone (typhoon) is huge, and its area may reach hundreds of square kilometers. There are oscillations similar to piston performing reciprocal movements. Moving up “the piston” generates microbaroms, moving down it produces microseisms at the bottom. The oscillations are coherent (co-phased).
The source mechanism theory was proved experimentally at Lamont Doherty Earth Observatory (Tabulevich, 1986). Microseisms from SWW generated by moving cyclones propagate at large distances. Seismic stations all over the world record them. For example, microseism from Atlantic cyclones are recorded not only by European stations but also by Asian stations at Tashkent and Ashkhabad, stations at western part of Kazakhstan (Longuet-Higgins, 1950), Siberian (Irkutsk and Novosibirsk) and others (Tabulevich et al., 2002). Combined analysis of meteorological, seismic and infrasound data was performed to check whether IS31 records microbaroms generated by SWW (Smirnov et al., 2011). Microbaroms and microseisms are generated when SWW have high power.$
Comparison between the observed backazimuths and microseism/microbarom energy distribution
Having this concept in mind, infrasound and seismic signals origination from the same region at the same time was investigated. The period chosen was July 1 – June 30, 2008. The detector used was PMCC between 0.07 and 0.5 Hz (Cansi, 1995). Analysis was performed for four Kazakh seismic arrays, namely Akbulak, Borovoe, Karatau and Makanchy, and by infrasound array IS31. Figure 12 shows the detections represented in a 2D histogram. There are some features in the azimuthal distribution of the signals that are common at all stations but there are also some individual peculiarities.
Results on Akbulak array have a clear trend. The array recorded signals mostly from 300-360°. Similar azimuths are typical for Borovoe array, however, in the summer time, the station records signals arriving from the south. Karatau array behaves in a similar way, and records signals originating in the east. Makanchy array recorded signals arriving from the south over a year except for the period between January and April. Makanchy recorded signals from the north-west from January to April. IS31 infrasound array recorded signal mostly from north-east, similar to Akbulak. There is only a small amount of detections from the south during the summer months. Therefore, all the stations recorded signals from the north-west.
Table of contents :
1 Introduction
1.1 Background
1.2 Analysis of historical detection bulletins at IMS stations
2 Observation network and methods
2.1 Observation system
2.1.1 Array configuration and errors
2.1.2 Power spectral density of the noise at seismic and infrasound arrays
2.1.3 Review of existing databases (continuous recordings, detection bulletins, seismicity catalogs)
2.2 History of the microbarom and microseism observations in Kazakhstan
2.2.1 Generation of microseisms and microbaroms
2.2.2 Comparison between the observed backazimuths and microseism/microbarom energy distribution
2.2.3 Ocean noise recorded at European arrays
2.3 Assessment of processing methods, propagation of seismic and infrasound waves
2.3.1 Signal detection: the PMCC method
2.3.2 Assessment of source models and their validity
2.3.3 Azimuth corrections from seismicity catalogs (ISC)
2.4 Reprocessing historical infrasound records
2.4.1 PMCC processing configuration for infrasound data
2.4.2 PMCC processing configuration for seismic data
2.5 Comparison between observations and predictions
2.5.1 Comparison for infrasound arrays
2.5.2 Comparison for seismic arrays
2.5.3 Metrics to compare observations with predictions
3 Results
3.1 Microbarom detections as recorded by infrasound subnetwork and simulation results
3.2 Microseism detections as recorded by seismic sub-network and simulation results
4 Discussions
4.1 Joint analysis of the detection results
4.1.1 Dominant direction of microbaroms
4.1.2 Common microbarom/microseism backazimuths throughout network
4.2 Spatio-temporal variability of microbarom signals
4.2.1 Sudden stratospheric warming
4.2.2 Comparison of the source location results with the IFREMER model
4.3 Use of the redefined static corrections
4.3.1 Explanation of the nature of the deviations in backazimuths for the Rayleigh waves
4.4 Joint analysis of the infrasound and seismic detections bulletins for IMS and national stations
4.5 Localization of the source region
4.6 Catalog of oceanic sources for the reconstruction of atmospheric model
4.7 Direct and inverse problem solution
4.8 Comparison between PMCC detections and effective sound speed ratio
5 Conclusions
6 Perspectives
7 References
8 Publications
