Earthquake Reaearch in China  2019, Vol. 33 Issue (2): 186-194     DOI: 10.19743/j.cnki.0891-4176.201902010
The Seasonal Variation of Large Volume Airgun Signals in Hutubi, Xinjiang
SU Jinbo1, WANG Qiong1, ZHANG Wenxiu1, WEI Yunyun1, CHEN Hao2, WANG Haitao3
1. Earthquake Agency of Xinjiang Uygur Autonomous Region, Urumqi 830011, China;
2. First Monitoring Center of China Earthquake Administration, Tianjin 300180, China;
3. China Earthquake Networks Center, Beijing 100045, China
Abstract: In order to study the seasonal variation of large volume airgun signals in Hutubi, Xinjiang, we analyzed 2, 936 signals of airgun source excitations during 2015-2016 received by a seismograph on the bank of the excitation pool. Firstly, the RMS value of the signal amplitude and the daily average temperature were compared after linearly superimposing the signal in days, to analyze the influence of the surface ice cover on the excitation energy release of the airgun source. The result shows that the ice cover will reduce the excitation energy, and the thicker the ice cover is, the more obvious the excitation energy reduces. Secondly, the time-frequency analysis method was used to analyze the influence of the surface ice cover on the signal frequency. It is concluded that the existence of the ice cover has little effect on the signal frequency, but it will affect the intensity of the signal around 4Hz between 1-2s after excitation. The cause of these phenomena is that the ice cover affects the bubble oscillation, which in turn affects the energy conversion. The study shows that when using the cross-correlation delay method to calculate the wave velocity, the signals can be divided into two periods according to the daily average temperature:with or without ice cover on the upper surface of the excitation pool. This can help eliminate the influence of the source variation and improve the accuracy of the monitoring results.
Key words: airgun source     seasonal variation     signal energy     signal frequency

INTRODUCTION

In recent years, seismologists have introduced airgun sources, which had been widely used in offshore oil exploration, into inland use. After many experiments and attempts, a new method has been explored that uses a large volume airgun source for monitoring regional underground medium changes and media structure detection in inland areas (Wang Baoshan et al., 2016; Chen Yong et al., 2007). A large number of experiments have shown that the large volume airgun source has the advantages of being high energy with good repeatability, rich low-frequency components, energy saving and is also environmentally friendly, making it a high-quality source for studying the internal medium change of the earth (Su Jinbo et al., 2016). Since April 2011, seismologists have built a number of large volume airgun source signal launchers in Binchuan in Yunnan, Hutubi in Xinjiang and Zhangye in Gansu. Continuous excitations of large volume airgun sources were carried out to dynamically monitor the regional medium changes underground, and good observation results were obtained (Wang Bin et al., 2016; Zhang Yuansheng et al., 2016; Wei Bin et al., 2016; Su Jinbo et al., 2016).

The seismic wave generated by the airgun source excitation comes from the high-pressure gas which instantaneous releases into the water and forms approximately spherical bubbles around the airgun. The bubble oscillates in the water to attenuate and rupture. The water body oscillation caused by the bubble oscillation is coupled with the earth around the water body and forms a series of pulses (Xia Ji et al., 2016a; Su Jinbo et al., 2017). In theory, the release of high-pressure air in the water body hardly causes damage to the near-field, and has no effect on the nature of the water. Therefore, each time the airgun excited, the surrounding water conditions are consistent, which makes the airgun source signal highly repeatable (Yang Wei et al., 2013).

However, in the actual airgun tests, many factors can affect the repeatability of the airgun signal, such as the excitation pressure, the water level of the excitation pool, and the synchronism of each airgun in the array when excited. The Hutubi large volume airgun source signal transmitting station (hereinafter referred to as transmitting station) is a large volume airgun source signal launching system based on an artificial excitation pool. The water level of the excitation pool is stable. Therefore, under normal circumstances, the source repeatability is good. However, in the northern Tianshan area where the transmitting station is located, the annual temperature difference is quite large, and the daily average temperature can reach as low as -25℃ (Fig. 1). At such low temperatures, ice cover will appear on the surface of the excitation pool. In this paper, the airgun source signal received by a seismograph by the excitation pool is used to analyze and discuss whether the ice cover will affect the signal excited by the airgun source when the surface of the excitation pool is frozen in winter.

 Fig. 1 Daily average temperature from January 1, 2013 to December 31, 2017 in the Hutubi area
1 BASIC SITUATION OF THE EXPERIMENT AND DATA

The transmitting station is equipped with 6 large volume airguns with a capacity of 2, 000in3 and an inverted cylinder-shaped excitation pool. The pool has a 100m top surface diameter, 15m lower surface diameter and a depth of 15m. The airguns are placed 10m below the water surface via the floating platform (Wei Bin et al., 2016), and the energy generated by the simultaneous excitation is equivalent to a natural earthquake of magnitude 0.9 (Yang Wei et al., 2013). The experiment mode is to carry out a continuous excitation experiment every week since August, 2013, and in each experiment, the large volume airgun source of Hutubi was continuously excited about 40 times.

The transmitting station is located in the ancient river beach wasteland in the northern foot of the northern Tianshan Mountains. It has a typical temperate continental climate. The lowest daily average temperature in the year can reach -25℃. The temperature record shows that among the 1825 days from January 1, 2013 to December 31, 2017, there are 571 days with an average daily temperature below 0℃, accounting for 31.29% (Fig. 1). When the average daily temperature is below 0℃, a layer of covered ice will appear on the upper surface of the excitation pool during the night. The lower the average daily temperature, the thicker the ice cover is (Fig. 2). However, in order to improve the signal-to-noise ratio, airgun experiments are always conducted at night with less ambient noise but lower temperatures (Su Jinbo et al., 2017). Therefore, nearly one-third of the time, the airgun experiments were carried out under the condition where the surface of the excitation pool was covered with ice.

 Fig. 2 General view of the ice cover of the excitation pool

In order to study the influence of seasonal variation on airgun source experiments, this paper selected 2, 936 airgun signal waveform data records from January 1, 2015 to December 31, 2016 for analysis. The data was recorded by a broadband seismic meter erected about 20m from the bank of the excitation pool (Seismograph Guralp CMG-40T; data collector Q330s+, hereinafter referred to as reference station). This station is close to the excitation pool, therefore, it can reflect the change of the airgun source intuitively, as the influence of external factors such as medium change and winter snow cover can almost be ignored.

2 ANALYSIS OF THE IMPACT OF SEASONAL VARIATION ON AIRGUN SIGNALS

The airgun signal consists of two parts: main pulse and bubble pulse (Fig. 3). The main pulse is the first positive pressure pulse generated by the high pressure air released from airgun into the water. It has a high frequency and short duration. The size of main pulse is mainly determined by airgun capacity and working pressure. The main frequency of the large volume airgun source of Hutubi is concentrated at 2-6Hz (Wei Bin et al., 2016; Yang Wei et al., 2013). These low-frequency signals are mainly generated by a series of pulses when the high-pressure air forms bubbles, and the bubbles oscillate continuously until they rupture in the water surface (Xia Ji et al., 2016). When the upper surface of the excitation pool is covered with a layer of ice, the liquid-gas interface where the bubble collapses becomes a liquid-solid interface, and the generation of the airgun bubble and the coupling of the airgun signal and the interface may be affected by the ice surface, thereby making some changes to the airgun energy and bubble oscillation frequency.

 Fig. 3 The vertical component signal of the airgun source is linearly superposed in days The shaded parts represent the main pulse and the bubble pulse signal, respectively
2.1 The Impact of Seasonal Variation on Airgun Energy Release

Firstly, the airgun signals are cut from the continuous waveform data recorded by the reference station, according to the exact time of the airgun excitation (unless otherwise stated, the signals used in this study are the vertical component signals recorded by the seismograph). And then, the airgun signals are linearly superimposed

In days (Fig. 3). The principle of linear superposition is to directly superpose the signals that have been aligned by the excitation time, and find the average (Wu Anxu et al., 2016):

 $X(t) = \frac{1}{N}\sum\limits_{j = 1}^N {{x_j}} (t)$ (1)

Linear superposition is a simple time domain superposition method (equation(1)). Using linear superposition to find the average of daily signals can reduce the impact of experimental errors, for example, the pressure difference of each excitation, on subsequent analysis. Finally, the RMS (root mean square) value of the superposed signal amplitude is calculated. The RMS value is used to represent the energy released by the airgun signal (Yamaoka K. et al., 2014) (Fig. 4).

 Fig. 4 The comparison of temperature and RMS of the vertical component signal amplitude of the airgun source Black dot represents the amplitude RMS value of the airgun signals superposed in days, blue cross represents the daily average temperature

The Hutubi airgun excitation environment is relatively stable (Su Jinbo et al., 2017), and the signals are superposed during data processing, which largely reduces the influence of the airgun excitation pressure difference on the signal energy and frequency. In theory, the signal repeatability should be at a high level. However, as seen from Fig. 3, after eliminating the influence of these external factors, there is still some degree of fluctuation in the consistency of the Hutubi airgun source energy.

As mentioned above, due to seasonal factors, nearly 30% of the airgun excitations are carried out in the case of icing on the surface of the excitation pool. From Fig. 4, it can be seen that there is a good correlation between the amplitude RMS that represents the airgun excitation energy and the daily average temperature change in the station area, and the signal amplitude RMS value (the shaded part in Fig. 4) when daily average temperature is below 0℃ is generally significantly smaller than the amplitude RMS value when daily average temperature is above 0℃. This indicates that the ice cover on the excitation pool affects the release of the airgun source energy. As the temperature decreases, the ice cover becomes thicker, and the RMS value also decreases. This indicates that a thicker ice cover will cause a larger impact on the energy release (as shown in red dotted line in Fig. 4).

2.2 The Impact of Seasonal Variation on the Frequency of Airgun Source Signal

According to the daily average temperature, the intercepted airgun signals are classified into two categories (Fig. 3):signals with a daily average temperature lower than 0℃, and signals in other situation. In order to minimize the influence of the excitation pressure differences and water level differences at each excitation when analyzing the impact of the ice cover on the airgun signal frequency, it is necessary to linearly superpose the two types of signals to obtain the airgun source signals with or without ice cover on the surface of the excitation pool (Fig. 5). It can be seen from Fig. 5 that the signal amplitude with ice cover on the surface of the excitation pool is significantly smaller than that of the excitation pool without ice cover, which is consistent with the conclusions obtained above. It can be found that the amplitude difference of the main pulse signal A is small, indicating that the ice cover has little impact on the main pulse. However, the amplitude difference of the bubble pulse signal B is significant, and the difference becomes larger as the bubbles continue to oscillate and rise. This indicates that the ice cover on the surface of the excitation pool affects the air bubble oscillation mode, thereby affecting the airgun signal.

 Fig. 5 Comparison of the vertical component of airgun signals with or without ice cover

The time-frequency analysis of the two signals is carried out (Fig. 6). Since the main purpose is to analyze the frequency difference, the two signals are normalized respectively according to their maximum amplitudes, so that the interference of the maximum amplitude difference can be excluded. The comparison shows that the signal-frequency diagrams between 0-1s have high similarity between the two signals, and the difference gradually occurs between 1-2s. When there is no ice cover on the surface of the excitation pool, the frequency component of the signal is more abundant during the 1-2s period, but not significantly different from the other signal. Around 4Hz, the signal without ice cover is stronger than the signal with ice cover. According to the study of Xia Ji et al. (Xia Ji et al., 2016b), the airgun excitation energy can be expressed as below during bubble oscillation:

 Fig. 6 (a) Time-frequency diagram of the vertical component of the airgun signal when the surface of the excitation pool is covered with ice (b) Time-frequency diagram of the vertical component of the airgun signal when the surface of the excitation pool is not covered with ice
 $\mathit{signature}{\rm{ }} = R\left({{P_a} - {P_\infty } + \frac{{{\rho _\infty }_{{{\dot R}^2}}}}{2}} \right)$ (2)

Where R is the bubble radius; ${\dot R}$ is the bubble wall velocity; Pa is the internal pressure of bubble; P is the hydrostatic pressure; ρ is the density of water. In equation (2), Pa-P indicates the potential energy under the internal pressure of bubble and the hydrostatic pressure. ${\frac{{{\rho _\infty }_{{{\dot R}^2}}}}{2}}$ indicates the kinetic energy of the bubble wall when the bubble oscillates. When there is ice cover on the surface of the excitation pool, the bubbles generated by the airgun that oscillated to the water surface will continue to rise and do some work to overcome the pressure of the ice cover, and the kinetic energy will be partially converted into other energy such as heat. Therefore, the bubble oscillation signal with a frequency of about 4Hz between 1-2s is stronger without ice cover than with ice cover.

3 CONCLUSIONS AND DISCUSSION

The airgun source is an artificial source which is very suitable for monitoring the variation of regional underground medium. Its greatest advantage is the repeatability of the source signal, but this advantage is often affected by external factors, such as the water level of the excitation pool (Wang Bin et al., 2016), the shape of excitation water body (Hu Jiupeng et al., 2017), and the ice cover on the surface of excitation pool as mentioned in this paper. These factors are inevitable in airgun experiments. Therefore, when using the airgun signals to monitor the velocity variation of the underground medium, we need to eliminate the effects of these factors that will cause the source changes.

In theory, the reference source signal can be approximated as the source function s(t). The station recorded signal u(t) and the source function can be level deconvoluted in the frequency domain to eliminate the influence of the source change (Helmberger D. et al., 1971).

 $u(t)=s(t) * g(t)$ (3)
 $G(\omega) = \frac{{U(\omega)S*(\omega)}}{{\{ S(\omega)S*(\omega), \mathit{max}\{ S(\omega)S*(\omega)\} c\} }}$ (4)

Where * represents the convolution operator; g(t) represents the time function of the source to the station; G(ω) represents the frequency spectrum of g(t); U(ω) represents the frequency spectrum of u(t); S * (ω) represents the complex conjugate of S(ω); c represents the level factor. However, this method is not ideal for eliminating the source changes from stations with a long distance to airgun sources and with strong noise signals (Luan Yi et al., 2016). Monitoring the variation of underground mediums by airgun source is mainly realized by monitoring the velocity variation. The most commonly used method is the cross-correlation delay detection technology (Wei Yunyun et al., 2016; Liu Zifeng et al., 2015), i.e. using the cross-correlation function of the airgun signals recorded by two stations to calculate the time delay of the target phase. Since the variation of the underground medium is often very weak, the annual relative variation of wave velocity is usually only about 10-3-10-2 (Wang Baoshan et al., 2010). Therefore, a high stability of the source is required when monitoring the long-term dynamic change of underground medium.

By analyzing the impact of seasonal variations on the airgun signal energy and frequency stability, we can draw the following conclusions.

(1) With the change of season and daily average temperature, when the surface of the excitation pool is covered with ice, the energy generated by airgun excitation will be less than normal, and as the ice cover thickens, its energy release will decrease gradually. This is because the ice cover on the excitation pool affects the bubble oscillation mode and thus affects the airgun energy release.

(2) When there is ice cover on the surface of the excitation pool due to seasonal reasons, the frequency of the airgun signals will appear to be different in a certain way: in situations with no ice cover, the airgun signals between 1-2s after the airgun excitation possesses more energy than other situations, while the difference in the frequency component of the signal is not significant. This difference is mainly because when the bubbles have oscillated and risen to the ice cover, and continues to rise, it will overcome the resistance of the ice cover and do work, converting part of the kinetic energy of the bubble oscillation into other energy such as heat energy. As a result, when there is ice cover, the left energy is less, and the generated seismic wave signal is less than the signal when there is no ice cover.

In order to improve monitoring accuracy, we suggest that when using the Hutubi large volume airgun source to monitor the regional medium wave velocity variation, the signals recorded by the station shall be divided into two periods: with ice cover and without ice cover, according to the daily average temperature. And the wave velocity variations shall be calculated separately, which can to some extent reduce the impact on the measurement of wave velocity variations due to the source changes caused by seasonal variations.

ACKNOWLEDGEMENT

Thanks to the two reviewers for their valuable comments. My sincere thanks also goes to Professor Wang Baoshan from the University of Science and Technology of China and Dr. Ji Zhanbo from the Institute of Geophysics, China Earthquake Administration, for providing a lot of helps in the writing of the paper.

REFERENCES
 Chen Yong, Wang Baoshan, Ge Hongkui, Xu Ping, Zhang Wei. Proposed of transmitted seismic stations[J]. Advances in Earth Science, 2007, 22(5): 441-446 (in Chinese with English abstract). Helmberger D., Wiggins R.A. Upper mantle structure of midwestern United States[J]. Journal of Geophysical Research, 1971, 76(14): 3229-3245. DOI:10.1029/JB076i014p03229 Hu Jiupeng, Wang Baoshan, Chen Yong. The influence of water shape on land airgun triggered signals[J]. Journal of Seismological Research, 2017, 40(4): 543-549 (in Chinese with English abstract). Liu Zifeng, Su Youjin, Wang Baoshan, Wang Bin, Yang Jun, Li Xiaobin. Study on analysis method of travel time variations of seismic wave of active source in Binchuan[J]. Journal of Seismological Research, 2015, 38(4): 591-597 (in Chinese with English abstract). Luan Yi, Yang Hongfeng, Wang Baoshan. Large volume air-gun waveform data processing (Ⅰ):Binchuan, Yunnan[J]. Earthquake Research in China, 2016, 32(2): 305-318 (in Chinese with English abstract). Luo Guichun, Ge Hongkui, Wang Baoshan, Hu Ping, Mu Hongwang, Chen Yong. Fired models of air gun source[J]. Earthquake Research in China, 2007, 23(3): 225-232 (in Chinese with English abstract). Su Jinbo, Wang Qiong, Wang Haitao, Shi Yongjun, Chen Xiangjun, Feng Lei, Chen Hao, Zhang Wenxiu. Research on large volume airgun source signal reception in Hutubi, Xinjiang using seismic network data[J]. Earthquake Research in China, 2016b, 30(3): 326-332. Su Jinbo, Wang Haitao, Wang Qiong, Zhang Wenlai, Ji Zhanbo, Wei Yunyun, Chen Hao, Zhang Wenxiu. A study on the repeatability of large volum arigun seismic source in Hutubi, Xinjiang using cluster analysis method[J]. Journal of Seismological Research, 2017, 40(4): 613-618 (in Chinese with English abstract). Wang Baoshan, Yang Wei, Yuan Songyong, Guo Shijun, Ge Hongkui, Xu Ping, Chen Yong. An experimental study on the excitation of large volume airguns in a small volume body of water[J]. Journal of Geophysics and Engineering, 2010, 7(4): 388-394. DOI:10.1088/1742-2132/7/4/005 Wang Baoshan, Ge Hongkui, Wang Bin, Wang Haitao, Zhang Yuansheng, and Cai Huiteng. Practices and advances in exploring the subsurface structure and its temporal evolution with repeatable artificial sources[J]. Earthquake Research in China, 2016, 30(3): 284-297. Wang Bin, Li Xiaobin, Liu Zifeng, Yang Jun, Yang Runhai, Wang Baoshan and Yang Wei. The source and observation system of binchuan earthquake signal transmitting seismic station and its preliminary observation results[J]. Earthquake Research in China, 2016, 30(3): 316-325. Wei Bin, Su Jinbo, Wang Haitao, Zheng Liming, Wang Qiong, Zhang Wenlai, Yuan Shun, Wei Yunyun, Chen Hao. Site selection and construction of Hutubi airgun source signal transmitting seismic station and its characteristic of source[J]. Earthquake Research in China, 2016, 32(2): 222-230 (in Chinese with English abstract). Wei Yunyun, Wang Haitao, Su Jinbo, Chen Xiangjun, Wang Qiong. The preliminary study on travel time abnormal variation of reflection wave phase of air-gun in Xinjiang before two earthquakes with MS5.0[J]. Earthquake Research in China, 2016, 32(2): 270-281 (in Chinese with English abstract). Wu Anxu, Ye Beng, Li Hong, Li Yun, Zheng Yiwen, Zhang Hongyan, Li Yan. The preliminary analysis of reliability on weak signal extraction for air-gun source surveys[J]. Earthquake Research in China, 2016, 32(2): 319-330 (in Chinese with English abstract). Xia Ji, Jin Xing, Cai Huiteng and Xu Jiajun. The time-frequency characteristic of large volume air-gun source wavelet and its influencing factors[J]. Earthquake Research in China, 2016a, 32(2): 249-260 (in Chinese with English abstract). Xia Ji, Jin Xing, Cai huiteng and xu jiajun. The time-frequency characteristic of a large volume airgun source wavelet and its influencing factors[J]. Earthquake Research in China, 2016, 30(3): 364-379. Yamaoka K., Miyamachi H., Watanabe T., Kunitomo T., Michishita T., Ikuta R., Iguchi M. Active monitoring at an active volcano:amplitude-distance dependence of ACROSS at Sakurajima Volcano, Japan[J]. Earth, Planets and Space, 2014, 66: 32. DOI:10.1186/1880-5981-66-32 Yang Wei, Wang Baoshan, Ge Hongkui, Wang Weitao, Chen Yong. The active monitoring system with large volume airgun source and experiment[J]. Earthquake Research in China, 2013, 29(4): 399-410 (in Chinese with English abstract). Zhang Yuansheng, Guo Xiao, Qin Manzhong, Liu Xuzhou, Wei Congxin, Shen Xuzhang, Yan Wenhua, Zou Rui, Yin Liang, Wang Yahong, Sun Dianfeng, Feng Hongwu, Zhang Lifeng, Guo Yingxia. The construction of active source repeated monitoring in the Qilian Mountains of Gansu Province[J]. Earthquake Research in China, 2016, 32(2): 209-215 (in Chinese with English abstract).