The increasing number of marine engineering construction, such as offshore wind farm projects, cross-sea bridges, and submarine tunnels, has attracted public attention to its environmental impact. In particular, the sound emanating from these anthropogenic activities has been shown to have a wide range of potential effects on marine life (Nowacek et al., 2007; Kight and Swaddle, 2011; Southall et al., 2019). The impact of underwater noise generated in different construction periods of marine engineering projects on marine life cannot be ignored. To minimize the impact of underwater noise on marine life, it is essential to establish controls on the acoustic characteristics of noise sources to meet the exposure criteria for different animals. Sound exposure criteria are the sound levels, based on acoustic response thresholds, above which sounds may have negative effects on specified animals (Hawkins et al., 2020).

In the past decades, many studies have been carried out to address the impact of underwater noise on marine mammals (National Research Council [NRC], 2003; Southall et al., 2007; National Marine Fisheries Service [NMFS], 2018). There is growing concern about the effects of anthropogenic noise on fishes in recent years, and more studies are necessary to address the issue (Popper and Hawkins, 2016; Popper and Hawkins, 2019; Hawkins et al., 2020). Sound is used for communication, reproduction, the detection of prey and predators, orientation and migration, and habitat selection (Webb et al., 2008). Therefore, anything that biologically interferes with how fish live can have a negative effect on them. However, there are still substantial knowledge gaps in the potential effects of sound on some fishes, such as the large yellow croaker (Pseudosciaena crocea). The large yellow croaker, which has significant economic value, is one of the important aquaculture fish species in China. Large yellow croakers are known to produce sound, the acoustic characteristics of which have been widely studied in recent years (Ramcharitar et al., 2006; Ren et al., 2007; Zhou et al., 2022). Moreover, most croakers are sensitive to sound through their otoliths and swim bladder (Zhang et al., 2021). Wang et al. (2017) studied the noise field distribution of underwater blasting and evaluated its impact on the large yellow croaker. The results suggested that for a 155-kg charge, a juvenile yellow croaker requires a safe range of approximately 2,500 m, while young fish and adult fish require a range of 1,600 and 900 m, respectively. Lin et al. (2019) designed two laboratory experiments to study the impacts of ship noise on the growth and immunophysiological response in the juveniles of two Sciaenidae species, Larimichthys crocea and Nibea albiflora. The results showed that the physiological indices of both L. crocea and N. albiflora increased sharply within 3 h due to ship noise stimulation, but after a month of noise stimulation, the growth and immune indices decreased significantly. However, the effects of underwater noise on the species have rarely been investigated (Horodysky et al., 2008).

Pile driving is a construction method generally used to provide foundation support for buildings and structures including offshore wind turbines, bridges, harbor facilities, and offshore oil and gas production structures (Reyff, 2012). There are mainly two types of pile driving based on mechanical principle: impact and vibratory. Impact pile driving occurs during the installation in construction projects using high-energy impact hammers, which creates an intense, impulsive, and sharp sound that radiates into the surrounding environment (Amaral et al., 2020). Many studies indicated that impact pile driving noise has adverse effects on marine life, including marine mammals (Nehls et al., 2007; Kastelein et al., 2013; Leunissen and Dawson, 2018; Leunissen et al., 2019) and fish (Casper et al., 2013; Bagocius, 2015; Hawkins and Popper, 2017). Unlike impact pile driving, vibratory pile driving describes the process in which the pile is vibrated into the sediment rather than being hammered in (Popper et al., 2022). The sound produced by vibratory pile driving is nonimpulsive and continuous, which is different from impact pile driving (Dahl et al., 2015; Jiménez-Arranz et al., 2020). Vibratory pile driving has been recommended as a quieter alternative to impact pile driving in some cases. However, only a few studies have been conducted to investigate the effects of vibratory pile driving on marine life (Wang et al., 2014; Branstetter et al., 2018), which focused on cetaceans. No studies about the effects of vibratory pile driving on fish have been conducted. Assessments of the potential impacts of sound exposures are typically used to distinguish between continuous sounds and impulse sounds. Because different kinds of sounds have different attributes, they may have very different effects on animals. Assessments should consider the intensity of the sound at the moment of exposure, the duration of individual exposure events, the integration of all exposure events, and the time interval between repeated exposure events (Hawkins et al., 2020).

The acoustic characteristics of the pile driving noise may be related to the local ocean environment. Therefore, the sound features and differences between impact and vibratory pile driving noise produced during the same marine engineering and marine environment were investigated. The sound data on received levels at different sites were collected to fit noise propagation for the research area. Finally, the effects of pile driving noise on populations of the large yellow croaker are also evaluated in this paper by the field observation of the behavioral response of yellow croakers at each site.

The study was conducted within the Sandu Bay, Ningde City, Fujian Province, China. The large yellow croaker is the largest sea-cage culture fish in China, and more than 80% of large yellow croakers are produced in Ningde City (Chen et al., 2018). Measurements of underwater noise, including impact and vibratory pile driving noise, were made during the construction of the Dong-Wu-Yang cross-sea bridge in April and September 2022 (26.66°N, 119.94°E; Figure 1A ), at water depths of approximately 40–60 m. The measurement of impact pile driving noise was carried out simultaneously at six locations on two range transects (blue solid circles in Figure 1A ) during the installation of a steel casing pile (blue five-pointed star in Figure 1A ), with a diameter of 2.5 m and a length of 80 m, on 1 April 2022. The pile was driven into the seabed using a hydraulic impact hammer (IHC-800, IHC, Kinderdijk, Netherlands Figure 1B ) with an energy rating of 800 kJ. Underwater acoustic measurement for vibratory pile driving was conducted simultaneously at five locations on two range directions (red solid circles in Figure 1A ) during the installation of five steel casing piles (red five-pointed stars in Figure 1A ) from 7 to 14 September 2022. The diameter and length of five steel piles were 4.4 and 79 m, respectively. The piles were driven using a hydraulic vibratory piling hammer (YZ-800B, Yongan, Wenzhou, China) with a centrifugal force of 11,000 kN. The distances from the steel piles to the measurement locations were measured using a GNSS equipment (Global Navigation Satellite System (GNSS) N6, Sino, Guangzhou, China). To investigate the propagation attenuation of sound levels with distance for impact and vibratory pile driving noise, measurements were made at 80, 598, 664, 1,530, 3,563, and 4,573 m from the pile for impact pile driving and at 120, 717, 1,137, 1,484, and 1,933 m (averaged) for vibratory pile driving ( Figure 1A ).

All monitoring stations were equipped with the autonomous, low-power underwater acoustic recorders (USR2000, IOACAS) with hydrophones at a depth of 5 m. During the measurements, the deployment depth was recorded by a depth sensor (Duo-500, RBR) positioned 0.5 m above each hydrophone. The omnidirectional hydrophone has a flat frequency response (± 2 dB) between 20 Hz and 20 kHz. During the measurement of impact pile driving, recorders with a sampling frequency of 48 kHz were used in three stations with a distance of less than 1 km from each other, and the other three stations used hydrophones with an effective receiving sensitivity of −220 and −170 dB re 1 V/μPa. The recorders were sampled at 16 kHz and the hydrophones’ effective receiving sensitivity was −170 dB re 1 V/μPa for vibratory pile driving. Prior to measurements, all hydrophones were calibrated by the Hangzhou Institute of Applied Acoustics in Hangzhou, China. Water column sound speed measured by SVP (Minos X, AML Oceanographic) during the two measurements were 1,498 and 1,540 m/s, respectively. A portable depth sounder (SM-5, Speedtech, Great Falls, America) was used to measure the bathymetry of the study area. The average water depth at the pile position was approximately 55 m, which decreased subtly to approximately 40 m at 4,500 m to the northeast and approximately 30 m at 2,000 m to the southwest. The sediment layer in the study area consists of coarse sand and clay via sample analysis. In the same area, 10 min of ambient noise was measured when no pile driving occurred.

During the entire pile driving duration, field observation of the behavioral response of yellow croakers in a normal aquaculture cage at each site was also conducted. The size of the cage is 5 m × 5 m, with a depth of 8 m, which contains approximately 100,000 adult fishes. The sound exposure level in the cage was recorded while the behavioral response of croakers in the cage was observed. The average sound exposure level in multiple observations was estimated as the behavioral response threshold of croakers in this paper. In this paper, behavioral response is defined as the substantial change in the behavior of an animal population (the croakers, in this case) such as fleeing quickly, moving away from the sound source, and jumping out of the water.

The pile driving signals were detected and calculated by custom analysis scripts written in MATLAB R2019a. For impact pile driving, it can be characterized by using peak-to-peak sound pressure level (SPL_pp) and sound exposure level (SEL), which indicate the maximum peak-to-peak pressure of the impulse signal and the total energy for the duration of a single pulse, respectively. The waveform of vibratory pile driving appears as a continuous signal with a low SPL_pp; thus, the cumulative SEL (SEL_cum) in 1 min is calculated to characterize exposure energy. These are given by the following formulas:

where p(t) is the measured pressure signal. p _ref is the reference value of sound pressure (equal to 1 μPa) and t _ref is the reference value of time (equal to 1 s). t ₁ and t ₂ are the start and end points of time window, respectively, for a single exposure duration. The time interval is bounded by the times when the cumulative signal energy exceeds 5% of the total signal energy and ends when it reaches 95% for impulse signals (Southall et al., 2007). N is the number of signals.

The frequency spectrum of pile driving noise and ambient noise can be expressed in pressure spectral density in units of μPa²/Hz, which is computed in constant-width bands of 1 Hz.

Underwater noise from impact and vibratory pile driving was measured simultaneously at different distances during the construction of the Dong-Wu-Yang cross-sea bridge. The SPL_pps and SEL_sss of impact pile driving were measured at six positions at the range of 80–5,000 m, and the sound source levels were also estimated based on the measured values. In the same marine project, the measurements from vibratory pile driving were also taken simultaneously at five positions at the range of 120–2,000 m. The SPL_pps values of underwater noise from vibratory pile driving were lower than those from impact pile driving; thus, cumulative sound exposure source levels in 1 min were calculated by linear regression analysis. Based on the linear regressions, the average SEL_ss of impact pile driving and SEL_cum in 1 min of vibratory pile driving were predicted to be approximately 208.1 dB re 1 μPa²s at 1 m and 207.5 dB re 1 μPa²s at 1 m, respectively. The frequency spectrum calculated over a given bandwidth, generally, 1 Hz or one-third octave is also important. As different animals have different frequency responses, it is important to indicate the frequency bandwidth (Popper and Hawkins, 2019). The averaged narrow-band (resolution, 1 Hz) pressure spectral densities for impact and vibratory pile driving were different. The acoustic energy from impact and vibratory pile driving was concentrated between 100 and 1,000 Hz and below 100 Hz, respectively. The overall sound levels during impact and vibratory pile driving were higher than ambient noise levels.

The propagation properties of noise were determined by linear fitting regression to analyze the effect of underwater noise on marine animals. The propagation loss model is usually defined by N log r, where N is the spreading loss constant and r is the distance in meters. The regressive results showed that N was 20.4 (SPL_pp) and 18.5 (SEL_ss) for impact pile driving, and 19.2 (SEL_cum) for vibratory pile driving, which were consistent with spherical spreading transmission loss (20 log r). However, in the same marine construction project, the propagation attenuation coefficients of the two kinds of pile driving noise are different. The difference is reasonable because the coefficients are related to water column sound speed. During the measurement of vibratory pile driving, the sound speed is obviously higher than that of impact pile driving. For impact pile driving, although a previous study indicated that a relatively rapid energy loss with increasing distance was observed because of the strong acoustic interaction with the seafloor of Mach cone wave sequence radiating upwards and downwards (Han and Choi, 2022), the propagation attenuation in our study did not increase significantly with distance. The possible reason is that the hydrophone in measurements is close to the sea surface and far away from the seafloor. Three-dimensional (3D) effects can vastly affect acoustic propagation in a complex shallow water environment. Underwater sound wave is affected by a series of geological features and physical oceanographic processes and can produce horizontal reflection, refraction, and diffraction (Oliveira et al., 2021). Because variation in water depth and geological features in the study area is small, the 3D sound propagation effect is ignored in the present study. To improve the accuracy of sound source level prediction, an underwater sound propagation model should be selected to calculate transmission loss in the future.

Liu et al. (2014) investigated the peak sound pressure level safe threshold for the large yellow croaker through the acoustic stimulation experiment in the laboratory. However, sound exposure time and population effects were not considered in the experiment. Unlike marine mammals, it is more important to focus on population effects than individuals for fishes (Popper and Hawkins, 2016; Pirotta et al., 2018). The sound exposure level should be used to evaluate its effect on large yellow croaker populations. Based on field observation, the use of single-pulse SEL as assessment criteria for impact pile driving is suggested. The underwater noise from vibratory pile driving has a low sound pressure level; therefore, the SEL_cum over a given period of time is recommended. The accumulative period should be carefully detailed. The SEL_cum may be defined over a standard period or for the duration of an activity, or over the entire period that the animal will be exposed (Popper et al., 2014). In addition, the distribution and changes in the magnitude of sound events within that period also need to be considered. In the present study, the cumulative exposure time selected for the period with the highest amplitude is 1 min. However, choosing the cumulative time still needs to be investigated in the future when we can better understand the effects of anthropogenic noise on fishes.

Because not all anthropogenic noise can have a negative effect on fish, impact criteria must be regulated by how fishes respond to sound exposures. The effects on fishes mainly include death and injuries, physiological effects, and changes in behavior. Behavioral responses will be especially detrimental if fishes are more exposed to predators, are displaced from feeding or spawning grounds, have their migrations affected, or experience disruption of communication between individuals (Hawkins et al., 2020). However, these behavioral characteristics are difficult to observe for cage-cultured larger yellow croakers. It is more appropriate to consider the population effects. A criterion currently recommended by the National Marine Fisheries Service (NMFS) for behavioral response is 150 dB (Stadler and Woodbury, 2009); however, whether the value is a peak or root mean square (rms) level is not indicated. Through observation of the behavioral response in the field experiment, the criterion is not suitable to evaluate the effects on the larger yellow croaker. The sound exposure level should be used to evaluate its effect on the large yellow croaker.

Finally, based on sound propagation attenuation and the behavioral response thresholds, the range of behavioral response for adult large yellow croakers is calculated to be 4,798 m for impact pile driving and 1,779 m for vibratory pile driving. For noise due to underwater blasting of a 155-kg charge, adult large yellow croakers require a safe range of 900 m (Wang et al., 2017). The influence of pile driving noise on the large yellow croaker is larger than that of a small charge of underwater blasting. However, the influence of underwater blasting increased with increasing blasting charge. The accuracy of measurement and assessment results in the study is verified by simultaneous field observation of the behavior of the large yellow croaker. It is obvious that the influence range given in this paper was only used as a reference value for pile driving noise due to lack of sufficient test data. It is very difficult to set an acoustic response threshold for croakers because it is dependent on a suite of factors, such as individual differences, densities, and circumstances. As human activities in the ocean have increased, it is therefore important to assess the noise impact, including the measurements of pile driving noise levels, and investigate their propagation properties as a function of distance. The purpose of the present study is to enhance the understanding of the potential effect of pile driving on the large yellow croaker and provide reference for the conservation of croaker.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

Ethical review and approval was not required for the study of animals in accordance with the local legislation and institutional requirements. The study investigates the behavioral response of the large yellow croaker. No auditory or physical damage is caused to the large yellow croaker. Therefore, ethical approval is not required.

FN: investigation, methodology, formal analysis, and writing—original draft and review. JX and RX: data curation and acoustic data analysis. XZ, BC, and ZL: supervision, validation, and writing—review and editing. YY: methodology and writing—review and editing. All authors contributed to the article and approved the submitted version.