| Spotted seal(Phoca largha)is the only pinniped species that produceds in Chinese waters.Liaodong Bay(LB),NE China,is the southernmost geographic breeding population of this species.The population in LB is genetically distinct from other populations.The spotted seal is listed as a Grade 2 State Protection Species by China's Wild Animal Protection Law,enacted in 1988.Spotted seals are highly vocal druing the breeding season and use sounds for communication.However,there is currently no published information available on wild spotted seals underwater vocalizations.This study recorded and analyzed underwater vocalizaions from wild spotted seals in LB using passive acoustic methods.Firstly,I investigated time and freqeucny parameters and source levels of vocalizations,and the relationship between different parameters.The vocalizations were relatively brief(median 65.7 ms,interquartile range 24.2 to 157.5 ms,95%energy duration),low frequency(peak frequency<1.8 kHz)and narrow band(15?271 Hz,-3 dB bandwidth;283?728 Hz,-10 dB bandwidth).The median root-mean-square(rms)source level of spotted seal vocalizations was 150 dB rel ?Pa@lm(interquartile range 144?156 dB re 1 ?Pa@lm).These results indicated that vocalization of spotted seals in LB were similar in frequency to other seal species,but much shorter in duration.The spotted seals had similar source levels with other similar body size seals.The frequency parameters of spotted seal vocalizaitons showed a positive relationship with the bandwidth parameters but negative with time parameters.Secondly,I quantitatively classified the spotted seal underwater vocalizaitons into five major call types(81.9%of all vocalizations recorded)using multivariate statistical analyses of ten acoustic parameters;knock,growl,drum,sweep and chirp.Frequency variables(centroid freuqncey and top two peak frequencies)were the primary descriptors used to differentiate the call types.The spotted seals exhibited an extensive underwater vocal repertoire but very limited complexity.Knock,growl,drum,and sweep were pulsed sounds with different pulse repetition rate,while only chirp was a kind of tonal sound.Knocks were the most commom vocalizations produced by spotted seals under water.There were significantly difference in all sound parameters among the five call types(Kruskal-Wallis Test,p<0.05).Based on my field observation and the previous behavior studies of spotted seal in other waters,I deduced the purpose of each call type.The growls and drums maybe associated with their underwater mating behavior;sweep often heard before the seals' diving,suggesting they may use sweep to alarm other seals;chirps were likely applied by seals to maintain mother-pup associations;and spotted seals used knocks in series of repetitive knocks,which were probably generated by mals seal to set up their territories.Thirdly,I investigated how environmental variables,including captivity setting and daylight,affected spotted seal underwater vocalizations.Six acoustic parameters were selected to compare their call structures.Statistically,captive spotted seals produced underwater calls with longer duration,higher peak and center frequency,broader-3 dB bandwidth and lower source level but similar Q compared to wild spotted seals,possibly due to an adaptation to the shallow,acoustically complex habitat.These results revealed the captive seals' vocalizations were significantly different from those of wild seals,indicating that the captivity setting has pronounced effects on the seal vocalizations.As to the daylight,there was a significant difference in acoustic activity with seals producing more calls in the morning compared to the evening.Spotted seals produced more knock,growl and chirp calls during the morning but there was no difference in drum and sweep calls between morning and evening.These findings suggest that spotted seals were acoustically active in the morning.I also found that the temporal and spectral characters of morning and evening calls were varied with call types.There was no significant difference in the acoustic features(duration,peak frequency,centroid frequency,rms bandwidth and Qrms)between morning and evening chirps.Sweeps had lower peak frequencies in the evening.This characteristic would improve the seals'communication efficiency when light is limited.Finally,the frequency parameters of the growl,knock and drum calls were significantly different between the two periods,possibly because the three call types are associated with mating behaviour involving complex social interactions.Finally,I estimated the potential masking effect of temporally varying noise from pile driving on the detection of underwater knocks by spotted seal.Underwater sound generated by pile driving was recorded during the Hong Kong-Zhuhai-Macau Bridge(China)construction in 2014.Signal analysis showed the waveform of a pile driving resembles a series of short pulses,composes of many different frequencies with main frequency energy distribution from 100 Hz to 2 kHz,the maximum value occurs at about 200 Hz.The best-fit shallow water sound propagation model gave a root-mean-square source level of 208.2 dB re 1?Pa @1 m,a transmission spreading coefficient of 17.76,and an absorption coefficient of 0.015 dB/m.The value of transmission loss is in good agreement with the results calculated by the Marsh and Schulkin(M&S)semi-empirical expressions above 1 kHz,but is higher than the M&S results under 1 kHz,which may be due to the muddy seabed and turbid sea water in our study area.The noise generation mechanisms are complex,and many factors influence the noise radiated into the water column,including the water depth(exposing a different amount of the surface area of the pile),the seabed properties,and the penetration depth into the seabed by the pile,the pile dimensions,and the hammer energy.The duration of a single pile driving pulse was much longer than the mean duration of spotted seal knocks.To calculate the spectral and temporal differences between pile driving pulses received at various distances,testing was conducted with noise recorded both close to(100 m)and far from(1 km)the pile driving site.Sliding analysis windows(duration = knock mean duration,step size = 1/sampling rate)were used to measure the 1/3 octave band sound pressure levels of pile driving noise at 630 Hz(knock dominant frequency).Then I compared these results with the mean 1/3 octave band sound pressure level of knocks and background noise.The background noise were selected from the interval of two adjacent pile driving pulses.The comparison showed the knocks would be masked by the pile driving noise at 100 m;the spotted seals could not use knocks to communicate with each other at this close distance even during the pile driving pulse interval.Contrarily,the pile driving noise could not mask the knocks at 1 km,where the seals were able to detect the knocks and communicate with others in the pile driving pulse interval.This study represents the first report on underwater vocalizations recorded from spotted seals in wild.Knowledge of these is important for understanding the species behavior and for planning future acoustic surveys of its distribution and occurrence.The methods used in this study has critical reference significance to research and protect other marine mammals in the future. |
PDF Full Text Request