lable at ScienceDirect

Polar Science xxx (2016) 1e7

Contents lists avai

Polar Science

journal homepage: http: / /ees.e lsevier .com/polar/

Ice berg cracking events as identified from underwater ambient noisemeasurements in the shallow waters of Ny-Alesund, Arctic

M. Ashokan*, G. Latha, A. Thirunavukkarasu, G. Raguraman, R. VenkatesanNational Institute of Ocean Technology, Ministry of Earth Sciences, Govt. of India, Pallikaranai, Chennai, Tamil Nadu, 600100, India

a r t i c l e i n f o

Article history:Received 16 December 2015Received in revised form29 March 2016Accepted 4 April 2016Available online xxx

Keywords:KongsfjordenAmbient noiseIce bergCalvingBobbing

* Corresponding author.E-mail addresses: ashokan@niot.res.in (M.

(G. Latha), vathiru@niot.res.in (A. Thirunavu(G. Raguraman), venkat@niot.res.in (R. Venkatesan).

http://dx.doi.org/10.1016/j.polar.2016.04.0011873-9652/© 2016 Elsevier B.V. and NIPR. All rights r

Please cite this article in press as: Ashokan, Mshallow waters of Ny-Alesund, Arctic, Polar

a b s t r a c t

This paper presents the work carried out on the analysis of preliminary underwater ambient noisemeasurements in the shallow waters of Kongsfjorden fjord, Arctic in the summer season, in which the iceberg cracking noise is identified. In the summer period, the melting of ice cover is fast and hence the icebergs are free to move and float in the ocean. Underwater ambient noise has been acquired in theKongsfjorden fjord, Arctic sea on 19th July 2015 at 5 m water depth, where the ocean depth is 50 m. Dueto the tensile cracks at the surface of the sea ice by thermal expansion, ice berg calving and bobbingoccurred near the experiment site. Analysis of power spectra shows that ice berg calving noise falls in thefrequency band 100 Hze500 Hz and the ice berg bobbing noise falls in the frequency band 200 Hze400 Hz.

© 2016 Elsevier B.V. and NIPR. All rights reserved.

1. Introduction

The studies in the Arctic region have received a lot of attentionin the last two decades, since the sea level rise also depends on themelting of ice bergs (Church et al., 2013). The Arctic is an exclusiveenvironment, which is the least understood ecosystem on theearth.

This area is undergoing radical changes and global warming isexpected to cause a drastic reduction in sea-ice in the Arctic Oceanin 30e40 years (Wang and Overland, 2009; Serreze and Barry,2011; Roth et al., 2012). Acoustic observation can offer a valuableevidence on variations in the Arctic sea, together with the seasonalcirculation and behaviour of marine mammals in these waters(Klinck et al., 2012; Heard et al., 2013). Hence, study of the under-water noise in Arctic fjords, predominantly those that are sur-rounded by glaciers and ice berg, has gained importance. In order todetermine the dynamics of the ice flow, a continuous long termobservation is required (Moore and Huntington, 2008; Wenz, 1962,1972). Passive acoustic monitoring gives the opportunity to surveythe unreachable zones of Arctic regionwithout dependingmuch onthe meteorological parameters. Sea ice is a complex matter with

Ashokan), latha@niot.res.inkkarasu), ragoo@niot.res.in

eserved.

., et al., Ice berg cracking evenScience (2016), http://dx.doi

physical properties that depends on its salinity and temperature.Diachok (1976) have described three separate physical methodsnamely, iceberg calving, ice melting and freshwater discharge,which are involved in the ambient noise field. Usually the ice bergin the Arctic sea is thick (in the order of 3 m) and its thermalconductivity is very low. Icebergs create noise, when they arecracking. This noise is produced by the leakage of air, which isstored in the vacuoles of ice (Urick, 1984; Deane et al., 2014). Also,there is always a quick variation in the sea surface air temperatureand it generates a transitory temperature gradient between the airand the water. It causes the ice berg surface to crack. This crackproduces a highly impulsive underwater short burst noise and itspreads to a few meters by horizontal propagation, since verticalpropagation is attenuated more due to ice reflections (Harlandet al., 2005). Aaron Thode et al. (2010) proved that in the Arcticsea, underwater sound can spread horizontally with marginalbottom interaction that causes low transmission loss.

Due to its sea-ice cover, the Arctic Ocean has an exceptionalunderwater noise, in which the noise depends on the ice bergcracking as well as on the wind speed. Milne and Ganton (1964)have carried out an underwater ambient noise experiment in icecovered Arctic seas under a variety of situations. Marsh and Mellen(1963) revealed that the speed of sound in the Arctic sea increasesgradually and it is a function of depth. Milne et al. (1967) showedthat the cracking of ice berg depends on the sea surface air tem-perature. The cracks happening close to the surface of the ice are

ts as identified from underwater ambient noise measurements in the.org/10.1016/j.polar.2016.04.001

M. Ashokan et al. / Polar Science xxx (2016) 1e72

the effect of radiative cooling during the periods of fluctuating airtemperatures. Ganton andMilne (1965) showed that the ice surfacecracks occurring in day time during summer season produces im-pulse of sound in a low frequency band less than 1000 Hz.

The main objective of this paper is to present the identificationand analysis of iceberg calving as well as iceberg bobbing noisewhich is the least observed, from the preliminary measurements ofunderwater ambient noise in the Arctic region.

2. Experimental description and methodology

National Centre for Antarctic and Ocean Research (NCAOR),Ministry of Earth Sciences, Government of India have installed anArctic research station, namely Himadri nearer to the InternationalArctic Research base, which is located at Spitsbergen, Svalbard,Norway. National Institute of Ocean Technology (NIOT), Ministry ofEarth Sciences jointly with NCAOR, has deployed a mooring systemcalled “IndArc” in the Kongsfjorden fjord, Arctic for long termmeasurements. IndArc system consists of sensors for sub surfacemeasurements such as CTD (Conductivity, Temperature & Depth),ADCP (Acoustic Doppler Current Profiler), Current meter and PAR(Photosynthetically Active Radiation) sensor along with a hydro-phone (for ambient noise data) and data acquisition system for thetime series measurements for one year. The system was deployedon 19th July 2015. The principal objective of the acoustic moni-toring in the Arctic region is to study the ice flow dynamics and thebehaviour of marine species such as whales. Prior to deployment ofIndArc system, a preliminary survey of underwater ambient noisewas conducted at the same location for one day on 19th July 2015.The measurements and analysis described in this research paper isthe initial study and prelude towards the long term acoustic mea-surements and analysis for the noise in the Arctic. The experimentalsite is shown in Fig. 1 (Source: Google Earth). The hydrophone was

Fig. 1. Location of experiment s

Please cite this article in press as: Ashokan, M., et al., Ice berg cracking evenshallow waters of Ny-Alesund, Arctic, Polar Science (2016), http://dx.doi

placed at 5 m water depth, where the ocean depth is 50 m. ACetacean make omni direction hydrophone of frequency range8 Hze100 kHz has been used for this experiment. The preamplifiergain is 20 dB and the transducer sensitivity is �185 dB re 1 V/uPa.Noise data sets have been acquired at a sampling rate 50 kHz for aduration 60 s with 16 bit resolution. The sensor was calibrated atthe Underwater Acoustic Test Facility of NIOT, which is accreditedby National Accreditation Board for Testing and Calibration Labo-ratories (NABL) in India. The sensor was deployed from a Norwe-gian Polar Institute's research and expedition vessel, RV Lance andit was anchored till the end of the experiment. The sub surfaceparameters at the sensor depth such as underwater temperature,conductivity and salinity were also logged by using a CTD sensor.The underwater temperature was observed about 3 �C. The con-ductivity and salinity was observed about 32 milli-mho and 34 psm(practical salinity units) respectively.

The surface parameters such as wind speed and ambient airtemperature were also recorded by using hand-held thermo-anemometer. Wind speed was observed about 0.08 m/s (zero seastate) and the temperature was 9�Celsius. There was no snow fall atthe time of measurement in the experiment site and hence zeroprecipitation was observed. The environmental factors such as thecondition of ship generator (ON/OFF), echo-sounder frequency andpinging interval, man-made noise and ship propeller conditionwere noted. The ship echo-sounder (Simrad make EK60 seriesmodel) frequency was found to be 18 kHz.

3. Data processing methodology

The time series raw data has the background noise prevailing atthe location, which is due to echo sounder (self-noise), dieselgenerator noise, shipping noise and the ice berg noise. In order toanalyse the ice berg noise, the other sources need to be filtered and

ite (Source: Google earth).

ts as identified from underwater ambient noisemeasurements in the.org/10.1016/j.polar.2016.04.001

Fig. 3. Power spectrum (Frequency vs Noise level) of the time series raw data.

Fig. 4. Spectrogram (Frequency vs Time) of the time series raw data.

M. Ashokan et al. / Polar Science xxx (2016) 1e7 3

the processing methodology is given below.The time series of raw data (Fig. 2) clearly shows the ship echo-

sounder noise interference. Fig. 3 and Fig. 4 show the powerspectrum and the spectrogram of time series raw data respectively.It is clearly observed that the ship echo sounder frequency occurs inthe band 13 kHze23 kHz. At the low end of the frequency spectrum,the background noise is predominantly due to the ice movementand originates from nearby to the extended distances. Utilizing thebutter-worth filter algorithm, the frequency band higher than1 kHz was removed from the time series raw data, as it is not ofinterest. Fig. 5 shows the echo-sounder noise removed time seriesdata. Fig. 6 and Fig. 7 show the power spectrum and the spectro-gram of the echo-sounder noise filtered time series data respec-tively. The power spectrum shown in Fig. 6 exposes a small signal inthe frequency band less than 500 Hz, which is created by ice bergnoise. The original ice berg noise is not only buried with the echosounder noise but also suppressed by the environmental noise suchas diesel generator noise (a combination of diesel engine andelectric generator) and ship noise, which fall in the same frequencyband of ice-berg noise, i.e., less than 500 Hz. Ice berg calving and iceberg bobbing noise influenced with the environmental factors areclearly seen in the spectrogram, i.e., Fig. 8 (enlarged portion till1000 Hz of Fig. 7). Hence it is necessary to remove these noise foranalysing the ice berg calving and ice berg bobbing noise. Since thebroadband frequencies are difficult to differentiate, time seriespattern recognition technique was utilized for filtering the noise.Initially the echo sounder noise filtered time series data wassegmented into hundreds of samples and the nature of acousticsignals pertaining to the diesel generator noise and ship noise wasexamined (Spiegel et al., 2011; Matsumoto et al., 2014).

The examined noise was isolated from the time series data. Thetime series pattern matching with the isolated noise was thenremoved from the entire echo sound filtered time series data. Fig. 9shows the time series of (a) original raw data and (b) final data afterremoving echo sounder noise and environmental noise, whichclearly shows the information about ice berg calving noise and iceberg bobbing noise.

4. Results and discussions

It is understood from Fig. 9b that the noise having burstspersistent for a few milli-seconds was generated by the rubbingand bouncing of ice masses. It also shows that the ice berg cracknoise is horizontally propagating and hence there is no harmonics

Fig. 2. Time series (Time vs Noise level) of 60 s of original raw data a

Please cite this article in press as: Ashokan, M., et al., Ice berg cracking evenshallow waters of Ny-Alesund, Arctic, Polar Science (2016), http://dx.doi

observed. A sequence of approximately 8 numbers of crack eventsare observed in the time series (Fig. 9b) in which three are of highintensity and five are of low intensity noise. The stronger burst atthe 38th second is due to the ice berg calving noise. It is observedthat the entire ice berg calving noise falls in the frequency band100e500 Hz and ice berg bobbing noise in the band 200e400 Hz.At the experiment site, during zero sea state (calm condition on thesea surface, Knudsen et al., 1948), the ambient noise level increasedby 20e30 dB from normal values in the same frequency band.Since, when an ice movement happens, the noise level increases to20 dB or above, which is greater than the Knudsen's zero sea state(Knudsen et al., 1948; Dahl et al., 2007; Kahl et al., 1992). The noiseis made by separate cracks and the noise amplitude peaks arespecifically non-Gaussian (Calderon, 1964). This is clearly observedfrom Fig. 9b showing time series filtered data. It is witnessed byexamining the noise data sets through goodness-of-fit tests such as,Jarque-Bera test (jbtest), Lillieforst test (lillietest) and One-sampleKolmogorov-Smirnov test (kstest). These are utilized from theinferential statistical analysis methods (Brockett et al., 1987;Keenan and Merriam, 1991). Fig. 10 and Fig. 11 show the powerspectrum and the spectrogram of the time series filtered data. The

cquired on 19th July 2015 at 12:54:54 (HH:MM:SS) UTC± 00:00.

ts as identified from underwater ambient noise measurements in the.org/10.1016/j.polar.2016.04.001

Fig. 5. Time series (Time vs Noise level) of echo sounder noise filtered data.

Fig. 6. Power spectrum (Frequency vs Noise level) of the echo sounder noise filteredtime series data.

Fig. 7. Spectrogram (Frequency vs Time) of the echo sounder noise filtered time seriesdata.

Fig. 8. Spectrogram (Frequency vs Time) of the echo so

M. Ashokan et al. / Polar Science xxx (2016) 1e74

Please cite this article in press as: Ashokan, M., et al., Ice berg cracking evenshallow waters of Ny-Alesund, Arctic, Polar Science (2016), http://dx.doi

sea water temperature and salinity were found to be 3 �C and 34psm respectively. This may change the stratification and the rate ofice berg melting, because of the transfer of heat from sea water toice berg. During this experiment, visual observations of the icecracking on the surface of ice-berg were made. Cracking noise wasaudible to the unassisted ear. The noise signature considerablyvaried in between bangs and crashes with their echoes tending toirregular cracking. Mellen and Marsh (1963) found that the lowfrequency sound signatures were spawned by the lengthier cracksthat cause the ice berg bobbing buoyant force, which is observed inthe current data set too as shown in Fig. 11 (spectrogram of ice bergcalving and ice berg bobbing noise).

The spectral noise level grasped an extensive maximum in thefrequency band 100e500 Hz (Fig. 11). It is understood from thespectrogram of the time series chopped data (Fig. 12) that icecracking noise occurs in the order of milli seconds (Garrison et al.,1983).

5. Conclusions

This research work reveals that in the Arctic summer season,there is an increase in the underwater ambient noise with most of

under noise filtered time series data, till 1000 Hz.

ts as identified from underwater ambient noisemeasurements in the.org/10.1016/j.polar.2016.04.001

Fig. 9. (a) Time series (Time vs Noise level) of original raw data (b) Time series (Time vs Noise level) of data after removing echo sounder noise and environmental noise.

Fig. 10. Power spectrum (Frequency vs Noise level) of the environmental noise filtered time series data.

M. Ashokan et al. / Polar Science xxx (2016) 1e7 5

Please cite this article in press as: Ashokan, M., et al., Ice berg cracking events as identified from underwater ambient noise measurements in theshallow waters of Ny-Alesund, Arctic, Polar Science (2016), http://dx.doi.org/10.1016/j.polar.2016.04.001

Fig. 11. Spectrogram (Frequency vs Time) of the environmental noise filtered time series data.

Fig. 12. Spectrogram (Frequency vs Time) of the environmental noise filtered time series data, till 1 s

M. Ashokan et al. / Polar Science xxx (2016) 1e76

the energy being generated at frequencies less than 500 Hz. Thenoise sources are at or nearby the surface of the ice berg and aremainly due to the ice berg calving and bobbing, which are causedby the thermal expansion in the ice. So far, field data is inadequateto analyse the association between the sea surface air temperatureand the time rate of incidence of ice cracking events under variousseasons. The hydrophone deployed for a longer period at the site formonitoring the underwater noise field would enable understand-ing of the ice flow dynamics.

Acknowledgements

The authors gratefully acknowledge the Director, NationalInstitute of Ocean Technology for his encouragement in carryingout this work. The authors express immense thanks to the fieldteam of Ocean Observation System (OOS), NIOT in measuring theocean ambient noise in the Arctic region. Special thanks to NCAORteam and crew of Norwegian Polar Institute's research and expe-dition vessel, RV Lance for facilitating data collection at the Arctic.Also the authors are thankful to Dr.M.C.Sanjana &Mrs.A. Malarkodifor their suggestions in this work.

Please cite this article in press as: Ashokan, M., et al., Ice berg cracking evenshallow waters of Ny-Alesund, Arctic, Polar Science (2016), http://dx.doi

References

Brockett, Patrick L., Hinich, Melvin, Wilson, Gary R., 1987. Nonlinear and non-Gaussian ocean noise. J. Acoust. Soc. Am. 82 (4), 1386e1394.

Calderon, M.A., 1964. Probability density analysis of ocean ambient and ship noise.Res. Dev. Rep, 1248, California.

Church, J.A., Clark, P.U., Cazenave, A., Gregory, J.M., Jevrejeva, S., Levermann, A.,Merrifield, M.A., Milne, G.A., Nerem, R.S., Nunn, P.D., Payne, A.J., Pfeffer, W.T.,Stammer, D., Unnikrishnan, A.S., 2013. In: Stocker, T.F., Qin, D., Plattner, G.K.,Tignor, M., Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P.M.(Eds.), Sea Level Change in Climate Change 2013: the Physical Science Basis.Contribution of Working Group I to the Fifth Assessment Report of the Inter-governmental Panel on Climate Change. Camb. Univ. Pres., United Kingdom andNewyork.

Dahl, Peter H., Miller, James H., Cato, Douglas H., Andrew, Rex K., 2007. Underwaterambient noise. Acoust. Tod 23e33.

Deane, Grant B., Glowacki, Oskar, Tegowski, Jaroslaw, Moskalik, Mateusz,Blondel, Philippe, 2014. Directionality of the ambient noise field in an Arcticglacial bay. J. Acoust. Soc. Am. Exp. Let 136 (5), 350e356.

Diachok, O.I., 1976. Effects of sea-ice ridges on sound propagation in the ArcticOcean. J. Acoust. Soc. Am. 59 (5), 1110e1120.

Ganton, J.H., Milne, A.R., 1965. Temperature and wind-dependent ambient noiseunder midwinter Pack ice. J. Acoust. Soc. Am. 38 (3), 406e411.

Garrison, G.R., Francois, R.E., Early, E.W., Wen, T., 1983. Sound absorption mea-surements at 10-650 kHz in arctic waters. J. Acoust. Soc. Am. 73 (2), 492e501.

Harland, E.J., Jones, S.A.S., Clarke, T., 2005. SEA 6 Technical Report: UnderwaterAmbient Noise. Qinetiq, Hampshire, UK.

Heard, Garry J., Pelavas, Nicos, Pecknold, Sean P., Lucas, Carmen, Martin, Bruce,2013. Arctic ambient noise measurements in support of the Northern Watch

ts as identified from underwater ambient noisemeasurements in the.org/10.1016/j.polar.2016.04.001

M. Ashokan et al. / Polar Science xxx (2016) 1e7 7

Project. J. Acoust. Soc. Am. Proc. Meet. Acoust 19, 1e9.Kahl, Jonathan D., Serreze, Mark C., Shiotani, Spencer, Skony, Suzanne M.,

Schnell, Russell C., 1992. In situ meteorological sounding archives for arcticstudies. Bull. Am. Met. Soc. 73 (11), 1824e1830.

Keenan, Ruth Eta, Merriam, Lynn Renee Lineback, 1991. Arctic abyssal Tphases:coupling seismic energy to the ocean sound channel via under-ice scattering.J. Acoust. Soc. Am. 89 (3), 1128e1133.

Klinck, Holger, Nieukirk, Sharon L., Mellinger, David K., Klinck, Karolin,Matsumoto, Haruyoshi, Dziak, Robert P., 2012. Seasonal presence of cetaceansand ambient noise levels in polar waters of the North Atlantic. J. Acoust. Soc.Am. Exp. Let 132 (3), 176e181.

Knudsen, V.O., Alford, R.S., Emling, J.W., 1948. Underwater ambient noise. J. Mar.Res. 7, 410e429.

Marsh, H.W., Mellen, R.H., 1963. Underwater sound propagation in the Arctic Ocean.J. Acoust. Soc. Am. 35 (4), 552e563.

Matsumoto, H., Bohnenstiehl, DelWayne R., Tournadre, Jean, Dziak, Robert P.,Haxel, Joseph H., Lau, T.K.A., Fowler, Matt, Salo, Sigrid A., 2014. Antarctic ice-bergs: a significant natural ocean sound source in the Southern Hemisphere.Geochem. Geophy. Geosys 15 (8), 3448e3458.

Mellen, R.H., Marsh, H.W., 1963. Underwater sound reverberation in the ArcticOcean. J. Acoust. Soc. Am. 35 (10), 1645e1648.

Milne, A.R., Ganton, J.H., 1964. Ambient noise under arctic sea ice. J. Acoust. Soc. Am.36 (5), 855e863.

Milne, A.R., Ganton, J.H., McMillin, D.J., 1967. Ambient noise under sea ice and

Please cite this article in press as: Ashokan, M., et al., Ice berg cracking evenshallow waters of Ny-Alesund, Arctic, Polar Science (2016), http://dx.doi

further measurements of wind and temperature dependence. J. Acoust. Soc.Am. 41 (2), 525e528.

Moore, Sue E., Huntington, Henry P., 2008. Arctic marine mammals and climatechange: impacts and Resilience. Eco. App 18 (2), 157e165.

Roth, Ethan H., Hildebrand, John A., Wiggins, Sean M., 2012. Underwater ambientnoise on the Chukchi Sea continental slope from 2006e2009. J. Acoust. Soc. Am.131 (1), 104e110.

Serreze, Mark C., Barry, Roger G., 2011. Processes and impacts of Arctic amplifica-tion: a research synthesis. Glob. Plan. Chan 77, 85e96.

Spiegel, Stephan, Gaebler, Julia, Lommatzsch, Andreas, De Luca, Ernesto,Albayrak, Sahin, 2011. San Diego. In: Pattern Recognition and Classification forMultivariate Time Series. Proc. Fifth Inter. Workshop on Knowledge Discoveryfrom Sensor Data.

Thode, Aaron, Kim, Katherine H., Greene Jr., Charles R., Roth, Ethan, 2010. Longrange transmission loss of broadband seismic pulses in the Arctic under ice-freeconditions. J. Acoust. Soc. Am. Exp. Let 128 (4), 181e187.

Urick, R.J., 1984. Ambient Noise in the Sea. Undersea warfare technology office,Washington DC.

Wang, Muyin, Overland, James E., 2009. A sea ice free summer Arctic within 30years? Geophy. Res. Let 36 (7), 1e5.

Wenz, Gordon M., 1962. Acoustic ambient noise in the ocean: spectra and sources.J. Acoust. Soc. Am. 34 (12), 1936e1956.

Wenz, Gordon M., 1972. Review of underwater acoustics research: noise. J. Acoust.Soc. Am. 51 (3), 1010e1024.

ts as identified from underwater ambient noise measurements in the.org/10.1016/j.polar.2016.04.001