UNDERWATER SOUND RECEIVED FROM SOME DEFENCE ACTIVITIES IN SHALLOW OCEAN REGIONS Adrian D. Jones(1) and Paul A. Clarke() (1) Program Office, DSTO Edinburgh, SA 5111, Australia () Maritime Oerations Division, DSTO Edinburgh, SA 5111, Australia Abstract The Environmental Protection and Biodiversity Conservation (EPBC) Act 1999 became effective on 16 July. This Act laces requirements on the Deartment of Defence in regard to actions hich are likely to have an imact on the environment anyhere in the orld. As a consequence, it is essential for Defence to be aare of the environmental imlications of its activities. In the area of Defence maritime oerations, relevant issues include the radiation of acoustic energy, articularly in regard to sonar systems, and the acoustic signal levels hich are exected to be incident uon marine fauna. This aer revies some of the rogress in relevant studies carried out by DSTO, ith articular emhasis made on the signals received in shallo ocean areas as a result of detonations of small exlosives knon as SUS charges. Data dislayed in this aer, for a articular shallo troical ater location north of Australia, suggest that received eak levels, in articular, may be substantially less than ublished eak shock theory suggests. Based on the circumstances of these measurements, reasons for the discreancies beteen measured and redicted eak levels are suggested. Nomenclature IL received level, db re 1 Pa Hz (sectrum) EL equivalent energy level db re 1 Pa s Hz P t eak ressure amlitude of SUS ulse, Pa time constant for SUS aveform, seconds Introduction There is strong interest ithin the Australian Defence Force (ADF), and ithin defence forces internationally, in their role in the elfare of the maritime environment. In articular, it is a desire of the ADF that it has the caability to conduct its maritime oerations and maintain its related equiment in an environmentally resonsible manner both ithin Australian ocean aters and orldide. For this reason, the Directorate of Environmental Steardshi has a requirement that relevant henomena are investigated and essential rinciles are established. Relevant issues include the radiation of acoustic energy, articularly in regard to sonar detection systems. DSTO is roviding suort to the ADF via relevant scientific advice. This aer outlines rogress in a study of the extent to hich Defence maritime activities insonify the underater ocean environment and overvies techniques used to quantify the levels of underater sound generated. In articular, this aer shos some recent rogress in evaluating the insonification caused in shallo ocean aters in the Australian region by the detonation of small underater exlosives knon as Signals Underater Sound (SUS). The latter ork is artly in resonse to the estimations of high eak ressure levels for SUS resented ithin a reort sonsored by Defence [1]. Signal and Noise Transmission and Data Processing Received level for signal and noise sources For sound sources located underater, the Source Level (SL) is a measure of the sound ressure level corresonding ith the acoustic intensity signature, on the axis of maximum outut, extraolated back to a osition 1 m from its acoustic centre []. For brevity, the term intensity level is used herein, in lace of sound ressure level corresonding ith intensity. For any sound source tye, the incident, or received level (IL) is herein defined as the acoustic intensity received at a oint of interest ithin the ocean. The IL is then a combination of SL and transmission loss TL, as IL SL TL, (1) here IL is incident, or received level, db re 1Pa (line), db re 1 Pa Hz (continuous sectrum) It is desirable to make estimates of IL, based on knon values of SL and TL, and this intensity-based calculation is routine for continuous tonal signals and continuous broadband signals. For transient signals, hoever, the temoral nature of the radiated aveform and sound transmission imact on the determination of radiated and Received intensity and must be considered. Acoustic Intensity As shon by, for examle, Urick section 1.5 [], instantaneous acoustic intensity I is defined as I Watts m () c
here is instantaneous acoustic ressure, Pa; is density of acoustic medium (sea ater), kg 3 m ; c is seed of sound in acoustic medium (sea ater), m s. For ractical uroses, a measurement of IL must be based on an average of intensity I, obtained over a duration T, hich follos from equation () as T t 1 I dt Watts m. (3) T c Clearly, the averaging time T is significant in determination of I, and must be sufficient for the determination of a true average of squared sound ressure. Note that IL is based on an average of intensity, but is determined as follos: density is Eref Pa c Joules m 1 units of EL are stated as re 1 Pa s re 1 Pa s Hz, and the db or db deending on hether tonal or sectrum values are considered. IL 1 log 1 db re 1Pa (4) ref here is received mean-square sound ressure, Pa; is reference sound ressure, 1 1 6 Pa (1 Pa). ref Received intensity for imulsive signal sources If the sonar signal source is imulsive, as from an exlosion or imlosion, the received signal must be integrated over the signal duration. For signal transients, Urick (section.6 []) suggests a descrition of the received signal in terms of an energy flux density, E, as E 1 c t dt Joules m. An examle of a received transient resulting from an exlosive SUS source is the acoustic ressure aveform shon in Figure 1. Clearly, the integration in equation (5) need be carried out for the duration of the received signal, only. Alying a limit of time T to the integration, and making reference to equation (3), gives 1 E c T t dt Joules m I T hich indicates the relationshi beteen intensity and energy flux density. In a similar ay that a reference for the db form of intensity (the intensity level) is made to the intensity resulting from a mean-squared ressure of ref ie. 1 Pa, the energy flux density level, as equivalent energy level EL in db, may be referenced to the energy flux density resulting from the equivalent of a mean-squared ressure of 1 Pa integrated over one second. That is, the reference value of energy flux (5) (6) Figure 1. Received signal from Mk 64 SUS at 197 m range in a shallo ocean The equivalent energy level, EL, received at a given location and due to an imulsive signal, may then be defined as EL 1 log 1 T 1 log1 T 1 Pa s Hz T ref t dt ref T db re 1 Pa s Hz here is mean-square sound ressure determined over duration T of signal, Pa. The concet and use of the equivalent energy level, or its equivalent, is considered by Urick (section 4.4 []) and is used in studies conducted by the Centre for Marine Science and Technology at Curtin University ([3], [4]). By comaring equation (7) for T = 1 second, ith equation (4), e see that the exressions for received level IL and equivalent energy level EL are the same, numerically, as the term in T disaears. Thus, it follos that an exression of a brief transient as EL in db re is equivalent to a determination of the received level, IL, by averaging over 1 second SUS Charge Noise Source Signal levels generated by underater exlosions have been studied extensively, eg. see section 4.4 of Urick [] and Richardson et al [5] section 6.7. There are several tyes of underater exlosives of interest to the ADF including the SUS charge. (7)
The radiated aveform is dominated by a very brief, high level sound ressure sike, hich is radiated from the very high ressure mass of gas hich is generated on detonation of the charge. This mass of gas exands raidly, and then subsequently collases. Loer level eaks of reducing amlitude are radiated by the subsequent non-linear resonant exansion and contraction of the bubble mass. The aveform has the aroximate aearance as shon in Figure. initial sike t.1 s bubble ulses Figure. Exlosive source aveform (idealized) Mk 64 SUS As revieed by Urick, for a TNT charge, the initial sike consists of a shar rise to a eak of amlitude P folloed by an exonential dro to near ambient ressure. The ressure sike then is aroximately t t t P o e for < t (8) here t is time t for SUS aveform ressure to dro to P e. 368P, seconds. The magnitude of the ressure sike and the time constant are deendent uon the charge eight and range from the exlosion [] as 1.13 1 3 7 5.4 1 P Pa, and (9) r. 1 3 5 1 3 9.5 1 t s (1) r here is exlosive charge eight, kg; r is range from exlosion, m. In the case of the SUS aveform, an estimate for the broadband EL at 1 m may be based on an integration of the sound ressure trace formed by the initial sike and the exonential decay, as shon by Urick (. 9 []). Using equation (7), this integration may be shon to be 1 log P t EL 1 db re 1 Pa s. (11) ref More recise determinations of EL may be made by taking into account a more accurate SUS aveform as shon in Figure, as aroriate for the detonation deth (see, for examle, Gasin and Shuler [6]). Values for EL at 1 m distance for Mk 61 and Mk 64 SUS, as determined from equation (11), and as determined from a more accurate SUS aveform for a SUS deth of 18.3 m (6 ft) are shon in Table 1 the former in arenthesis. It is noteorthy that broadband values determined using Equation (11) are very close to those determined by the use of a more detailed aveform, and the aroximate result is adequate for all ractical uroses. Table 1 also shos eak ressure values and time constant values as determined by Equations (9) and (1). The values of EL at 1 m distance in Table 1 may be regarded as SL values for determination of EL at longer range, if transmission is aroximately the same at each frequency. SUS Tye Table 1. SUS signal source characteristics at 1 m TNT Charge Weight kg Peak Pressure P at 1 m Pa Time Constant t seconds at 1 m Broadband Equivalent Energy Level db re 1 Pa s at 1 m Mk 64.31 1.4 1 7.37 ms 17 (16) Mk 61.83 4.9 1 7.88 ms 31 (3) Determination of Received Level Multi-ath transmission in a shallo ocean has the result that the transient is received along many aths, each ith a articular time delay and a articular alteration to the shae of the ulse. The determination of the duration over hich the multi-ath signals arrive, and their relative amlitude and shae change, is by no means straight-forard and, in general, is unknon. In all cases, hoever, the many arrivals form an imulse enveloe resonse in the shae of a decay hich ersists until the higher order arrivals become indistinguishable above the noise. The IL, for a signal received from an imulse or short transient, must be based on an integration of the arrivals over the time that the significant arrivals are received. In ractice, the imulse ill not be greater than 1 second, so the IL may be determined using a 1 second eriod, this then being numerically the same as the EL. Received eak ressure for exlosive sources As outlined by Richardson et al, Page 15 [5], Gasin derived a limiting range r for alicability of equations 1 3 (8) through (1), as given by r 4.76 m. For ranges greater than r, Richardson et al state that Rogers derived the folloing exressions for the eak ressure P and time constant t of an exonential ave from eak shock theory: r L lnr r r L lnr r 1 1 1 P r P ( r ) Pa (1)
t r t r 1 r L lnr 1 s (13) r 3 r P r for L c t o here P r is eak ressure at limiting range r, Pa; tr is time constant at limiting range r, s; is dimensionless constant of value 3.5. Values of eak ressure and time constant for ranges of relevance, as determined by use of Equations (1) and (13) are shon in Table. Also shon are estimates of the broadband equivalent energy EL for these ranges, as based on the values of P r and t r determined from Equations (1) and (13) and the use of equation (11). SUS Tye Table. SUS signal characteristics at large range Range m Peak Pressure P (r ), Equ (1) Pa Time Constant t (r ), Equ (13) seconds Broadband Equivalent Energy db re 1 Pa s Equ (11) Mk 64 1 8.95 ms 155 db 5 15.15 ms 141 db 1, 74.19 ms 135 db Mk 61 1 5,.6 ms 169 db 5 47.9 ms 155 db 1, 3.3 ms 149 db It is noteorthy that the values of time constant shon in Table are greater than for a 1 m distance shon in Table 1 indicating a slight sreading ith range. As shon in Table, this sreading increases very sloly beyond close range. The authors acknoledge that the recise aveform of a direct imulse from a SUS, or exlosive, is a subject of active research. The above exressions from eak shock theory ere used in the resent study, as the same aroach had been used in an earlier study sonsored by Defence [1]. As the reort from that study [1] had been distributed to some organizations external to Defence, a comarison of eak values from eak shock redictions and from measurements as desirable. Measured Data DSTO s Maritime Oerations Division has measured underater signals received from SUS charges along a number of tracks ithin continental shelf aters in the Australian region. Most of this data as obtained from horizontal ranges as close as a kilometre, aroximately, to about 3 km or more. For each track, the in-situ details obtained include ater temerature versus deth for at least one oint along the track (from hich sound seed versus deth has been obtained). The recordings of the data resented belo ere obtained using a measurement system hich as designed ith the exectation that SUS eak levels ould be as indicated by equation (9). All signals selected for analysis ere examined for the resence of overload and any exceeding system criteria ere rejected [7]. Track A Data for Track A is based on Mk 64 SUS detonated at 18.3 m (6 ft) deth and received at 18.3 m. The data samling rate for these recordings as khz, giving a maximum frequency, taking account of anti-alias filtering, of 8kHz. (The effective data samling rate then becomes 16, Hz.) The bathymetry along the track shoed a near constant deth at 56 m. The sound seed rofiles obtained at each of the start and end of the track, are shon in Figure 3. An acoustic ray diagram to 3 km along the Track is shon in Figure 4. Received ressure time series data are shon in Figures 5 to 7. Measured and redicted data are shon in Table 3. The received time series data shon in Figure 1 is, in fact, the full aveform for range 197 m along Track A. Deth (m) 1 3 4 5 Start End 1538 1539 154 1541 154 1543 1544 1545 Sound seed (m/s) Figure 3. Sound Seed Profile for Track A Figure 4. Ray lot for Track A
Figure 5. Received sound ressure time series, Track A, initial.1 s, at. km The time series detail in Figure 5 shos a comlex series of arrivals. By looking at the ray lot, Figure 4, it is sufficiently clear that at 18.3 m deth and about. km range (conditions of measurement for data in Figs. 1 and 5), the received signal ill include a direct arrival, a surface reflection and a bottom reflection. This ray lot contains 11 rays launched at angles evenly saced beteen ½ degrees. This ray lot is range indeendent and uses the sound seed rofile at the start of the Track. Clearly, if rays at steeer angles ere included, ray arrivals at the receiver at.km range ould include those ith combinations of surface and bottom bounces. Figure 6. Received sound ressure time series, Track A, full ulse, 5.1 km Figure 7. Received sound ressure time series, Track A, full ulse, 1.1 km Table 3. SUS signal characteristics measured and redicted along Track A Peak ressure (measured) Peak ressure Equ (1) Time constant Equ (13) Broadband EL db re 1 Pa s (meas) Broadband EL db re 1 Pa s (data from Equ (1), (13) in Equ (11)) Horizontal Range. km 5.1 km 1.1 km 597 Pa 68 Pa 146 Pa 356 Pa 149 Pa 73 Pa.1 ms.1 ms.11 ms 157 db 15 db 144 db 148 db 14 db 135 db Note that the theoretical derivations for the received eak ressure, as determined by Equation (1), and for the received EL, as determined by Equation (11), are for a single arrival of the SUS detonation time series along a direct ath. In a realistic shallo ocean scenario, the arrival structure ill be highly multi-ath in nature, as shon in Figures 1, 6 and 7, and ill be comrised of many arrivals suerimosed. If a log r reduction is alied to the Source Level values, that is the data at 1 m, shon in Table 1 for Mk 64 SUS, the data in Table 4 may be obtained (here, the more accurate value of EL used). It is noteorthy that the values obtained using Source Level data and simly alloing for sherical sreading are, as a first order estimate, close to the theoretical derivations for eak ressure and the EL ( nd and last ros of Table 3, resectively).
Table 4. SUS signal characteristics redicted along Track A, based on 1 m data Peak ressure Table 1 log r Broadband EL db re 1 Pa s, value at 1 m log r Discussion Horizontal Range. km 5.1 km 1.1 km 6,36 Pa 75 Pa 139 Pa 15 db 143 db 137 db Equivalent Energy Level EL The data in the last to ros of Table 3 sho that the measured received values of EL are indicative of about 1 db less loss than values of EL redicted for the single direct ath arrival of the SUS ulse received at each resective range. This may be seen as indicating that the broadband TL, taking account of shallo ater multiath henomena, and the details of surface and bottom loss, is about 1 db less than for sherical sreading. The data obtained for Track A for EL received at large range is then, roughly, in accord ith exectations for certain realistic conditions. Received Peak Pressure The data in the first to ros of Table 3 sho that the amlitude of the received ressure eak, as measured at each range value, is much less than the eak redicted from Equation (1) for the direct ath arrival at the corresonding range. This discreancy is of the order of a factor of over 5, that is, about 15 db. A conceivable reason for this discreancy may be that the time delays beteen arrivals are so small that the surface reflection, hich is negative, simly cancels the direct and bottom bounce ositive arrivals. This has been investigated, very briefly, for arrivals at the closest range for hich data exists. km, as exlained belo. Assuming straight ray ath transmission, the first three arrivals are as shon in Figure 8. source d = 18.3 m r = 197 m D = 56 m Figure 8. Ray ath for first 3 arrivals, Track A, isovelocity (idealized for straight rays) It may be shon, that the ath difference beteen the direct and surface reflected aths is very nearly d r metres. For range. km, a source/receiver deth 18.3 m, this gives a ath difference of.3 m, and, for isovelocity, an arrival time difference of.19 ms. From Table 3, the theoretically-derived time constant is.1 ms. Whether the surface reflection ill cancel the direct ath eak ill be very deendent uon exact arrival times. The data in Figure 3 indicate that the average sound seed is slightly higher for the surface reflection, such that the signal ill travel slightly faster along the surface reflected ath than along the direct. If the surface reflected ath is assigned an average sound seed.3 m/s greater, it follos that the surface reflected arrival recedes the direct ath arrival by.8 ms, and ould cause some measure of cancellation of the direct arriving eak. The time series data in Figure 5 is inconclusive, but does sho that the first arrival has negative ressure, thus identifying it as the surface reflected arrival. For straight ray aths, it may be shon that the bottom reflected ath is 1.9 m longer than the direct, and for isovelocity ill be exected to arrive.8 ms later. Based on data in Figure 3, an average sound seed of about 154 m/s may be guessed for the bottom bounce ath, giving a corrected delay of.7 ms. Data in Figure 5 shos a ositive eak about.5 ms after the first ositive eak, but a second ositive eak exists about.5 ms later not exlained by this simle ray analysis. For secular reflection from a reflective sea surface or seafloor, the reflected arrival is exected to have an amlitude of the order of the direct arrival. Figure 5 shos that the first eaks have similar amlitudes of about 4 Pa, but each is much less than the redicted eak value of 356 Pa shon in Table 3. This may be indicative of one or more of the folloing: lack of coherence in transmission due to medium irregularities; lack of coherence of reflection due to surface irregularities; chance cancellation of ositive ulse by corresonding surface reflected ath. In fact, a lack of coherence along each ath ould be associated ith some time sreading, enhancing the ossibility of ulse cancellation. It must be acknoledged that the exact henomena are unknon. The main oint, hoever, is that the measured eal amlitude is less than the theory. The measured eak ressure data shon in Table 3 indicates a decrease ith range, hich is not unexected. At the longer range values of 5.1 km and 1.1 km, it may be shon that no direct ath exists and every arrival has combinations of surface and bottom reflections. In each case the measured eak ressure is much less than that redicted for a direct arrival, hoever, this may ell be exected to be a result of reflection losses and coherence losses on reflection. Data samling issues The rate of data samling used in the descrition of the time series has imlications in the ability to follo the imulse aveform and cature the eak amlitude P. If the aveform is a sudden rise folloed by an exonential decay, a maximum error may be ostulated for a digital samling system based on missing the eak value. For the resent data, assuming an effective data samling rate of 16, Hz, for the time constant given by Equation (13) for a Mk 64 SUS gives a
maximum error (orst case of missing the eak) of about 5 db at ranges beteen 1 m and 1, m. Data Integrity It is the authors understanding that the data acquisition system as secifically designed for the receit of SUS signals and that all received data as checked for overloading and that data exceeding re-set criteria ere rejected. Data ere obtained using SSQ- 41B sonobuoys for hich the electronics ere modified for a flat frequency resonse and for the attenuation of exected high amlitude levels [7]. It is the authors understanding that the eak aveform data discussed in this aer are ithin the design limits of the sonobuoy system. Data for other Tracks Data available for to other tracks in different shallo ocean regions sho received eak ressure values similarly about 15 db less than values redicted by Equation (1) for the direct ath arrival. [4] Duncan, A. J. and McCauley, R. D., Modelling Seismic Survey Noise Exosure in the Timor Sea, CMST Reort C99-, 17 February 1999 [5] Richardson, W.J.; Greene, C.R. Jr.; Malme, C.I. and Thomson, D.H., Marine Mammals and Noise, Academic Press, Inc., 1995 [6] Gasin, J.B. and Shuler, V.K. Source Levels of Shallo Underater Exlosions, Nav. Ordnance Lab. Re. NOLTR, 71-16, 13 October 1971, AD 734381 [7] Valentine Flint, S. and Larence, M.W. Bottom bounce roagation on SEAMAP Pacific routes: Data acquisition and recording system, MRL-TN- 587, 199 Acknoledgements The authors ish to thank Mr. J. Exelby for rearation of the ray lots, Dr. M. Hall for sulying the measured data, and Dr. D. Cato for his suggestion of the study of received SUS eak ressure, based on his on rior ork in hich a reduced received eak ressure had been observed. Conclusions Data resented above sho that levels of eak sound ressure received in a articular shallo ocean at ranges greater than about km from a small underater exlosion are much less than redicted by eak shock theory for a direct arrival ulse. In this study of data obtained along several tracks ithin the Australian region, the measured eak values are about 15 db less than those redicted from eak shock theory. The reasons for this discreancy are still under active study, hoever it is believed that loss of coherence in in-ater transmission, and loss of coherence on reflection from ocean boundaries are the most likely causes. References [1] Anon., Environmental Imact Assessment of Underater Sonar Oerations and Mitigation Procedures, Reort reared by PPK Environment & Infrastructure Pty. Ltd., June 1 [] Urick, Robert J. Princiles of Underater Sound, 3rd edition, McGra-Hill, 1983 [3] McCauley, R. D. et al, Marine Seismic Surveys: Analysis and Proagation of Air-Gun Signals; and Effects of Air-Gun Exosure on Humback Whales, Sea Turtles, Fishes and Squid, CMST Reort R99-15, August