Stéphane G. Conti, David A. Demer, and Andrew S. Brierley. Introduction - PDF

ICES Journal of Marine Science, 62: 956e965 (2005) doi: /j.icesjms Broad-bandwidth, sound scattering, and absorption from krill (Meganyctiphanes norvegica), mysids (Praunus flexuosus

Please download to get full document.

View again

of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.


Publish on:

Views: 10 | Pages: 10

Extension: PDF | Download: 0

ICES Journal of Marine Science, 62: 956e965 (2005) doi: /j.icesjms Broad-bandwidth, sound scattering, and absorption from krill (Meganyctiphanes norvegica), mysids (Praunus flexuosus and Neomysis integer), and shrimp (Crangon crangon) Stéphane G. Conti, David A. Demer, and Andrew S. Brierley Conti, S. G., Demer, D. A., and Brierley, A. S Broad-bandwidth, sound scattering, and absorption from krill (Meganyctiphanes norvegica), mysids (Praunus flexuosus and Neomysis integer), and shrimp (Crangon crangon). e ICES Journal of Marine Science, 62: 956e965. Sound scattering and absorption by Northern krill (Meganyctiphanes norvegica) were measured over the acoustic bandwidth of 30e210 khz and compared with similar scattering measurements for Antarctic krill (Euphausia superba). The measurements of total target strength (TTS; energy scattered in all directions, averaged over all angles of incidence) match the SDWBA model (stochastic distorted-wave Born approximation) recently developed for Antarctic krill, indicating its validity for other euphausiid species with similar size and shape. However, the TTS of crustaceans with markedly different shapes are not well predicted by SDWBA derived with the generic krill shape and scaled to animal length (L). Therefore, crustacean target strength (TS) may not be estimated accurately by a linear function of log 10 (L), irrespective of shape, questioning the validity of the current TS relationship used for Antarctic krill derived from data measured from multiple crustaceans. TTS and TS are dependent upon both L and shape, and different crustaceans have significantly different shapes and width-to-length relationships. In contrast, modelled TTS and TS spectra for gravid and non-gravid krill appear to have differing amplitudes, but similar shapes. Additionally, measurements of absorption spectra from decapods indicate that the absorption cross-section increases with the volume of the animal. Ó 2005 International Council for the Exploration of the Sea. Published by Elsevier Ltd. All rights reserved. Keywords: acoustic scatter, Born approximation model, distorted-wave, modelling, stochastic, total cross-section, total target strength. Received 1 November 2004; accepted 10 January S. G. Conti and D. A. Demer: Southwest Fisheries Science Center, 8604 La Jolla Shores Drive, La Jolla, CA 92037, USA. A. S. Brierley: Gatty Marine Laboratory, University of St Andrews, Fife KY16 8LB, Scotland, UK. Correspondence to S. G. Conti: tel: C ; fax: C ; Introduction Knowledge of the abundance and distribution of species is among the most basic requirements for understanding ecosystem function and for the management of living resources. In the marine environment, pelagic organisms such as commercially and ecologically important fish, micronekton, and zooplankton can be sampled with nets, but acoustic-survey techniques are advantageous because they permit continuous and non-destructive sampling over larger areas in shorter periods of time. The abundance and distribution of Antarctic krill (Euphausia superba), for example, are routinely determined by acoustic surveys (e.g. Brierley et al., 1999; Hewitt et al., 2004). Accurate target strength (TS) information is fundamental to the interpretation of acoustic-survey data. TS describes the portion of incident acoustic energy that is scattered from an individual fish or zooplankton. It is dependent on acoustic frequency, and the animal s size (length and volume), shape, orientation, and material properties. Although TS is highly sensitive to these variables, they have of necessity been largely ignored in the analysis of acoustic-survey data. TS can be estimated in a number of ways including the direct ensonification of tethered individuals, i.e. confined within the usually narrow echosounder beam (Stanton et al., 1998), in situ observations (Warren et al., 2001), and modelling (Foote, 1991; Ona, 1999). Although there have been numerous recent /$30.00 Ó 2005 International Council for the Exploration of the Sea. Published by Elsevier Ltd. All rights reserved. Broad-bandwidth, sound scattering, and absorption from krill, mysids, and shrimp 957 advances in the direct acoustic observation of zooplankton TS (e.g. McGehee et al., 1998; Brierley et al., 2004), for logistical reasons most studies have been restricted to one species and narrow ranges of frequency, organism size, and maturity stage. For example, the TS/length relationship for Antarctic krill presently used (Greene et al., 1991; Hewitt et al., 2004) does not take account of the maturity stage of the animal. A gravid adult female of 55 mm standard length has a mass 9.5% greater than a mature male of the same length (Morris et al., 1988) but both are assumed to have the same TS. This shortcoming contributes to uncertainty in acoustic estimates of krill abundance. Recently, a method has been developed for estimating the total scattering cross-section s T from scatterers in motion, like ÿ marine organisms (total target strength TTSZ10 log st 10 4p ; de Rosny and Roux, 2001). TTS represents the acoustic energy scattered from an object in all directions, averaged over all angles of incidence. TS is closely related to TTS, but while TS is a function of many different variables (e.g. animal size, shape, orientation), this measurement technique conveniently provides accurate and precise measurements of broad-bandwidth TTS (Demer et al., 2003). Application of the technique to live, freeswimming organisms can provide improved TS information leading to substantial improvements in the at-sea acoustic identification and assessment of ecologically and commercially important pelagic species (Conti and Demer, 2003; Demer and Conti, 2003a). This is possible because physicsbased scattering models validated using TTS measurements can be used to predict TS at different angles of incidence, or averaged over an applicable distribution of angles (Demer and Conti, 2005). In this study, the TTS technique is used to measure the broad-bandwidth sound scatter from Northern krill (Meganyctiphanes norvegica) and compared with similar measurements made on Antarctic krill (Euphausia superba). These euphausiid species have similarities in their shapes and maturity stages, have overlapping length distributions, and are key species in the North Atlantic and Southern Ocean ecosystems, respectively. Proof of equivalence in their broad-bandwidth sound scattering characteristics will expand the opportunities to improve the accuracy of TS estimation and aid in the acoustic identification and survey of both species. For example, convenient access to the shorter-lived M. norvegica could allow broad-bandwidth TTS measurements to be made rapidly over a variety of maturity stages. This, in turn, could lead to a reappraisal of the TS relationship used internationally for Antarctic krill, and could thus provide an indication of the likely error resulting from the effective dismissal of shape and maturity stage differences in conventional euphausiid TS (Greene et al., 1991). We also present TTS measurements for mysids and decapod shrimp. Data from these taxa were used by Greene et al. (1991) in their determination of TS for Antarctic krill. However, their different gross anatomical shapes make the use of these data perhaps questionable. Also presented are the first measurements of absolute sound absorption from marine crustaceans. The absorption measurement technique employed here was first described by Conti et al. (2004) and led ÿ to the measurement ÿ of the absorption cross-section s a ABSZ10 sa log10 4p : These measurements add to the study of sound propagation in the presence of krill. Methods From 1 to 8 August 2003, TTS and ABS measurements of krill (Meganyctiphanes norvegica), mysids (a mix of Praunus flexuosus and Neomysis integer), and shrimp (Crangon crangon) were made at the Scottish Association of Marine Sciences Dunstaffnage Marine Laboratory, Oban, Scotland. Details of the general processing steps are outlined in de Rosny and Roux (2001), Demer et al. (2003), Demer and Conti (2003a), and Conti et al. (2004). Equipment and analysis steps specific to these experiments follow. The krill were captured in Loch Etive, Scotland, in the deep holes to the southeast of Bonawe quarries using a 1-m diameter hoop net. The net was towed horizontally from RV Seol Mara at various depths ranging from approximately 120 m during the daytime to approximately 60 m at night. The mysids were captured in a shallow embayment to the south of the Marine Laboratory in Dunstaffnage Bay using a hand net. The shrimp were captured in Tralee Bay at low tide by hand-dragging a 2-m long beam trawl. The krill and shrimp were kept alive and transferred to the laboratory in 20-l insulated containers and stored temporarily in aquaria. The mysids were captured immediately prior to making the measurements and stored briefly in buckets. For each experiment, a 20-l stainless-steel bucket was filled with 17 l (G1%) of ambient seawater at temperatures ranging from 13.8 to 14.2(C. Groups of live swimming animals were then added sequentially (Figure 1). Four sets of frequency-modulated pulses (1.0 V p-p ;18ms pulse length) with overlapping bandwidths (30e90, 80e130, 120e170, and 160e210 khz, centre frequencies f c of 60, 105, 145, and 185 khz, respectively) were generated (Hewlett Packard 33120A arbitrary waveform generator), amplified 20 db (Krohn-Hite 7500 power amplifier), and transmitted once per second using an omni-directional, broad-bandwidth emitter (ITC 1042) and received bi-statically with two omni-directional, broad-bandwidth receivers (ITC 1042 and Reson TC4013). The emitter and two receivers were suspended in the tank via aluminium rods attached to the top of the bucket (Figure 1). For each bandwidth, ensembles of 100 pulses were transmitted, and the reverberation time-series h k (t) (where k is the pulse number) were recorded over 70 ms and digitized at 500 khz, using a 12-bit analog-todigital converter (National Instruments DAQPad 6070E). 958 S. C. Conti et al. To reduce noise and align the time of origin for all frequencies, the h k (t) were match-filtered via crosscorrelations with their respective transmitted signals. The coherent intensity ðch k ðtþh kc1 ðtþdþ in the 100-pulse ensembles identified sound scattered from the echoic bucket. Because the positions of the animals were uncorrelated from ping-to-ping, the incoherent intensity ðch k ðtþ 2 DÞ described sound scattering from the animals. The ratio of coherent and incoherent intensities decayed exponentially: # Ch k ðtþh kc1 ðtþd SðtÞZ Ch k ðtþ 2 D exp ÿtc ns t v ; ð1þ where s t is the total scattering cross-section of one animal, v the volume of the cavity, n the number of animals, c the speed of sound in seawater, CD designated the average taken over the pings, and [] the average taken over the experiments for the same group of animals (i.e. runs with different positions of the emitter and receivers). The exponential decay of S(t) was estimated for each 100-pulse ensemble by low-pass filtering the numerator and denominator (f cutoff Z 0.016! f sampling /2), and fitting a slope (d ln(s(t))/dt) in the least-squares sense, while requiring 5 % t % 20 ms and forced through the origin (ln(s(t Z 0)) Z 0). Knowing the volume of the cavity (v), the number of animals (n), and the speed of sound in seawater (c), an estimate of s t was made for each group of animals and bandwidth: s t ÿ v cn dlnðsðtþþ : ð2þ dt The s T calculated from the data was then converted to TTS as a function of frequency. Thus, TTS(f) measurements were made from the reverberation sensed at two receiver locations, and during multiple 10-min runs. In total, measurements were made of two aggregations of 115 and 135 non-gravid swimming krill; five aggregations of 50e75 gravid and non-gravid shrimp (Table 1); and three aggregations of 1055, 1479, and 2256 gravid and non-gravid mysids. In order to prevent animals aggregating at the side of the tank, which they did in daylight presumably in an attempt to avoid light (Figure 1), the measurements were performed in the dark. Following the measurements of each aggregation, the length of each animal L, viz. from the anterior tip of the rostrum and the posterior end of the uropods, excluding their terminal setae (Figure 2), was measured to the nearest millimetre before it was preserved in sample jars with 4% buffered formaldehyde. To predict the empirical estimates of s t, the SDWBA model (Demer and Conti, 2003b) was run to obtain SDWBA TTS. The computations are detailed in Demer and Conti (2003a). Parameters include a shape with width-to-length ratio 1.4 times that of the generic krill shape from McGehee et al. (1998), cz 1455 m s ÿ1, the non-dimensional speed of sound and density contrasts (h Z and g Z ) from Foote et al. (1990) and Foote (1991), respectively. The random phase was chosen from a normal distribution (4 Z N[0, 40.5(]) from Demer and Conti (2003a). The generic krill shape, which was derived from a starved krill with L Z mm, was scaled proportionately to represent the larger krill in these experiments (average L Z mm). Thus, SDWBA TTS was evaluated from 30 to 210 khz. The absorption cross-sections (s a ) of the krill, shrimp, and mysids were measured using the same experimental setup, and the technique outlined in Conti et al. (2004). Prior to the introduction of the animals in the bucket, the reverberation time-series for the empty bucket h k empty (t) were recorded multiple times for the same position of the emitter and receivers. Then the animals were introduced into the bucket without moving the emitter and receivers, and the reverberation time-series h k (t) were recorded with the animals swimming. The incoherent intensity for the empty bucket and the bucket with the animals swimming are compared, and the ratio decays exponentially: Table 1. The parameters of the groups of shrimp for the absorption measurements. Number of shrimp L min L max Maturity Figure 1. Northern krill swimming in a 17 G 1%-l stainless-steel bucket (Vollrath 58200). Aluminium rods were used to suspend the omni-directional emitter (ITC 1042), and two omni-directional receivers (ITC 1042 and Reson TC4013), and a thermocouple near the centre of the water volume Mixed Mixed Mixed Mixed Gravid Broad-bandwidth, sound scattering, and absorption from krill, mysids, and shrimp 959 Figure 2. Length distributions of the mysids (a), 115 krill (b), 135 krill (c), combined krill (d), and the shrimp (e). Chk ðtþ 2 D S a ðtþz Ch kempty ðtþ 2 D exp ÿt cns a v : ð3þ Using the same low-pass filter as in the total scattering cross-section measurement, the absorption cross-section was obtained knowing the volume of the cavity (v), the number of krill (n), and the speed of sound in seawater (c): s a ÿ v dlnðs a ðtþþ cn dt ð4þ sa ABSZ10 log 10 : ð5þ 4p The addition of the animals slightly changes the volume of the tank but the modification is small. The overall absorption from the tank remains identical. However, the arrival times of the main echoes from the cavity may shift when the total volume changes between the recordings without and with the animals. This could lead to problems in estimating the ratio S a (t) because the times of zeroes from the main echoes for Ch k ðtþ 2 D and Ch k empty ðtþ 2 D may not correspond. This problem is compensated by using multiple recordings for the tank without and with animals to average the zeroes, and smoothing both Ch k ðtþ 2 D and Ch k empty ðtþ 2 D: The reverberation time-series recorded for each of the four experimental bandwidths provided the average values of the total scattering and absorption cross-sections over each bandwidth. The total scattering spectra s T (f) were obtained by filtering the reverberation time-series in narrow bands. Between 30 and 90 khz, the spectra were measured with a 6-kHz resolution. For the three bands spanning 80 to 210 khz, the frequency resolution was 5 khz. For the absorption measurements, only the average values for each of the four experimental bandwidths were obtained. Results The mean standard lengths of the two groups of 115 and 135 Northern krill were 36.5 (s.d. Z 2.1) and 36.2 960 S. C. Conti et al. (s.d. Z 1.9) mm, respectively. The mean TTS spectra of each group was measured (Figure 3). The TTS measurement precision (s.d.) was estimated from multiple measurements at each frequency as G14% and G12%, respectively. Over the entire measured bandwidth, the TTS measurements for the krill matched the SDWBA TTS model to better than 26% (1 db). Multiple additional measurements were attempted with groups of 26e75 krill, but discarded because of low signal-to-noise ratios. Some measurement series were also discarded due to disturbances of the water/air interface, low signal-to-noise ratio, and water-temperature fluctuations during the acquisition. The TTS measurements with 115 and 135 krill were obtained from animals with similar lengths (Figure 2). Therefore, the TTS spectra normalized to one animal were similar in both cases. They compared favourably with the theoretical predictions derived and validated for Antarctic krill (Demer and Conti, 2003a). The TTS measurements for the mysids and shrimp were made from groups of animals with different size distributions (Figure 2). The mysids were smaller than the krill, and the size distribution of the shrimp was broader. Neither the TTS spectra for the mysids nor the shrimp matched the SDWBA predictions derived with the krill shape and adjusted for length (Figure 4). This result is not surprising since each of the three species has significantly differing length-to-width relationships (Figure 5, Table 2) with those of shrimp and mysids being larger than that of the krill. Consequently, their TTS are approximately 100% and 200% higher than the theoretical predictions derived using a generic krill shape, respectively. The measurement precision for TTS of mysids and shrimp was estimated as G20%. The dorsal-aspect views and shapes of gravid and nongravid krill show differences (Figure 6, and unpublished data from Tarling et al. d see acknowledgements). The abdomen for a gravid krill is larger than for a non-gravid one. Using the SDWBA model and the shapes of gravid and non-gravid northern krill both TTS and TS predictions were computed (Figure 7). These predictions for TTS show that the spectrum for a gravid krill is higher than for a non-gravid one, but despite the amplitude difference, the discrepancies in the animal shape do not modify the frequency response significantly. Compared with the predictions for Antarctic krill, the spectra are similar in amplitude, and a difference can be observed at low frequencies where the spectrum for Antarctic krill drops faster than for Northern krill mainly because the Northern krill are thinner than the Antarctic krill in the modelling. The TS show some discrepancies for the dorsal aspect, but these discrepancies tend to be smoothed when the TS is averaged for the different distribution of orientation angles found in the literature for krill (Kils, 1981; Endo, 1993; Demer and Conti, 2005). Figure 3. Mean TTS of M. norvegica measured from aggregations totalling 115 and 135 animals (bold dashed lines in a and b, respectively). For each run, the spectra were recorded with two hydrophones at different positions in the bucket. Standard deviations were calculated from three or four measurements from each aggregation and plotted (light dashed lines). The SDWBA TTS predictions (solid lines), were computed with g Z , h Z , generic shape, and the overall krill-length frequency distribution (see Figure 2). Broad-bandwidth, sound scattering, and absorption from krill, mysids, and shrimp 961 Figure 4. Broad-bandwidth mean TTS measurements of mysids, krill, and shrimp (green, red, and blue dots, respectively) with standard deviation (dashe
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks