The new fisheries multibeam echosounder ME70: description and expected contribution to fisheries research

The new fisheries multibeam echosounder ME70: descriptionand expected contribution to fisheries research

Verena M. Trenkel, Valerie Mazauric, and Laurent Berger

Trenkel, V. M., Mazauric, V., and Berger, L. 2008. The new fisheries multibeam echosounder ME70: description and expected contribution tofisheries research. – ICES Journal of Marine Science, 65: 645–655.

Recently, Simrad in collaboration with Ifremer developed a calibrated, multibeam, vertical echosounder (ME70) for fisheries research.We describe its capabilities and technical limitations. The ME70 has up to 45 beams with distinct frequencies in the range 70 –120 kHz, spanning at most 1508. All beams are stabilized in vessel roll and pitch. It has reduced side-lobe levels, up to 270 dB(two-way) instead of the 225 dB (one-way) of conventional systems. We outline research areas for which the ME70 mightprovide new types of information and hence lead to novel insights. We illustrate the potential contributions with datasets collectedin the English Channel and on the continental-shelf break of the Bay of Biscay. Finally, future research and developments using the newsystem are outlined.

Keywords: abundance, acoustics, behaviour, three-dimensional.

Received 25 July 2007; accepted 25 February 2008.

V. M. Trenkel: Ifremer, Departement EMH, rue de l’Ile d’Yeu, BP 21105, 44311 Nantes Cedex 03, France. V. Mazauric and L. Berger: Ifremer,Departement NSE, Plouzane, France. Correspondence to V. M. Trenkel: tel: þ33 240 374000; fax þ33 240 374075; e-mail: [email protected].

IntroductionThe use of acoustic methods to estimate fish abundance has ahistory of .50 years (see review in Fernandes et al., 2002, fortheir use within the ICES Area). Since the early 1950s, substantialprogress has been made at all stages of the process, from measure-ment instruments to signal extraction and interpretation(Simmonds and MacLennan, 2005). Vertically orientated, single-beam echosounders operating at frequencies ranging from 38 to200 kHz are most commonly used. Quantitative estimates weremade possible by the introduction of standardized calibrationprotocols (Foote et al., 1987).

Several physical factors influence the utility of acoustic methodsfor abundance estimation. Beam width in combination withdepth, and also pulse duration and bottom topography, determinethe extent of the dead zone near the seabed for which no infor-mation is available (Ona and Mitson, 1996). The acoustic obser-vations of different species are influenced to varying degrees bythis dead zone. For example, a large proportion of cod (Gadusmorhua), haddock (Melanogrammus aeglefinus), and redfish(Sebastes spp.) is thought to reside in it for most of the time(Aglen et al., 1999). The detection range for single targetsdepends on echosounder performance, fish target strength (TS),and ambient- and vessel-noise levels, so may differ betweenvessels (Mitson and Knudsen, 2003). As beam width determinesthe diameter of the volume or surface insonified (sampled) at agiven depth, for vertical echosounders the backscattering strengthsfrom fish schools smaller than this diameter will be underesti-mated and their dimensions overestimated; therefore, the bias inevaluating small schools increases with depth (Diner, 2001,2007). For oblique beams, the backscattering strength decreaseswith increasing angle (Melvin et al., 2003; Zedel et al., 2005).

The acoustic-backscattering properties of individual fish resultfrom the interplay between physical and biological factors. Theyare a function of signal frequency and depend not only on themorphology of the species, its schooling behaviour, and an indi-vidual’s orientation and body shape and size (Misund, 1997),but also physiological state (Ona, 2003). Fish reactions to noiseand light emitted by a surveying vessel bias abundance estimates(Olsen et al., 1983; Misund, 1997), probably because of changesin fish tilt angle as they dive in front of a survey vessel (Gerlottoand Freon, 1992; Handegard and Tjøstheim, 2005; Cutter andDemer, 2007).

In addition to the standard method of abundance and biomassestimation by echo-integration, in combination with fishing haulscarried out for species and size identification (see description inSimmonds and MacLennan, 2005), acoustic information hasalso been used to study schooling behaviour (ICES, 2000) andthe potential impact of fishing on spatial distribution patterns(Wilson et al., 2003). For this, standard protocols for extractingmorphological school descriptors from vertical echosoundershave been proposed (ICES, 2000; Reid et al., 2000), providingschool characteristics in two dimensions (depth and alongship),because no information perpendicular to the survey track is avail-able. Therefore, neither horizontal (athwartship) avoidance reac-tions nor their expected impact on school morphology can bestudied using single-beam echosounders. To overcome this limit-ation, horizontal and vertical scanning multibeam sonars havebeen employed (Misund and Aglen, 1992; Gerlotto et al., 1999;Soria et al., 2003). Multibeam sonars have also been used increas-ingly to study fish behaviour in the wild, including the determi-nation of swimming speeds, aggregation dynamics, and spatialschool characteristics (Misund and Aglen, 1992; Mayer et al.,

# 2008 International Council for the Exploration of the Sea. Published by Oxford Journals. All rights reserved.For Permissions, please email: [email protected]

645

by guest on Decem

ber 4, 2015http://icesjm

s.oxfordjournals.org/D

ownloaded from

http://icesjms.oxfordjournals.org/

2002; Gerlotto and Paramo, 2003; Brehmer et al., 2006), andpredator–prey interactions (Benoit-Bird et al., 2004).

To propose an instrument that might be useable in situationswith mainly small schools where beam width is a strongly limitingfactor, Ifremer and Simrad joined forces in 2001. They developedthe first calibrated multibeam echosounder (ME70) for fisheriesresearch and installed it on the French FRV “Thalassa”, amodern, noise-reduced fisheries research vessel (Mitson andKnudsen, 2003) in summer 2005.

The ME70 has two principal operational modes for observingorganisms in the water column (fisheries-research mode), andfor bathymetry and habitat classification (bathymetry mode).Our aim here is to describe the fisheries-research mode of theME70, to illustrate its expected contribution, and to presentsome preliminary results. We also provide a detailed technicaldescription of the new system and a discussion of its advantagesand limitations, then illustrate its expected contributions tofuture fisheries research. The last have been grouped under threeheadings in relationship to the limitations of acoustic measure-ments: schools of small size, mixed-species associations, andindividual and school behaviour. Areas for which the ME70 willnot offer any improvements are not considered. Conclusions aredrawn, and directions for future research are then proposed.

Characteristics of the ME70 in fisheries-researchmodeThe ME70 is a calibrated, multibeam echosounder for obtainingquantitative information on organisms in the water column(Andersen et al., 2006). It can be configured with up to 45 beamsspanning at most 1508; the athwartship centre angle of the fancan be adjusted from þ458 to 2458. Each beam has a unique fre-quency in the range 70–120 kHz and a unique steering angle, exceptfor one special configuration in which all beams have the samesteering angle. The spread of steering angles through the fan iseither spaced linearly or optimized for side-lobe reduction. Beamwidth can be selected from 2.28 to 208 for the central beam, andthe width of the other beams are then adjusted depending on fre-quency and steering angle. Additionally, there are two specificbeams, referred to as reference beams, which can be configured sep-arately and freely in terms of frequency, steering angle, and beamwidth. This allows comparison with single-beam echosounders orto increase the sampling volume by adding them as wide beams

(208 opening), steered at 508. All beams can be operated as splitbeams. Beam emission is in groups of 1–4 beams. As with single-beam vertical echosounders used in fisheries research, the instru-ment provides values of volume-backscattering strength (Sv) andTS when the beams are configured as split beams.

The innovative design of the ME70 is based on a matrix of 800transducer elements that are individually controlled, functioningboth at transmission and reception. The challenges of real-timeprocessing of huge amounts of data were met, and the resultingtool is promising and allows a high degree of user-controlledflexibility. The ME70’s matrix transducer offers up to 270 dBside lobes (two way) compared with about 225 dB (one way)for “T” array geometries of conventional multibeam systems(Figure 1a). Hence, the ME70 is a multifrequency echosoundercapable of collecting data throughout the whole fan, in particularoutside the spherical volume (Figure 1b), where seabed backscat-tering through the side lobes (Figure 1a) strongly limits fish detec-tion in conventional multibeam systems. All beams are stabilizedwith respect to a vessel’s roll and pitch to keep the steeringangles of all beams fixed between transmission and reception.

The ME70 has many configuration options, each a trade-offbetween several factors, and their choice will depend on the objectivesof the study (Table 1). Figure 2 shows a typical beam configuration(“V” configuration): Figure 2a is a schematic view of the athwartshipbeam distribution, the alongship distribution is shown in Figure 2b,and the steering angle and frequency of each beam in the configur-ation is depicted in Figure 2c. When selecting a beam configuration,various issues need to be considered, as covered below.

Selection of the number of beamsand the pulse lengthAs each beam of the fan has a distinct frequency, the number ofbeams depends on the selected pulse length and the frequency band-width, and vice versa. With the maximum frequency bandwidthranging from 70 to 120 kHz and the maximum number of beamsbeing 45, the minimum pulse length available is 2048 ms for a con-tinuous wave (CW) signal. If the number of beams is reduced to 21,the minimum pulse length is reduced to 1024 ms.

Selection of the number of beamsto be transmitted simultaneouslyThe options range from 1–4 beams per transmission group. Thechoice affects the detection range, because the source level is

Figure 1. Definition of acoustic-sampling volumes: (a) outer beam with side lobes; beam detection of A is polluted by bottom detectionthrough side lobe at point B; (b) definition of sampling volumes of multibeam echosounder.

646 V. M. Trenkel et al.

by guest on Decem



ownloaded from


12 dB higher for separate transmissions (a group of one) than withtransmissions by groups of four. The extent of the blind zoneunderneath the vessel is also affected, because it is inverselyproportional to the number of frequencies per group (see rule inTable 1).

Selection of the frequency-distribution patternThe frequency distribution in the fan can be set up in differentways. Three options have been evaluated. To maximize the detec-tion range, the “V” configuration has the lowest frequency in thecentre and the highest frequencies in the most-steered beams.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 1. Parameter choices for the multibeam fisheries echosounder ME70, and the resulting technical performance.

Choice Input parameters Technical performance

Characteristics ofmultibeam fan

Frequency bandwidth 70–120 kHz Determines minimum pulse length

Number of beams 3–45 Trade-off between frequency bandwidth, number ofbeams, and pulse lengthType of signal CW, or linear-frequency

modulated

Pulse length 64–5120 ms

Number of frequencies emittedper group

1–4 Range detection is longer when fewer frequencies aretransmitted together Blind zone extends � c T/2([Nfan/Nfreq]þ1), where c is the sound velocity, T thepulse length, Nfan the number of beams in the fan, andNfreq the number of frequencies emitted per group. []denotes the integer part of the quantity

Order of transmission Increasing or decreasingsteering angles

Determines the shape of the blind zone

Distribution infrequency

Shape V configuration Longer detection range in centred beams. Beamopening narrower in outer beams than in centredbeams

L configuration Longer detection range in outer beams. Beam openingnarrower in centred than in outer beams

I configuration Suitable for multifrequency analysis

Frequency spacing Linear Suitable for I configuration

Optimized Reduces leakage between frequencies

Distribution inspace

Alongship beam width for thecentred beam

2.2–208 Determines the alongship and athwartship sampledvolume (beyond blind zone)

Athwartship beam width atthe centred beam

2.2–208 maximum Realizable beam opening range depends on frequency

Alongship steering of the fan 08 default (58 maximum)

Athwartship steering of the fan 08 default (458 maximum)

Beam spacing Linear (in 8)

Optimized (21 to 24 dB)

Figure 2. Schematic view of beam patterns of the ME70 sounder: (a) front view of multibeam fan; (b) across-track view of multibeam fan;(c) frequencies and angular directions of beams in “V” configuration. Vertical bars indicate frequency bandwidth for each beam.

The new fisheries multibeam echosounder ME70 647

by guest on Decem



ownloaded from


To maximize the angular resolution, the “L” configuration has thehighest frequencies in the centre, and the lowest frequencies in theouter beams. To explore species discrimination using multifre-quency analysis, all frequencies (beams) in the “I” configurationpoint, not necessarily vertically, in the same direction. Thechoice between “V” and “L” configurations will be guided bythe objectives of the scientific survey in terms of angular resolutionwhen there is the possibility of a narrower beam opening at thevertical for the “L” configuration and detection range, given thelonger range at the vertical with the “V” configuration. Whennarrow angular resolution and long detection range are bothrequired at the vertical, the “L” configuration combined with areduced number of frequencies per transmission group can be asuitable option, but to the detriment of increasing the length ofthe blind zone underneath the vessel.

Selecting the sampling volume (fan width) perpendicularto the vessel trackThree series of parameters relating to the prevailing role of thecentred beam in the definition of the fan’s characteristics requiredetermination: (i) along- and athwartship beam width of thecentre beam, (ii) athwartship beam spacing, and (iii) along- andathwartship steering angle of the centre beam.

In technical terms, an appropriate weighting of all cells in thearray is automatically calculated when the beam width of thecentred beam is set to a given value. As a consequence, the beamwidths of all other beams are a function of the frequency distri-bution, the beam-steering angles, and this pre-calculated weight-ing. However, it is somewhat restrictive to describe the fan’scharacteristics only through the sampled volume and, remember-ing the originality of the ME70 in terms of side-lobe reduction,valuable information can be collected with a given beam configur-ation when the resulting side-lobe levels are included in the con-sideration. The results shown in Figure 3 emphasize that it canbe judicious to lower slightly the angular resolution, to increasedata quality and to reduce noise outside the spherical volume.As beam width increases from 2.38 in Figure 3a to 2.88 inFigure 3b, the corresponding weighting on the array reduces theside-lobe level for each beam and leads to reducedartefacts outside the spherical volume in the outer beams (2338steering angle in Figure 3).

The high available angular resolution (minimum 2.28 for abeam centred on 120 kHz) and reduced side lobes of the ME70

can reduce the dead zone compared with traditional echosounderswith 78 or 118 beam widths. Unfortunately, the ME70 currentlysuffers from a limitation in beam-forming, which affects datacollected in a fixed layer of 75 cm above the seabed visible inall subplots of Figure 3. Despite this limitation, the improvedbottom resolution of the ME70 for an uneven seabed is evidentwhen comparing data obtained with an ER60 sounder (78,120 kHz; Figure 4a) with data from the central beam of theME70 (2.88, 117 kHz; Figure 4b).

In addition to collecting information with the multibeam fan,the ME70 offers the possibility of duplicating data collected withtraditional vertical echosounders. For this, two additional referencebeams can be used, whose frequency, beam opening, and steeringangle can be set freely by the user. Only the pulse length remainsthe same as in the fan beams. The two reference beams are alwaystransmitted after the fan beams, to reduce frequency leakage withthe latter. Although the reference beams increase the extent of theblind zone underneath the vessel by about one or two pulselengths, depending on whether they are in the same orseparate emission groups, they can be useful for comparisons withconventional vertical echosounders operating simultaneously.

Once a particular configuration has been selected, the systemcan be run with default system gains calculated automaticallyfrom stored values. However, it is recommended that each con-figuration be calibrated. The ME70 calibration procedure issimilar to that for other echosounders (Foote et al. 1987), particu-larly for the ER60 echosounders. To adjust the nominal gain set bydefault, the two parameters “gain adjustment” and “Sa correctionadjustment” must be measured for each beam. If we define thegain as the ratio of the intensity measured in a given beam tothat for an omni-directional transducer, the first parameter con-cerns the offset between the theoretical gain and that measuredfor the calibration sphere. The second parameter corresponds tothe offset between the theoretical contribution of the shape ofthe pulse on the integrated response of the sphere and themeasured one. A typical view of the fan in the calibration interfaceis presented in Figure 5a for a configuration with 19 beams and asequence of 150–350 sphere TS measurements per beam. Theresulting calibration parameters are displayed in Figure 5b.Calibration results for a beam configuration with only 15 beamsand a wider athwartship sector are also presented, includingstandard deviations for the gain-adjustment estimates. Thecalibration sphere used for this test was a 25-mm diametersphere made from tungsten carbide with a 6% cobalt binder,

Figure 3. Impact of beam width and side lobes on seabed backscattering outside the spherical volume for a beam steered at 2338.(a) Optimal beam width of 2.38 (one-way) obtained with the “L” configuration; (b) the same for 2.88 beam width. Echograms are displayedwith non-calibrated Sv data and a visualization threshold of 265 dB (using calibration results in Figure 5b, the threshold after calibration isexpected to correspond to about 263 dB).


by guest on Decem



ownloaded from


Figure 4. Dead zone for an uneven seabed. (a) ER60 (120 kHz, 78), (b) ME70 centred beam (117 kHz, 2.88 beam width); visualization threshold262 dB.

Figure 5. Calibration of ME70 with two “L” configurations. The first configuration has 19 beams, a 658 athwartship sampling volume, and a2.88 (one-way) beam width at the vertical, and the second configuration has 15 beams, a 1008 athwartship sampling volume, and a 58 beamwidth at the vertical. (a) Spatial distribution of hits in the fan with 19 beams; (b) calibration parameters.


by guest on Decem



ownloaded from


provided with the ME70. Owing to the small weight of this sphere,an additional weight was suspended below it to stabilize the cali-bration target, but this required careful attention when calibratingthe outer beams, to avoid detecting the echo from the weight.Future calibrations may use larger and heavier spheres, asdescribed in Foote (2006). Their use alleviates the need for anadditional weight and their high TS values improve detections ina noisy environment. A disadvantage is that two spheres insteadof one are needed to calibrate the whole fan, requiring furtherdevelopments of the calibration procedure.

To take fullest advantage of the ME70 and the five frequencyER60s installed on “Thalassa”, all are configured and controlledusing HERMES software, specifically designed for the purpose.HERMES stores the data from all echosounders (ME70 andER60s) in a single file, using the HAC standard format detailedin ICES (2005).

Potential research applications of ME70Some of the challenges in the acoustic method identified in theintroduction may be addressed with data from the ME70. Thisis illustrated here with two datasets. Data from herring (Clupeaharengus) schools were collected during March 2006 in shallowwaters of the English Channel (Table 2). Data were also collectedduring March 2007 from mixed species, mainly small schools,on the continental-shelf break of the Bay of Biscay. The ME70 con-figurations used for each dataset are listed in Table 2. MOVIESsoftware (Berger et al., 2005) was used to perform echo-integrationand extraction of shoal parameters.

Small schoolsThe size, e.g. the horizontal diameter, of single or multispeciespelagic fish schools depends on species, season, time of day,depth, and geographic location (Scalabrin and Masse, 1993;Soria et al., 2003). Schools are often grouped into clusters(Petitgas, 2003). In the Bay of Biscay, nearest-neighbour horizontaldistances between pelagic schools within clusters range from 400to 3300 m (Petitgas, 2003), and horizontal school length is often�10 m. Little is known about school width (perpendicular tosurvey track), but investigations with vertical-scanning multibeamsonar have shown that pelagic fish schools seem to be elongated,with school length up to twice the width for Sardinella aurita(Gerlotto and Paramo, 2003). Based on geometric considerationsand simulations, Diner (2001) proposed an algorithm forcorrecting school-length measurements, which are increasinglyoverestimated with depth because of the increased samplevolume. The correction is valid if the true school length is atleast 1.5� the sampled beam diameter (e.g. school length mustbe �12 m at 100-m depth for a 4.58 beam width, and �18-mlong if beam width is 78). Hence, a small school is one whoselength is less than 1.5� the beam diameter at the given depth.

The definition of “small” depends both on the depth of theschool and the beam width. Similar, non-negligible depth effectsmight exist for abundance estimates based on echo-integration,owing to negatively biased backscattering cross sections of smallschools, as defined above, at greater depths (Diner, 2007).

The reduced beam angles (minimum 2.28) of the ME70 areexpected to provide several benefits, including improved estimatesof school length and backscattering cross-section estimates forsmaller schools, i.e. those too small for the conventional 78 or118 beams. Of course, using a smaller beam width only reducesthe number of schools that fit the aforementioned definition of“small”. For the mixed species data, in the central beam ofthe ME70 (beam opening 4.38), �30% of encountered shoalswere ,1.5� the beam width at the depth of the shoal. For the78 single-beam measurements at 70 kHz, the correspondingvalue is �65%. In that dataset, most schools were near theseabed at depths of around 200 m.

Based on the use of several beams, an additional estimate ofschool width, and therefore the entire school geometry, isobtained. The mixed example shows the effect of increasing thespatial resolution on observed school structure (Figure 6). It alsoillustrates the effect of beam stabilization for the ME70.Although the image from the ER60 is smeared perpendicular tothe survey track, the image reconstructed from the three centralbeams of the ME70 is much sharper.

A drawback of using small beam widths is that the ping ratemight have to be increased or the vessel slowed to maintain anoverlap between subsequent samples. The effect of non-overlapping, along-track samples appears for the herring schoolsituated in shallow water (Figure 7). For large schools, unsampledareas along the survey track are not a problem, but for smallschools, this might lead to biased estimates of school-morphologyparameters.

Mixed species associationsPelagic species form mixed or closely located schools in many shelfseas. For example, in the Bay of Biscay, anchovy (Engraulisencrasicolus), horse mackerel (Trachurus trachurus), mackerel(Scomber scombrus), sprat (Sprattus sprattus), and pilchard(Sardina pilchardus) are commonly found together in varyingspecies combinations related to geographic location (Masse, 1996).Tropical pelagic species are also often found in mixed schools,with individuals of similar size but different species occurringtogether (Freon, 1984). Similar observations have been made inthe Bay of Biscay, where small horse mackerel were caught togetherwith large anchovy of similar size (VMT, unpublished data).

Many attempts have been made to identify monospecific schoolsbased on morphological (school length, perimeter, density, etc.) andcontextual information (location, depth, time of day, etc.; Scalabrinet al., 1996; LeFeuvre et al., 2000; Lawson et al., 2001). Although this

. . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . .

Table 2. Details of the datasets used in this study and of the ME70 configuration used.

Dataset Start date and time Depthrange (m)

Distance(nauticalmiles)

Vesselspeed(knots)

Numberof beams

Frequencyset-up

Beam width (88888) Analysisthreshold(dB)

Herringschool

25 February 2006 04:30 30 2 4.5 17 V+ 408 5 central; 4 outer 260 dB

Mixedspecies

03 March 2007 14:00– 18:00 80 –200 15 �6 15 V+ 258 4.2 central; 3.3 outer 260 or280 dB


by guest on Decem



ownloaded from


approach has allowed species classifications for particular datasets,its general applicability seems to be hampered by the fact thatschool morphology varies with environmental characteristics

(Scalabrin and Masse, 1993; Muino et al., 2003; Soria et al.,2003), fish size (Iglesias et al., 2003), and maturation state(Mackinson, 1999), perhaps even more so than species.

Figure 6. Single-beam views and projected two-dimensional view for three central beams of the ME70 compared with the ER60 (70 kHz)results for schools in a mixed-species dataset. Units on the two-dimensional views are in metres.

Figure 7. Herring schools. (a) Three-dimensional view; (b) lateral view of the three-dimensional school. Direction of travel from left to right inboth panels. Black axes represent the coordination system.


by guest on Decem



ownloaded from


At least two alternative strategies have been proposed toadvance acoustic-species classification: inclusion of the thirddimension for morphological-school description (Fernandeset al., 2002), and use of multifrequency information (see reviewin Misund, 1997). So far, broad groups, e.g. plankton, fish withand without swimbladders, have been distinguished successfullybased on differential sound-scattering over a range of discreteechosounder frequencies (e.g. 18, 38, 120, and 200 kHz;Korneliussen and Ona, 2003), or using the response differencebetween 38 and 120 kHz (Woodd-Walker et al., 2003).

The ME70 is not only multibeam (multi-angle), but also multi-frequency in the range 70–120 kHz. This frequency range may beuseful for collecting information needed for species identification.Multifrequency data from the vertical and steered beams are froma single transducer array and do not suffer from the commonproblem of horizontally displaced transducers not samplingexactly the same scene (Korneliussen and Ona, 2002; Korneliussenet al., 2008). Tank measurements have shown that anchovy andsardine TS values do not vary much between 70 and 120 kHz(flat frequency response), although the absolute levels differbetween the two species (Conti and Demer, 2003). This might bea general result. Simmonds and Armstrong (1990) found that itwas possible to distinguish between herring, mackerel, andgadoids in the range 27–54 kHz, although they did not investigatehigher frequencies. From the literature, we might expect the fre-quency response of many species to be flat over the range coveredby the ME70 (70–120 kHz), which would be favourable forabundance estimation, but ineffective for species classification.Therefore, the new information provided by the ME70 might bemost effective if used in combination with data over a widerrange of frequencies from single-beam echosounders. A subset (3nautical miles) of the mixed-species dataset illustrates the potentialof combining information from the ME70 and ER60s (Figure 8).The example includes both diffuse and dense shoals. The frequencyresponse derived from ER60 data could be combined with

information from the ME70 on fish-density distribution withinshoals and shoal morphology. The three-dimensional shape of theflat layer on the right side of Figure 8, though unidentified, maybe indicative of a single species. However, frequency responses, par-ticularly in shallow water, are often variable because of the reducedoverlap of the volume sampled by the different frequency ER60s,which is caused by the position of the transducers on the hullof the ship, maximum distance 1 m on “Thalassa”, but alsovariable beam widths, when considering 18 kHz. In deeper water,the difference in steering angle of the transducers (maximum18 on “Thalassa”) can also lead to some variability in thefrequency-response curve. The roll-and-pitch-stabilized three-dimensional image obtained with the ME70 will be valuable indetermining which echoes were well sampled by all frequencies,and consequently for filtering the data to obtain more accurate fre-quency response curves. In addition, improved in situ single-targetmeasurements, from the narrower beam widths and the simul-taneous use of different frequencies, as proposed by Demer et al.(1999), might also help species identification.

Individual and school behaviourThe potential impact of avoidance behaviour on abundanceestimates derived from acoustic measurements has long been aconcern (Olsen et al., 1983; Gerlotto and Freon, 1992; Vabøet al., 2002). Reactions to an approaching survey vessel havebeen observed using multibeam echosounders (Gerlotto et al.,1999; Gerlotto and Paramo, 2003; Soria et al., 2003). In contrast,Fernandes et al. (2000) compared acoustic measurements fromthe FRV “Scotia”, which was built to ICES reduced-noise stan-dards, with those from an autonomous underwater vehicle, andconcluded that herring did not seem to avoid noise-reducedsurvey vessels. However, Ona et al. (2007) found that such avessel induced stronger vertical-diving reactions in herring thana conventional survey vessel. Using such a conventional surveyvessel, Gerlotto and Paramo (2003) found that fish schools near

Figure 8. Three-dimensional view of mixed shoals of mackerel, horse-mackerel, and other species in the mixed-species dataset.


by guest on Decem



ownloaded from


a vessel track, most likely round sardinella, were nearly twice aslong as they were wide, whereas those located �30 m from thetrack line had similar width and length. Vertical reactions (i.e.diving responses) have also been observed for anchovetta(Gerlotto and Freon, 1992), herring (Vabø et al., 2002), and cod(Handegard and Tjøstheim, 2005).

The ME70 may contribute to behavioural studies of individualfish and schools, be it natural or in reaction to a vessel. Aspreviously demonstrated, three-dimensional measures of fishschools as a function of horizontal distance from the vessel trackallow behaviour to be visualized and quantified. Consider againthe herring-school example (Figure 7). The school might havereacted to the vessel, because individuals on the vessel track werelower in the water than those to the side (Figure 7a). Moreover,as the vessel progressed over the school, the depths of individualsincreased (Figure 7b). As all ME70 beams can be configured assplit beam, the movements of individual fish along and acrossthe track, i.e. within and across beams, can be measured andtracked. Studies of fish movements might facilitate progress oncharacterizing fish-reaction behaviour and the remote identifi-cation of species or age groups. For example, Kang et al. (2006)found that the variability of swimming angles within schools ofwalleye pollock (Theragra chalcogramma) was related to thedominant age group present in the school.

The configuration of steering angles of the ME70 fan is some-what flexible, ranging from all beams (with different frequencies)pointing in the same direction (“I” configuration, see above), toeach beam having a separate angle and also a separate frequency.A configuration with multiple “I”s, i.e. the same range of frequen-cies at different steering angles, is not possible, but multiple,slightly tilted “I”s, with non-overlapping frequency ranges, arepossible. Such a configuration would allow exploration ofwhether changes in tilt angle as a result of avoidance reactionslead to changes in backscattering, which differ between steeringangles, as derived theoretically by Cutter and Demer (2007). Intheir simulation study, in which all beams had the same frequency,those authors found that sideways and diving responses of fishschools were visible as a difference in backscattering strengthbetween central and steered beams (.308). As the ME70 beamshave different frequencies, the overlaying frequency responsemight blur the picture somewhat.

Further, the potential biases in abundance estimates attribu-table to school reactions can be evaluated more directly. Thevalues calculated for the shoals in Figure 7 were at least fourtimes as high in the outer beams (athwartship angle+ 208) thanin the more central beams.

Trawl efficiency is another topic that might be evaluated fromME70 observations, for example using observations similar to

those shown in Figure 9. When the survey vessel passed over afishing trawl in action, individual fish were clearly visible in thetrawl opening (Figure 9b). Therefore, assessing the number offish inside and outside the trawl might allow trawl efficiency tobe estimated.

The next stepsAs shown here, the calibrated multibeam echosounder ME70is a powerful complement to other sampling devices, such astraditional multifrequency echosounders. However, to achieveits maximum potential, a few technical challenges need to besurmounted. For example, reliable bottom detection must beachieved in all beams, a problem particularly acute for the outerbeams where the non-oblique, incidence angle reduces bottomcontrast. Moreover, tools for visualizing the ME70 data in two,three, and four dimensions are essential (Mayer et al., 2002).

The dependence of Sv and TS on frequency and incidence angleneed further exploration and quantification. Hazen and Horne(2003) found that tilt-angle variation was the most importantfactor for changes in simulated TS values, even when comparedwith body length or depth, and its importance was confirmed byexperimental work on krill (McGehee et al., 1998).

The current ME70 provides three-dimensional views forfish schools. However, research is needed to develop pertinentdescriptors for the amorphous shapes of fish schools which arenot well described by spherical or cylindrical bodies (Gerlottoand Paramo, 2003); school width and length metrics are not suffi-cient to capture the entire geometry.

Algorithms also need to be developed to connect shoals andindividual fish across ER60 and the ME70 beams. Improvedprecision for tracking of individual fish pursuing the approachproposed by Handegard (2007), who uses several types of infor-mation simultaneously, e.g. phase angles, echo intensity, ranges,and times, will provide further insights into fish behaviour andits effects on biomass estimation. It will be necessary to track indi-vidual fish not only within a beam but also across beams, whichmight be complicated by the potential impact of observationangle and frequency on TS values.

The ME70 is operational for studying the pelagic zone.However, a noise problem currently limits its application in thelayer extending to 75 cm above the seabed. This problem may besurmounted with enhanced beam-forming techniques and fastercomputers for real-time processing.

AcknowledgementsWe thank Simrad for fruitful collaboration, and Noel Diner forinitiating the development of the ME70. The development of theME70 was supported by the European funding scheme IFOP.

Figure 9. Acoustic image of commercial pelagic trawl crossed by survey vessel: (a) three-dimensional view; (b) lateral view.


by guest on Decem



ownloaded from


We are also grateful to Niels Olav Handegard, Dave Demer, ChrisWilson, and an anonymous referee for useful comments on anearlier version of this manuscript.

ReferencesAglen, A., Engas, A., Huse, I., Michalsen, K., and Stensholt, B. K. 1999.

How vertical fish distribution may affect survey results. ICESJournal of Marine Science, 56: 345–360.

Andersen, L. N., Berg, S., Gammelster, O. B., and Lunde, E. B. 2006.New scientific multibeam systems (ME70 and MS70) forfishery-research applications. Journal of the Acoustical Society ofAmerica, 120: 3017.

Benoit-Bird, K. J., Wursig, B., and McFadden, C. J. 2004. Duskydolphin (Lagenorhynchus obscurus) foraging in two different habi-tats: active acoustic detection of dolphins and their prey. MarineMammal Science, 20: 215–231.

Berger, L., Durand, C., and Marchalot, C. 2005. Moviesþ UserManual Version 4.3. Ifremer. 72 pp.

Brehmer, P., Lafont, T., Georgakarakos, S., Josse, E., Gerlotto, F., andCollet, C. 2006. Omnidirectional, multibeam, sonar-monitoring:applications in fisheries science. Fish and Fisheries, 7: 165–179.

Conti, S. G., and Demer, D. A. 2003. Wide-bandwidth acousticalcharacterization of anchovy and sardine from reverberationmeasurements in an echoic tank. ICES Journal of MarineScience, 60: 617–624.

Cutter, G. R., and Demer, D. A. 2007. Accounting for scattering direc-tivity and fish behaviour in multibeam-echosounder surveys. ICESJournal of Marine Science, 64: 1664–1674.

Demer, D. A., Soule, M. A., and Hewitt, R. P. 1999. A multiple-frequency method for potentially improving the accuracy andprecision of in situ target-strength measurements. Journal of theAcoustical Society of America, 105: 2359–2376.

Diner, N. 2001. Correction on school geometry and density: approachbased on acoustic-image simulation. Aquatic Living Resources, 14:211–222.

Diner, N. 2007. Evaluating uncertainty in measurements of fish-shoal,aggregate-backscattering cross section caused by small shoal sizerelative to beam width. Aquatic Living Resources, 20: 117–121.

Fernandes, P. G., Brierley, A. S., Simmonds, E. J., Millards, N. W.,McPhail, S. D., Armstrong, F., Stevenson, P., et al. 2000. Fish donot avoid survey vessels. Nature, 404: 35–36.

Fernandes, P. G., Gerlotto, F., Holliday, D. V., Nakken, O., andSimmonds, E. J. 2002. Acoustic applications in fisheries science:the ICES contribution. ICES Journal of Marine Science, 215:483–492.

Foote, K. G. 2006. Optimizing two targets for calibrating a broadband,multibeam sonar. Oceans, 2006: 1–4.

Foote, K. G., Knudsen, H. P., Vestnes, G., MacLennan, D. N., andSimmonds, E. J. 1987. Calibration of acoustic instruments for fish-density estimation: a practical guide. ICES Cooperative ResearchReport, 144. 57 pp.

Freon, P. 1984. La variabilite des tailles individuelles a l’interieur descohortes et des bancs de poissons. I: Observations et interpretation.Oceanologica Acta, 7: 457–468.

Gerlotto, F., and Freon, P. 1992. Some elements on vertical avoidanceof fish schools to a vessel during acoustic surveys. FisheriesResearch, 14: 251–259.

Gerlotto, F., and Paramo, J. 2003. The three-dimensional morphologyand internal structure of clupeid schools as observed usingvertical-scanning, multibeam sonar. Aquatic Living Resources,16: 113–122.

Gerlotto, F., Soria, M., and Freon, P. 1999. From two dimensions tothree: the use of multibeam sonar for a new approach in fisheriesacoustics. Canadian Journal of Fisheries and Aquatic Sciences,56: 6–12.

Handegard, N. O. 2007. Observing individual fish behaviour in fishaggregations: tracking in dense fish aggregations using a split-beamechosounder. Journal of the Acoustical Society of America, 122:177–187.

Handegard, N. O., and Tjøstheim, D. 2005. When fish meet a trawlingvessel: examining the behaviour of gadoids using a free-floatingbuoy and acoustic split-beam tracking. Canadian Journal ofFisheries and Aquatic Sciences, 62: 2409–2422.

Hazen, E. L., and Horne, J. K. 2003. A method for evaluating the effectsof biological factors on fish target strength. ICES Journal of MarineScience, 60: 555–562.

ICES. 2000. Report on echo-trace classification. ICES CooperativeResearch Report, 238. 115 pp.

ICES. 2005. Description of the ICES HAC standard data exchangeformat, version 1.60. ICES Cooperative Research Report, 278. 86pp.

Iglesias, M., Carrera, P., and Muino, R. 2003. Spatio-temporal patternsand morphological characterization of multispecies, pelagic-fishschools in the north-western Mediterranean Sea. Aquatic LivingResources, 16: 541–548.

Kang, M., Honda, S., and Oshima, T. 2006. Age characteristics ofwalleye-pollock school echoes. ICES Journal of Marine Science,63: 1465–1476.

Korneliussen, R. J., Diner, N., Ona, E., Berger, L., and Fernandes, P. G.2008. Proposals for the collection of multifrequency acoustic data.ICES Journal of Marine Science, 65 (in press).

Korneliussen, R. J., and Ona, E. 2002. An operational system for pro-cessing and visualizing multi-frequency acoustic data. ICES Journalof Marine Science, 59: 293–313.

Korneliussen, R. J., and Ona, E. 2003. Synthetic echograms generatedfrom the relative frequency response. ICES Journal of MarineScience, 60: 636–640.

Lawson, G. L., Barange, M., and Freon, P. 2001. Species identificationof pelagic fish schools on the South African continental shelf usingacoustic descriptors and ancillary information. ICES Journal ofMarine Science, 58: 275–287.

LeFeuvre, P., Rose, G. A., Gosine, R., Hale, R., Pearson, W., and Khan,R. 2000. Acoustic-species identification in the Northwest Atlanticusing digital-image processing. Fisheries Research, 47: 137–147.

Mackinson, S. 1999. Variation in structure and distribution of pre-spawning Pacific herring shoals in two regions of BritishColumbia. Journal of Fish Biology, 55: 972–989.

Masse, J. 1996. Acoustic observations in the Bay of Biscay: schooling,vertical distribution, species assemblages and behaviour. ScientiaMarina, 60(Suppl. 2): 227–234.

Mayer, L., Li, Y., and Melvin, G. 2002. 3D visualization for pelagic fish-eries research and assessment. ICES Journal of Marine Science, 59:216–225.

McGehee, D. E., O’Driscoll, R. L., and Martin Traykovski, L. V. 1998.Effects of orientation on acoustic scattering from Antarctic krill at120 kHz. Deep Sea Research II, 45: 1273–1294.

Melvin, G. D., Cochrane, N. A., and Li, Y. 2003. Extraction and com-parison of acoustic backscatter from a calibrated multi- and single-beam sonar. ICES Journal of Marine Science, 60: 669–677.

Misund, O. A. 1997. Underwater acoustics in marine fisheries and fish-eries research. Reviews in Fish Biology and Fisheries, 7: 1–34.

Misund, O. A., and Aglen, A. 1992. Swimming behaviour of fishschools in the North Sea during acoustic surveying and pelagictrawl sampling. ICES Journal of Marine Science, 49: 325–334.

Mitson, R. B., and Knudsen, H. P. 2003. Causes and effects of under-water noise on fish-abundance estimation. Aquatic LivingResources, 16: 255–263.

Muino, R., Carrera, P., Petitgas, P., Beare, D. J., Georgakarakos, S.,Haralambous, J., Iglesias, M., et al. 2003. Consistency in the corre-lation of school parameters across years and stocks. ICES Journal ofMarine Science, 60: 164–175.


by guest on Decem



ownloaded from


Olsen, K., Angell, J., and Løvik, A. 1983. Quantitative estimations ofthe influence of fish behaviour on acoustically determined fishabundance. FAO Fisheries Report, 300: 139–149.

Ona, E. 2003. An expanded target-strength relationship for herring.ICES Journal of Marine Science, 60: 493–499.

Ona, E., Godø, O. R., Handegard, N. O., Hjellvik, V., Patel, R., andPedersen, G. 2007. Silent research vessels are not quiet. Journalof the Acoustical Society of America, 121: 145–150.

Ona, E., and Mitson, R. B. 1996. Acoustic sampling and signal proces-sing near the seabed: the deadzone revisited. ICES Journal ofMarine Science, 53: 677–690.

Petitgas, P. 2003. A method for the identification and characterizationof clusters of schools along the transect lines of fisheries-acousticsurveys. ICES Journal of Marine Science, 60: 872–884.

Reid, D., Scalabrin, C., Petitgas, P., Masse, J., Aukland, R., Carrera, P.,and Georgakarakos, S. 2000. Standard protocols for the analysis ofschool-based data from echosounder surveys. Fisheries Research,47: 125–136.

Scalabrin, C., Diner, N., Weill, A., Hillion, A., and Mouchot, C. 1996.Narrowband acoustic identification of monospecific fish schools.ICES Journal of Marine Science, 53: 181–188.

Scalabrin, C., and Masse, J. 1993. Acoustic detection of the spatial andtemporal distribution of fish shoals in the Bay of Biscay. AquaticLiving Resources, 6: 269–283.

Simmonds, E. J., and Armstrong, D. A. 1990. A wideband echosoun-der: measurements on cod, saithe, herring and mackerel from 27to 54 kHz. Rapports et Proces-Verbaux des Reunions du ConseilInternational pour l’Exploration de la Mer, 189: 183–187.

Simmonds, E. J., and MacLennan, D. N. 2005. Fisheries Acoustics.Theory and Practice, 2nd edn. Blackwell, Oxford. 437 pp.

Soria, M., Bahri, T., and Gerlotto, F. 2003. Effect of external factors(environment and survey vessel) on fish-school characteristicsobserved by echosounder and multibeam sonar in theMediterranean Sea. Aquatic Living Resources, 16: 145–157.

Vabø, R., Olsen, J., and Huse, I. 2002. The effect of vessel avoidance ofwintering Norwegian spring spawning herring. Fisheries Research,58: 59–77.

Wilson, C. D., Hollowed, A. B., Shima, M., Walline, P., and Stienessen,S. 2003. Interactions between commercial fishing and walleyepollock. Alaska Fishery Research Bulletin, 10: 61–77.

Woodd-Walker, R. S., Watkins, J. L., and Brierley, A. S. 2003.Identification of Southern Ocean acoustic targets using aggregationbackscatter and shape characteristics. ICES Journal of MarineScience, 60: 641–649.

Zedel, L., Patro, R., and Knutsen, T. 2005. Fish behaviour andorientation-dependent backscatter in acoustic Doppler profilerdata. ICES Journal of Marine Science, 62: 1191–1201.

doi:10.1093/icesjms/fsn051


by guest on Decem



ownloaded from


Documents

The new fisheries multibeam echosounder ME70: description and expected contribution to fisheries research