High-resolution Multibeam Sonar Research Articles

Seafloor Observations Eliminate a Landslide as the Source of the 1918 Puerto Rico Tsunami

ABSTRACT The 11 October 1918 devastating tsunami in northwest Puerto Rico had been used as an example for earthquake-induced landslide tsunami hazard. Three pieces of evidence pointed to a landslide as the origin of the tsunami: the discovery of a large submarine landslide scar from bathymetry data collected by shipboard high-resolution multibeam sonar, reported breaks of submarine cable within the scar, and the fit of tsunami models to flooding observations. Newly processed seafloor imagery collected by remotely operated vehicle (ROV) show, however, pervasive Fe–Mn crust (patina) on the landslide walls and floor, indicating that the landslide scar is at least several hundred years old. C14 dates of sediment covering the landslide floor verify this interpretation. Although we have not searched the region systematically for an alternative tsunami source, we propose a possible source—a two-segment normal-fault rupture along the eastern wall of Mona rift. The proposed fault location matches the published normal faults with steep bathymetry and is close to the International Seismological Center–Global Earthquake Model catalog locations of the 1918 mainshock and aftershocks. The ROV observations further show fresh vertical slickensides and rock exposure along the proposed fault trace. Hydrodynamic models from an Mw 7.2 earthquake rupture along the eastern wall of the rift faithfully reproduce the reported tsunami amplitudes, polarities, and arrival times. Our analysis emphasizes the value of close-up observations and physical samples to augment remote sensing data in natural hazard studies.

Geomorphological drivers of deeper reef habitats around Seychelles

Mesophotic (30–150 m) and rariphotic (150–300 m) deeper reef habitats are important from an ecological and conservation perspective, yet remain understudied. Key knowledge gaps exist on the environmental patterns and processes that drive and shape their geographical distributions. Understanding these is particularly important for regions as the Western Indian Ocean, where deeper reefs are poorly known but support food security and host economically important species. Spatial predictive models of assemblage occurrences, using terrain variables as predictors, offer a solution to address knowledge gaps around deeper reef distributions. We identified relationships between seafloor geomorphology, quantified at multiple scales, and sessile benthic assemblages in four atoll seascapes in Seychelles using terrain models derived from high-resolution multibeam sonar and underwater video surveys. Using random forests and boosted regression trees, we demonstrated that terrain derivatives extracted over multiple scales perform as reliable predictors of deeper reef assemblages. The most influential environmental predictors were depth, distance to shore, topographic complexity, slope and curvature and substrate characteristics. The relative importance of predictors was explained by assemblage functional characteristics. Assemblage–environment relationships were used to produce probability distribution maps that showed similar distributional patterns for identified assemblages across locations, with high occurrence probabilities linked to complex geomorphological structures. Our results help contribute to a consistent baseline understanding of the relationship between seascape structure and mesophotic reef ecosystems in this area. Complex geomorphological structures, including terraces and paleoshorelines, supported high densities of mesophotic assemblages and could be considered priority habitats for management.

Short-Range Movement Pattern of Amphidromous Lagoon Fish Schools: Ecological Applications

High-resolution multibeam sonar allows estimating movements of pelagic fish schools at short range. Taking advantage of this methodology, we calculated a Straightness Index (SI) to quantify the proportion of schools migrating actively from those residents in lagoon channels. This information enhances our knowledge of both fish school displacements and migration processes, which are essential to improve our understanding of ecosystem functioning. Most fish schools (65%) exhibited a SI value demonstrating oriented swimming behavior through the channels displayed by schools reaching the sea during fall migration. This trend appears as an intrinsic property of school movements, allowing monitoring of the school migration process in a channel to provide information for manager vs. fishing regulation measures or lagoon planning. The result strengthens the ‘multi-transit’ hypothesis, as 35% of schools show sinuous trajectories representative of schools staying in the channel or displaying high exploratory behaviors. Lastly, the fish school Exploration Swimming Speed (ESS) was tested as a fishery-independent sampling method to evaluate the proportion of different fish species monitored using hydroacoustics. This approach demonstrates the interest in using swimming behavioral characteristics of fish schools for ecological and management purposes.

Heterogeneity of rhodolith beds expressed in backscatter data

Rhodoliths play a major role as benthic habitats and are important calcium carbonate bio-factories. Several studies have acoustically mapped rhodolith beds as a homogeneous habitat but spatial changes in nodule coverage, or density, commonly occur. Here, the objective is to acoustically recognize heterogeneities in rhodolith bed structure using the response of backscatter to map changes in nodule coverage and density. High-resolution multibeam sonar (MBSS) data were acquired on April 2018, in the Marine Protection Area (MPA) Costas das Algas, located in the Espírito Santo continental shelf, Eastern Brazilian coast. Three areas were selected based on available data, totalizing 73km2 of data acquisition. Ground truth was obtained at 80 stations, using a drop camera system. Seabed classification was determined by applying an image segmentation process on multibeam sonar backscatter mosaic. Ground truth data was used to statistically show the relation between backscatter and seabed types. Results identified three classes related to rhodoliths in different coverage densities: low rhodolith coverage (inferior to 25% of rhodoliths over a determined area), moderate rhodolith coverage (between 25% and 35% of rhodoliths), and high rhodolith coverage (>35%). Classes associated with sediment, bioconcretions and red algae genus Peyssonelias were also identified. A map showing changes in rhodolith coverage across the shelf depth gradient were produced. The seabed classes had a statistically significant correlation with the segmented backscatter classes. In general, higher rhodolith coverage correlates with higher backscatter. The backscatter mosaic over the area is marked by an extensive presence of the red algae Peyssonelias, resulting in a lower correlation index, when compared to the areas with no Peyssonelia. The seabed mapping approach used in this work provided an efficacious tool to map benthic habitats and to detail nodules distribution across a rhodolith bed. Moreover, the recognition of heterogeneities in rhodolith bed structure provides important information for monitoring and protecting these vulnerable marine ecosystems. The main driver of the coralligenous algae nodule coverage, in the research area, seems to be the morphology of the area and the influence of mesoscale circulation.

Potential links between Baltic Sea submarine terraces and groundwater seeping

Abstract. Submarine groundwater discharge (SGD) influences ocean chemistry, circulation, and the spreading of nutrients and pollutants; it also shapes sea floor morphology. In the Baltic Sea, SGD was linked to the development of terraces and semicircular depressions mapped in an area of the southern Stockholm archipelago, Sweden, in the 1990s. We mapped additional parts of the Stockholm archipelago, areas in Blekinge, southern Sweden, and southern Finland using high-resolution multibeam sonars and sub-bottom profilers to investigate if the sea floor morphological features discovered in the 1990s are widespread and to further address the hypothesis linking their formation to SGD. Sediment coring and sea floor photography conducted with a remotely operated vehicle (ROV) and divers add additional information to the geophysical mapping results. We find that terraces, with general bathymetric expressions of about 1 m and lateral extents of sometimes >100 m, are widespread in the surveyed areas of the Baltic Sea and are consistently formed in glacial clay. Semicircular depressions, however, are only found in a limited part of a surveyed area east of the island of Askö, southern Stockholm archipelago. While submarine terraces can be produced by several processes, we interpret our results to be in support of the basic hypothesis of terrace formation initially proposed in the 1990s; i.e. groundwater flows through siltier, more permeable layers in glacial clay to discharge at the sea floor, leading to the formation of a sharp terrace when the clay layers above seepage zones are undermined enough to collapse. By linking the terraces to a specific geologic setting, our study further refines the formation hypothesis and thereby forms the foundation for a future assessment of SGD in the Baltic Sea that may use marine geological mapping as a starting point. We propose that SGD through the submarine sea floor terraces is plausible and could be intermittent and linked to periods of higher groundwater levels, implying that to quantify the contribution of freshwater to the Baltic Sea through this potential mechanism, more complex hydrogeological studies are required.

Hand gesture recognition from multibeam sonar imagery

Abstract: Divers perform demanding tasks in a complex and hazardous underwater environment, which prevents them from carrying special devices that may allow them to communicate with their robotic diving buddies. In this world of natural human-robot interaction in the underwater environment, envisioned by the FP7 Cognitive Robotics project CADDY, hand detection and gesture interpretation is a prerequisite. While hand gesture recognition is most often performed with cameras (mono and stereo), their use in the underwater environment is compromised due to water turbidity and lack of sunlight at greater depths. This paper deals with this lack of performance by introducing the concept of using high resolution multibeam sonars (often referred to as acoustic cameras) for diver hand gesture recognition. In order to ensure reliable communication between the diver and the robot, it is of great importance that the classification precision is as high as possible. This paper presents results of hand gesture recognition which is performed by using two approaches: convex hull method and the support vector machine (SVM). A novel approach that fuses the two methods is introduced as a way of increasing the precision of classification. The results obtained on more than 1000 real sonar samples show that the precision using the convex hull method is around 92%, and using the SVM around 94%, while fusing the two approaches provides around 99% classification precision.

Late Wisconsinan grounding-zone wedges, northwestern Gulf of St Lawrence, eastern Canada

Grounding-zone wedges (GZWs) are formed on high-latitude continental shelves by the accumulation of subglacial sediments from fast-flowing ice at the grounding zone of an ice sheet during temporary stillstands (e.g. Mosola & Anderson 2006; Dowdeswell & Fugelli 2012; Batchelor & Dowdeswell 2015). GZWs are therefore recognized to record phases of episodic ice-sheet retreat (Dowdeswell et al. 2008). The morphology and distribution of a series of GZWs mapped using high-resolution multibeam sonar and airgun seismic data in the northwestern Gulf of St Lawrence, eastern Canada, are described here (Fig. 1a–d). These ice-contact submarine landforms were deposited when the marine-based southeastern margin of the Laurentide Ice Sheet stabilized for a few decades to centuries during its rapid retreat through the Gulf of St Lawrence. Fig. 1. Grounding-zone wedges (GZW) in the northwestern Gulf of St Lawrence, eastern Canada. ( a ) Location of study area (red box; map from GEBCO_08). ( b ) Oblique view of grounding-zone wedge at 180 m water depth (GZW 1) illustrating circular depressions along its steeper down-ice slope. ( c ) Seismic-reflection profile with two GZWs. Seismic profile shows bedrock surface which consists of gently dipping monoclinal Middle Ordovician sedimentary rocks of the St Lawrence Platform (Sanford 1993). Profile is located …

Enigmatic ridges in Lake Vättern, Sweden

The latest generation of high-resolution multibeam sonars reveal ever smaller subaqueous landforms, particularly in shallow marine and lake environments. High-resolution bathymetric data from the floor of Lake Vattern, Sweden, reveal fine-scale landforms on a bouldery sediment surface, which we attribute to ice-related processes. We explore subglacial, grounding-zone, crevasse infill and push/scour processes of grounded and floating ice in an attempt to explain the origin of these enigmatic features. Lake Vattern in south-central Sweden is a deep, elongate rift basin which lies amid regional landform and sediment assemblages produced during the retreat of the last Fennoscandian Ice Sheet (FIS). The regional retreating ice margin terminated in the large proglacial Baltic Ice Lake, to which the Vattern Basin was connected before its eventual drainage (Lundqvist & Wohlfarth 2001). A recent high-resolution multibeam survey of sectors of the Vattern lake floor revealed landforms associated with marginal ice grounding and retreat. On a mid-lake shoal at 30–40 m depth, a suite of fine-scale, enigmatic landforms have been formed in stiff sediment (Fig. 1). Fig. 1. ( a ) Multibeam bathymetry from Lake Vattern showing 1–2 m amplitude sediment ridges that are formed on a shoal, which comprise both parallel, linear components (L, L*) and sinuous ridges (S, S*) of …

High-Resolution 3D Laser Imaging Advances to Subcentimeter Precision

This article, written by Editorial Manager Adam Wilson, contains highlights of paper OTC 23006, ’SS MTS: Subsea Monitoring - High-Resolution 3D Laser Imaging for Inspection, Maintenance, Repair, and Operations,’ by C. Embry, M. Hardy, and B. Nickerson, 3D at Depth; N. Manning, CDL; D. Goodyear, UTEC Survey; and D. Richardson, SPE, and J. Pappas, SPE, RPSEA, prepared for the 2012 Offshore Technology Conference, Houston, 30 April-3 May. The paper has not been peer reviewed. 3D laser imaging is a powerful data-collection system that provides 3D information for a specific region of interest. Developing the technology to provide high-definition subsea laser imaging enables the deepwater industry to use the current state of the art in 3D metrology and related best practices developed for the terrestrial market. Terrestrial laser scanners commonly produce centimeter spatial and range accuracy at the several-hundred-meter range. Because of the absorption of water, realizable deepwater systems are limited to tens-of-meters range, depending on the target and water conditions. Here, subcentimeter accuracy from greater than 6 m is demonstrated for an underwater laser system. Introduction Sonar systems have been the main tools for measurement and survey for many years and have become very sophisticated. Many companies have developed high-resolution multibeam sonar systems that can produce angular resolutions as small as 0.2°. However, the resolution of these systems is fundamentally limited by the underlying physics; a wave-length of light is more than 108 times smaller than ultrasonic wavelengths and can achieve angular resolutions of less than 100 Μrad (0.006°). Therefore, light-based sensors inherently have higher spatial resolutions than sonar systems. The underwater 3D laser sensor presented here has the fundamental advantage of providing subcentimeter precision from greater than 7 m, a capability that, to the authors’ knowledge, has not been demonstrated previously. The Sensor The optical head [laser imaging unit (LIU)] consists of the transmitter subsystem (laser and accompanying optics), the receiver subsystem (detector and accompanying optics), and the scanner. These subsystems can be optimized for different operational scenarios and environ-mental conditions. For this prototype system, we used a diode-pumped, passively Q-switched, Nd:YAG laser with an external nonlinear crystal to perform the frequency doubling to 532-nm laser light. The receiver is a high-speed silicon photo diode, and the scanning subsystem is a two-axis galvanometer scanner to allow programmable scanning in two orthogonal directions. Along with the LIU are multiple electronic control systems. These include the control subsystem for controlling the LIU and monitoring-system health; the central processing unit and bus for general system control and operation; the data subsystem for data collection, transformation, and processing; the data-storage system for housing data locally; the power and environmental monitoring system to convert the host-vehicle power to the various voltages required by the sensor; and the input/output (communication) subsystem. All of these systems are housed within the subsea sensor (Fig. 1). The prototype sensor is 8.2 in. in diameter and 28 in. long and weighs 76.1 lbm in air.

Infaunal biodiversity and ecological function on a remote oceanic island: The role of biogeography and bio-physical surrogates

Understanding and predicting the processes determining biological assemblages and marine biodiversity is critical to managing and conserving marine ecosystems. Infauna are a substantial component of shelf biodiversity and important contributors to ecological function. To examine the bio-physical relationships structuring infaunal assemblages across the Lord Howe Island (LHI) shelf, we mapped broad-scale physical and habitat-related features of the seabed using high-resolution multibeam sonar and sampled sediments and infauna across the shelf using a Smith-McIntyre grab. Broad-scale biogeographic processes were important determiners of the infaunal shelf assemblage. Shelf sediments supported an impoverished infaunal assemblage, with a high proportion of rare species, and many new and endemic species. This, combined with the numerical dominance of brooding endemics relative to lower densities of mainland taxa suggests that much of the islands infaunal diversity reflects transient dispersers rather than self-recruiting island populations. Local-scale physical processes, such as disturbance and current-borne nutrients, were also important in structuring the infaunal assemblage. Three geomorphic zones (drowned lagoon, relict reef and outer shelf) were strong predictors (or surrogates) of infaunal community and trophic structure over broad-scales (the extent of the shelf), while sediment composition (e.g. sorting and mean grain size) and seafloor structure (e.g. topography) were important predictors of population abundances and finer-scale (within-zone) community patchiness. Species richness and abundance were highest offshore on the outer shelf. Here, topographically raised sites subjected to oceanic currents supported high densities of suspension feeders and the highest levels of infaunal diversity. In contrast, the dynamic rippled sediments of the drowned lagoon supported an impoverished assemblage with a reduced trophic structure indicative of harsh physical environments. While biogeographic isolation in part explains low localised densities of non-endemic species, overall numbers per species indicate that many individuals are successfully arriving at LHI, but that sediment conditions – particularly in the drowned lagoon and to a lesser extent across the relict reef – may be too dynamic to provide suitable habitat for many of these species. These results suggest that interactions between shelf topography, physical wave disturbance and nutrient supply from oceanic currents are important factors structuring oceanic shelf assemblages. This study highlights that seabed habitat mapping when explicitly integrated with fine-scale bio-physical surveys can provide substantial insight into the spatial distribution and ecological function of benthic assemblages, and provides an essential ecological and spatial framework to predict the future structure of these populations and assemblages in response to natural and anthropogenic change.

Neuro-fuzzy Classification of Submarine Lava Flow Morphology

This paper presents a novel approach to semi-automated classification of volcanic morphology on the seafloor using high-resolution multibeam sonar bathymetry and side-scan sonar backscatter imagery. The classification methodology combines a fuzzy inference system and neural network theory in an adaptive neuro-fuzzy inference system (ANFIS) and is capable of rapidly classifying submarine lava morphology based on bathymetry-derived surface geometry and backscatter-derived attributes of acoustics and texture. The system has been applied to a study area on a seafloor spreading ridge, the Galapagos Spreading Center (GSC), in order to quantify the distribution and relative abundance of lava flow types, which can be used to indicate variations in eruption and emplacement dynamics. A detailed assessment shows the classification has an overall accuracy of almost 90 percent with a kappa coefficient of 0.84. The neuro-fuzzy method described here is shown to be an efficient and reliable tool for classification of submarine lava morphology.

Search for Bermuda’s deep water caves

The mid-Atlantic islands of Bermuda harbor one of the richest and most diverse anchialine communities known from anywhere on Earth. However, all known anchialine caves in Bermuda (maximum depth—26 m) were dry during the last glacial period extending from approximately 9,000 to 115,000 years ago when glacial sea levels were as much as 127 m lower. Since it is highly unlikely that Bermuda’s endemic cave species evolved since the caves were flooded by sea level rise, alternate deeper habitats must have existed to shelter anchiane fauna for prolonged periods of lower sea level during the Pleistocene. In order to systematically search for such now deep water cave habitats, high-resolution multibeam sonar and remotely operated vehicles were used to map and explore the seafloor off Bermuda in 60–200 m depths along the outer shelf break edge of the submarine escarpment surrounding the Bermuda Platform and an adjacent seamount. Specific goals were to discover deep water cave and/or crevicular habitats and to characterize the nature, geological stratification and composition, and sea level history of the platform margin, in particular focusing on features directly relating to Pleistocene low sea stand events. During this sea floor survey, clearly defined paleo-shoreline features generated by wave and current erosion were found to encircle the Bermuda seamount and Challenger Bank at 60 and 120 m depths.

Three‐dimensional internal spatial structure of young‐of‐the‐year pelagic freshwater fish provides evidence for the identification of fish school species

Active acoustic detection and characterization in three dimensions (3D) with multibeam sonars is a powerful technique for ecological studies of schooling fish. The alpine Lake Annecy provides ideal conditions for sampling fish with active acoustic methods: it has calm water, low species diversity, and the density of pelagic fish schools is high. We carried out investigations on the internal 3D morphological characteristics of young‐of‐the‐year (Y‐O‐Y) pelagic fish schools, using high resolution multibeam sonar in the vertical plane. The objective was to discriminate between the two fish species that school in the lake: perch (Perca fluviatilis) and roach (Rutilus rutilus). To discriminate between these Y‐O‐Y fish species, we used only one characteristic of their internal spatial structure: the total number of vacuoles, defined as empty volumes inside the school, relative to the school volume. Two distinct linear relationships between school volume and the total number of vacuoles were determined. These two types were consistent with the proportions of the percentage occurrence of the two different species as obtained from pelagic trawl sampling in 2004. In 2008, perch was dominant (93%) in the lake, and only one relationship was detected: this relationship corresponded closely to the one in 2004 attributed to perch. Thus, we assume that the schooling behavior is a phenotypic expression that could be used for remote fish species identification.

Integration and Testing of Multi-Purpose Platform Technologies System and High Resolution Multi-Beam Sonar on ROV Holland I

Abstract This paper describes the integration and testing of multi-purpose platform technologies system (MPPT-Ring), high resolution multi-beam sonar and state-of-the-art navigation system with the Ireland national work class ROV Holland I. In addition, the paper presents results of high resolution sea bed mapping along with the results of MPPT-Ring augmented reality support tools in offshore applications.

Discriminating between Posidonia oceanica meadows and sand substratum using multibeam sonar

Abstract Di Maida, G., Tomasello, A., Luzzu, F., Scannavino, A., Pirrotta, M., Orestano, C., and Calvo, S. 2011. Discriminating between Posidonia oceanica meadows and sand substratum using multibeam sonar. – ICES Journal of Marine Science, 68: 12–19. High-resolution, multibeam sonar (MBS) (455 kHz) was used to identify two typologies of seabed 8 m deep: Posidonia oceanica meadow and sandy substratum. The results showed that the heterogeneity of the architecture of the P. oceanica canopy and the relatively simple morphology of a sandy substratum can be detected easily by statistical indices such as standard deviation or range-of-beam depth. Based on these indices, an automated classification was performed for seabed mapping. The overall classification accuracy was as high as 99 and 98% in October and January, respectively. The probability that P. oceanica in situ was omitted on the map was <7%, whereas the probability that an area classified as P. oceanica on the map did not correspond to the seagrass in situ was consistently negligible. Based on these results, high-resolution MBS can be considered to be an accurate tool for mapping P. oceanica and sand substrata, and its discriminating power seems to be independent of season (autumn or winter).

High-Resolution Mapping of Mines and Ripples at the Martha's Vineyard Coastal Observatory

High-resolution multibeam sonar and state-of-the-art data processing and visualization techniques have been used to quantify the evolution of seafloor morphology and the degree of burial of instrumented mines and mine-shapes as part of the U.S. Office of Naval Research (ONR, Arlington, VA) mine burial experiment at the Martha's Vineyard Coastal Observatory (MVCO, Edgartown, MA). Four surveys were conducted over two years at the experiment site with a 455-kHz, Reson 8125 dynamically focused multibeam sonar. The region is characterized by shore-perpendicular alternating zones of coarse-grained sand with 5-25-cm-high, wave orbital-scale ripples, and zones of finer grained sands with smaller (2-5-cm-high) anorbital ripples and, on occasion, medium scale 10-20-cm-high, chaotic or hummocky bedforms. The boundaries between the zones appear to respond over periods of days to months to the predominant wave direction and energy. Smoothing and small shifts of the boundaries to the northeast take place during fair-weather wave conditions while erosion (scalloping of the boundary) and shifts to the north-northwest occur during storm conditions. The multibeam sonar was also able to resolve changes in the orientation and height of fields of ripples that were directly related to the differences in the prevailing wave direction and energy. The alignment of the small scale bedforms with the prevailing wave conditions appears to occur rapidly (on the order of hours or days) when the wave conditions exceed the threshold of sediment motion (most of the time for the fine sands) and particularly during moderate storm conditions. During storm events, erosional ldquowindowsrdquo to the coarse layer below appear in the fine-grained sands. These ldquowindowrdquo features are oriented parallel to the prevailing wave direction and reveal orbital-scale ripples that are oriented perpendicular to the prevailing wave direction. The resolution of the multibeam sonar combined with 3-D visualization techniques provided realistic looking images of both instrumented and noninstrumented mines and mine-like objects (including bomb, Manta, and Rockan shapes) that were dimensionally correct and enabled unambiguous identification of the mine type. In two of the surveys (October and December 2004), the mines in the fine-grained sands scoured into local pits but were still perfectly visible and identifiable with the multibeam sonar. In the April 2004 survey, the mines were not visible and apparently were completely buried. In the coarse-grained sand zone, the mines were extremely difficult to detect after initial scour burial as the mines bury until they present the same hydrodynamic roughness as the orbital-scale bedforms and thus blend into the ambient ripple field. Given the relatively large, 3-D, spatial coverage of the multibeam sonar along with its ability to measure the depth of the seafloor and the depth and dimensions of the mine, it is possible to measure directly, the burial by depth and burial by surface area of the mines. The 3-D nature of the multibeam sonar data also allows the direct determination of the volume of material removed from a scour pit.

Herring schooling manoeuvres in response to killer whale attacks

The antipredator behaviour of overwintering Norwegian spring-spawning herring (Clupea harengus L.) was investigated during repeated attacks by killer whales (Orcinus orca L.) in Tysfjord in northwestern Norway. The observations were made using a high-resolution (455 kHz) multibeam sonar. Ten different types of predator-prey interactions were recorded during 54 observed events (an average of one antipredator event every 3.9 min). Antipredator responses included "split," "hourglass," "vacuole," "bend," "dive," "herd," and "fountain." Large attacked schools demonstrated a different repertoire of antipredator manoeuvres than small ones and were less likely to be attacked. Despite being located in the vicinity of the whales, herring schools with a cross section exceeding 460 m2 were not attacked by killer whales. Attacked schools were significantly more circular (p < 0.0001) and had higher relative densities (p < 0.05) than schools that were not attacked.