Elsevier

Fisheries Research

Volume 243, November 2021, 106067
Fisheries Research

Full length article
Use of manned submersible and autonomous stereo-camera array to assess forage fish and associated subtidal habitat

https://doi.org/10.1016/j.fishres.2021.106067Get rights and content

Abstract

Forage fish and fish associated with particular benthic habitats (e.g., rockfishes, sand eels, sand lances) may be particularly difficult to assess through standard survey methodologies. Stereo-cameras, video, and automated visual data may serve as useful complementary tools to provide insight into the dynamics of these species. Visual methods may be used not only to estimate abundance and distribution, but also to inform important biological metrics and life history attributes. We explored the application of these methods to assess Pacific sand lance (Ammodytes personatus), a forage fish associated with benthic sediments, using a combination of directed observations from a manned submersible and quantitative analysis of fixed image footage obtained with a stereo-camera. This research provides a better understanding of how in situ observations and automated image analysis might complement other methods to estimate fish abundance, distribution, habitat, and behavior. Visual data were compared to data collected via directed sampling using physical extraction methods at the same site in the same year. Submersible observations provided new insights on the physical conditions and habitat. Visual observations confirmed wavefield morphologies previously identified through multibeam acoustic imagery and measured attributes relevant to the physical oceanography of the water column above this benthic habitat feature. Visual observations also informed understanding of light penetration, relevant to diurnal cues for seasonal progression and diel vertical migration and foraging. Submersible observations provided insights into abundance, schooling dynamics, and behavioral attributes, including avoidance in response to physical disturbance and aggregation in presence of artificial light. Quantitative analysis of stereo-camera data in center-edge and north-south transects determined that fish abundance and length distribution was relatively uniform throughout this particular benthic habitat. Estimates of measurement error associated with stereo-cameras were calculated and correction factors identified. Mean lengths estimated in visual data and in physical specimens were closely matched, though variance in visual data measurements was far greater. This error was reduced when filtering data on the basis of orthogonal position or incidence angle relative to the camera. Our research provides important insights to the presence, distribution, abundance, and movement of Pacific sand lance within benthic sand wavefield habitats. Our research also provides insight to the applications, opportunities, and constraints to observation-based sampling methods, including the use of manned submersibles and automated stereo-cameras.

Introduction

Effective monitoring of pelagic marine fish populations may require a variety of assessment methods (Boldt et al., 2017, 2018, 2019; Baker et al., 2018; Moriarty et al., 2020). Standardized abundance indices based on catch and effort indices and fishery-dependent data are a fundamental input to stock assessments (Hilborn, 1979; Maunder and Punt, 2004; Bishop, 2006), but fishery-dependent data may be of limited utility in monitoring non-target species or non-target areas (Thorson and Ward, 2014; Thorson et al., 2016). Surveys and other traditional fishery-independent assessment methods (Hilborn and Walters, 2013) provide more comprehensive indices of system biomass (Sainsbury et al., 2000; Koslow and Davison, 2016), species distribution (Baker and Hollowed, 2014; Moriarty et al., 2020) and species dynamics (Gaichas et al., 2010; Karp et al., 2019). These methods, however, may be spatially limited (Link et al., 2011), biased in their target or design (Thorson et al., 2016), ineffective in certain habitats (Baker et al., 2019a) or constrained in accurately assessing certain types of fishes, particularly pelagic and forage fishes (Fréon and Misund, 1999; Alheit and Peck, 2019). Estimating fish biomass is particularly challenging where catchability or availability of fish to survey gear is limited (Ward, 2008). Estimating biomass and abundance is further complicated where the distribution of the species is restricted to specific habitats (Millar and Methot, 2002) or the species of interest is characterized by patchiness in spatial or temporal distribution (Thorson et al., 2011; Boyd et al., 2015).

Fish are temporally and spatially variable in their abundance and distribution, particularly forage fish (Fréon and Misund, 1999; Greene et al., 2015; Baker, 2021). This may restrict their availability to surveys (McGowan et al., 2019). Fish species with specific habitat-specific associations, such as rockfish, also pose a particular challenge (Spencer and Ianelli, 2014). Often alternative methods are required (Clarke et al., 2009; Rooper et al., 2010; Honkalehto et al., 2011; Hanselman et al., 2012). Remotely operated vehicles (ROVs; Brodeur, 2001; Auster et al., 2003), autonomous underwater vehicles (AUVs; Tolimieri et al., 2008) and human-occupied submersibles (Stein et al., 1992; Starr et al., 1996; Yoklavich et al., 2000; Nasby-Lucas et al., 2002; Rodgveller et al., 2011; Pacunski et al., 2008, 2013) have demonstrated utility in assessing fish abundance and fish habitat. We used a manned submersible (Fig. 1a) and an attached stereo-camera system (Fig. 1b) to enumerate, measure, and observe habitat interactions for an unassessed sand-associated North Pacific forage fish at a deep-water sand wavefield in the Salish Sea.

Our study focuses on Pacific sand lance (PSL; Ammodytes personatus, Orr et al., 2015), an ecologically important forage fish distributed throughout the North Pacific Ocean (Appendix, Fig. A-1). Relatively little is known about this species in contrast to commercially valuable North Pacific forage fish such as Pacific herring, sardines, and anchovies (Liedtke et al., 2013). PSL are also distinguished from other common northern latitude forage fish species in their reliance on bottom sediments for refuge (Bizzarro et al., 2016). The San Juan Archipelago in the Salish Sea has a complex bathymetry influenced by previous glaciation (Greene and Barrie, 2011) and provides habitat for what is likely a very significant number of PSL (Greene et al., 2011, 2020; Baker et al., 2021b, In Review). The region is characterized by strong currents, significant oceanic inputs and upwelling (Thomson and Ware, 1996) and is also important habitat to hundreds of species of birds, mammals, and fishes, many of which rely on PSL as a prey resource (Gaydos et al., 2008; Gaydos and Pearson, 2011; Pietsch and Orr, 1999).

An extensive sampling effort conducted by Selleck et al. (2015) found PSL along 82 % of sampled shoreline in the San Juan Islands and the Strait of Georgia. In addition to nearshore populations, Greene and Pacunski (unpublished; 2004) discovered a subtidal sand wave field, during a Washington Department of Fish and Wildlife (WDFG) remotely operated vehicle (ROV) video survey. This San Juan Channel sand wave field has been extensively studied and determined to be an important habitat for PSL (Greene et al., 2017, 2020; Baker et al., 2019b; Baker et al., 2021b, In Review). It is estimated to provide benthic habitat for as many as 100 million PSL (Sisson and Baker, 2017) ages 0–4 years (Matta and Baker, 2020). Many benthic sediments with similar sediment features in the area have since been identified as potential PSL habitat (Greene et al., 2011; Baker, unpublished data).

Since 2010, the Pelagic Ecosystem Function research apprenticeship [http://courses.washington.edu/pelecofn/index.html] at the University of Washington Friday Harbor Laboratories, WA, USA has been focused on studies of PSL at this site (Newton et al., 2018, 2019). Sampling has largely involved sampling PSL with a Van Veen grab. This method has been useful for securing fish and answering questions about sediment association (Baker et al., 2021b, In Review), length-at-age analyses (Matta and Baker, 2020), and demographics and annual condition (Baker et al., 2019b). While successful and efficient in securing fish at known benthic sites (Høines and Bergstad, 2001; Hassel et al., 2004; Greenstreet et al., 2010), sampling by means of Van Veen has limitations. The probability of Van Veen closure is restricted to certain sediment types because large-sized sediment may prevent closure of the device. Additionally, Van Veen grabs are limited to the top layer of sediment (< 22 cm). These methods are also restricted to periods of time when the fish are dormant in sediments, rather than active in the water column. Most importantly, these methods do not allow for observation of fish behavior, response to disturbance, movement between the benthos and the water column, schooling dynamics, physical dynamics related to sediment movement, or analysis of environmental conditions at depth.

In many pelagic fishes, abundance and distribution are effectively monitored and analyzed through acoustic methods (Horne, 2000; Gauthier and Horne, 2004). Unlike most pelagic fishes, however, sand lance and sand eels (Ammodytes spp.) lack swim bladders and their acoustic properties are very different than other pelagic fish species (Mosteiro et al., 2004). Specifically, these fish have relatively low target strength values (Forland et al., 2014), which are necessary for accurate acoustic measurements. Acoustic approaches have been applied in the Atlantic (Hassel et al., 2003; Mackinson et al., 2005; Johnsen et al., 2009); however, as sand lance and sand eel form compact schools with low reflectance (weak acoustic backscatter) and are often distributed near-bottom, acoustic methods are limited in their effectiveness and resolution (Ona and Mitson, 1997). In the Northeast Atlantic, where sand eel support one of the most important commercial fisheries by volume (ICES, 2018), pelagic trawls and dredges are used to measure relative densities of these fish in the seabed (Jensen, 2001; van der Kooij et al., 2008). Here too, there are issues of catchability, both in survey trawls (Fraser et al., 2007) and dredge tows (Mackinson et al., 2005; van Deurs et al., 2012). There is international demand for fishery-independent data to improve abundance estimation (Kubilius and Ona, 2012) and to inform management of forage species (ICES, 2008). Current limitations to effective sampling of these fish motivate our efforts to explore and apply new methods.

As an alternative assessment method, stereo-camera surveys have been successfully applied to assess abundance and distribution and to conduct fish length measurements (Harvey et al., 2003; Watson et al., 2005; Shortis et al., 2009; Williams et al., 2016b; Boldt et al., 2018). Fish measurements obtained from stereo-camera imagery have been shown to be accurate (Harvey et al., 2003; Seiler et al., 2012). In addition to developing estimates for fish abundance and morphological metrics, fish behavior can be observed (Somerton et al., 2017). Stereo-camera imagery allows for continuous sampling through space and time as stereo-camera systems can be used on fixed stationary platforms. Operated from submersibles, stereo-cameras may be combined with direct observation to quantify observed dynamics, metrics, and behaviors. Used either in isolation or in combination with other approaches, stereo-cameras can provide a unique perspective and insights, particularly on species difficult to assess through alternative methods.

We applied stereo-camera data from submersible surveys in concert with Van Veen grab sampling methods to integrate and contrast these approaches and gain greater insight to sources of error associated with each approach. By applying multiple methods at a common site, we provide insight to the behavior, abundance, and attributes of an important, but poorly understood forage fish in the North Pacific. We also introduce potential approaches and correction factors to address bias in stereo-camera morphometric measurement estimates related to fish orientation and suggest improvements in the application of stereo-camera arrays and submersible survey efforts.

Section snippets

Study site

PSL were collected or observed at the San Juan Channel (SJC) sand wavefield (48° 31′ N, 122° 57′ W; Fig. 2) in the Salish Sea, Washington, USA. The sand wavefield covers an area of approximately 600,000 m2 and is oriented north-south. The sand wavefield is approximately 0.74 km wide (east-west) and 1.88 km long (north-south) at a depth of 60 m in the north and 80 m at the southern extent. This wavefield contains bedforms with wavelengths up to 100 m and heights of approximately 1–4 m within its

Submersible transects and Van Veen sampling

OceanGate Cyclops I was used to conduct two dives at the SJC sand wave field (Fig. 2, left panel). Sample locations for a series of Van Veen sediment grabs between September 10, 2018 and December 2, 2018 are indicated (Fig. 2, right panel).

Physical system

The high abundance and widespread distribution of PSL throughout the San Juan Channel sediment wave field confirm this as an extensive sub-tidal deep-water habitat feature for PSL, an important forage fish in the Salish Sea and throughout the North Pacific.

Value of observational data in submersible surveys

These submersible surveys enabled observations of PSL in situ that provided insights into pelagic schooling dynamics, entry and exit from benthic sediments, interactions with currents, sand wave morphology, light, and disturbance. It also enabled visual surveys of the full extent of an important benthic habitat. This served to groundtruth observations and inferences previously made from MBES bathymetry and soundings. While past research had focused on Van Veen grab sampling for fish and

Supplemental information

Additional information on this research effort is available through OceanGate.

https://oceangate.com/expeditions/salish-sea-survey-expedition.html and the SeaDoc Society https://www.seadocsociety.org/submersible. Additional videos are available at the following links:

https://youtu.be/wsEnDkoPFS8; https://youtu.be/h-UdTQsvIYc.

CRediT authorship contribution statement

M.R. Baker: supervision, conceptualization, methodology, data analysis, visualization, writing. K. Williams: conceptualization, methodology, data analysis, visualization, writing, technical support. H.G. Greene: supervision, conceptualization, methodology, writing. C. Greufe: data analysis, visualization. H. Lopes: data analysis, visualization. J. Aschoff: visualization. R. Towler: technical support.

Declaration of Competing Interest

The authors report no declarations of interest.

Acknowledgments

We greatly appreciate the efforts of Joe Gaydos and the SeaDoc Society, Orcas Island, WA to secure and facilitate the use of the OceanGate Cyclops I to conduct these underwater observations. We also greatly appreciate the work of OceanGate Inc. leadership and the OceanGate engineers, plan teams and dive teams to dedicate their time, personnel, and equipment to support this research, particularly Stockton Rush, Wendy Rush, Russell McDuff, Tony Nissen, Chris Howard, Neil McCurdy, and Mikayla

References (121)

  • S. Mackinson et al.

    Using acoustics to investigate changes in efficiency of a sandeel dredge

    Fish. Res.

    (2005)
  • M.N. Maunder et al.

    Standardizing catch and effort data: a review of recent approaches

    Fish. Res.

    (2004)
  • D.W. McGowan et al.

    Influence of environmental factors on capelin distributions in the Gulf of Alaska

    Deep. Sea Res. Part Ii Top. Stud. Oceanogr.

    (2019)
  • F.D. McIntyre et al.

    Visual surveys can reveal rather different ‘pictures’ of fish densities: comparison of trawl and video camera surveys in the Rockall Bank, NE Atlantic Ocean

    Deep-Sea Research

    (2015)
  • O.A. Ormseth et al.

    Introduction to understanding ecosystem processes in the Gulf of Alaska, volume 2

    Deep Sea Res. Part II

    (2019)
  • M.D. Robards et al.

    Changes in proximate composition and somatic energy content for Pacific sand lance (Ammodytes hexapterus) from Kachemak Bay, Alaska relative to maturity and season

    Journal of Experimental Marine Biology and Cology

    (1999)
  • C.N. Rooper et al.

    Effect of underwater lighting on observations of density and behavior of rockfish during camera surveys

    Fish. Res.

    (2015)
  • C.N. Rooper et al.

    Estimating habitat-specific abundance and behavior of several groundfishes using stationary stereo still cameras in the southern California bight

    Fish. Res.

    (2020)
  • J. Seiler et al.

    Assessing size, abundance and habitat preferences of the Ocean Perch (Helicolenus percoides) using a AUV-borne stereo camera system

    Fish. Res.

    (2012)
  • J.T. Thorson et al.

    Accounting for vessel effects when standardizing catch rates from cooperative surveys

    Fish. Res.

    (2014)
  • J. Alheit et al.

    Drivers of dynamics of small pelagic fish resources: biology, management and human factors

    Mar. Ecol. Prog. Ser.

    (2019)
  • P.J. Auster et al.

    Use of sand wave habitats by silver hake

    J. Fish Biol.

    (2003)
  • M.R. Baker

    Contrast of warm and cold phases in the Bering Sea to understand spatial distributions of Arctic and sub-Arctic gadids

    Polar Biol.

    (2021)
  • M.B. Baker et al.

    Influence of climate and environmental variability on pelagic and forage species

  • M.R. Baker et al.

    Model of trawlable area using benthic terrain and oceanographic variables—informing survey design and habitat maps in the Gulf of Alaska

    Fish. Oceanogr.

    (2019)
  • M.R. Baker et al.

    Intra-seasonal and inter-annual patterns in the demographics of sand lance and response to environmental drivers in the North Pacific

    Mar. Ecol. Prog. Ser.

    (2019)
  • M.R. Baker et al.

    Diel vertical migration in a pelagic forage fish (Ammodytes personatus) associated with benthic substrates

    Behav. Ecol.

    (2021)
  • M.R. Baker et al.

    Analyses of sediment association and distribution of Pacific sand lance (Ammodytes personatus) in a benthic sand wave field

    Cont. Shelf Res.

    (2021)
  • J. Bishop

    Standardizing fishery-dependent catch and effort data in complex fisheries with technology change

    Rev. Fish Biol. Fish.

    (2006)
  • J.J. Bizzarro et al.

    Burrowing behavior, habitat, and functional morphology of the Pacific sand lance (Ammodytes personatus)

    Fish. Bull.

    (2016)
  • J.L. Boldt et al.

    Methods and techniques for sampling and assessing small pelagic fish populations

  • J.L. Boldt et al.
    (2019)
  • J.Y. Bouguet

    Camera Calibration Toolbox for Matlab

    (2014)
  • C. Boyd et al.

    Bayesian posterior prediction of the patchy spatial distributions of small pelagic fish in regions of suitable habitat

    Can. J. Fish. Aquat. Sci.

    (2015)
  • L.L. Britt et al.

    Visual pigments in the early life stages of Pacific northwest marine fishes

    J. Exp. Biol.

    (2001)
  • M.E. Clarke et al.

    Using the seabed AUV to assess populations of groundfish in untrawlable areas

  • A. De Robertis et al.

    Fish avoidance of research vessels and the efficacy of noise-reduced vessels: a review

    Ices J. Mar. Sci.

    (2013)
  • C. Denney et al.

    Development of new method for quantifying fish density using underwater stereo-video tools

    J. Vis. Exp.

    (2017)
  • T.N. Forland et al.

    Broad bandwidth acoustic backscattering from sandeel—measurements and finite element simulations

    Ices J. Mar. Sci.

    (2014)
  • H.M. Fraser et al.

    Taking account of catchability in groundfish survey trawls: implications for estimating demersal fish biomass

    Ices J. Mar. Sci.

    (2007)
  • P. Fréon et al.

    Dynamics of Pelagic Fish Distribution and Behavior: Effects on Fisheries and Stock Assessment

    (1999)
  • S.K. Gaichas et al.

    Using food web model results to inform stock assessment estimates of mortality and production for ecosystem-based fisheries management

    Can. J. Fish. Aquat. Sci.

    (2010)
  • S. Gauthier et al.

    Acoustic characteristics of forage fish species in the Gulf of Alaska and Bering Sea based on Kirchhoff-approximation models

    Can. J. Fish. Aquat. Sci.

    (2004)
  • J.K. Gaydos et al.

    Birds and mammals that depend on the Salish Sea: a compilation

    Northwest. Nat.

    (2011)
  • J.K. Gaydos et al.

    Top 10 principles for designing healthy coastal ecosystems like the Salish Sea

    EcoHealth

    (2008)
  • H.G. Greene et al.

    Potential Marine Benthic Habitats of the San Juan Archipelago. Geological Survey of Canada Marine Map Series, 4 Quadrants, 12 Sheets, Scale 1:50,000; Geological Survey of Canada: Sidney, BC, Canada

    (2011)
  • H.G. Greene et al.

    Construction of digital potential marine benthic habitat maps using a coded classification scheme and its application

    Mapp. Seafloor Habitat Charact. Can. Geol. Assoc. Spec. Pap.

    (2007)
  • H.G. Greene et al.

    Deep-Water Pacific Sand Lance (Ammodytes hexapterus) Habitat Evaluation and Prediction for the Northwest Straits Region; Final Report to Northwest Straits Commission

    (2011)
  • C. Greene et al.

    Forty years of change in forage fish and jellyfish abundance across greater Puget Sound, Washington (USA): anthropogenic and climate associations

    Mar. Ecol. Prog. Ser.

    (2015)
  • H.G. Greene et al.

    Characteristics and Dynamics of a Large Sub-tidal Sand Wave Field – Habitat for Pacific Sand Lance (Ammodytes Personatus)

    (2017)
  • Cited by (11)

    • Oil spill assessment maps of the central Salish Sea – Marine seafloor &amp; coastal habitats of concern – A tool for oil spill mitigation within the San Juan Archipelago, Washington State, USA

      2023, Continental Shelf Research
      Citation Excerpt :

      Shipping lanes were mapped to correspond to high volume tanker traffic based on year 2017 identified Automated Identification Systems (AIS) data. Data from a comprehensive sediment sampling effort of the central Salish Sea prior to 2011, using Van Veen and Poner grab samplers are used in groundtruthing the maps (Greene et al., 2011a) as is sampling data obtained by Matt Baker of Friday Harbor Labs and the Tombolo Mapping Lab (Baker et al., 2021; Greene et al., 2017, 2021). Sediment analyses were undertaken with a RoTap™ sieving machine using screens that sorted sediment ranging from silt to coarse gravel at 1φ sieve intervals (see Table 1 of Blott and Pye, 2001).

    • Hazards evaluation of a valuable vulnerable sand-wave field forage fish habitat in the marginal Central Salish Sea using a submersible

      2023, Oceanologia
      Citation Excerpt :

      In this paper we describe the observations of a sediment wave-field with dynamic bedforms supporting a large population of PSL in order to evaluate the potential impacts of infrastructure development on this valuable prey resource. This is a qualitative (subjective) and not a quantitative study that adds value to previous quantifying investigations; a more statistical analysis is underway and will be reported upon at a later date (Baker et al., 2021). The MBES bathymetric images and marine benthic habitat maps along with marine acoustical geophysical and geological data are used in this study for seafloor characterization and identification of seafloor geomorphology.

    • Spatial distribution of arctic sand lance in the Chukchi Sea related to the physical environment

      2022, Deep-Sea Research Part II: Topical Studies in Oceanography
      Citation Excerpt :

      These species also typically exhibit a distinct diel pattern (Baker et al., In Review), such that, in the summer season, these fish emerge in daylight hours to forage and remain buried in sediments at night (Winslade, 1974a; Freeman et al., 2004; Engelhard et al., 2008). This reliance on sediments for refuge and rest results in a highly particular choice of sediment to facilitate rapid burrowing and respiration while buried (Greene et al., 2020, 2021; Baker et al., 2021). Sand lance prefer well-sorted coarse-grain substrates with no mud or silt (Meyer et al., 1979; Pinto et al., 1984; Holland et al., 2005).

    View all citing articles on Scopus
    View full text