Elsevier

Ocean & Coastal Management

Volume 174, 15 May 2019, Pages 108-115
Ocean & Coastal Management

Reliability of marine faunal detections in drone-based monitoring

https://doi.org/10.1016/j.ocecoaman.2019.03.008Get rights and content

Highlights

  • We developed and tested drone-based surveillance procedures for sharks and other fauna.

  • Drone-based field detections of marine fauna can be prone to significant perception error.

  • Error rates of field detections depended mostly on the type of fauna.

  • Perception error of field detections of fauna were not influenced by environmental factors.

Abstract

An increase in shark bites, declining shark populations, and changing social attitudes, has driven an urgent need for non-destructive shark monitoring. While drones may be a useful tool for marine aerial surveillance, their reliability in detecting fauna along coastal beaches has not been established. We developed a drone-based shark surveillance procedure and tested the reliability of field-based fauna detections and classifications against rigorous post-analysis. Perception error rates were examined across faunal groups and environmental parameters. Over 316 shark surveillance flights were conducted over 12 weeks, out of a possible 360, with adverse weather preventing most flights. There were 386 separate sightings made in post-analysis, including 17 sightings of shark, 125 of dolphin, 192 of ray, 19 of turtle, 15 of baitfish school, and a further 18 sightings of other fauna. When examining error rates of field-based detections, there were large differences found between fauna groups, with sharks, dolphins, and baitfish schools having higher probabilities of detection. Some fauna, such as turtles, were also more difficult to classify following a detection than other groups. The number of individuals in a sighting, was found to have significant but relatively subtle effects, whilst no environmental covariates were found to influence the perception error rate of field-based sightings. We conclude that drones are an effective monitoring tool for large marine fauna off coastal beaches, particularly if the seabed can be distinguished and post-analysis is performed on the drone-collected imagery. Where live field-based detections are relied upon, such as for drone-based shark surveillance, the perception error rate might be reduced by machine-learning software assistance, such as neural network algorithms, or by utilising a dedicated ‘observer’ watching a high-resolution glare-free screen.

Introduction

The perceived threat of shark attacks along coastal beaches is a global problem and can cause significant social and economic consequences for local communities (Chapman and McPhee, 2016; Lemahieu et al., 2017; Pepin-Neff and Wynter, 2018). Although various factors may contribute to the increase in reported unprovoked shark attacks over the last few decades (Chapman and McPhee, 2016), it is likely that under the current management regimes, the rate of shark attacks will continue to rise as human coastal populations and subsequent beach visitation increases (West, 2011; McPhee, 2014). Consequently, there is increasing pressure to expand existing shark control programs to encompass more coastline (Curtis et al., 2012; Lemahieu et al., 2017).

Currently, many shark control programs aimed at reducing shark attack risk still employ cull-based technologies involving mesh net or drum line methods that have been used since as early as the 1930s (Reid et al., 2011; Neff and Yang, 2013; Gibbs and Warren, 2015). While mesh net and drum line methods are often considered effective, there is also increasing debate about the ethical issues and potential ecological consequences, such as contributing to declining shark populations and adverse impacts on various bycatch (Myers et al., 2007; Ferretti et al., 2010; Friedrich et al., 2014; Gibbs and Warren, 2015). This is driving a shift towards the development of alternative shark management strategies that not only spare marine life, but also appease public perceptions and improve beach safety (Chapman and McPhee, 2016; Gray and Gray, 2017; Pepin-Neff and Wynter, 2018).

Aerial shark surveillance has been one of the most publicly accepted, non-destructive shark management strategies, but its use is often restricted by cost-inefficiencies, and its effectiveness at detecting sharks is thought to have been overestimated by stakeholders (Crossley et al., 2014; Robbins et al., 2014). Like other marine aerial surveys, shark surveillance by manned aircraft is known to suffer sightability errors, resulting in shark sighting rates as low as 12.5% and 17.5% from fixed-wing aircraft and helicopters respectively, which was estimated using submerged shark silhouettes (Robbins et al., 2014). Sightability errors comprise ‘availability’ and ‘perception’ errors, which can reduce the data reliability in all marine aerial surveys and surveillance operations of submerged fauna (Rowat et al., 2009; Colefax et al., 2017; Brack et al., 2018). Availability errors arise when an unknown number of target animals in a survey area are not in the visible portion of the water column at the time of observation. Whereas perception errors arise when an animal is residing in the ‘available’ section of the water column but is missed or misclassified nonetheless (Marsh and Sinclair, 1989; Robbins et al., 2014; Lubow and Ransom, 2016; Brack et al., 2018).

Drone technology (also ‘UAV’ or ‘RPAS’; see Chapman, 2014 and Chabot, 2018 for justification to use the term drone) is rapidly gaining popularity as an alternative aerial survey platform for conducting fauna surveys in general, but research investment has also been trialling their utility in marine environments (Goebel et al., 2015; Christie et al., 2016; Hodgson et al., 2016; Kiszka et al., 2016). They may offer increased utility over many current ground, boat and air -based methods (Ivosevic et al., 2015; Adame et al., 2017; Raoult et al., 2018), and offer reduced disturbance towards wildlife and habitats from data collection if executed appropriately (Christiansen et al., 2016; Amerson, 2018; Ramos et al., 2018). Drones also have the potential to offer improved data quality and reliability than manned aircraft because of digital capture methods and the ability to sample areas more intensively, but empirical comparisons between manned aircraft and drones are currently few (Johnston et al., 2017; Angliss et al., 2018; Ferguson et al., 2018). Regardless, it is plausible to expect that drones used for sighting submerged fauna are subject to the same sightability error constraints and may suffer similar perception biases as arise in an equivalent manned aircraft operation (Hodgson et al., 2013; Colefax et al., 2017; Brack et al., 2018).

Sightability errors are typically reduced by sampling in periods of relatively clear and calm sea conditions, and by modelling the expected portion of time an individual of a species is likely to reside in the visible portion of the water column (Pollock et al., 2006; Fuentes et al., 2015; Hodgson et al., 2017). For example, drone-based surveys of submerged fauna, such as on sharks and rays, are conducted in shallow environments and are timed to coincide with favourable weather and water conditions that allow clear view of the entire water column, rendering zero sightability error (Kiszka et al., 2016; Raoult et al., 2018; Rieucau et al., 2018). Where the bottom cannot always be distinguished, availability bias corrections can be applied to the data to improve abundance estimates and effectively reduce some availability error, which has been done in drone-based whale surveys (Hodgson et al., 2017).

Conversely, perception errors are usually a human error component, or limitations of detection software (Brack et al., 2018). Perception errors, once identified, may be reduced or remedied by altering operational procedures, such as having multiple observers that are appropriately trained and skilled in manned aircraft surveys (Pollock et al., 2006; Rowat et al., 2009). However, the empirical examination of perception error rates, or even the adoption of a multiple observer approach in drone-based surveys has been rarely undertaken (Vermeulen et al., 2013; Kiszka et al., 2016; Brack et al., 2018).

It is expected that drone-based shark surveillance will start to be adopted in more areas as public pressure for alternative shark management strategies continues to grow (Gibbs and Warren, 2015; Engelbrecht et al., 2017; Lemahieu et al., 2017). This may also provide unprecedented opportunity to obtain high-resolution assemblage data of large marine fauna at coastal beaches along stretches of coastlines (Kelaher et al., 2019). However, because the use of drones in marine surveys is still largely at its infancy, methods have not yet been properly established nor standardised for sighting submerged fauna (Brack et al., 2018). Although quantifying availability error and devising methods to improve this for deep or poor water clarity scenarios is important (Schoonmaker et al., 2011; Robbins et al., 2014), determining the extent of perception error from field-based detections is also important for establishing and validating new protocols in shallow coastal environments where real-time detections are relied upon. Here, we (i) outline and test a method for performing shark surveillance using cost-effective very-small (<2 kg) drones, (ii) examine the operational efficacy of using drones across environmental conditions, and (iii) quantify the perception error associated with field detections and classifications of various faunal groups during an extensive drone-based shark surveillance experiment.

Section snippets

Shark surveillance methodology and locations

Drone-based shark surveillance surveys were conducted at five beaches in New South Wales, Australia, including Lennox Head (28.78467 °S, 153.59376 °E), Ballina (28.86821 °S, 153.59262 °E), Evans Head (29.11149 °S, 153.43413 °E), Redhead (33.01532 °S, 151.71758 °E), and Kiama (34.67779 °S, 150.85594 °E) (see Fig. 1). These sites have a recent history of shark bite incidences, are in vicinity to surf life-guard clubs, and encompass different beach morphologies to represent a broad range of

Results

Out of a potential 360 flights over the 12-week period at the five locations, 316 flights were completed, with weather (rain, wind, and both wind and rain), military operations, technical faults, and hazardous shark sightings being responsible for pilots either cancelling or not completing 12.2% of flights. Of the flights cancelled or aborted due to reasons pertaining to weather (the vast majority), in 72% of cases, no people were sighted in the water. When wind gusts started to exceed 35 km h

Discussion

The drone-based survey methodology trialled would be effective for detecting and identifying large marine life along coastal beaches with relatively high precision providing video footage is post-analysed. Accordingly, the drone-based survey methodology could be used successfully to quantify assemblages of large marine fauna in near-shore coastal environments. However, although drone-based surveys are often considered highly effective (Kiszka et al., 2016; Kelaher et al., 2019), the results

Conflicts of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Acknowledgments

Project funding and support was provided by the New South Wales Department of Primary Industries (NSW DPI) and associated NSW Shark Management Strategy, Southern Cross University, and the Paddy Pallin Foundation. Surveillance flights were made under NSW DPI and NSW Office of Environment & Heritage scientific permits (Ref. P01/0059; MWL000102746). We wish to thank the dedication of the commercial drone pilots from Hover UAV, Scout Aerial, MJ Visual Media, and Vision Media for their involvement.

References (67)

  • D. Rowat et al.

    Aerial survey as a tool to estimate whale shark abundance trends

    J. Exp. Mar. Biol. Ecol.

    (2009)
  • J.A. Thomson et al.

    Correcting for heterogeneous availability bias in surveys of long-diving marine turtles

    Biol. Conserv.

    (2013)
  • K. Adame et al.

    Detectability and categorization of California sea lions using an unmanned aerial vehicle

    Mar. Mamm. Sci.

    (2017)
  • A. Amerson

    Six-word Lessons for Drone Pilots and Outdoor Enthusiasts: 100 Lessons to Make Drone Flights Safe, Ethical and Green for Wildlife and Humans (The Six-word Lesson Series)

    (2018)
  • D.R. Anderson

    The need to get the basics right in wildlife field studies

    Wildl. Soc. Bull.

    (2001)
  • R.P. Angliss et al.

    Comparing manned to unmanned aerial surveys for cetacean monitoring in the Arctic: methods and operational results

    J. Unmanned Veh. Syst.

    (2018)
  • D. Bates et al.

    Fitting linear mixed-effects models using lme4

    J. Stat. Softw.

    (2015)
  • E. Bevan et al.

    Unmanned aerial vehicles (UAVs) for monitoring sea turtles in near-shore waters

    Mar. Turt. Newsl.

    (2015)
  • I.V. Brack et al.

    Detection errors in wildlife abundance estimates from Unmanned Aerial Systems (UAS) surveys: synthesis, solutions and challenges

    Meth. Ecol. Evol.

    (2018)
  • D. Chabot

    Trends in drone research and applications as the Journal of Unmanned Vehicle Systems turns five

    J. Unmanned Veh. Syst.

    (2018)
  • D. Chabot et al.

    Evaluation of an off-the-shelf unmanned aircraft system for surveying flocks of geese

    Waterbirds

    (2012)
  • D. Chabot et al.

    Computer-automated bird detection and counts in high-resolution aerial images: a review

    J. Field Ornithol.

    (2016)
  • D. Chabot et al.

    Object-based analysis of UAS imagery to map emergent and submerged invasive aquatic vegetation: a case study

    J. Unmanned Veh. Syst.

    (2017)
  • A. Chapman

    It's okay to call them drones

    J. Unmanned Veh. Syst.

    (2014)
  • F. Christiansen et al.

    Noise levels of multi-rotor unmanned aerial vehicles with implications for potential underwater impacts on marine mammals

    Front. Mar. Sci.

    (2016)
  • K.S. Christie et al.

    Unmanned aircraft systems in wildlife research: current and future applications of a transformative technology

    Front. Ecol. Environ.

    (2016)
  • A.P. Colefax et al.

    The potential for unmanned aerial vehicles (UAVs) to conduct marine fauna surveys in place of manned aircraft

    ICES (Int. Counc. Explor. Sea) J. Mar. Sci.

    (2017)
  • R. Crossley et al.

    Public perception and understanding of shark attack mitigation measures in Australia

    Hum. Dimens. Wildl.

    (2014)
  • T.H. Curtis et al.

    Responding to the risk of white shark attack: updated statistics, prevention, control methods, and recommendation

  • J.W. Durban et al.

    Photogrammetry of killer whales using a small hexacopter launched at sea

    J. Unmanned Veh. Syst.

    (2015)
  • T. Engelbrecht et al.

    Shark Spotters: successfully reducing spatial overlap between white sharks (Carcharodon carcharias) and recreational water users in False Bay, South Africa

    PLoS One

    (2017)
  • M. Ferguson et al.

    Performance of manned and unmanned aerial surveys to collect visual data and imagery for estimating arctic cetacean density and associated uncertainty

    J. Unmanned Veh. Syst.

    (2018)
  • F. Ferretti et al.

    Patterns and ecosystem consequences of shark declines in the ocean

    Ecol. Lett.

    (2010)
  • Cited by (65)

    View all citing articles on Scopus
    View full text