Beach safety: can drones provide a platform for sighting sharks?

Abstract

ContextA series of unprovoked shark attacks on New South Wales (Australia) beaches between 2013 and 2015 triggered an investigation of new and emerging technologies for protecting bathers. Traditionally, bather protection has included several methods for shark capture, detection and/or deterrence but has often relied on environmentally damaging techniques. Heightened environmental awareness, including the important role of sharks in the marine ecosystem, demands new techniques for protection from shark attack. Recent advances in drone-related technologies have enabled the possibility of real-time shark detection and alerting. AimTo determine the reliability of drones to detect shark analogues in the water across a range of environmental conditions experienced on New South Wales beaches. MethodsA standard multirotor drone (DJI Inspire 1) was used to detect shark analogues as a proxy during flights at 0900, 1200 and 1500 hours over a 3-week period. The 27 flights encompassed a range of environmental conditions, including wind speed (2–30.0kmh−1), turbidity (0.4–6.4m), cloud cover (0–100%), glare (0–100%), seas (0.4–1.4m), swells (1.4–2.5m) and sea state (Beaufort Scale 1–5 Bf). Key resultsDetection rates of the shark analogues over the 27 flights were significantly higher for the independent observer conducting post-flight video analysis (50%) than for the drone pilot (38%) (Wald P=0.04). Water depth and turbidity significantly impaired detection of analogues (Wald P=0.04). Specifically, at a set depth of 2m below the water surface, very few analogues were seen by the observer or pilot when water turbidity reduced visibility to less than 1.5m. Similarly, when water visibility was greater than 1.5m, the detection rate was negatively related to water depth. Conclusions The present study demonstrates that drones can fly under most environmental conditions and would be a cost-effective bather protection tool for a range of user groups. ImplicationsThe most effective use of drones would occur during light winds and in shallow clear water. Although poor water visibility may restrict detection, sharks spend large amounts of time near the surface, therefore providing a practical tool for detection in most conditions.

Variation in cownose ray Rhinoptera neglecta abundance and group size on the central east coast of Australia

“…Although most of the environmental variables we tested did not explain significant variation in the abundance of R. neglecta , wind had a significant negative effect on the number of individuals per transect. Wind can influence the detectability of marine wildlife during aerial surveys (Butcher et al ., ; Lubow & Ransom, ) and limiting surveys to days with favourable weather conditions has been recommended to avoid this issue (Kelaher et al ., ). The negative association between wind and cownose rays may also have been caused by R. neglecta moving offshore into deeper water during periods of strong winds.…”

Bay watch: Using unmanned aerial vehicles (UAV’s) to survey the box jellyfish Chironex fleckeri

“…Sea state parameters, such as visibility, wind speed and wind direction, had no effect on rates of animal detection. These findings, when considered alongside the significant body of literature supporting role these environmental factors on UAV survey success, seem somewhat counterintuitive [ 7 , 8 , 10 , 20 ]. However, these results reflect core ecological characteristics unique to the model system.…”

“…A large number of studies have reported this phenomenon, with similar levels of discrepancy, and have accredited this trend to many different factors [ 8 ]. These range from (i) the challenge for drone pilots to split their attention between flying and animal detection to (ii) technical difficulties involving telemetry screen resolution and glare [ 8 , 20 ]. However, unlike previous studies, all piloting for this project was carried out using first ‘person-view DJI goggles.…”

“…Cloud cover was positively related to animal detectability with higher levels of cloud increasing the detection rate of jellyfish. Sun glare is a major confounding factor for detectability in marine environments [ 8 , 20 ]. This is due to water’s high capacity to reflect light which generates glare.…”

“…All flights were completed using a DJI Mavic 2 zoom drone. The inbuilt 12 mega pixel camera was equipped with a ND16 lenses polariser and video was recorded at a 70degree angle to reduce sea surface glare which, synonymous to other similar studies, optimised footage for animal detection [ 8 ]. During each transect the camera was set to record the entirety of the flight, with footage in full high definition (4k) at 30 frames s -1 in the widest field of view setting.…”