Dugongs (Dugong dugonare) fairly hefty marine mammals found across coastal waters of east Africa and the Western Pacific. Feeding almost wholly on seagrass, and occasionally algae, these critters have been known to live up to 73 years old. Whilst the places dugongs in the ‘vulnerable’ category (the lowest of the ‘threatened with extinction categories’), dugong have been subject to a long history of human exploitation and many of the populations that once existed are now gone. The largest population is found in Australia, with a range stretching from Shark Bay in Western Australia to Moreton Bay in Queensland. The dugongs are faced with several threats, primarily gill netting, human settlement, agricultural pollution, and subsistence hunting. Monitoring this population is important for implementing effective management, but the problem facing all researchers and conservationists looking at the remaining populations is that they aren’t all that easily accessible. Even within Australia, only the ‘Urban Queensland’ section of the range has been studied and monitored to any depth. As we move into the Northern Territories and Western Australia, we know less and less about the dugongs and the particular threats they face in those zones, nor how many there are in different places so effective catch limits for sustainable fishing can be implemented. Red List
Surveying these populations by air is the quickest and simplest solution, but the problems with surveying are many. First there is the cost. Planes, pilots, accommodation, training is not cheap. Then, once you have all that in place, if the weather turns sour, or the sea is too rough to be able to see into it, then you can’t do the survey. Even with a plane, some sites are just too far from an airfield to realistically be surveyed in any detail with the small survey craft. Then there’s the risk of misidentification of both the animal and the exact location. And believe it or not, there is also a safety concern. Some 11 marine mammal researchers have reportedly been killed in plane crashes undergoing aerial surveys. You can see why Amanda Hodgson from Murdoch University and colleagues from CSIRO and the Australian Marine Mammal Centre decided to give drones a try instead.
The researchers focused their trial on one area in particular – Shark Bay located on the north-western tip of Australia. Using a ScanEagle UAV, a digital SLR camera was attached, and then flew the drone over 1.3 square kilometres seven times, and at different altitude. The drone was able to capture some 6,243 images which were analysed for dugongs, and assessed for sea state, turbidity, and sun glitter. What they found was quite positive, with 95% of the dugongs sighted were classed as unmistakably dugong. It seems that the images from the drones weren’t heavily impacted by the altitude the drone flew at, and sun glitter could largely be compensated for by overlapping the images that were taken at fixed intervals. Turbidity and sea state may still reduce positive identification, but not to the same extent as using surveyors on board an aircraft. And as a boon, the risk of double counting was heavily reduced, with individuals being identified and removed from further count. The researchers were also able to successfully ID a whole host of other species in Shark Bay – particularly dolphins and turtles. Having a permanent record of these images means they could be shared among different interest groups, reducing the costs of needing multiple fly-overs, or multiple surveyors on an aircraft. It also seems that using drones is more environmentally friendly. A standard aircraft used for these dugong surveys in Australia uses some 75 – 90 L of fuel an hour. The drone needed just 330ml of fuel an hour. The researchers were even able to view the SLR screen from their ground base in real time – handy for confirming that the images were indeed being captured in situ.
One of the downsides of this technology was the time it took for the researchers to analyse the images after they had obtained them. From this relatively small survey area, the drone generated some 6243 images that had to be manually checked. In future, it may be possible to develop some software to identify species from aerial photographs. Developing this sort of software isn’t that straight forward for the marine environment because there are all sorts of factors that can alter the clarity of the image, such as light, white caps, and changes in the sea floor. Here, the human eye might be a little better.
The next step for researching the use of these drones is to directly compare it to a manned aerial survey to make sure that the drone is able to at least pick up the same proportion of dugongs from an image as human eyes in the sky can, but so far the technology looks very promising.
The paper is published in the open access journal _PLoS ONE – you can access it here dx.doi.org/10.1371/journal.pone.0079556
Image: A Dugong near Marsa Alam (Egypt). Credit Julien Willem/Wikimedia (CC BY-SA 3.0)