Eye in the sky: drones as tools for conservation biologists

Monitoring wildlife populations is not a straightforward task. Individuals can be dispersed over wide areas of inaccessible terrain and can move across the landscape and avoid detection. However, a new era in wildlife biology appears to be dawning as research scientists are increasingly turning to small, un-manned aerial vehicles (UAVs or drones) to aid data collection. These devices can be fitted with an array of sensors including cameras for capturing images or video, thermal imaging devices and multispectral sensors that are able to measure the health of vegetation. By flying lower and slower than conventional aircraft, UAVs allow the collection of data at very fine spatial scales. This makes monitoring achievable for many species that would be impossible to monitor using aeroplanes or satellites. These advantages have seen UAVs used for a range of conservation applications including surveying for orangutan nests (see video footage here), as a rhinoceros anti-poaching tool in Africa, and marine applications such as counting sea turtles in waters surrounding nesting beaches.

Some of the seabirds monitored by UAVs. Photo: Rowan Mott

Some of the seabirds monitored by UAVs. Photo: Rowan Mott

The remarkable footage obtained by UAVs and the relative ease with which data are obtained has led many to predict big things to come for the use of UAVs for research and conservation. These predictions are based on the assumptions that data quality from UAV surveys is higher than traditional methods or that benefits to cost effectiveness and collection efficiency outweigh any drawbacks. Despite rapidly increasing use, the quality of UAV-collected data compared to data from traditional methods remained un-tested.

Our research group is involved in the monitoring of several large seabird colonies. This typically involves a team of counters making twice-a-year boat voyages to remote islands where upwards of 100, 000 seabirds breed. Using binoculars, spotting scopes, and a notepad and pencil we undertake the daunting task of counting every last bird. But this could be about to change. On recent visits to these breeding colonies we took along a new tool – a small, remote controlled quadcopter. We flew this UAV over a number of colonies collecting aerial images of the seabirds below. At the same time we had ground counters count the seabirds as we normally would. Back in the laboratory a dedicated team of volunteers counted the birds in the images and then we compared the counts made from UAV-imagery with those made by our team of expert ground counters.

An image of the seabird colony taken by the team's UAVs. Image: Jarrod Hodgson

An image of the seabird colony taken by the team's UAVs. Image: Jarrod Hodgson

The results were astounding. Counts of the imaged birds made by volunteers were consistently closer to each other than those made by ground counters. What is even more remarkable is that some of the volunteers that counted UAV-imagery had never seen the species they were counting, whereas the team of ground counters each had years of bird-counting experience. We don’t know for certain which count method resulted in an estimate of colony size closest to the true number of birds in each colony. For all we know both methods could be under- or over-estimating how many birds were breeding. So why then is this result important?

When monitoring a wildlife population, biologists are hoping to be able to detect population trends that could indicate the population is at risk of being wiped out or is recovering if a threat has been alleviated. If, when the same colony is counted by two different people, the different counts vary widely from each other there can be little certainty as to what the true number of individuals is. By minimising the variance between separate counts the confidence in those estimates increases and population fluctuations of smaller magnitude are more likely to be recognised. This will have a big advantage in situations where early intervention can achieve the desired conservation outcome more easily than if action is delayed.

Our research shows that the predictions of big things to come for UAV technology are not unfounded. The quality of data collected using UAVs can be higher than that collected using traditional methods. Yet there remain many important questions relating to the use of UAVs in wildlife monitoring. Not least of these is the question of how close counts made from UAV-imagery are to the true number and we hope to have an answer to this in the near future.

Photo: Rowan Mott

Photo: Rowan Mott

Also, there are ethical considerations that need to be addressed. Do UAVs result in greater disturbance levels to the wildlife populations that are being monitored than traditional methods? Although seabirds rarely fall victim to aerial predators, the same cannot be said for many ducks and shorebirds that otherwise seem ideal candidates to be monitored using UAVs. If the unfamiliar silhouette of an UAV is perceived to pose a threat similar to a fast-approaching falcon or eagle then the use of UAVs in the survey of these species is likely to be inappropriate. Currently, in the United States, it is illegal to fly UAVs over protected species, so until issues relating to the ethics of UAV use are fully explored, their full potential will not be met. These factors aside, the surge in uptake of UAVs for conservation looks set to continue and biologists across the world will be finding new ways in which this exciting new tool can further their research.


Check out the latest paper by Rowan and his colleagues here, and at the citation below. 

Hodgson, J., Baylis, S.M., Mott, R.M., Herrod, A. and Clarke, R.H. (2016). Precision wildlife monitoring using unmanned aerial vehicles. Scientific Reports doi:10.1038/srep22574

Cover image taken by Shane Baylis