Impact case study database

In UR 1 we showed that drones can be used to count orangutan nests and determine orangutan distribution and relative density using standard visual spectrum cameras. The aerial counts can therefore be an alternative to ground counts of orangutan nests and reduce the time to collect data by at least a factor two.

This led a number of Non-Governmental Organisations (NGOs) in Indonesia (from 2016 onwards) to count orangutan nests for surveys on orangutan distribution and density in the areas that they work (Sumatran Orangutan Conservation Program [S1], Orangutan Information Centre [S2]) and an NGO in Tanzania (Greater Malale Ecosystem Research and Conservation) to use drones to count chimpanzee nests [S3].

Mr. Hadisiswoyo, Founder of the Orangutan Information Centre: “In terms of orangutan survey work I estimate that the drone usage has cut the time needed for surveys and analyses in half which is a huge time and cost saving.” [S2].

This research has also influenced research centres and NGOs (such as WWF) to start using drones for monitoring of animal species.

Dr. Goossens, Danau Girang Field Centre lead researcher: “Drones are a key emerging technology that we are keeping a keen eye on and the research by Prof. Serge Wich has influenced our thinking about usage of drones for our conservation efforts. Serge has visited our site several years ago to demonstrate the usage of drones and provide advice on their usage for conservation projects.” [S4].

Dr. van der Hoeven, WWF Netherlands: “The conservation research using drones that has been conducted by Prof. Serge Wich has been very influential on getting us started with drones for various of our projects. As a result of his papers and further discussions with him we are ow using drones in several of our conservation projects and are planning to extend this. Drones have in several cases facilitated our conservation efforts….” [S5].

The studies that used drones equipped with a thermal infrared sensor to detect and classify animals and poachers through manual and automated methods (funded by STFC for £411,988) with a thermal sensor (UR 2 and 4) led to a number of collaborations with conservation NGOs (GMERC, WWF UK/Sabah, Durrell Wildlife Conservation Trust, Endangered Wildlife Trust, ConMonoMaya), universities (e.g. University of Vera Cruz),). During those collaborations drones with thermal cameras were used to successfully detect species such as orangutans, spider monkeys, poachers and riverine rabbits.

Our collaboration with GMERC in Tanzania has shown that through using drones with a thermal camera and machine learning we can detect poachers 17x faster in images than through a manual method.

Dr Piel, Lead researcher of the Greater Mahale Ecosystem Research and Conservation Project: “… drones coupled with machine learning can detect poachers 17x faster than when images are examined manually.” [S3].

One of these collaborations led to a paper itself (UR 3) that indicated that drone surveys result in better counts of large spider monkey sub-groups and faster surveys (S5). As a result drones are now being used to count spider monkeys in Mexico.

Dr Spaan, Population Monitoring Coordinator of the Mexican Conservation NGO ‘ConMonoMaya’: “… the drone allows for much faster coverage of an area than the [traditional] ground method does.” [S6]

The work drones and machine learning with EWT in South Africa has led to 10-15x faster sorting of images than through manual methods which saves time and money.

Dr Theron, Head of Drylands Conservation Program at the Endangered Wildlife Trust: “The research with drones and camera traps coupled with machine learning has shown that we can now detect Riverine Rabbits much faster in images than through our default method which would require human analysts going through tens of thousands of images. We have seen by way of example that manually sorting images will take 10 to 15 times the amount of time required to simply upload the images to the machine learning pipeline. This represents a significant amount of skilled resources that can be redirected to more productive conservation activities. The direct cost saving amounts to approximately £ 1500 per large camera trap survey (80 or more cameras) or more than 20 working days in one team members time.” [S7].

This drone work has also led WWF to implement drones in various of their conservation programs in Malaysia, Nepal, and Brazil [S5 & S8].

The fire detection study that was funded by GCRF and led to UR5 indicated that peat fires above and below ground could be detected with thermal sensors. On the basis of this started a collaboration with the Borneo Nature Foundation and the University of Palangkaraya to use drones for fire detection and monitoring and directly have impact. In 2019 the LJMU team went to Borneo to test the drone equipment during the 2019 fire season and train the local counterparts on the usage of the equipment. This has led to drones with thermal and visual spectrum sensors being used by the Borneo Nature Foundation to increase the efficiency of the fire fighting teams during the fires of 2019 [S9].

In his letter, Bernat Capilla, Director of Programs at BNF, states, “… the overwhelmingly positive feedback from fire-fighting teams and their desire for thermal drones to attend all possible fire events, attests to this positive impact. This derives from an increased ability and speed of detecting fires (compared to checking on foot, which is dangerous and may take many hours for a large fire, we estimate that use of a thermal drone allows fire hotspot location mapping to be completed around at least 10x quicker and with almost no safety risk), which in turn improves the ability of teams to monitor the current fire situation and deploy team members as necessary to prevent fire spread and extinguish hotspots. … The technology is also particularly useful in surveying fire sites post-initial fire-fighting, to confirm that all hotspots are extinguished and it is safe to leave the site, again increasing the speed of this by at least 10x. This is vitally important, as peat fires can smoulder beneath the surface, where they are difficult to detect without the aid of thermal imaging technology, and then re-emerge after fire-fighters have gone home, causing more damage.” This can be further summarized in two impacts. 1) Increased efficiency in finding/extinguishing fires: The fire-fighters estimate that compared to traditional methods, our drone system accurately identifies the location of the fires over 10x quicker, helps extinguish them 50% quicker and makes it over 10x quicker to confirm fires are extinguished. 2) Improved health of firefighters and surrounding population: the increased efficiency in finding/extinguishing the fires led to (i) virtually zero exposure to the previous safety risks finding fires; (ii) 200 days less smoke inhalation for the fire fighters, the effects of which are known to be extremely harmful; (iii) a corresponding drop in the concentration of toxic haze which had to be endured by the ~15 million Kalimantan province residents. In 2020 we intended to bring additional drone systems with new automated fire-detection capability to additional fire teams in Indonesia. We believe this would have significantly increased the above impacts. However, COVID-19 travel restrictions meant it was not possible to deliver the systems.

The research conducted in UR2-5 led to a grant from UKRI to directly impact conservation and economic growth in Madagascar in which LJMU, the Durrell Wildlife Conservation Trust, and Government of Madagascar are collaborating to use drones to protect biodiversity but also to stimulate economic growth through drone technology in other sectors of the economy in Madagascar. An initial survey in Madagascar has indicated that Bandro lemurs can be detected 100x faster with drones equipped with a thermal imaging camera compared to the traditional surveys by canoe and saves in operating costs [S10]. Dr Hudson, Head of Research, Durrell Wildlife Conservation Trust: “… the team then flew the system over key areas of marsh that have been inaccessible to previous surveys. In 3 x 20-minute flights, the drone system reproduced lemur densities estimated from previous surveys requiring 24 person-weeks of effort. Scaling-up this programme to cover the entire lake represents an approximately 100 times increase in survey efficiency. … The adoption of this research technology will revolutionise the scale and frequency with which DWCT and our partners in Madagascar, including the Madagascar National Parks Service, can conduct surveys.”

Dr Hudson: our drone system represents “…a saving of £8,000, annually. In the field of conservation where funding for ‘routine’ actions such as monitoring is hard won, this saving is a fantastic achievement and will help to guarantee the sustainability of our field programmes, long-term.” [CS3]

The research on habitat mapping in UR 6 has also allowed for much faster mapping of orangutan habitat.

Mr. Hadisiswoyo, Founder of the Orangutan Information Centre: “The drones have also allowed us to do habitat mapping much faster than if we were to do this on foot. I estimate that data collection is approximately 9 times faster and yields data that we would not be able to collect from the ground.” [S2]