10
ZSL
Big Data In The Zoo And To Protect Animals
Background
As well as running the world-famous London Zoo, ZSL are responsible for far-reaching conservation efforts around the world, in an effort to tackle the ongoing extinction threats that many species are facing.
Traditionally, most conservation work has been fieldwork carried out on the ground by zoologists and other scientists, manually tracking the movement of animal populations or the spread of vegetation using tracking devices or their own eyes.
However, with the increasing sophistication of data collection and analytics technology, along with the increasingly urgent need to take action to save an ever-growing number of species from extinction, new methods are being constantly developed to aid in the monitoring and tracking of wildlife.
Last year, ZSL together with other research organizations, including NASA and the European Commission’s Joint Research Council, held the first international symposium on “remote sensing” for conservation. In essence, remote sensing involves combining the latest high-resolution satellite imagery with zoological, demographic and geographical data and advanced computer modelling and predictive analytics, to better understand the effects that human activity is having on animal and plant populations.
What Problem Is Big Data Helping To Solve?
The threat to biodiversity from human activity is one of the biggest challenges facing us as a species. The ecosystem which we all depend on for survival is reliant on a complex balance of organisms that has developed over millions of years and makes the planet suitable for sustaining life.
The rise of humans as the dominant species on earth has had disastrous effects on the extinction rates of other species. As it is thought that as little as 15% of the plant, mammals, reptiles, fish and insect species on earth have been identified, the vast majority of these extinctions go unnoticed – it is estimated that the planet is losing as many as 140,000 species a year.
There is no way of knowing for sure what the long-term impact of this decrease in biodiversity will be. Plant and animal ecosystems interact with each other and with human life in an incredibly complex range of ways, from the food chain to the nitrogen cycle that keeps the air breathable. Creating imbalance in these processes could have potentially disastrous consequences for us as a species as well as all life on the planet.
Conservation efforts like those established by ZSL through their Institute of Zoology are essential to understanding the consequences of the damage that has already been done, and to working to mitigate it.
How Is Big Data Used In Practice?
In recent years, thanks to huge advances in technology, computer processing power and analytical science, focus has moved to developing methods of tracking, quantifying and understanding animal populations through remote sensing.
Conservation work is hugely expensive, and vastly underfunded considering the scale of the problem. To find ways to overcome this, last year ZSL brought together experts from a wide range of scientific establishments, NGOs and charities. Graeme Buchanan of the Royal Society for the Protection of Birds (RSPB) told the symposium: “It is established that the funding available to conservation falls well short of the amount needed to address the current extinction crisis.
“Consequently, conservationists need to target funding to where it is most needed. To set priorities, the conservation community needs information on what is where and how is or might be changing. While in situ field data is invaluable, remote sensing can provide a valuable tool for delivering this information.”
Plans discussed as part of the programme include using satellite imagery to track populations from space in order to track animal movement and the impact of human activity such as deforestation and urbanization.
This data can then be used with predictive modelling algorithms to anticipate future movements as well as to highlight geographical areas where animal populations are particularly at risk, or where urgent human intervention could prevent the loss of invaluable biodiversity from the planet through extinction.
Understanding animal population movement and dynamics is thought to be key to predicting the effect that human activity will have on the species we share the planet with, and data analytics is proving to be an effective tool for developing that understanding.
What Were The Results?
Through the application of these ideas, many new frameworks have been developed that can be used by scientists around the globe to study and predict the movement of animals, and the effect manmade changes to their environment are likely to have.
Thanks to this, conservation groups, charities and political lobbyists can focus efforts on making change where it will be most effective in halting the ongoing extinction of life on earth.
What Data Was Used?
Very high resolution (VHR) satellite imaging technology has reached the stage where images are detailed enough to show individual animals (and people). This data can then be fed into counting algorithms that quantify the size of a particular population in a particular area.
As well as quantities, migration patterns can be captured and assessed. This allows for modelling of likely migration pathways in other locations, based on data extrapolated from the observed populations.
As well as this, more down-to-earth data is also collected from camera traps, observers in the field and, increasingly, drone aircraft fitted with camera equipment.
One programme involves monitoring photographs posted by tourists and the wider public to social media, which can be scanned with image-recognition software. This software can be programmed to recognize animal or plant life, determine the location through the photo’s metadata and use it to build another reference database of biodiversity within a given area.
Biological information, existing data on species distribution and human demographic data are all also used to predict and assess animal populations and how they are being affected by outside influences.
Satellite data from systems used by NASA to monitor forest fires can also be incorporated into these programmes, to monitor the effects of deforestation by fire. LiDAR technology, which replaces radar’s radio waves with light waves, is used to determine the height and biomass density of vegetation in a particular area, allowing for more accurate predictions of the variety and volume of animal species inhabiting it.
What Are The Technical Details?
Datasets collected by ZSL’s migration-tracking programmes and used to inform international population indexes is hosted on Amazon Web Services (AWS) and Microsoft Azure frameworks. The society’s analysts also heavily use the open-source H20 analytics platform, which allows complex analytics to be run on the distributed datasets and results to be fed back over a Web browser interface.
Dr Robin Freeman, head of indicators and assessments at the society, tells me: “Almost everyone I work with uses R [the statistical program] to some extent.
“The things you need to learn as a graduate or research student at a zoo are shifting more towards understanding statistical methods, and machine-learning programming, because it is becoming increasingly certain that you will come up against Big Data in your research.”
Any Challenges That Had To Be Overcome?
In conservation work, the biggest challenge is prioritization. With so many species disappearing at such a fast rate, it is essential that methods be developed for identifying those most at risk. This enables the efficient and effective deployment of resources, as well as the opportunity to instigate or campaign for societal changes, such as regulation, necessary to prevent loss of biodiversity.
Using data gathered through remote-sensing methods, the Wildlife Conservation Society have established a group of researchers from scientific establishments, governments and NGOs committed to identifying the 10-most-pressing problems currently facing conservationists. These include predicting future deforestation, identifying hotspots where habitat change is leading to high levels of extinction and the expense of accessing accurate data.
These priorities are used to inform global conservation efforts such as those undertaken by ZSL in order to help ensure the most effective steps are being taken.
What Are The Key Learning Points And Takeaways?
Conservation work is vital to the future of life on earth, and Big Data analytics are an essential ingredient. Access to more accurate and timely data is improving our ability to understand and anticipate the effects human activity has on the global wildlife population, and how those changes will inevitably come back to bite us.
Data gathered through remote sensing cuts down on the need for exp-ensive, time-consuming and sometimes dangerous fieldwork zoologists need to carry out. While ground-based sensors and manmade observations will still provide reliable data, increasingly it can be inferred from satellite imagery combined with geographic, biological and demographic data to produce accurate models and predictions.
As analytics technology becomes more advanced, we will be able to gain an increasingly clear picture of where our priorities should lie if we want to mitigate the damage we have already caused to the ecosystem.
REFERENCES AND FURTHER READING
For more information about ZSL and Big Data, visit: