Scientists develop model to improve biodiversity protection using AI technology

Release date: 24 March 2022

  • New model using artificial intelligence (AI) helps identify areas for urgent conservation
  • Scientists argue ‘videogame-like’ AI simulations can outperform current conservation tactics
  • The model is hoped to efficiently address biodiversity loss and save species from extinction

In a new paper, published today in journal Nature Sustainability, scientists from the Royal Botanic Gardens, Kew and partners highlight how artificial intelligence can be used to improve conservation policies and detail a novel framework that could be used to prioritise areas for protection. This model comes at a crucial point for the planet as plants are being lost at a rate unprecedented in human history.

The international team of scientists, from RBG Kew, the University of Fribourg in Switzerland, the University of Gothenburg in Sweden and London-based company Thymia Ltd, drew evidence from biology, environmental economics, and computer science to develop a new model to determine where to most efficiently establish protected areas in a region or country.

AI holds great promise for improving the conservation and sustainable use of biodiversity and ecosystem services in a rapidly changing and resource-limited world. AI-optimised models were found by the authors to outperform alternative conservation policies that look at simply protecting the most species-rich areas or the largest extent possible with limited budgets – strategies that often result in preventable species loss. For example, focusing conservation efforts on protecting the largest extent of area will prioritise based on the cost of land, rather than on the complementarity among protected areas, resulting in a less efficient conservation policy.

Professor Alexandre Antonelli, Director of Science at the Royal Botanic Gardens, Kew says:Increasingly urgent calls for action to halt biodiversity loss are ultimately constrained by the financial resources that governments are able to allocate for conservation initiatives. Finding a way to efficiently protect nature is further complicated by increasing widespread human pressure on ecosystems and the escalating climate crisis. This presented to us an opportunity to develop a conservation tool that can monitor biodiversity efficiently and prioritise areas that are in most need of protection.

To do this, the scientists created a software named CAPTAIN (Conservation Area Prioritisation Through Artificial INtelligence) to integrate biodiversity data, conservation budgets, climate change and human pressure – for example, based on land use changes captured from satellite images. The framework feeds data into a neural network and quantifies trade-offs between the costs and benefits of area and biodiversity protection, exploring multiple biodiversity metrics.

CAPTAIN protects substantially more species from extinction than areas selected randomly or naively, such as areas based solely on species richness, and outperforms commonly-used conservation tools. An analysis of hundreds of simulations saw CAPTAIN outperform popular conservation planning software Marxan in 64 to 77 per cent of cases, depending on the settings, boasting an average improvement of prevented species loss between nine and 18 per cent.

AI-based solutions to conservation have been sparingly used in the past, but the study’s authors believe CAPTAIN is the first of its kind to implement reinforcement learning (RL). Models trained through RL can highlight areas of interest for policymakers, based on biodiversity simulations through time and in response to external pressures such as climate change and human activity. The resulting conservation tool can inform on-the-ground decision-making in areas facing biodiversity loss.

The tool was tested on a real-world database of more than 1,500 trees native to Madagascar. The data had been previously used in a conservation planning experiment using Marxan. CAPTAIN was found to consistently outperform Marxan within the defined budget constraints, increasing the average fraction of protected range per species by 50 per cent and even exceeding the set conservation targets.

Daniele Silvestro, computational biologist at the University of Fribourg and lead author of the paper says:To optimize our AI models, we simulate an artificial world that includes many species exposed to human pressure – for example, direct exploitation or land use changes - and climate change. We then let the algorithm play policymaker like in a videogame, where the reward is how many species were spared from extinction at the end of the game. The program plays the game many times, after which it learns how to best place protected areas in this simulated world. After this training phase, the algorithm is ready to be applied to real-world data.

Antonelli says:Given that not a single of the 20 Aichi biodiversity targets internationally agreed upon in 2010 has been fully met, it is clear that we need to rethink how effective and realistic conservation policies are drawn up. In the run-up to COP15 in China this year, where new biodiversity targets will be set for the coming decades, we think AI can be a game-changing tool to help policymakers to make the best use of available data and halt irreversible biodiversity loss. A key learning from our study is that quality is more important than quantity: just increasing the amount of protected land and sea isn’t enough, we need to protect the right regions”.

The Aichi biodiversity targets represent a global commitment to address the underlying causes of biodiversity loss and to reduce the pressures impacting nature today. These include 20 specific targets, such as increasing protective measures over biodiverse areas, that governments have pledged to meet by 2020 – but have failed to achieve.

ENDS

Notes to Editors

For high-res images, please download from the following link and credit as named: https://we.tl/t-C3iojKmSy6

About the Royal Botanic Gardens, Kew

The Royal Botanic Gardens, Kew is a world-famous scientific organisation, internationally respected for its outstanding collections as well as its scientific expertise in plant and fungal diversity, conservation and sustainable development in the UK and around the world. Kew Gardens is a major international and a top London visitor attraction. Kew Gardens’ 132 hectares of landscaped gardens, and Wakehurst, Kew’s Wild Botanic Garden, attract over 2.5 million visits every year. Kew Gardens was made a UNESCO World Heritage Site in July 2003 and celebrated its 260th anniversary in 2019. Wakehurst is home to Kew's Millennium Seed Bank, the largest wild plant seed bank in the world. The Kew Madagascar Conservation Centre is Kew’s third research centre and only overseas office. RBG Kew receives approximately one third of its funding from Government through the Department for the Environment, Food and Rural Affairs (Defra) and research councils. Further funding needed to support RBG Kew’s vital work comes from donors, membership and commercial activity including ticket sales.

About the University of Fribourg

The University of Fribourg in Switzerland encompasses five faculties where people study, teach and research. These are Arts and Humanities, Science and Medicine, Management, Economics and Social Sciences, Law and Theology. As well as these there are numerous interdisciplinary institutes and centres of excellence. The approximately 10,000 students in the Bachelor, Masters and PhD programmes receive first-class support from over 800 professors, lecturers and research assistants.

About Thymia Ltd

Thymia is a London-based mental health tech start-up. Utilising the latest advances in the fields of Neuropsychology, Linguistics and Ethical AI, they analyse speech patterns, visual cues and the output of specially designed video games in order to assess depression. Their online platform offers a comprehensive suite of tools used by clinicians and patients before, during and after each appointment, providing a way for clinicians and patients to objectively benchmark and monitor the progress of specific depressive symptoms treatment response as well as treatment response more generally.