ETH Zurich team takes home a quarter million in biodiversity prize

In the heart of the rainforest, the hum of insects, the calls of tropical birds, and the rustling of leaves creates a natural orchestra. Here every leaf, every drop of water, every species is part of a delicate balance of biodiversity that sustains the world in ways we are only just beginning to understand.

Conserving rainforests requires a greater understanding of these complex ecosystems. Rising to this challenge, team ETH BiodivX joined an initial 300 contenders for XPRIZE Rainforest - a global, 5-year competition to develop new technologies that expedite real-time monitoring of tropical diversity. Announced ahead of the G20 Summit in Rio de Janeiro, XPRIZE awarded the ETH BiodivX team the Bonus prize valued at a quarter of a millions USD, the proceeds of which will be dedicated, at least in part, to the development of technologies to monitor biodiversity and support local communities.

Biodiversity sensing drones and canopy rafts

Newly minted ETH Zurich professors, Stefano Mintchev and Kristy Deiner launched the ETH BiodivX team in the early stages of the pandemic. Combining their respective expertise in robotics and environmental DNA (eDNA), they conceptualized a method to automate the entire biodiversity monitoring pipeline minimizing human intervention into protected ecosystems. Over the past three years, the team collaborated with Zoo Zurich to conduct weekly field tests in the its humid Masoala rainforest biome to optimize its remote sensing and drone sampling probe.

As one of six teams to qualify for the finals, ETH BiodivX simplified and scaled an array of robust and reliable technologies to include: satellite- and drone-based remote sensing, probes that descend from a drone to collect eDNA surface and water samples, drone-deployed canopy rafts equipped with light and sticky traps, a camera to collect images of insects and the surrounding canopy, and, finally, audio sensors to record bioacoustics data.

Stefano Mintchev reflects on the final round of the competition that took place in July 2024 near Manaus, Brazil, ¡°We had the opportunity to come together and compete in the finals. With just 24 hours to survey and collect eDNA in a 100-hectare plot in the Amazon, I think we really performed well together.¡±

During this time the team flew four-drones for a total of 54 flights deploying and retrieving the canopy rafts, sampling eDNA from branches, foliage, and water, and conducting remote sensing flyovers. With only a further 48 hours to analyse and identify the eDNA collected the team worked feverishly to infer as much as possible about the biodiversity of the ecosystem.

Backpack lab key for rapid DNA sequencing

¡°Complex societal challenges take a transdisciplinary team,¡± says Kristy Deiner.¡± DNA analysis can take days, even months to yield a result, but one of the goals of the XPRIZE Rainforest competition was to provide near real-time biodiversity insights. To achieve this feat, Deiner worked with ETH Zurich professor, Lo?c Pellissier to devise a portable, field-ready backpack lab and a protocol for analysis which significantly reduced the time from a typical 100 hours to a mere 17 hours from sample to insight. The team also created a barcode sampling system and took advantage of rapid PCR testing technology, the development of which accelerated during the pandemic.

ETH BiodivX¡¯ revealed an abundance of life in the Amazon rainforest. They collected eDNA by swabbing vegetation over 740 times and by filtering 42 litres of water in the 24-hour collection period. From these samples, they identified over 240 species and many more that could not be matched to a scientific name using multiple detection methods. The team¡¯s findings include the first known drone footage of the rare bird ¨C the male Pompadour Cotinga (Xipholena punicea). Deiner notes, ¡°While the samples resulted in millions of DNA sequences, the team was only able identify a small percentage of the sequences, others remain unmatched, and some are likely unknown to science.¡±

While robot technology automated the collection process, AI and machine learning implemented by ETH Zurich Senior Researcher, David Dao sped up the analyses of images collected from satellite data and the thousands of sounds detected. He also supported the team by aggregating interactive dashboards to make the biodiversity data accessible.

It takes a global village to fulfil a dream

Dao, who is a co-founder of the non-profit GainForest, was initially an XPRIZE competitor. Prior to the Singapore semi-finals in 2023, Dao merged his team with Deiner and Mintchev¡¯s to form and co-lead ETH BiodivX. He facilitated the team¡¯s work with Indigenous Peoples and local communities in the Amazon region. He encouraged Marina Mura, an Indigenous person from S?o Sebasti?o do Uatum?, to log her experiences working with the team. Mura writes, ¡°What we have learned most throughout our ancestry is to hear and feel the forest, imagine being able to put this on a computer and identify every sound that nature makes, for me it was a dream coming true.¡±

Dao used AI chatbots to digitize Indigenous knowledge and wisdom about the forest and the properties of medicinal plants. He also laid the groundwork to incentivize conservation by setting up a web3-based, eco-economy payment platform. Indigenous and local communities in the Amazon became executive members of the ETH BiodivX team that encompassed 51 talented, multidisciplinary researchers working in scientific, economic, journalistic, and artistic knowledge fields. Within the team 14 countries and institutions were represented. The team also included ETH Zurich professors Stefano Mintchev, Kristy Deiner, Lo?c Pellissier, and Tom Crowther, three ETH Spinoffs - Restor, Diaxxo, and SimplexDNA, as well as numerous supporters and sponsors.