Opportunity seeds

Conserving ecosystems into the future requires a predictive ecology revolution. Foundational AI models are able to surpass the limitations of existing predictive models by encoding high complexity across global multi-sensor environmental data into a single digital representation, an ‘embedding’, which can be used as data in downstream analyses.

This project aims to produce the first embeddings explicitly trained on future climate projections across the land-sea interface up to 2100. The team will be focusing their approach on UK seabirds, a highly threatened yet functionally critical and mobile group spanning marine and terrestrial ecosystems, validated by tracking seabird's movements at the UK's warming edge on the Isles of Scilly. This provides a scalable testbed for extending predictive ecosystem engineering to other vulnerable taxa and regions.

The overall aim is to create a globally scalable toolkit for predicting future habitat suitability, identifying populations at risk of local extinction, and highlighting places where targeted interventions could improve resilience.

Freshwaters are the largest natural methane source to the atmosphere, contributing around a third of global methane emissions, yet there are few practical tools to reduce emissions in these systems. The high oxygen and methane content of rivers make suitable habitats for methane oxidising bacteria (MOB), microorganisms that can sequester methane by converting it to CO2 and biomass.

This project looks to enrich MOB communities on moss to create nature-based biofilters to deploy in stretches of river with high methane concentrations. This project will identify optimal conditions that balance the growth and methane sequestering capabilities of the moss-bacteria composite, develop models to predict stretches of rivers with high methane concentrations where intervention is needed, and apply methane monitoring and prototype biofilters to identified stretches. Due to the native range of the moss, these biofilters could provide a natural in-stream carbon management tool that can be applied across the temperate northern hemisphere.

Invasive species are a major global challenge, damaging biodiversity, disrupting ecosystems, and costing millions of pounds to manage. One such species is Rhododendron ponticum, which is widespread across the UK. R. ponticum stifles understory growth, outcompeting all other species, and threatening forest regeneration, particularly oak. Removal is the only control strategy currently effective, but the costs are prohibitive.

This project will survey R. ponticum populations in North Wales to measure how frequently it is eaten by native insect species. Using lab experiments and genomic analyses, the team aims to understand the mechanisms that allow these insects to feed on the toxic plant. By doing so, this project will shed light on how native insects can utilise invasive species as new host plants, and potentially lead us to genetic variants that may help control R. ponticum in the future.

Vegetation has a three-dimensional structure that shapes how ecosystems function. It affects light, temperature, habitats, water and carbon cycling, and how well ecosystems recover after disturbance. Yet this structural information is still rarely used in day-to-day environmental monitoring, because collecting it has usually required expensive laser scanners and specialist teams.

This project aims to change that by developing a low-cost, repeatable way to capture detailed 3D measurements of trees, enabling more frequent monitoring of ecosystem structure at a larger scale. By combining new low-cost LiDAR sensing technology, tailored workflows, and automated software, the project will turn specialist 3D surveying into a practical tool for routine monitoring.

This approach will allow users to produce reliable ‘3D structural fingerprints’ of trees and forests, helping them detect change earlier, assess whether management actions are working, compare sites more consistently over time and reveal new structural indicators.

Zebra mussels (Dreissena polymorpha) are among the most damaging freshwater invaders in the UK, altering food webs, degrading water quality, and driving native species decline. Current responses rely largely on chemical or biological interventions that act at ecosystem scale, are difficult to reverse, and risk unintended consequences.

PRIZM will explore whether autonomous robotic systems can enable a new class of direct, reversible, and targeted ecological intervention by using advanced sensing and precise mechanical tools to detect and physically remove invasive species.

Specifically, the team will develop a robotic system capable of navigating dark, confined, turbid, and flowing environments to detect, detach, and collect invasive mussels from submerged infrastructure under live operating conditions. This will establish the feasibility of closed-loop robotic intervention for invasive species in real-world infrastructure, presenting a new opportunity for ecosystem management.

To protect and restore threatened species and their habitats, we need both accurate and timely species-level data, and data proxies for poorly known taxonomic groups and ecosystem functions. Nature provides a solution: birds as roving sensor platforms – an ‘Internet of Birds (IoB)’.

Birds are critically important in nutrient cycling, pollination, and seed dispersal, and also serve as valuable bioindicators for climate change, pollution, water quality, and biodiversity. However, current bird monitoring methods have significant limitations in terms of spatio-temporal coverage, effort vs. data-return, and tag size, weight, cost, and battery life.

This project will develop bird leg-rings that are small, light, inexpensive, solar powered, and continuously globally tracked. This approach will eliminate species size or tag cost restrictions and frequency-related data bias. The data could inform individual species conservation, environmental impact mitigation for infrastructure, zoonotic disease modelling, and interventions to halt biodiversity loss and protect ecosystem functions.

While genetic biobanks have become an essential component of conservation strategy, they remain fundamentally archival: frozen samples preserve genetic diversity, but they do not allow us to use it to understand resilience, vulnerability, or adaptive potential. By ‘immortalising’ the full genome of an individual in a self-renewing cellular state, induced pluripotent stem cells (iPSCs) have the potential to transform genetic diversity from a static, archival resource into a living, experimentally tractable tool.

To make iPSC technology a broadly applicable tool for ecosystem science, this project aims to create a portable programming capability by engineering next-generation transcription factors through directed evolution and deep mutational scanning. Such transcription factors would enable experimental interrogation of genetic diversity before it is lost, support functional testing of resilience strategies in contained settings, and create a bridge between genomics, cellular biology, and ecosystem-level decision-making.

To strengthen ecosystem resilience, conservationists need fast feedback on whether animals are coping with heat, hunger, disease and other pressures. Today, most monitoring tells us where animals are or how many remain, but provide little insight into their health.

This project aims to create a new class of remote, non-contact health scanner for wildlife: a way to read vital signs, energy reserves, body condition and thermal stress from a distance, without capture, sedation or tagging.

Working with engineers, ecologists and the UK zoo network, the team aims to build and validate a prototype that turns external measurements of an animal into estimates of hidden physiological state. The central challenge is that coats, movement and environmental conditions distort biological signals. This approach combines complementary sensing with physics-informed models so that the animal’s true condition can be estimated robustly in real-world settings.

Pollinator corridors are vital for restoring biodiversity, but in many of the UK’s most degraded, compacted and space-constrained landscapes, these fragile systems are difficult to establish and even harder to sustain.

This project will develop an autonomous rover to create pollinator ecosystems that are more resilient, self-stabilising and lower maintenance over time by precisely shaping the subsurface environment around plant roots, deploying beneficial soil biology and combining this with spatially intelligent planting of UK-native flowering species.

New autonomous innovations open up the possibility to create pollinator corridors in fundamentally new ways, building the hidden biological and structural conditions that help habitats take hold and endure, season after season.

Invasive species pose severe threats to biodiversity; in a UK context this might be mice and rats attacking nesting birds on island sanctuaries, grey squirrels, signal crayfish and more. Genetics-based approaches to control the population of species like disease-carrying mosquitoes have been under development for years, and could in theory be applied to invasive species for conservation purposes. However, these approaches are species-specific and time-intensive to develop – they also rely on micro-injection of embryos, which isn't feasible for many species.

This project aims to develop a new method that can be more easily transferred from one invasive pest species to another by harnessing a form of extrachromosomal inheritance. In principle, this system could be adapted to counter both disease-bearing organisms and also invasive species in ecologically sensitive areas, with far fewer off-target effects than present methods that are reliant on the use of chemical toxins.

As global climate instability accelerates, the restoration of complex biological networks is vital to ensuring ecosystem resilience. Reforestation can be a key tool for this, but success requires more than just planting trees. Wood ants (Formica rufa group) are keystone ecosystem engineers that support forests by driving multi-trophic biodiversity. However, these ants are poor at dispersing to new habitats naturally, so their colonies need to be translocated to target forests – a process whose success depends on a canopy-based ant-aphid mutualism.

Located up to 40m high, this ‘biological engine’ is inaccessible to traditional ground-based monitoring. To address this, the ARISE project will develop a first-of-its-kind diagnostic pipeline that integrates drone-enabled micro-sampling, nutritional metabolomics, and hyperspectral imaging with machine learning to remotely quantify aphid health. Field-validated through ongoing translocations, this embodied-AI-enabled solution aims to deliver precision ecology in a scalable format, providing the data-driven insights required to engineer resilient, biodiverse forests.

Resilient ecosystems need more than just enough individuals of each species – populations in ecosystems also need to be genetically diverse so they can survive long term and adapt to environmental change. Predicting genetic diversity for ecosystems is complex and computationally demanding, and as a result is under-considered when protecting or restoring habitats.

This project will develop new machine learning tools to help design landscapes that better support genetic diversity across multiple species. Using a combination of individual-based simulations and machine learning, the team will explore how the arrangement, size, and connectivity of habitats influence genetic variation for different species.

The approach will allow rapid exploration of many possible landscape designs, identification of those that perform best, and discovery of habitat features that consistently promote genetic resilience. By revealing general principles that link landscape structure to genetic diversity, this work aims to make genetic diversity a practical consideration in conservation planning.

Currently, applying beneficial microbes to natural environments is hindered by an inability to effectively guide, track, and recover them once released. To overcome this, this project aims to develop a novel ‘magnetobiotics’ platform. By magnetically labelling microbes, researchers can track their movement using ultra-sensitive quantum sensors based on nitrogen-vacancy centres in diamond. This technology will be integrated into a portable chip, allowing scientists to monitor microbial migration.

As a proof-of-concept, the team will target the deadly chytrid fungus, a pathogen responsible for severe global declines in amphibian populations. Using magnetic gradients, they plan to deploy a library of anti-fungal bacteria directly to infection hotspots and safely recover them afterwards. This controllable, targeted therapy promises to provide a surgical alternative to environmentally damaging chemical treatments, establishing a new paradigm for ecological conservation.

A healthy ocean is essential to life on Earth, supporting extraordinary biodiversity and providing critical ecosystem services. Yet many marine areas remain poorly mapped and infrequently monitored whilst undergoing rapid alteration from human activity and climate change.

This project will develop new methods to facilitate and upscale marine ecosystems monitoring using automated pipelines built using cutting-edge technology. Autonomous underwater vehicles (AUVs) deployed from shore will collect acoustic, image, and bathymetry data, which will then be analysed using AI to identify habitats and explore how they relate to patterns of biodiversity. The work will generate a proof of concept for more efficient marine monitoring, an open dataset from a UK marine protected area, and a methodology that can be expanded to other marine environments. By making it easier to map and assess underwater habitats, the project aims to support better protection and management of ocean biodiversity.