Environment

The focus of the environment theme is to seek methodological innovations that can transform our understanding and management of the natural environment. This is a major cross-disciplinary challenge requiring a close collaboration between environmental scientists, computer sciences, statisticians, social scientists, and many others.

The Environment theme aims to develop new understanding and innovative solutions to the dual crises of climate change and biodiversity loss, which are inextricably linked. This time-critical mission requires close cross-disciplinary collaboration between ecologists, environmental scientists, computer sciences, statisticians, social scientists, and many others.

Biodiversity is under unprecedented threat due to habitat destruction, climate change, pollution, and other human-driven pressures. Understanding and addressing these challenges requires innovative approaches that can make sense of complex ecological systems, large yet fragmented datasets, and rapid environmental changes.

At the intersection of data science, artificial intelligence, and environmental research, we seek to develop and apply cutting-edge methods that:

Capture and analyse the complexity, interconnectedness, and uncertainty inherent in ecosystems.
Overcome challenges posed by sparse and irregular data from extreme environmental events and biological organisms.
Integrate and interpret heterogeneous biodiversity and environmental datasets over space, time and taxa.
Detect critical shifts in ecosystems, such as species population declines or habitat degradation, across scales.

Emerging technologies, including acoustic telemetry, photogrammetry, remote sensing, bioacoustics, automated species identification, and cloud and exascale computing, have the potential to revolutionise biodiversity monitoring. These technologies enable real-time, high-resolution data collection on species movements, environmental conditions, and ecosystem dynamics, while cloud and exascale computing provides the computational power to integrate and analyse vast datasets across disciplines and regions.

AI and machine learning further enhance our ability to process and interpret these complex datasets, offering predictive models for biodiversity loss, early warning systems for ecosystem collapse, and data-driven strategies for conservation and restoration.

By leveraging expertise across environmental sciences, computer science, statistics, and social sciences, we aim to position data science and AI at the forefront of tackling biodiversity loss and broader environmental challenges. Our goal is to establish the Data Science Institute as a global leader in harnessing digital innovation for a nature positive future.

Theme Lead

Dr Sally Keith

Senior Lecturer in Marine Biology

Centre for Sustainable Soils, Centre of Excellence in Environmental Data Science, DSAIL- Environment, Ecology and Conservation

enable JavaScript to view email address A25, A - Floor, LEC 3

A woman with a glucose monitor in her arm

A Breath of Fresh Air

Monitors that identify air-pollution sources enable industry and governments to act appropriately to improve health, compliance and public confidence. However, some existing monitors lack directional information, making it difficult to distinguish pollution sources. Lancaster researchers have invented a simple monitor that resolves wind directions and so identifies pollution sources and trends.

The research team of Professors Duncan Whyatt, Kevin Jones and Roger Timmis brought together expertise in pollution sampling and spatial analysis along with expertise in end-users’ needs from the Environment Agency (EA). This research resulted in two patents for passive (unpowered) directional samplers of air quality.

These cover a range of pollutants including nitrogen dioxide (traffic), fugitive particles (steelworks), benzene (refineries), dusts (waste and recycling processes) and ammonia (intensive farming). The samplers have been patented internationally and exclusively licensed to SGS Galson for commercial applications.

Lancaster’s samplers were deployed at the Harsco Metals and Tarmac/Lafarge sites adjoining Scunthorpe Steelworks and raised site operators’ awareness of dust issues, leading to the adoption of new management strategies. The EA site inspector confirmed an 80% reduction in the total dust signal. A risk of fines of £30 million from the EU was avoided.
Lancaster’s novel methods featured in a report by DEFRA’s Air Quality Expert Group. The EA also used the methods to investigate dust complaints at the European Metal Recycling works in Newhaven, East Sussex.
US Environmental Protection Agency legislation (EPA325) requires fence-line monitoring of the carcinogen benzene. SGS are the world’s leading inspection and certification company and read about the success at Scunthorpe; they swiftly acquired exclusive licenses to both sampler patents and started manufacturing under the name ‘AIHR Shark’.
There are about 180 oil refineries in the US and ‘Sharks’ have been deployed at the most problematic. Refineries must minimise on-site electrical equipment, so Lancaster’s unpowered DPAS technique was particularly suitable. Conventional sampling failed to identify the sources of emissions but ’Sharks’ pinpointed them after only 10 days. Operators called it “a simply genius technology”. SGS Galson highlight that the sampler has “yet to fail to identify an emission source to date”.
The EPA 325 compliance monitoring market in the US is already worth several million dollars. This market value reflects the costs of non-compliance, as evidenced by a $4.64 million fine that the EPA served on Valero Energy in October 2020. Sharks have also been deployed at mining sites in the US.

Lancaster Environment Centre research helps reduce uncertainty in flood risk assessments

Significant flood events between 2000 and 2020 cost up to £4.2 billion in economic damages, claimed 27 lives and flooded 90,000 households. Environmentalists from Lancaster worked with government to improve understanding of flood risk and demonstrate where adaptations can have the most impact, protecting life and billions in assets. As a result, the way flooding is represented in the UK risk register was improved.

Improved flood risk modelling based on risk and uncertainty research by Distinguished Professor Keith Beven and Professor Rob Lamb has been applied across the UK to provide the Government with a more realistic understanding of inland flood risk, from catchment to national scales – protecting life and assets – and resulting in improvements to the UK National Risk Register.

The application of novel statistical methods (developed with Professor Jonathan Tawn at Lancaster) to model extreme river flows and rainfall improved the Government’s appreciation of flood resilience and was cited in the National Security Risk Assessment, underpinning government emergency planning.
After damaging floods in the UK between 2013 and 2016, the Government established a (NFRR) based on evidence from a Scientific Advisory Group, of which Lamb was a member and contributed research. The NFRR committed the Government to £12.5 million on new mobile flood defences, raising their number four-fold, as announced in Parliament and by the BBC. This in addition to supporting the Government’s on-going spend of £2.3 billion to protect 300,000 homes.
The Construction Industry Research and Information Association produced Guide C721 based on the research – good practice in the assessment of uncertainty in flood risk mapping.
Environmental “models of everywhere” was adopted by industry partner JBA to improve national flood risk maps. The Environment Agency procured the maps and Lamb and Beven’s interdisciplinary research has influenced the 2nd generation of the National Flood Risk Assessment, to help identify and alleviate the risk of flooding that currently affects more than 5.2 million properties in England.