Project highlights

  • Multidisciplinarity – This project combines bioinformatics, climate science modelling and artificial intelligence to develop user-friendly tools for the monitoring and forecasting of biodiversity loss in different climate and pollution scenarios.
  • Big data science – This project uses advanced computational tools and biostatistics to: 1) identify molecular taxonomic units (MOTUs) that delivery ecosystem functions; 2) rank environmental pollutants based on their adverse effect on biodiversity and ecosystem functions. These actions guide conservation and mitigation interventions.
  • Link to policy – The co-design and co-supervision of the PhD by the UK Environment Agency ensures the translation and dissemination of intervention mechanisms to transform biodiversity science and environmental practice. The analytical tools of this proposal will be developed using Data Visualization Technology for direct applications by regulators.

Overview

Lack of understanding of the interlinked processes underpinning ecosystem services has led to mismanagement, with negative impacts on the environment, the economy and our own wellbeing. Managing biodiversity whilst ensuring the delivery of ecosystem services is a complex problem because of limited resources, competing objectives and the need for economic profitability [1]. Protecting every species is impossible. We need a whole-system, evidence-based approach in order to make the right decisions in the future.

Biodiversity loss happens over many years and is often caused by the cumulative effect of multiple environmental threats [2]. Only by quantifying biodiversity, during and after pollution and climatic events, can the causes of biodiversity and ecosystem service loss be identified [3].

Continuous temporal data, including palaeoecological, chemical and environmental data collected from sedimentary archives will used to establish past correlations that inform ecological process-based models. We will develop Emulators that ‘learn’ from past correlations, and forecasts of biodiversity and ecosystem functions with measured uncertainties. Interactions among several biotic and abiotic variables will be used to reduce uncertainty in forecasts (e.g. [4] [4]). Generating predictions that account for all these variables in different scenarios (e.g. projected IPCC climate scenarios) is computationally intensive and time consuming. Emulators can provide robust predictions with calculated uncertainties across multiple scenarios while reducing computational cost and time. An ‘emulator’ is a low-order, computationally efficient model which emulates the specified output of a more complex model in function of its inputs and parameters. Emulators are widely applied in big data science, such as i) climate science to generate predictions under different socio-economic scenarios in long-range simulations (e.g. [5] and references therein); ii) ecology to predict the status of ecological processes in changing environments using e.g. long-term remote sensing data [4]; and iii) environmental science to predict e.g. the hydrological status of water reservoirs [6]. To overcome adoption barriers by stakeholders, the applicant will develop an AI-based Emulator dashboard, accessible to regulators and policy makers through data visualizations techniques. These tools can be adapted for probabilistic predictions of ecosystem services to aid decision-making and socio-economic trade-offs.

CENTA Flagship

This is a CENTA Flagship Project

Case funding

This project is suitable for CASE funding

Host

University of Birmingham

Theme

  • Climate and Environmental Sustainability

Supervisors

Project investigator

Co-investigators

  • Dr Luisa Orsini, School of Biosciences, University of Birmingham and The Alan Turing Institute, British Library, 96 Euston Road, London NW1 2DB, UK ([email protected])
  • Dr Scott Hosking, The Alan Turing Institute, British Library, 96 Euston Road, London NW1 2DB, UK ([email protected]) and British Antarctic Survey, Natural Environment Research Council, Cambridge, CB3 0ET, UK
  • Drs Glenn Watts and Kerry Walsh Environment Agency ([email protected]; [email protected])

How to apply

Methodology

Biocomputing. Metabarcoding or marker gene sequencing, stoichiometric and chemical data have been collected from sediment cores in shallow lakes. The DR will learn to analyse these data using bioinformatics, biogeochemistry approaches and multivariate statistics.

Climate data analysis. The DR will learn how to collect and analyse climate data from weather stations and public databases. These data, together with gene sequencing, stoichiometric and chemical data will be used to explain past biodiversity trends.

Forecasting with AI. Sparse Canonical Correlation Analysis (sCCA) will be employed to regress measures of biodiversity attributes on biotic and abiotic factors in standard, multiple regressions using generalized linear models (GLM). sCCA is ideal for discovering complex, group-wise patterns between high-dimensional datasets.

Translation. To create long-lasting impacts beyond the project, the DR will have placements at the UK EA, during which they will engage with the research team, ensure transfer of knowledge and drive translation of research findings into environmental practice.  

Training and skills

The DR will receive multisciplinary training spanning from evolutionary biology and biodiversity science (Orsini), climate science (Hosking), and AI (Zhou). The DR will benefit from access to some of the most up to date high performance computing facilities in the country, including UoB and Alan Turing facilities. The DR will spend up to 6 months at the UK Environment Agency to learn the skills of translational science.

Partners and collaboration

This is a CASE application in which the UKEA offers funding at £1K per year, and placements for the DR in the applied research team at the UKEA. Dr Watts works in the Evidence Directorate of the Environment Agency, leading a research team of 11 specialising in climate change and resource efficiency. Dr Walsh specializes in molecular approaches for understanding species presence in freshwater ecosystems. She has been responsible for introducing operational DNA approaches at the UKEA. Both have co-designed the proposed project and will offer co-supervision and training to the DR.

Further details

For inquiries please contact [email protected] and [email protected]

If you wish to apply to the project please visit: https://sits.bham.ac.uk/lpages/LES068.htm

Possible timeline

Year 1

Collection of climate data and correlation among biotic and abiotic changes through time and space. Student conference in Birmingham.

Year 2

Forecasting using AI. Prepare draft of first thesis chapter. Placement at the Alan Turing Institute to learn the use of machine Learning algorithms and artificial intelligence in model forecasts.

Year 3

Placement at the UK Environment Agency and development of an AI-based Emulator dashboard – develop predictive tools for regulators. Present thesis work in international conferences. Write thesis. Submit research article.

Further reading

  1. Jax, K. et al. (2018) Handling a messy world: Lessons learned when trying to make the ecosystem services concept operational. Ecosystem Services 29, 415-427.
  2. Nogues-Bravo, D. et al. (2018) Cracking the Code of Biodiversity Responses to Past Climate Change. Trends Ecol Evol 33 (10), 765-776.
  3. Eastwood, N. et al. (2021) The Time Machine framework: monitoring and prediction of biodiversity loss. Trends in Ecology & Evolution in press.
  4. Leeds, W.B. et al. (2014) Emulator-assisted reduced-rank ecological data assimilation for nonlinear multivariate dynamical spatio-temporal processes. Statistical Methodology 17, 126-138.
  5. Chantry, M. et al. (2021) Opportunities and challenges for machine learning in weather and climate modelling: hard, medium and soft AI. Philos Trans A 379, 20200083.
  6. Asher, M.J. et al. (2015) A review of surrogate models and their application to groundwater modeling. Water Resources Research 51, 5957–5973.

COVID-19

This PhD proposal uses empirical data that have already been collected from sediment cores and climate data that will be retrieved from public databases or weather stations. The computational nature of the project mitigates any risks associated with data collection and enables the DR to commence the project promptly. The supervisory team is trained to deliver remote training and supervision adopting tested approaches during the past two years of Covid-19 pandemic. The internship of the DR at the Environmental Agency is planned towards the end of the program. It can take place remotely if restrictions do not permit face to face interactions.