EI Conference 2021 – Registrations are now open

JCEEI partners - University of Exeter and Met Office
JCEEI partners - University of Exeter and Met Office

EI Conference 2021 Posters

Name: Mr Steve Adams – Product Futures Lead, Met Office & Miles Gabriel – Esri UK Head of Government Business Development & Disaster Response Program Lead

Theme(s): Harnessing Transformative Technologies to achieve Net Zero

Title: Applying Climate and Location Data to Make Informed Decisions

Short Abstract:

Forecasting climate change is the beginning; to maximise mitigation and adaption action we must enable users to understand contextual impact

Useful links: Case study: UK National Curriculum Interactive Education content: https://teach-with-gis-uk-esriukeducation.hub.arcgis.com/pages/climate

Name: Mr Timothy Lam – EI CDT, Phd Student, Exeter University

Theme(s): Harnessing Science in Practise to deliver Net Zero

Title: Teleconnections of Droughts in Indonesian Borneo

Short Abstract: We quantify causal pathways of teleconnections of droughts in Indonesian Borneo in the past and future using observational and CMIP6 data. 

Long Abstract

Teleconnections of droughts during boreal summer in Indonesian Borneo are analysed and quantified using causal inference theory and causal networks, and their possible changes under a warming climate are projected using CMIP6 models. Causal hypotheses are first developed based on climate model experiments in literature and then justified by means of partial regression analysis and Peter and Clark Momentary Conditional Independence (PCMCI) between time series of climate drivers and rainfall data in Indonesian Borneo. We find that El Niño Southern Oscillation (ENSO) is strongly associated with droughts in Indonesian Borneo, with the onset of El Niño event serving to directly influence droughts. The rainfall variability in Borneo and the associated teleconnections may amplify in the future under SSP2 and SSP5 scenarios, possibly enhancing the robustness of predictors of drought. 


Timothy Lam1, Jennifer Catto2, Alberto Arribas3,4, , Rosa Barciela5, Peter Challenor2, Anna Harper2

1Centre for Doctoral Training in Environmental Intelligence, University of Exeter, Exeter, UK

2College of Engineering, Mathematics and Physical Sciences, University of Exeter, Exeter, UK

3Department of Meteorology, University of Reading, Reading, UK

4Microsoft, Reading, UK

5Met Office, Exeter, UK

Name: Mr Charlie Kirkwood – Phd Student, Exeter University

Theme(s): Harnessing AI to achieve Net Zero

Title: Bayesian deep learning to fully utilise auxiliary information

Short Abstract: We share an effective approach for probabilistic modelling of environmental observations sampled over auxiliary grids (terrain, satellite).

Long Abstract:

I would like to submit a poster demonstrating my research in developing Bayesian deep learning for geostatistical applications i.e. probabilistic spatial and spatio-temporal interpolation. The key improvement that our approach brings over traditional methods is the ability to learn features automatically from the landscape itself, rather than having to engineer features manually (which is often largely a matter of guess work). This allows us to perform approximate inference over a much wider range of hypotheses for the generation of the observed data, and in doing so uncover more complex relationships than traditional approaches allow. The result is that we can generate high resolution, high quality predictive maps and simulations which adhere to Tilman Gneiting’s paradigm of “maximising the sharpness of the predictive distribution subject to calibration”. We achieve this by harnessing auxiliary information in an ‘optimal’ way. We will demonstrate the approach by interpolating a yet-to-be-decided environmental variable, but the key take away from a net-zero perspective will be that by providing a sharper well-calibrated predictive distribution, we enable more efficient downstream decision making for environmental management, because the knowledge the model provides is more precise, without sacrificing honesty.

Name: Mr Seb Hickman – Phd Student, Cambridge University

Theme(s): Harnessing AI to achieve Net Zero

Title: DetectreeRGB: delineating trees in tropical forests

Short Abstract: We deploy a deep learning algorithm, Mask R-CNN, to delineate trees and analyse their dynamics in tropical forests using RGB and LiDAR data.

Useful Links:  More information can be found here: https://twitter.com/Toby_D_Jackson/status/1456294193589063681?s=20.

Name: Kevin Donkers – Phd Student, Exeter University

Theme(s): Harnessing Science in Practise to deliver Net Zero / Harnessing AI to achieve Net Zero / Working with EI Next Generation to deliver Net Zero

Title: Modelling UK agroforestry: Net Zero vs food

Short Abstract: Planting trees to meet Net Zero targets threatens to increase already high food imports. UK agroforestry could be a solution.

Long Abstract: UK Government policy on land-use is in a period of drastic change. A combination of leaving the EU Common Agricultural Policy (CAP) and legally binding targets for achieving Net Zero by 2050 may lead to competition for future land uses. NetZeroPlus is a multi-stakeholder project at the University of Exeter assessing the potential for large-scale afforestation to achieve Net Zero in the UK. My PhD will focus on modelling agroforestry in order to assess the impacts of afforestation on agricultural production and determine whether certain agroforestry systems can mitigate trade-offs between sequestering carbon and producing food for people to eat. I am right at the start of my PhD so choices of techniques are likely to change. Potential of AI applications are using Bayesian statistics for uncertainty quantification or the use of Gaussian processes to emulate complex land-use models.

Name: Dr Pete Falloon – Climate Service Lead Food/Farming/Nat. Env., Met Office

Theme(s): Building interdisciplinary communities to deliver Net Zero

Title: Community building for UK food system resilience to climate extremes & Net Zero

Abstract: We outline how Environmental Intelligence can support UK food system resilience to climate extremes and our plans for community workshops.

Name: Miss Trish Novak – Phd Student, Exeter University

Theme(s): Harnessing AI to achieve Net Zero / Working with EI Next Generation to deliver Net Zero

Title: Wandering dusts

Short Abstract: Using Machine Learning to disentangle reasons for atmospheric mineral dust generation, transport and its impact on climate and environment.

NameDr. Emilie Vanvyve – Informatics Lab, Met Office 

Theme(s): Harnessing Science in Practise to deliver Net Zero / Harnessing AI to achieve Net Zero / Building interdisciplinary communities to deliver Net Zero 

Title: CLIMAR framework  

Short Abstract: A risk-modelling framework for quantifying and visualising the risks associated with environmental and climatic change. 

Long Abstract: 

CLIMAR acronym stands for Climate Impacts, Mitigation, Adaptation and Resilience. Through leveraging data science, it integrates different data streams providing information on hazard, vulnerability and exposure, and creates decision-ready information in a form that is accessible to a range of audiences, including policy makers, industry, general public. This ability is going to be crucial across a wide variety of sectors in designing pathways to Net Zero and enhancing our resilience to climate change. In addition to presenting the framework, we will present an example application of CLIMAR for heat resilience in a city council. 

JCEEI partners - University of Exeter and Met Office