PROJECT TITLE: Atmospheric data and machine learning for national emissions evaluation
DTP Research Theme(s): Changing Planet
Lead Institution: University of Bristol
Lead Supervisor: Prof. Matt Rigby, University of Bristol, School of Chemistry
Co-Supervisor: Dr. Pete Levy, Centre for Ecology and Hydrology
Co-Supervisor: Prof. Simon O’Doherty, University of Bristol, School of Chemistry
Project Enquiries: [Email Address Removed]
Project keywords: greenhouse gas, emissions, carbon dioxide, climate, machine learning
Figure 1: The UK’s carbon dioxide emissions
inventory.
Figure 2: Numerical simulation of the “air
history” of a measurement at the Ridge Hill
DECC network site (Met Office NAME
model).
Project Background
How do we know a country’s greenhouse gas (GHG) emissions? According to international treaties such as
the United Nations Framework Convention on Climate Change (UNFCCC), the requirement is only that
nations estimate and report their own emissions based on economic activity data. However, the UK has
pioneered an alternative approach that aims to improve transparency and accuracy: GHG concentrations
are measured in the air, and atmospheric models and Bayesian methods are employed to infer emissions
from the surrounding regions. The Bristol-led DECC network and DARE-UK project underpin the UK’s worldleading “top-down” emissions reporting system. This project will help develop and use next-generation
atmospheric modelling and machine learning tools for estimating national GHG emissions in the UK and
around the world.
Project Aims and Methods
Through projects such as DARE-UK, we are developing novel techniques to evaluate a country’s greenhouse
gas emissions (e.g., Figure 1). These methods have traditionally involved using atmospheric models to
simulate the flow of GHGs to ground based GHG measurement sensors (Figure 2). In the last few years, a
new generation of GHG satellites have come online, making space-based emissions inference possible.
These systems open the possibility of estimating emissions of carbon dioxide and methane across most of
the earth’s surface. However, the challenge is to develop the modelling tools that can effectively make use
of these very large datasets. In this project, you will develop machine learning approaches for the efficient
simulation of GHG atmospheric transport and Bayesian inference methods for national emissions
evaluation using dense datasets. To achieve this, you will work closely with ongoing projects funded by
NERC, BEIS and Google. The project can accommodate a range of interests in this field, and we encourage
students to contact the supervisory team to discuss options.
NERC GW4+ DTP Projects 2023
Candidate requirements
You must have, or expect a degree in physical sciences, mathematics, or computing. You must have
excellent communication and interpersonal skills. If you wish to be involved in making atmospheric
measurements, you must have previous laboratory experience. If you are primarily interested in developing
modelling approaches, a strong background in Mathematics or computing is essential, but no experience in
Chemistry is required. We welcome and encourage student applications from under-represented groups.
We value a diverse research environment.
Collaborative partner
The Centre for Ecology and Hydrology are world leading in the estimation of GHG fluxes using
measurements and bottom-up (process-based) models. Dr. Pete Levy has over 20 years’ experience in the
field and leads the UKCEH GHG Flux Network. The student will work with Dr. Levy on interpreting the UK
GHG emissions inventory and using atmospheric data to inform sector-level emissions estimates.
Training and external collaboration
You will be trained by Prof. Rigby and ACRG and Met Office staff to use the UK Met Office NAME model.
Training will be provided in Bayesian statistics and machine learning by Prof. Rigby and Dr. Levy. If desired,
you will be trained in cloud computing and can participate in the OpenGHG initiative. Students will be
encouraged to participate in measurement site visits and field campaigns under the supervision of Prof.
O’Doherty. You will have the opportunity to present your work at international conferences and
collaborate with our extensive network of national and international partners (e.g., through projects such
as DARE-UK, AGAGE, ICOS, etc.).
Background reading and references
Rigby, M., et al., Nature, 569(7757), 546–550, doi:10.1038/s41586-019-1193-4, 2019.
White, E. D. et al., Atmos. Chem. Phys., 19(7), 4345–4365, doi:10.5194/acp-19-4345-2019, 2019.
Useful links
http://www.bristol.ac.uk/chemistry/courses/postgraduate/
Bristol NERC GW4+ DTP Prospectus:
http://www.bristol.ac.uk/study/postgraduate/2023/doctoral/phd-great-western-four-dtp/
How to apply to the University of Bristol:
http://www.bristol.ac.uk/study/postgraduate/apply/
Please note: If you wish to apply for more than one project please contact the Bristol NERC GW4+ DTP
Administrator to find out the process for doing this.
The application deadline is Monday 9 January 2023 at 2359 GMT.
Interviews will take place during the period 22 February – 8 March 2023.
NERC GW4+ DTP Website:
For more information about the NERC GW4+ Doctoral Training Partnership please visit
https://www.nercgw4plus.ac.uk.
General Enquiries:
Bristol NERC GW4+ DTP Administrator
Email: [Email Address Removed]
Continue with Facebook
