PhD Studentship: A ML Approach to Deriving Station Present Weather

The Met Office currently have ‘present weather’ sensors at their Operational Land Surface sites. These instruments try to provide a present weather code as would traditionally be reported by a human at the station. However, the instruments are somewhat problematic (e.g. suspectable to spiders) and expensive.

Surface sites contain a whole range of additional information, a range of temperature, LCBR measurements, rain gauges etc. Also available is wider information from Weather Radar and NowCasting/Recent NWP output.

This project aims to use machine learning to assimilate this data together and use it to create a ‘virtual sensor’ for present weather. There will also be scope to add additional information, for example nearby roadside camera data, and proposing and trialing new instruments such as cameras.

Future Met Office efforts would then be required to operationalise any output from the project, but has the capacity to reduce the cost, complexity and maintenance of our Land Surface Stations whilst maintaining or even enhancing the ‘Present Weather’ capability.

This project represents the first step of developing a ‘Virtual Sensors’ methodology and there is potential to extend the approach to other problematic parameters (e.g. visibility and grass temperature).

The project will be supervised by Prof Lee Chapman. 

For further information on this project and details of how to apply to it please visit https://centa.ac.uk/studentship/2025-b35-a-ml-approach-to-deriving-station-present-weather/

Further information on how to apply for a CENTA studentship can be found on the CENTA website: https://centa.ac.uk/

Funding notes:

This project is offered through the CENTA3 DTP, with funding from the Natural Environment Research Council (NERC). Funding covers an annual stipend, tuition fees (at home-fee level) and Research Training Support Grant.

Our project-based studentships are open to all applicants who meet the academic requirements (at least a 2:1 at UK BSc level or at least a pass at UK MSc level or equivalent).

For further information please visit https://centa.ac.uk/.

UKRI allows international students to be eligible for studentships but only for a maximum of 30% of the cohort. Please be aware that CENTA funding does not cover any additional costs relating to moving to and residing in the UK. All international applicants must ensure they can fulfil the University of Birmingham’s international student entry requirements, which includes English language requirements.  For further information please visit https://www.birmingham.ac.uk/postgraduate/pgt/requirements-pgt/international/index.aspx.

References: Journal:

Zhang, Y., Wang, Y., Zhu, Y., Yang, L., Ge, L. and Luo, C., 2022. Visibility prediction based on machine learning algorithms. Atmosphere, 13(7), p.1125.

Ortega, L., Otero, L.D. and Otero, C., 2019, April. Application of machine learning algorithms for visibility classification. In 2019 IEEE International Systems Conference (SysCon) (pp. 1-5). IEEE.

Ellis, R.A., Sandford, A.P., Jones, G.E., Richards, J., Petzing, J. and Coupland, J.M., 2006. New laser technology to determine present weather parameters. Measurement science and technology, 17(7), p.1715.

Merenti-Välimäki, H.L., Lönnqvist, J. and Laininen, P., 2001. Present weather: comparing human observations and one type of automated sensor. Meteorological Applications, 8(4), pp.491-496.

Advert information