2025-B35 A ML approach to deriving station present weather

Project highlights

Work with real world data from the Met Office Observations National Capability.

Develop novel ‘virtual sensing; methods using machine learning to an increase observational capacity.

Have the potential to develop the proof of concept for a future operational service.

Overview

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).

Figure 1: A Met Office present weather sensor deployed in a surface enclosure.

Case funding

This project is suitable for CASE funding

Host

University of Birmingham

Theme

Climate and Environmental Sustainability

Supervisors

Project investigator

Lee Chapman, University of Birmingham: [email protected]

Co-investigators

Edmund Stone, Met Office: [email protected]

How to apply

Each host has a slightly different application process.
Find out how to apply for this studentship.
All applications must include the CENTA application form.
Choose your application route

Methodology

The project will require a review of the state-of-the-art for present weather monitoring both in terms of literature and data availability. The curation of available data will be a large component of the early methodology. This will be achieved by both primary research and interviews with the data experts and will be informed by the choice of machine learning techniques to be explored.

The virtual sensing methodology is a new approach and will require a good understanding of data science techniques to underpin algorithm selection and data partitioning (i.e testing, training and validation). Machine learning is proposed, alternative statistical methods and intercomparison approaches may also be explored.

Present weather is the initial target for this PhD, the main aim of the project is to demonstrate the feasibility of the virtual sensing approach to extend observational capacity across the UK and beyond.

Training and skills

DRs will be awarded CENTA Training Credits (CTCs) for participation in CENTA-provided and ‘free choice’ external training. One CTC can be earned per 3 hours training, and DRs must accrue 100 CTCs across the three and a half years of their PhD.

Training will be provided on methods of observations, including access to the Met Office College courses on such. This will also include infield training with visits to the Camborne Observatory and other sites as required.

Development of trials training, and interaction with operational change management.

If required, it is also expected that the DR would participate in data science / ML training.

Partners and collaboration

The University of Birmingham is a Met Office Academic Partnership University, this is the first CASE studentship related to the Advancing Observations Theme as introduced by the Met Office Research and Innovation Strategy.

Further details

For any enquiries related to this project please contact Lee Chapman, [email protected].

 To apply to this project: 

You must include a CENTA studentship application form, downloadable from: CENTA Studentship Application Form 2025.

You must include a CV with the names of at least two referees (preferably three) who can comment on your academic abilities.

Please submit your application and complete the host institution application process via: https://sits.bham.ac.uk/lpages/LES068.htm.   Please select the PhD Geography and Environmental Science (CENTA) 2025/26 Apply Now button. The CENTA Studentship Application Form 2025 and CV can be uploaded to the Application Information section of the online form.  Please quote CENTA 2025-B35 when completing the application form.

 Applications must be submitted by 23:59 GMT on Wednesday 8th January 2025.  

Possible timeline

Year 1

– Undertake a lit review of approaches to Present Weather, including intercomparisons (journal article)
– Undertake a visit and review of the surface Land station at Camborne Observatory (or similar) to understand the current limitations and available data.
– Develop test and training datasets.
– Review and start to test various ML methods using the datasets (conference paper)

Year 2

– Provide a gap analysis to enable trials to improve data coverage. This may include developing and trialling instruments to fill the gaps.
– Develop further test and training datasets
– Test ML methods using these new datasets (journal article for present weather)

Year 3

– Develop further test and training datasets
– Test ML methods using these new datasets
– Position paper on virtual sensors and potentially translation to other parameters