PhD - Seeing extreme winds: machine learning applied to video to better estimate site-specific wind hazard

European Geosciences Union e.V.; Loughborough; On-site; GBP 20,000 - 24,000; Full time

Job summary

Loughborough University invites applications for a 3.5-year PhD studentship commencing October 2027. The role focuses on seeing extreme winds through machine learning applied to video to better estimate site-specific wind hazard.

The project funds a tax-free stipend (£21,805 p.a.) and UK tuition fees, with a campus-scale living lab setup and interlinked data from ERA5. The research teams will combine optical flow from video with ML, laser scanning, and a digital twin of trees to map gusts and

Qualifications

• PhD studentship offers 3.5 years funded from October 2027 with a tax-free stipend (£21,805 p.a.) and UK-rate tuition fees (£5,238).
• Excellent international student eligibility with fee waiver for the difference Home vs International for one candidate.
• Training covers machine learning, image processing, Python/R, applied meteorology and fieldwork.

Responsibilities

• Set up a campus-scale living lab at Loughborough University with cameras and anemometers.
• Combine dataset with ML, laser scanning and a physics-based digital twin of trees.
• Link small-scale observations to ERA5 and other historical weather data.
• Support training in ML, image processing and fieldwork.

Skills

Machine learning

Image processing

Fieldwork

Tools

Python

R

Job description

PhD - Seeing extreme winds: machine learning applied to video to better estimate site-specific wind hazard

Employer

Loughborough University

Loughborough has been one of the UK's top ten universities for ten years now.

Atmospheric Sciences (AS)
Geosciences Instrumentation and Data Systems (GI)
Natural Hazards (NH)

Type

Full time

Level

Student / Graduate / Internship

o The studentship is funded for 3.5 years starting from October 2027 and provides a tax-free minimum stipend of £21,805 per annum (from 1 October 2026) for the duration of the studentship plus tuition fees at the UK rate (£5,238 in 2026/27)

Extratropical cyclones such as Storm Eunice inflict extreme winds on the UK, causing billions of pounds in insured losses. Yet the peak 3-second gusts that drive damage vary greatly over short distances in cities, and weather stations and models cannot capture this. This PhD tackles the problem with visual anemometry: extracting gust speeds from short video clips of trees.

Building on pilot work showing that optical flow tracking of tree motion captures peak gusts, you will set up a campus-scale “living lab” at Loughborough University, installing four to five camera and anemometer sets for at least a year. You will combine this dataset with machine learning, laser scanning and a physics-based digital twin of trees to test how well video measures wind, even in extreme conditions. You will then link these small-scale observations to historical weather data such as ERA5 and show how buildings shelter or expose sites.

The work collaborates with the ArboKlim project (Prof. Theresa Blume), which is instrumenting trees in Potsdam, Germany.

The aim is a low-cost, proof-of-concept hazard-mapping tool. It could scale nationally using CCTV networks and inform the (re)insurance sector.

Training covers machine learning, image processing, Python/R, applied meteorology and fieldwork.

• The studentship is funded for 3.5 years starting from October 2027 and provides a tax-free minimum stipend of £21,805 per annum (from 1 October 2026) for the duration of the studentship plus tuition fees at the UK rate (£5,238 in 2026/27) and a Research Training Support Grant (RTSG) of £8,000. Further guidance about eligibility is available at UKRI Terms and Conditions. Excellent international students (including EU) are eligible to apply and there will be a fee waiver in place covering the difference between Home and International fees for one successful candidate.