Title: Identifying physically-consistent worst case scenarios for extreme regional precipitation in Southeast Asia and improving prediction
Lead Supervisor: Tom Frame, Department of Meteorology, University of Reading
Email: t.h.a.frame@reading.ac.uk
Co-supervisors: Sam Ferrett, National Centre for Atmospheric Science; John Methven, Department of Meteorology, University of Reading; Carlo Cafaro, Met Office
UKRI funding only covers Home fees which increase annually. International students may still apply to this project, but will be required to meet the difference between the International and Home student fees themselves.
Southeast Asia is frequently affected by severe rainfall events leading to flash flooding and landslides, which can have devastating impacts on society. The region experiences unique weather phenomena such as Borneo vortices and cold surges (Howard, 2022), as well as impacts from large-scale eastward and westward propagating equatorial waves (Ferrett, 2020) and strong vertical circulations identified with local Hadley and Walker circulations. Recent work has shown that the most extreme precipitation events occur when several of these phenomena occur together. For example, Figure 1 shows the resultant rainfall from the co-occurrence of two equatorial waves. The reason co-occurrences lead to such heavy rainfall is not known, nor is the potential combination of factors that could result in the most extreme events. This project seeks to address these questions:
- What are the combinations of atmospheric factors that result in the worst-case severe rainfall scenarios?
- How likely are these combinations of factors?
- What are the key drivers that enhance the likelihood of these factors, and can they be harnessed to extend the range of prediction and early warning in the region?
By definition the circumstances giving rise to worst-case precipitation scenarios must be rare, occurring only a few times in the limited duration of the observational record in the region. Climate simulations lack the resolution required to explicitly represent the deep convection which is central to the weather systems and associated precipitation in the region. However, the Met Office have been running an 18-member convection-permitting ensemble forecast in the region, producing 5-day forecasts twice per day since 2018. This forecasting system was shown (Ferrett, 2021) to provide much higher skill in rainfall forecasts than models with parametrized convection and is already supporting weather services in the region, helping to improve early warnings of extreme events. This is the only convection-permitting ensemble to be run consistently in the Tropics for almost a decade, yielding a unique data set of many thousands of forecasts of physically-plausible extreme rainfall events: some will be events that materialised, but most will be scenarios which were possible but did not occur. This presents a timely opportunity to address questions 1 to 3.
In the first phase of the project, you will use the huge number of physically-consistent alternative realities represented by the ensemble forecasts to identify the most extreme precipitation events that could occur (called the UNSEEN method; Thompson, 2017). Once they have been identified, the dynamical factors interacting to produce the extreme precipitation will be characterised using the forecast model data.
In the second phase, you will investigate the predictability of the large-scale factors that combine to create extreme events in the region and how these are represented in the Met Office models. There is also scope to build on recent work using hybrid dynamical-statistical methods (Gonzalez, 2023; Wolf, 2024) to improve precipitation forecasts, potentially incorporating new predictors identified during the first phase.
The final phase will explore the capability of convection-permitting ensemble models to improve the prediction of high-impact weather risk and warnings in Southeast Asia using a dataset of severe weather cases, selected in collaboration with partners from Indonesia, Malaysia, Vietnam and the Philippines. There are several options open here, including experiments re-running the convection-permitting ensemble with other recent model configurations being developed under the Met Office K-Scale project, or evaluating the performance of the new AI ensemble prediction models.
This project offers a unique opportunity to work at the interface of cutting-edge numerical weather prediction, tropical meteorology, and applied forecasting science. It will be supported by a strong supervisory team from the University of Reading and the Met Office, with expertise in tropical weather systems, ensemble forecasting, and model evaluation. The student will benefit from close collaboration with international partners, access to state-of-the-art modelling tools and CASE sponsorship from the Met Office with placements within the regional modelling team.The outcomes of this research will contribute directly to improving weather prediction capabilities in Southeast Asia, with real-world benefits for disaster preparedness and resilience. It will also advance our understanding of tropical weather predictability and the role of high-resolution ensemble modelling in operational forecasting.
Training opportunities:
As a member of the Tropical Meteorology research group at Reading you will be able to learn about current research and methods. To get up to speed with using the Met Office model you will have the opportunity to join the MetUM training course and NCAS Climate Modelling Summer School. As part of the CASE sponsorship of this project you will have extended visits at the Met Office HQ in Exeter, when you will work with leading scientists in the Weather Science and Foundation Science divisions, as well as participating in the annual workshop of WCSSP Southeast Asia programme.
Student profile:
This project would be suitable for students with a degree in physics, mathematics or a closely related physical science. Programming skills are important for this project. Knowledge of meteorology and experience of processing large datasets are desirable but not required. UKRI funding only covers Home fees which increase annually. International students may still apply to this project, but will be required to meet the difference between the International and Home student fees themselves.
Co-Sponsorship details:
This project will receive a CASE award from the UK Met Office.
References:
• Ferrett, S., Frame, T. H. A., Methven, J., Holloway, C. E., Webster, S., Stein, T. H. M., & Cafaro, C. (2021). Evaluating convection-permitting ensemble forecasts of precipitation over Southeast Asia. Weather and Forecasting, 36(4), 1199–1217. https://doi.org/10.1175/WAF-D-20-0216.1
• Ferrett, S., Yang, G.-Y., Woolnough, S., et al. (2020). Linking extreme precipitation in Southeast Asia to equatorial waves. Quarterly Journal of the Royal Meteorological Society, 146(728), 665–684. https://doi.org/10.1002/qj.3699
• Gonzalez, P. L. M., Howard, E., Ferrett, S., Frame, T. H. A., Martínez-Alvarado, O., Methven, J., et al. (2023). Weather patterns in Southeast Asia: Enhancing high-impact weather subseasonal forecast skill. Quarterly Journal of the Royal Meteorological Society, 149(750), 19–39. https://doi.org/10.1002/qj.4378
• Howard, E., Thomas, S., Frame, T. H., Gonzalez, P. L., Methven, J., Martínez-Alvarado, O., et al. (2022). Weather patterns in Southeast Asia: Relationship with tropical variability and heavy precipitation. Quarterly Journal of the Royal Meteorological Society, 148(750), 747–769. https://doi.org/10.1002/qj.4227
• Thompson, V., Dunstone, N. J., Scaife, A. A., et al. (2017). High risk of unprecedented UK rainfall in the current climate. Nature Communications, 8, 107. https://doi.org/10.1038/s41467-017-00275-3
• Wolf, G., Ferrett, S., Methven, J., Frame, T., Holloway, C., Martínez-Alvarado, O., et al. (2024). Comparison of probabilistic forecasts of extreme precipitation for a global and convection-permitting ensemble and hybrid statistical–dynamical method based on equatorial wave information. Quarterly Journal of the Royal Meteorological Society, 150(759), 877–896. https://doi.org/10.1002/qj.4627
