Alison Peard

Deep learning methods for climate hazard modelling

Alison is working on how deep generative models can be used to make climate risk modelling more comprehensive while maintaining computational efficiency and usability. Her DPhil project is focussed on building models that facilitate the better prediction of the expected damages from compound hazards. She works primarily with GANs and ERA5 analysis data. She also does broader work in the context of the OPSIS, analysing drought indicators in the UK and climate threats to critical infrastructure in various countries.

• Adshead, D., Paszkowski, A., Gall, S.S. et al. (2024) Climate threats to coastal infrastructure and sustainable development outcomes. Nat. Clim. Chang. 14: 344–352
• Peard, A. and Hall, J. (2023) Combining deep generative models with extreme value theory for synthetic hazard simulation: a multivariate and spatially coherent approach. arXiv: 2311.18521.
• Murgatroyd, A., Peard, A., Becher, O., Coxon, G. et al (2023) Optimal drought indicators to predict water supply failure in England. Poster, AGU Fall Meeting 2023