Skip to main content


Photo of Mr Ernst Werner

Mr Ernst Werner




I am currently a doctoral student on an EPSRC funded PhD studentship entitled "Deep Learning and Bayesian Time Series Analysis for Probabilistic Weather Forecasting" - this project will investigate novel deep learning and Bayesian methods to improve the quality of future numerical weather predictions through more accurate probabilistic forecasts. This research is supervised by Dr Saptarshi Das (Department of Mathematics) and co-supervised by members of the weather science team at the UK Met Office. 

Main goals

  • Firstly, to improve the reliability of probabilistic forecasts through post-processing procedures such as (a) Bayesian time series techniques or by making use of (b) Deep-learning based algorithms.
  • Secondly, explore the use of generative models, to generate synthetic data from the forecast probability distribution.
  • Thirdly, correctly classifying the forecast data to generate weather symbols using deep learning-based classification algorithms to match weather observations.


Since the field of weather science typically involves handling several terabytes of streaming data per day from multiple forecast models and ensembles, one difficulty arises from exploiting the correlation information between multiple variables to better define the risks of different outcomes. Further, it is challenging to produce a range of consistent weather scenarios within a well-calibrated forecast probability distribution. One idea would be to incorporate something of the complex relationships between weather diagnostics and the orography.

Research interests

  • Data Analytics.
  • Dynamical Systems.
  • Machine Learning.
  • Statistical Modelling.

Previous qualification

  • MSc Data Analytics, University of Glasgow, United Kingdom.