Research Fellow in Weather Forecast Postprocessing with Machine Learning

Salary: Full-time starting salary is normally in the range £36,130 to £45,413, with potential progression once in post to £48,149.

Contract Type: Fixed Term Contract up to March 2028 

Background

This post will advance the application of Machine Learning (ML) in weather forecasting and hydrological prediction. The Research Fellow will develop ML methods for postprocessing numerical ensemble weather forecasts over India to improve the skill of precipitation predictions and to generate hydrological forecasts.

The RF will be part of a research environment with strong ML activities. The post is mainly part of the project ‘HEavy Precipitation forecast Post-processing over India with Machine Learning’ (HEPPI-ML), which is funded through the ‘Weather and Climate Science for Service Partnership’ (WCSSP) programme. It is also linked to the National Institute for Health and Care Research (NIHR) project ‘Improving primary health care for patients with non-communicable diseases during severe flooding in India’, and to several ML-based projects at the British Antarctic Survey (BAS). There are also strong connections to the Institute for Data and Artificial Intelligence (IDAI) at UoB. 

Role Summary

• Implement and test different ML architectures for postprocessing precipitation forecasts over India.
• Determine how to maximise information extracted from the raw forecasts and how to optimise postprocessing skill for heavy precipitation.

Main Duties

The responsibilities may include some but not all of the responsibilities outlined below.

• Implement and test different Artificial Neural Network (ANN) architectures, such as convolutional and encoder-decoder ANNs, for postprocessing ensemble precipitation forecasts over India from the National Centre for Medium Range Weather Forecasting (NCMRWF) global ensemble prediction system (NEPS-G).
• Develop innovative specifications of input and output of postprocessing that account for the stochastic nature of precipitation and for systematic location errors in the original forecasts.
• Apply Interpretable AI concepts to make the postprocessing transparent and to improve the understanding of processes during heavy precipitation events over India.
• Implement the ML postprocessing methods on high performance computing systems in a way that is suitable for operational use.

Person Specification

• PhD, or close to completion, in a relevant, quantitative field, e.g. meteorology, machine learning, climate science, physics, mathematics, statistics or related fields.
• Evidence of good understanding (or capacity to develop understanding) of statistics and ML. 
• Evidence of a good understanding (or capacity to develop understanding) of meteorological processes and numerical weather prediction, and preferably specific knowledge related to monsoon precipitation.
• Experience working with large meteorological datasets.
• Good programming skills in languages such as Python, MATLAB or R.
• Familiarity with UNIX/LINUX.
• High level analytical capability.
• Ability to communicate complex information clearly.
• Ability to assess resource requirements and use resources effectively.

Informal enquiries to Martin Widmann, email: m.widmann@bham.ac.uk 

To download the full job description and details of this position and submit an electronic application online please click on the 'Apply' button above.

Valuing excellence, sustaining investment
We value diversity and inclusion at the University of Birmingham and welcome applications from all sections of the community and are open to discussions around all forms of flexible working

Advert information