Dr Hilary Weller   <h.weller_AT_reading.ac.uk>   On Google Scholar
Orcid ID orcid.org/0000-0003-4553-7082   ResearcherID: A-1565-2011
Associate Professor in Meteorology and Computational Atmospheric Dynamics

Research

The main focus of my research is numerical methods for atmospheric modelling, including transport schemes, mesh and grid types, flow over orography and mesh adaptivity. I have also worked on multi-fluid modelling of atmospheric convection.

• I am working on advection (transport) schemes that enable long, stable time steps using implicit time stepping on arbitrary grids of the sphere.
  • I am working with PhD student Amber te Winkel to find accurate and efficient implicit advection schemes. Amber is co-supervised by Christian Kuehnlein at ECMWF and James Kent at the UK Met Office.
  • James Woodfield completed his PhD with me in 2023 on "Conservative Monotone Advection Schemes allowing Long Time Steps". James was co-supervised by Colin Cotter at Imperial College and was funded by the Mathematics for Planet Earth Centre for Doctoral Training
• I have worked on multi-fluid modelling of atmospheric convection in order to remove some of the long standing restrictions of mass-flux parameterisation of convection. This was part of the ParaCon project and was with:
  • Dan Shipley, who completed his PhD in 2021.
  • Will McIntyre who completed his PhD in 2020.
  • Peter Clark and Chris Holloway
  • John Thuburn at the University of Exeter.
• Between 2015 and 2018 I led the NERC project Moving Meshes for Global Atmospheric Modelling, with:
  • Chris Budd at Bath and Andrew McRae
  • Colin Cotter at Imperial and Jemma Shipton who is now at Exeter
  • Hiroe Yamazaki who was at Reading and is now at Imperial
  • Phil Browne who was at Reading and is now at ECMWF
• Between 2011 and 2016 I worked on the Met Office/NERC/STFC Gung Ho project to design and build the UK's next generation model of the global atmosphere.

I develop AtmosFOAM, which is a set of tools for atmospheric modelling using OpenFOAM.
These tools were used to find the first numerical solution of the Monge-Ampere equation on the sphere. Solutions of the Monge-Ampere equation generate optimally transported meshes which can adapt to flow features.