Computational Scientist
Computing Department, HPC Section, HPC Applications Group
Summary:
Sam joined ECMWF in October 2019 shortly after completing his PhD with Tim Palmer and Peter Dueben at the University of Oxford. Sam's work spans numerical methods and high-performance computing within numerical weather prediction. He is principally responsible for the implementation of single-precision arithmetic in the ocean model NEMO at ECMWF. He has also been closely involved with the efforts to port the ECMWF model, the IFS, to novel high-performance computing architectures, including Fugaku and Summit.
Professional interests:
- Floating-point arithmetic in high-performance numerical weather prediction
- Algorithmic design and efficient execution of spectral transforms
- Portability of large Fortran codebases across a variety of heterogeneous supercomputing architectures
- Open-source software in the geosciences
Career background:
DPhil in Environmental Research, University of Oxford (NERC Environmental Research Doctoral Training Partnership), 2015-2019
- Thesis title: Reduced-precision arithmetic in numerical weather prediction with an emphasis on data assimilation
- Supervisors: Prof. Tim Palmer, Dr. Peter Dueben
MSci in Physics, University of Bristol, 2010-2014
- First-class honours
- Final year project title: Knots in geometrically-confined polymers: nanochannels and other geometries
- Final year project advisor: Dr. Simon Hanna
- 2024
- Milan Klöwer, Maximilian Gelbrecht, Daisuke Hotta, Justin Willmert, Simone Silvestri, Gregory L Wagner, Alistair White, Sam Hatfield, Tom Kimpson, Navid C Constantinou, Chris Hill (June 2024) SpeedyWeather.jl: Reinventing atmospheric general circulation models towards interactivity and extensibility, Journal of Open Source Software. DOI: 10.21105/joss.06323
- 2023
- Valentine Anantharaj, Samuel Hatfield, Inna Polichtchouk, Nils Wedi (May 2023) Data challenges and opportunities from nascent kilometre-scale simulations. DOI: 10.5194/egusphere-egu23-13331
- 2022
- Milan Klöwer, Samuel Hatfield, Matteo Croci, Peter D. Düben, Tim Palmer (March 2022) Fluid simulations accelerated with 16 bits: Approaching 4x speedup on A64FX by squeezing ShallowWaters.jl into Float16. DOI: 10.5194/egusphere-egu22-3095
- Milan Klöwer, Sam Hatfield, Matteo Croci, Peter D. Düben, Tim N. Palmer (February 2022) Fluid Simulations Accelerated With 16 Bits: Approaching 4x Speedup on A64FX by Squeezing ShallowWaters.jl Into Float16, Journal of Advances in Modeling Earth Systems n. 2. DOI: 10.1029/2021MS002684
- 2021
- Simon T. K. Lang, Andrew Dawson, Michail Diamantakis, Peter Dueben, Samuel Hatfield, Martin Leutbecher, Tim Palmer, Fernando Prates, Christopher D. Roberts, Irina Sandu, Nils Wedi (October 2021) More accuracy with less precision, Quarterly Journal of the Royal Meteorological Society n. 741, pp. 4358-4370. DOI: 10.1002/qj.4181
- Matthew Chantry, Sam Hatfield, Peter Duben, Inna Polichtchouk, Tim Palmer (March 2021) Machine learning emulation of gravity wave drag in numerical weather forecasting. DOI: 10.5194/egusphere-egu21-7678
- Sam Hatfield, Matthew Chantry, Peter Dueben, Philippe Lopez, Alan Geer, Tim Palmer (September 2021) Building Tangent‐Linear and Adjoint Models for Data Assimilation With Neural Networks, Journal of Advances in Modeling Earth Systems. DOI: 10.1029/2021MS002521
- Samuel Edward Hatfield, Matthew Chantry, Peter Dominik Dueben, Philippe Lopez, Alan Jon Geer, Tim N Palmer (February 2021) Building tangent-linear and adjoint models for data assimilation with neural networks. DOI: 10.1002/essoar.10506310.1
- Sam Hatfield, Kristian Mogensen, Peter Dueben, Nils Wedi, Michail Diamantakis (March 2021) Operational Single-Precision Earth-System Modelling at ECMWF. DOI: 10.5194/egusphere-egu21-733
- Milan Klöwer, Sam Hatfield, Matteo Croci, Peter Düben, Tim Palmer (August 2021) Fluid simulations accelerated with 16 bit: Approaching 4x speedup on A64FX by squeezing ShallowWaters.jl into Float16. DOI: 10.1002/essoar.10507472.2
- 2020
- (April 2020) Single-Precision in the Tangent-Linear and Adjoint Models of Incremental 4D-Var, Monthly Weather Review. DOI: 10.1175/mwr-d-19-0291.1
- 2019
- (June 2019) Accelerating High-Resolution Weather Models with Deep-Learning Hardware, Platform for Advanced Scientific Computing Conference '19. DOI: 10.1145/3324989.3325711
- 2018
- Sam Hatfield, Peter Düben, Matthew Chantry, Keiichi Kondo, Takemasa Miyoshi, Tim Palmer (September 2018) Choosing the Optimal Numerical Precision for Data Assimilation in the Presence of Model Error, Journal of Advances in Modeling Earth Systems. DOI: 10.1029/2018MS001341
- 2017
- (December 2017) Improving weather forecast skill through reduced precision data assimilation, Monthly Weather Review. DOI: 10.1175/MWR-D-17-0132.1