-
High-level, high-resolution ocean modeling at all scales with Oceananigans
Authors:
Gregory L. Wagner,
Simone Silvestri,
Navid C. Constantinou,
Ali Ramadhan,
Jean-Michel Campin,
Chris Hill,
Tomas Chor,
Jago Strong-Wright,
Xin Kai Lee,
Francis Poulin,
Andre Souza,
Keaton J. Burns,
John Marshall,
Raffaele Ferrari
Abstract:
We describe the vision, user interface, governing equations, and numerical methods that underpin new ocean modeling software called ``Oceananigans''. Oceananigans is being developed by the Climate Modeling Alliance as part of a larger project to build a trainable climate model with quantifiable uncertainty. We argue that Oceananigans status as a popular, capable modeling system realizes a vision f…
▽ More
We describe the vision, user interface, governing equations, and numerical methods that underpin new ocean modeling software called ``Oceananigans''. Oceananigans is being developed by the Climate Modeling Alliance as part of a larger project to build a trainable climate model with quantifiable uncertainty. We argue that Oceananigans status as a popular, capable modeling system realizes a vision for accelerating progress in Earth system modeling that balances demands for model accuracy and performance, needed for state-of-the-art science, against accessibility, which is needed to accelerate development. This vision combines three cooperative elements: (i) a relatively simple finite volume algorithm (ii) optimized for high-resolution simulations on GPUs which is (iii) exposed behind an expressive, high-level user interface (using the Julia programming language in our case). We offer evidence for the vision's potential by illustrating the creative potential of our user interface, showcasing Oceananigans physics with example simulations that range from simple classroom problems to a realistic global ocean simulation spanning all scales of oceanic fluid motion, and describing advances in parameterization, numerical methods, and computational efficiency.
△ Less
Submitted 19 February, 2025;
originally announced February 2025.
-
Oceananigans.jl: A Julia library that achieves breakthrough resolution, memory and energy efficiency in global ocean simulations
Authors:
Simone Silvestri,
Gregory L. Wagner,
Christopher Hill,
Matin Raayai Ardakani,
Johannes Blaschke,
Jean-Michel Campin,
Valentin Churavy,
Navid C. Constantinou,
Alan Edelman,
John Marshall,
Ali Ramadhan,
Andre Souza,
Raffaele Ferrari
Abstract:
Climate models must simulate hundreds of future scenarios for hundreds of years at coarse resolutions, and a handful of high-resolution decadal simulations to resolve localized extreme events. Using Oceananigans.jl, written from scratch in Julia, we report several achievements: First, a global ocean simulation with breakthrough horizontal resolution -- 488m -- reaching 15 simulated days per day (0…
▽ More
Climate models must simulate hundreds of future scenarios for hundreds of years at coarse resolutions, and a handful of high-resolution decadal simulations to resolve localized extreme events. Using Oceananigans.jl, written from scratch in Julia, we report several achievements: First, a global ocean simulation with breakthrough horizontal resolution -- 488m -- reaching 15 simulated days per day (0.04 simulated years per day; SYPD). Second, Oceananigans simulates the global ocean at 488m with breakthrough memory efficiency on just 768 Nvidia A100 GPUs, a fraction of the resources available on current and upcoming exascale supercomputers. Third, and arguably most significant for climate modeling, Oceananigans achieves breakthrough energy efficiency reaching 0.95 SYPD at 1.7 km on 576 A100s and 9.9 SYPD at 10 km on 68 A100s -- the latter representing the highest horizontal resolutions employed by current IPCC-class ocean models. Routine climate simulations with 10 km ocean components are within reach.
△ Less
Submitted 14 October, 2024; v1 submitted 12 September, 2023;
originally announced September 2023.
-
Capturing missing physics in climate model parameterizations using neural differential equations
Authors:
Ali Ramadhan,
John Marshall,
Andre Souza,
Xin Kai Lee,
Ulyana Piterbarg,
Adeline Hillier,
Gregory LeClaire Wagner,
Christopher Rackauckas,
Chris Hill,
Jean-Michel Campin,
Raffaele Ferrari
Abstract:
We explore how neural differential equations (NDEs) may be trained on highly resolved fluid-dynamical models of unresolved scales providing an ideal framework for data-driven parameterizations in climate models. NDEs overcome some of the limitations of traditional neural networks (NNs) in fluid dynamical applications in that they can readily incorporate conservation laws and boundary conditions an…
▽ More
We explore how neural differential equations (NDEs) may be trained on highly resolved fluid-dynamical models of unresolved scales providing an ideal framework for data-driven parameterizations in climate models. NDEs overcome some of the limitations of traditional neural networks (NNs) in fluid dynamical applications in that they can readily incorporate conservation laws and boundary conditions and are stable when integrated over time. We advocate a method that employs a 'residual' approach, in which the NN is used to improve upon an existing parameterization through the representation of residual fluxes which are not captured by the base parameterization. This reduces the amount of training required and providing a method for capturing up-gradient and nonlocal fluxes. As an illustrative example, we consider the parameterization of free convection of the oceanic boundary layer triggered by buoyancy loss at the surface. We demonstrate that a simple parameterization of the process - convective adjustment - can be improved upon by training a NDE against highly resolved explicit models, to capture entrainment fluxes at the base of the well-mixed layer, fluxes that convective adjustment itself cannot represent. The augmented parameterization outperforms existing commonly used parameterizations such as the K-Profile Parameterization (KPP). We showcase that the NDE performs well independent of the time-stepper and that an online training approach using differentiable simulation via the Julia scientific machine learning software stack improves accuracy by an order-of-magnitude. We conclude that NDEs provide an exciting route forward to the development of representations of sub-grid-scale processes for climate science, opening up myriad new opportunities.
△ Less
Submitted 6 March, 2023; v1 submitted 23 October, 2020;
originally announced October 2020.
-
Synthesis of hydrogen- and methyl-capped long-chain polyynes by intense ultrashort laser pulse irradiation of toluene
Authors:
Ali Ramadhan,
Michal Wesolowski,
Tomonari Wakabayashi,
Haruo Shiromaru,
Tatsuya Fujino,
Takeshi Kodama,
Walter Duley,
Joseph Sanderson
Abstract:
Hydrogen- and methyl-capped polyynes were synthesized by irradiating pure liquid toluene with 35 femtosecond, 300 μJ laser pulses having a central wavelength of 800 nm, generated by a regeneratively amplified Ti:sapphire tabletop laser at a repetition rate of 1 kHz. Raman spectroscopy was used to confirm the presence of polyynes in the irradiated samples while high-performance liquid chromatograph…
▽ More
Hydrogen- and methyl-capped polyynes were synthesized by irradiating pure liquid toluene with 35 femtosecond, 300 μJ laser pulses having a central wavelength of 800 nm, generated by a regeneratively amplified Ti:sapphire tabletop laser at a repetition rate of 1 kHz. Raman spectroscopy was used to confirm the presence of polyynes in the irradiated samples while high-performance liquid chromatography was used to separate hydrogen-capped polyynes up to C18H2 and methyl-capped polyynes up to HC14CH3. These represent the first such methyl-capped polyynes and the longest hydrogen capped chains synthesized to date by the ultrafast laser based method. Furthermore our results show that choice of the starting solvent molecule directly influences the end caps of the polyynes which can be produced.
△ Less
Submitted 31 March, 2017; v1 submitted 1 December, 2016;
originally announced December 2016.
-
Ultrafast molecular dynamics of dissociative ionization in OCS probed by soft X-ray synchrotron radiation
Authors:
Ali Ramadhan,
Benji Wales,
Isabelle Gauthier,
Reza Karimi,
Michael MacDonald,
Lucia Zuin,
Joe Sanderson
Abstract:
Soft X-rays (90-173 eV) from the 3rd generation Canadian Light Source have been used in conjunction with a multi coincidence time and position sensitive detection apparatus to observe the dissociative ionization of OCS. By varying the X-ray energy we can compare dynamics from direct and Auger ionization processes, and access ionization channels which result in two or three body breakup, from 2+ to…
▽ More
Soft X-rays (90-173 eV) from the 3rd generation Canadian Light Source have been used in conjunction with a multi coincidence time and position sensitive detection apparatus to observe the dissociative ionization of OCS. By varying the X-ray energy we can compare dynamics from direct and Auger ionization processes, and access ionization channels which result in two or three body breakup, from 2+ to 4+ ionization states. We make several new observations for the 3+ state such as kinetic energy release limited by photon energy, and using Dalitz plots we can see evidence of timescale effects between the direct and Auger ionization process for the first time. Finally, using Dalitz plots for OCS$^{4+}$ we observe for the first time that breakup involving an O$^{2+}$ ion can only proceed from out of equilibrium nuclear arrangement for S(2p) Auger ionization.
△ Less
Submitted 15 August, 2016; v1 submitted 28 June, 2016;
originally announced June 2016.
