Oceananigans.jl: Fast and friendly geophysical fluid dynamics on GPUs

Ramadhan, Ali; Wagner, Gregory LeClaire; Hill, Chris; Campin, Jean‐Michel; Churavy, Valentin; Besard, Tim; Souza, Andre N.; Edelman, Alan; Ferrari, Raffaele; Marshall, John

doi:10.21105/joss.02018

Cited by 40 publications

(39 citation statements)

References 8 publications

(9 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The non-local effects are therefore important in transporting properties from one side of the boundary layer to the other as illustrated in figure 1. To further un- (Ramadhan et al, 2020) whereas the 2D simulations use a pseudo-spectral code (Burns et al, 2019). In both sets of simulations, the ratio of the penetrating shortwave heat flux to the diffusive heat flux in the deep thermocline is F/Φ = 10, so that the MLD is h = ln(F/Φ) ≈ 2.3.…”

Section: Discussionmentioning

confidence: 99%

“…Figure8: (Color figure) Comparison of the horizontally and temporally averaged fluxes and gradients of buoyancy and passive tracer in 2D and 3D simulations with the same boundary conditions. The 3D simulations are done using a finite volume code(Ramadhan et al, 2020) whereas the 2D simulations use a pseudo-spectral code(Burns et al, 2019). In both sets of simulations, the ratio of the penetrating shortwave heat flux to the diffusive heat flux in the deep thermocline is F/Φ = 10, so that the MLD is h = ln(F/Φ) ≈ 2.3.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Turbulent mixing of a passive scalar in the ocean mixed layer

2020

View full text Add to dashboard Cite

We study the 2D turbulent mixing of a passive scalar in the ocean mixed layer.As an example, we examine a steady-state convective mixed layer in which the boundary conditions are chosen so that the system reaches a dynamical equilibrium. In this idealized case, we parameterize the horizontally and temporally averaged fluxes as a functional of the horizontally and temporally averaged property gradients. Here, w c = − dz K(z|z )∂ c /∂z , where K(z|z ) is the eddy diffusivity kernel which describes the vertical transport by eddies at any vertical location z. The full kernel K(z|z ) is computed by adding passive scalars to a buoyancy-driven flow field in a 2D DNS of the ocean surface layer. This functional form of the eddy diffusivity highlights both local and non-local effects of the mixing of a passive scalar, and is based on an unapproximated representation of the idealized physics. This type of formulation can be further extended to other problems in turbulence concerning the mixing of a passive scalar to determine a parameterization based on an accurate representation of ocean physics.

show abstract

Section: Discussionmentioning

confidence: 99%

mentioning

confidence: 99%

Turbulent mixing of a passive scalar in the ocean mixed layer

2020

View full text Add to dashboard Cite

show abstract

“…• finite difference methods (DiffEqOperators.jl, a part of DifferentialEquations.jl (Rackauckas & Nie, 2017)) • finite volume methods (Oceananigans.jl (Ramadhan et al, 2020), Kinetic.jl (Xiao, 2021)) • spectral methods (ApproxFun.jl (Olver & Townsend, 2014), FourierFlows.jl (Constantinou & Wagner, 2021)) • finite element methods (Gridap.jl (Badia & Verdugo, 2020)) • discontinuous spectral element methods (Trixi.jl (Ranocha, Schlottke-Lakemper, et al, 2021;Schlottke-Lakemper et al, 2021 We are not aware of any open-source software library implementing all of the SBP classes using a unified interface or even several finite difference SBP operators on finite domains, which are usually heavily optimized (Mattsson et al, 2014(Mattsson et al, , 2018 and not available in other open source packages. Sometimes, restricted sets of coefficients are available online (Almquist, 2017;O'Reilly, 2019), but there is no other extensive collection of these methods.…”

Section: Related Research and Softwarementioning

confidence: 99%

SummationByPartsOperators.jl: A Julia library of provably stable discretization techniques with mimetic properties

Ranocha¹

2021

JOSS

View full text Add to dashboard Cite

SummationByPartsOperators.jl is a Julia library of summation-by-parts (SBP) operators, which are discrete derivative operators developed to get provably stable (semi-) discretizations, paying special attention to boundary conditions. Discretizations included in this framework are finite difference, Fourier pseudospectral, continuous Galerkin, and discontinuous Galerkin methods. The main aim of SummationByPartsOperators.jl is to be useful for both students learning the basic concepts and researchers developing new numerical algorithms based on SBP operators. Therefore, SummationByPartsOperators.jl provides a unified framework of all of these seemingly different discretizations. At the same time, the implementation is reasonably optimized to achieve good performance without sacrificing flexibility.

show abstract

“…Appendix A: Oceananigans.jl Oceananigans.jl (Ramadhan et al, 2020) is open source software for ocean process studies written in the Julia programming language (Besard, Churavy, et al, 2019;Bezanson et al, 2017). For the large eddy simulations (LESs) reported in this paper, Oceananigans.jl is configured to solve the spatially filtered, incompressible Boussinesq equations with a temperature tracer.…”

Section: A1 Subfilter Stress and Temperature Fluxmentioning

confidence: 99%

Uncertainty quantification of ocean parameterizations: application to the K-Profile-Parameterization for penetrative convection

Souza

Wagner

Ramadhan

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Parameterizations of unresolved turbulent processes often compromise the fidelity of large-scale ocean models. In this work, we argue for a Bayesian approach to the refinement and evaluation of turbulence parameterizations. Using an ensemble of large eddy simulations of turbulent penetrative convection in the surface boundary layer, we demonstrate the method by estimating the uncertainty of parameters in the convective limit of the popular "K-Profile Parameterization." We uncover structural deficiencies and propose an alternative scaling that overcomes them. Plain Language SummaryClimate projections are often compromised by significant uncertainties which stem from the representation of physical processes that cannot be resolved-such as clouds in the atmosphere and turbulent swirls in the ocean-but which have to be parameterized. We propose a methodology for improving parameterizations in which they are tested against, and tuned to, high-resolution numerical simulations of subdomains that represent them more completely. A Bayesian methodology is used to calibrate the parameterizations against the highly resolved model, to assess their fidelity and identify shortcomings. Most importantly, the approach provides estimates of parameter uncertainty. While the method is illustrated for a particular parameterization of boundary layer mixing, it can be applied to any parameterization.

show abstract

Oceananigans.jl: Fast and friendly geophysical fluid dynamics on GPUs

Cited by 40 publications

References 8 publications

Turbulent mixing of a passive scalar in the ocean mixed layer

Turbulent mixing of a passive scalar in the ocean mixed layer

SummationByPartsOperators.jl: A Julia library of provably stable discretization techniques with mimetic properties

Uncertainty quantification of ocean parameterizations: application to the K-Profile-Parameterization for penetrative convection

Contact Info

Product

Resources

About