publications

2026

1. Submesoscale and boundary layer turbulence under mesoscale forcing in the upper ocean

   Shirui Peng, Simone Silvestri, and Abigail Bodner

   2026

   Abs Bib PDF

   The interaction among quasi-geostrophic mesoscale eddies, submesoscale fronts, and boundary layer turbulence (BLT) is a central problem in upper ocean dynamics. We investigate these multiscale dynamics using a novel large-eddy simulation on a 100km-scale domain with meter-scale resolution. The simulation resolves BLT energized by uniform surface wind and convective forcing. A front interacts with BLT within a prescribed, spatially inhomogeneous mesoscale eddy field, representing a canonical eddy quadrupole. Using a triple flow decomposition, we analyze the dynamic coupling and kinetic energy budgets among the large-scale field, submesoscale field, and the resolved BLT. Our analysis reveals significant heterogeneity in the structure and intensity of submesoscales and BLT under varying mesoscale forcing.

   @misc{peng2026submesoscale,
     title = {Submesoscale and boundary layer turbulence under mesoscale forcing in the upper ocean},
     author = {Peng, Shirui and Silvestri, Simone and Bodner, Abigail},
     year = {2026},
     eprint = {2601.10441},
     archiveprefix = {arXiv},
     primaryclass = {physics.ao-ph},
   }

2025

1. Surface to Seafloor: A Generative AI Framework for Decoding the Ocean Interior State

   Andre N. Souza, Simone Silvestri, Katherine Deck, and 3 more authors

   2025

   Abs Bib PDF

   Understanding subsurface ocean dynamics is essential for quantifying oceanic heat and mass transport, but direct observations at depth remain sparse due to logistical and technological constraints. Here, we present a probabilistic framework based on score-based diffusion models to reconstruct three-dimensional subsurface velocity and buoyancy fields, including the energetic ocean eddy field, from surface observations. Using a 15-level primitive equation simulation of an idealized double-gyre system, we evaluate the skill of the model in inferring the mean circulation and the mesoscale variability at depth under varying levels of surface information. We find that the generative model successfully recovers key dynamical structures and provides physically meaningful uncertainty estimates.

   @misc{souza2025surfaceseafloorgenerativeai,
     title = {Surface to Seafloor: A Generative AI Framework for Decoding the Ocean Interior State},
     author = {Souza, Andre N. and Silvestri, Simone and Deck, Katherine and Bischoff, Tobias and Ferrari, Raffaele and Flierl, Glenn R.},
     year = {2025},
     eprint = {2504.15308},
     archiveprefix = {arXiv},
     primaryclass = {physics.geo-ph},
   }

2. High-level, high-resolution ocean modeling at all scales with Oceananigans

   Gregory L. Wagner, Simone Silvestri, Navid C. Constantinou, and 11 more authors

   2025

   Abs Bib PDF

   We describe the user interface, governing equations, and numerical methods underpinning the community ocean modeling software called Oceananigans. Oceananigans is written in the Julia programming language, which distinguishes it from usual software based on Fortran. Oceananigans can efficiently simulate all scales of ocean motion, ranging from millimeter-scale turbulence in a small box to planetary-scale ocean circulation. Oceananigans design combines (i) a basic structured finite volume algorithm (ii) optimized for high-resolution simulations on GPUs which is (iii) exposed behind a high-level, programmable user interface.

   @misc{wagner2025highlevelhighresolutionoceanmodeling,
     title = {High-level, high-resolution ocean modeling at all scales with Oceananigans},
     author = {Wagner, Gregory L. and Silvestri, Simone and Constantinou, Navid C. and Ramadhan, Ali and Campin, Jean-Michel and Hill, Chris and Chor, Tomas and Strong-Wright, Jago and Lee, Xin Kai and Poulin, Francis and Souza, Andre and Burns, Keaton J. and Marshall, John and Ferrari, Raffaele},
     year = {2025},
     eprint = {2502.14148},
     archiveprefix = {arXiv},
     primaryclass = {physics.ao-ph},
   }

3. A WENO finite-volume scheme for the evolution of potential vorticity in isopycnal ocean models

   Wenda Zhang, Yi-Hung Kuo, Simone Silvestri, and 3 more authors

   2025

   Bib

   @misc{zhang2025wenoMOM,
     title = {A WENO finite-volume scheme for the evolution of potential vorticity in isopycnal ocean models},
     author = {Zhang, Wenda and Kuo, Yi-Hung and Silvestri, Simone and Adcroft, Alistair and Hallberg, Robert and Griffies, Stephen M.},
     journal = {ESS Open Archive},
     year = {2025},
     doi = {10.22541/essoar.175380391.18723979/v1}
   }

4. NORi: An ML-Augmented Ocean Boundary Layer Parameterization

   Xin Kai Lee, Ali Ramadhan, Andre Souza, and 4 more authors

   2025

   Abs Bib PDF

   NORi is a machine-learned parameterization of ocean boundary layer turbulence that is physics-based and augmented with neural networks. NORi stands for neural ordinary differential equations (NODEs) Richardson number (Ri) closure. The physical parameterization is controlled by a Richardson number-dependent diffusivity and viscosity. The NODEs are trained to capture the entrainment through the base of the boundary layer, which cannot be represented with a local diffusive closure. NORi is designed for the realistic nonlinear equation of state of seawater and demonstrates excellent prediction and generalization capabilities.

   @misc{lee2025norimlaugmentedoceanboundary,
     title = {NORi: An ML-Augmented Ocean Boundary Layer Parameterization},
     author = {Lee, Xin Kai and Ramadhan, Ali and Souza, Andre and Wagner, Gregory LeClaire and Silvestri, Simone and Marshall, John and Ferrari, Raffaele},
     year = {2025},
     eprint = {2512.04452},
     archiveprefix = {arXiv},
     primaryclass = {physics.ao-ph},
   }

5. DJ4Earth: Differentiable, and Performance-portable Earth System Modeling via Program Transformations

   William S Moses, Gong Cheng, Valentin Churavy, and 17 more authors

   2025

   Bib

   @misc{moses2025james,
     title = {DJ4Earth: Differentiable, and Performance-portable Earth System Modeling via Program Transformations},
     author = {Moses, William S and Cheng, Gong and Churavy, Valentin and Gelbrecht, Maximilian and Kl\"{o}wer, Milan and Kump, Joseph and Morlighem, Mathieu and Williamson, Sarah and Apte, Dhruv and Berg, Paul and Giordano, Mos\`{e} and Hill, Chris and Loose, Nora and Montoison, Alexis and Narayanan, Sri Hari Krishna and Pal, Avik and Schanen, Michel and Silvestri, Simone and Wagner, Gregory LeClaire and Heimbach, Patrick},
     journal = {ESS Open Archive},
     year = {2025}
   }

6. A GPU-Based Ocean Dynamical Core for Routine Mesoscale-Resolving Climate Simulations

   Simone Silvestri, Gregory L. Wagner, Navid C. Constantinou, and 7 more authors

   Journal of Advances in Modeling Earth Systems, 2025

   e2024MS004465 2024MS004465

   Abs Bib PDF

   We describe an ocean hydrostatic dynamical core implemented in Oceananigans optimized for Graphical Processing Unit (GPU) architectures. On 64 A100 GPUs, equivalent to 16 computational nodes in current state-of-the-art supercomputers, our dynamical core can simulate a decade of near-global ocean dynamics per wall-clock day at an 8-km horizontal resolution; a resolution adequate to resolve the ocean’s mesoscale eddy field. Such efficiency, achieved with relatively modest hardware resources, suggests that climate simulations on GPUs can incorporate fully eddy-resolving ocean models. This removes a major source of systematic bias in current IPCC coupled model projections, the parameterization of ocean eddies, and represents a major advance in climate modeling.

   @article{Silvestri2025,
     author = {Silvestri, Simone and Wagner, Gregory L. and Constantinou, Navid C. and Hill, Christopher N. and Campin, Jean-Michel and Souza, Andre N. and Bishnu, Siddhartha and Churavy, Valentin and Marshall, John and Ferrari, Raffaele},
     title = {A GPU-Based Ocean Dynamical Core for Routine Mesoscale-Resolving Climate Simulations},
     journal = {Journal of Advances in Modeling Earth Systems},
     volume = {17},
     number = {4},
     pages = {e2024MS004465},
     doi = {https://doi.org/10.1029/2024MS004465},
     url = {https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2024MS004465},
     note = {e2024MS004465 2024MS004465},
     year = {2025}
   }

7. Formulation and Calibration of CATKE, a One-Equation Parameterization for Microscale Ocean Mixing

   Gregory LeClaire Wagner, Adeline Hillier, Navid Constantinou, and 8 more authors

   Journal of Advances in Modeling Earth Systems, 2025

   Abs Bib PDF

   We describe CATKE, a parameterization for fluxes associated with small-scale or microscale ocean turbulent mixing on scales between 1 and 100 m. CATKE uses a downgradient formulation that depends on a prognostic turbulent kinetic energy (TKE) variable and a diagnostic mixing length scale that includes a dynamic convective adjustment (CA) component. To calibrate CATKE, we use Ensemble Kalman Inversion to minimize the error between 21 large eddy simulations (LESs) and predictions of the LES data by CATKE-parameterized single column simulations at three different vertical resolutions. We find that CATKE makes accurate predictions of both idealized and realistic LES compared to microscale turbulence parameterizations commonly used in climate models.

   @article{Wagner_2025,
     author = {Wagner, Gregory LeClaire and Hillier, Adeline and Constantinou, Navid and Silvestri, Simone and Souza, Andre and Burns, Keaton and Ramadhan, Ali and Hill, Christopher and Campin, Jean-Michel and Marshall, John and Ferrari, Raffaele},
     title = {Formulation and Calibration of CATKE, a One-Equation Parameterization for Microscale Ocean Mixing},
     journal = {Journal of Advances in Modeling Earth Systems},
     volume = {17},
     number = {4},
     pages = {e2024MS004522},
     doi = {https://doi.org/10.1029/2024MS004522},
     url = {https://agupubs.onlinelibrary.wiley.com/doi/abs/10.1029/2024MS004522},
     year = {2025}
   }

2024

1. A New WENO-Based Momentum Advection Scheme for Simulations of Ocean Mesoscale Turbulence

   Simone Silvestri, Gregory Wagner, Jean-Michel Campin, and 4 more authors

   Journal of Advances in Modeling Earth Systems, 2024

   Abs Bib PDF

   We propose a new vector-invariant momentum advection scheme for ocean modeling based on a Weighted Essentially Non-Oscillatory (WENO) reconstruction of the velocity components. The scheme is designed to reduce the spurious features associated with the standard advection schemes used in ocean general circulation models, particularly at eddy-permitting resolutions. We test the scheme in idealized and realistic configurations, showing that it significantly reduces grid-scale noise and improves the representation of mesoscale turbulence.

   @article{Simone_silvestri_WENO,
     author = {Silvestri, Simone and Wagner, Gregory and Campin, Jean-Michel and Constantinou, Navid and Hill, Christopher and Souza, Andre and Ferrari, Raffaele},
     title = {A New WENO-Based Momentum Advection Scheme for Simulations of Ocean Mesoscale Turbulence},
     journal = {Journal of Advances in Modeling Earth Systems},
     volume = {16},
     number = {7},
     pages = {e2023MS004130},
     keywords = {numerical modeling, mesoscale ocean turbulence, eddy-permitting simulations},
     doi = {https://doi.org/10.1029/2023MS004130},
     year = {2024},
   }

2. SpeedyWeather.jl: Reinventing atmospheric general circulation models towards interactivity and extensibility

   Milan Klöwer, Maximilian Gelbrecht, Daisuke Hotta, and 8 more authors

   Journal of Open Source Software, 2024

   Abs Bib PDF

   SpeedyWeather.jl is an atmospheric general circulation model written in Julia that prioritizes interactivity and extensibility. It implements the primitive equations on a sphere using a spectral transform method, supporting multiple numerical precisions and architectures. SpeedyWeather.jl is designed to be easy to use, modify, and extend for research and education in atmospheric science.

   @article{SpeedyWeatherJOSS,
     doi = {10.21105/joss.06323},
     url = {https://doi.org/10.21105/joss.06323},
     year = {2024},
     publisher = {The Open Journal},
     volume = {9},
     number = {98},
     pages = {6323},
     author = {Kl\"{o}wer, Milan and Gelbrecht, Maximilian and Hotta, Daisuke and Willmert, Justin and Silvestri, Simone and Wagner, Gregory L. and White, Alistair and Hatfield, Sam and Kimpson, Tom and Constantinou, Navid C. and Hill, Chris},
     title = {{SpeedyWeather.jl: Reinventing atmospheric general circulation models towards interactivity and extensibility}},
     journal = {Journal of Open Source Software},
   }

3. A Modified Bisecting K-Means for Approximating Transfer Operators: Application to the Lorenz Equations

   Andre N. Souza, and Simone Silvestri

   2024

   Abs Bib PDF

   We investigate the convergence behavior of the extended dynamic mode decomposition for constructing a discretization of the continuity equation associated with the Lorenz equations using a nonlinear dictionary of over 1,000,000 terms. The primary objective is to analyze the resulting operator by varying the number of terms in the dictionary and the timescale. The dictionary comprises piecewise constant functions through a modified bisecting k-means algorithm and can efficiently scale to higher-dimensional systems.

   @misc{souza2024modifiedbisectingkmeansapproximating,
     title = {A Modified Bisecting K-Means for Approximating Transfer Operators: Application to the Lorenz Equations},
     author = {Souza, Andre N. and Silvestri, Simone},
     year = {2024},
     eprint = {2412.03734},
     archiveprefix = {arXiv},
     primaryclass = {physics.comp-ph},
   }

4. AI-driven emulation of ocean dynamics on sub-seasonal scales

   Suyash Bire, Jean Kossaifi, Simone Silvestri, and 4 more authors

   In ICLR Workshop on Climate Change AI, May 2024

   Bib

   @inproceedings{bire2024,
     title = {{AI}-driven emulation of ocean dynamics on sub-seasonal scales},
     author = {Bire, Suyash and Kossaifi, Jean and Silvestri, Simone and Kovachki, Nikola and Azizzadenesheli, Kamyar and Hill, Chris N and Anandkumar, Anima},
     booktitle = {ICLR Workshop on Climate Change AI},
     month = may,
     year = {2024}
   }

2023

1. Oceananigans.jl: A model that achieves breakthrough resolution, memory and energy efficiency in global ocean simulations

   Simone Silvestri, Gregory Wagner, Christopher Hill, and 10 more authors

   May 2023

   Abs Bib PDF

   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, the authors report several achievements: First, a global ocean simulation with breakthrough horizontal resolution – 488m – reaching 15 simulated days per day. Second, Oceananigans simulates the global ocean at 488m with breakthrough memory efficiency on just 768 Nvidia A100 GPUs. 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. Routine climate simulations with 10 km ocean components are within reach.

   @misc{silvestri2023oceananigansjl,
     title = {Oceananigans.jl: A model that achieves breakthrough resolution, memory and energy efficiency in global ocean simulations},
     author = {Silvestri, Simone and Wagner, Gregory and Hill, Christopher and Ardakani, Matin Raayai and Blaschke, Johannes and Campin, Jean-Michel and Churavy, Valentin and Constantinou, Navid and Edelman, Alan and Marshall, John and Ramadhan, Ali and Souza, Andre and Ferrari, Raffaele},
     year = {2023},
     eprint = {2309.06662},
     archiveprefix = {arXiv},
     primaryclass = {physics.ao-ph},
   }

2. OceanBioME.jl: A flexible environment for modelling the coupled interactions between ocean biogeochemistry and physics

   Jago Strong-Wright, Si Chen, Navid C. Constantinou, and 3 more authors

   Journal of Open Source Software, May 2023

   Abs Bib PDF

   OceanBioME.jl (Ocean Biogeochemical Modelling Environment) is a flexible modelling environment written in Julia for simulating the coupled interactions between ocean biogeochemistry, carbonate chemistry, and physics. OceanBioME can be used as a stand-alone box model, or coupled with Oceananigans.jl to run as a complete biogeochemical ocean model.

   @article{Strong-Wright2023,
     doi = {10.21105/joss.05669},
     url = {https://doi.org/10.21105/joss.05669},
     year = {2023},
     publisher = {The Open Journal},
     volume = {8},
     number = {90},
     pages = {5669},
     author = {Strong-Wright, Jago and Chen, Si and Constantinou, Navid C. and Silvestri, Simone and Wagner, Gregory LeClaire and Taylor, John R.},
     title = {OceanBioME.jl: A flexible environment for modelling the coupled interactions between ocean biogeochemistry and physics},
     journal = {Journal of Open Source Software},
   }

2021

1. Turbulence modulation in thermally expanding and contracting flows

   Simone Silvestri, and Rene Pecnik

   Journal of Fluid Mechanics, May 2021

   Abs Bib PDF

   We use direct numerical simulations to investigate how turbulence is modulated in heated and cooled channel flows with large temperature-dependent variations of density and viscosity. We show that the semi-local scaling framework, which accounts for local mean property variations, successfully collapses the first- and second-order turbulence statistics across a wide range of thermodynamic conditions. The analysis reveals that thermal expansion and contraction of the fluid create asymmetric turbulence modulation between the hot and cold walls.

   @article{silvestri_pecnik_2021,
     title = {Turbulence modulation in thermally expanding and contracting flows},
     volume = {926},
     doi = {10.1017/jfm.2021.726},
     journal = {Journal of Fluid Mechanics},
     author = {Silvestri, Simone and Pecnik, Rene},
     year = {2021},
     pages = {A30},
     dimensions = {true}
   }

2. Modelling turbulent heat flux accounting for Turbulence-Radiation Interactions

   Simone Silvestri, Dirk J.E.M. Roekaerts, and Rene Pecnik

   International Journal of Heat and Fluid Flow, May 2021

   Abs Bib PDF

   We investigate the modelling of turbulent heat flux in participating media where turbulence-radiation interactions (TRI) play a significant role. Using data from Direct Numerical Simulations of turbulent channel flows with radiative heat transfer, we assess the performance of standard gradient-diffusion models for the turbulent heat flux and propose corrections that account for the effects of TRI on the temperature variance and turbulent Prandtl number.

   @article{SILVESTRI2021108728,
     title = {Modelling turbulent heat flux accounting for Turbulence-Radiation Interactions},
     journal = {International Journal of Heat and Fluid Flow},
     volume = {89},
     pages = {108728},
     year = {2021},
     issn = {0142-727X},
     doi = {https://doi.org/10.1016/j.ijheatfluidflow.2020.108728},
     author = {Silvestri, Simone and Roekaerts, Dirk J.E.M. and Pecnik, Rene},
   }

2020

1. Mixed convection and radiation heat transfer in porous media for solar thermal applications

   Simone Silvestri, and Dirk J.E.M. Roekaerts

   May 2020

   Abs Bib PDF

   This chapter reviews the state of the art in modelling mixed convection and radiation heat transfer in porous media, with a focus on applications to solar thermal energy systems. We discuss the volume-averaged equations governing fluid flow, heat transfer, and radiative transport in porous structures, and review the available solution methods for the radiative transfer equation in participating porous media.

   @book{bookchapter,
     author = {Silvestri, Simone and Roekaerts, Dirk J.E.M.},
     title = {Mixed convection and radiation heat transfer in porous media for solar thermal applications},
     booktitle = {Convective heat transfer in porous media},
     publisher = {CRC Press},
     doi = {10.1201/9780429020261},
     year = {2020},
     editor = {Mahmoudi, Yasser and Hooman, Kamel and Vafai, Kambiz},
     pages = {227-262},
     chapter = {12},
   }

2019

1. A fast GPU Monte Carlo radiative heat transfer implementation for coupling with direct numerical simulation

   Simone Silvestri, and Rene Pecnik

   Journal of Computational Physics, May 2019

   Abs Bib PDF

   We present a fast GPU-based Monte Carlo method for radiative heat transfer designed for coupling with Direct Numerical Simulation of turbulent flows. The implementation exploits the massive parallelism of GPU architectures to trace millions of photon bundles simultaneously, achieving speedups of two orders of magnitude compared to CPU-based implementations. The method is validated against benchmark solutions and applied to turbulent channel flows with participating media.

   @article{SILVESTRI2019100032,
     title = {A fast GPU Monte Carlo radiative heat transfer implementation for coupling with direct numerical simulation},
     journal = {Journal of Computational Physics},
     volume = {3},
     pages = {100032},
     year = {2019},
     issn = {2590-0552},
     doi = {10.1016/j.jcpx.2019.100032},
     url = {http://www.sciencedirect.com/science/article/pii/S2590055219300484},
     author = {Silvestri, Simone and Pecnik, Rene},
   }

2. Assessing turbulence-radiation interactions in turbulent flows of non-gray media

   Simone Silvestri, Dirk J.E.M. Roekaerts, and Rene Pecnik

   Journal of Quantitative Spectroscopy and Radiative Transfer, May 2019

   Abs Bib PDF

   We assess turbulence-radiation interactions (TRI) in turbulent channel flows of non-gray participating media using Direct Numerical Simulation coupled with a Monte Carlo method for radiative heat transfer. The spectral properties of the medium are modelled using a line-by-line approach. We quantify the relative importance of TRI for different optical thicknesses and identify the conditions under which TRI effects are significant for the overall heat transfer.

   @article{SILVESTRI2019134,
     title = {Assessing turbulence-radiation interactions in turbulent flows of non-gray media},
     journal = {Journal of Quantitative Spectroscopy and Radiative Transfer},
     volume = {233},
     pages = {134-148},
     year = {2019},
     issn = {0022-4073},
     doi = {10.1016/j.jqsrt.2019.05.018},
     url = {http://www.sciencedirect.com/science/article/pii/S0022407318307234},
     author = {Silvestri, Simone and Roekaerts, Dirk J.E.M. and Pecnik, Rene},
   }

2018

1. Turbulence-radiation interaction in channel flow with various optical depths

   Simone Silvestri, Ashish Patel, Dirk J.E.M. Roekaerts, and 1 more author

   Journal of Fluid Mechanics, May 2018

   Abs Bib PDF

   We perform Direct Numerical Simulations of turbulent channel flow coupled with radiative heat transfer to investigate the interaction between turbulence and radiation at various optical depths. The radiative transfer equation is solved using a Monte Carlo method. We show that radiation significantly alters the mean temperature profile and the turbulent heat flux, and that these effects depend strongly on the optical thickness of the medium. The turbulence-radiation interaction is quantified by analysing the correlations between temperature and radiative source term fluctuations.

   @article{silvestri_patel_roekaerts_pecnik_2018,
     title = {Turbulence-radiation interaction in channel flow with various optical depths},
     volume = {834},
     doi = {10.1017/jfm.2017.738},
     journal = {Journal of Fluid Mechanics},
     publisher = {Cambridge University Press},
     author = {Silvestri, Simone and Patel, Ashish and Roekaerts, Dirk J.E.M. and Pecnik, Rene},
     year = {2018},
     pages = {359-384},
   }