SearcharxivSearch

arXiv subjects

Mark R. Petersen

Publications and source records attributed to Mark R. Petersen.

11 recordsLinked to original sources

Viability of Tensor Train Methods for Geophysical Fluid Dynamics

Tensor train (TT) methods have recently gained popularity for accelerating the solving of systems of PDEs. Here, we evaluate the performance of TT methods in the context of geophysical fluid dynamics (GFD) using the shallow water equations and a discretization scheme employed by the ocean component of the Energy Exascale Earth System Model (E3SM). Through a suite of four test cases of increasing complexity, we evaluate TT methods in terms of how much TT is able to compress the model state, the error incurred by the TT approximation, and the speedup obtained by TT versus an optimal standard non-TT implementation in a representative subproblem. We show that though TT is able to effectively compress and speed up simple flows, it struggles to efficiently represent more complex states that are common in realistic GFD applications.

physics.flu-dyn

Local Time-Stepping for the Shallow Water Equations using CFL Optimized Forward-Backward Runge-Kutta Schemes

The Courant-Friedrichs-Lewy (CFL) condition is a well known, necessary condition for the stability of explicit time-stepping schemes that effectively places a limit on the size of the largest admittable time-step for a given problem. We formulate and present a new local time-stepping (LTS) scheme optimized, in the CFL sense, for the shallow water equations (SWEs). This new scheme, called FB-LTS, is based on the CFL optimized forward-backward Runge-Kutta schemes from Lilly et al. (2023). We show that FB-LTS maintains exact conservation of mass and absolute vorticity when applied to the TRiSK spatial discretization (Ringler et al., 2010), and provide numerical experiments showing that it retains the temporal order of the scheme on which it is based (second order). In terms of computational performance, we show that when applied to a real-world test case on a highly-variable resolution mesh, the MPAS-Ocean implementation of FB-LTS is up to 10 times faster than the classical four-stage, fourth-order Runge-Kutta method (RK4), and 2.3 times faster than an existing strong stability preserving Runge-Kutta based LTS scheme (LTS3). Despite this significant increase in efficiency, the solutions produced by FB-LTS are qualitatively equivalent to those produced by both RK4 and LTS3.

math.NA

Application of Continuous Data Assimilation in High-Resolution Ocean Modeling

We demonstrate a formulation of the Azouani-Olson-Titi (AOT) algorithm in the MPAS-Ocean implementation of the primitive equations of the ocean, presenting global ocean simulations with realistic coastlines and bathymetry. We observe an exponentially fast decay in the error before reaching a certain error level, which depends on the terms involved and whether the AOT feedback control term was handled implicitly or explicitly. A wide range of errors was observed for both schemes, with the implicit scheme typically exhibiting lower error levels, depending on the specific physical terms included in the model. Several factors seem to be contributing to this wide range, but the vertical mixing term is demonstrated to be an especially problematic term. This study provides insight into the promises and challenges of adapting the AOT algorithm to the setting of high-resolution, realistic ocean models.

math.AP

CFL Optimized Forward-Backward Runge-Kutta Schemes for the Shallow Water Equations

We present the formulation and optimization of a Runge-Kutta-type time-stepping scheme for solving the shallow water equations, aimed at substantially increasing the effective allowable time-step over that of comparable methods. This scheme, called FB-RK(3,2), uses weighted forward-backward averaging of thickness data to advance the momentum equation. The weights for this averaging are chosen with an optimization process that employs a von Neumann-type analysis, ensuring that the weights maximize the admittable Courant number. Through a simplified local truncation error analysis and numerical experiments, we show that the method is at least second order in time for any choice of weights and exhibits low dispersion and dissipation errors for well-resolved waves. Further, we show that an optimized FB-RK(3,2) can take time-steps up to 2.8 times as large as a popular three-stage, third-order strong stability preserving Runge-Kutta method in a quasi-linear test case. In fully nonlinear shallow water test cases relevant to oceanic and atmospheric flows, FB-RK(3,2) outperforms SSPRK3 in admittable time-step by factors between roughly between 1.6 and 2.2, making the scheme approximately twice as computationally efficient with little to no effect on solution quality.

math.NA

Storm Surge Modeling as an Application of Local Time-stepping in MPAS-Ocean

This paper presents the first scientific application of local time-stepping (LTS) schemes in the Model for Prediction Across Scales-Ocean (MPAS-O). We use LTS schemes in a single-layer, global ocean model that predicts the storm surge around the eastern coast of the United States during Hurricane Sandy. The variable-resolution meshes used are of unprecedentedly high resolution in MPAS-O, containing cells as small as 125 meters wide in Delaware Bay. It is shown that a particular, third-order LTS scheme (LTS3) produces sea-surface height (SSH) solutions that are of comparable quality to solutions produced by the classical four-stage, fourth-order Runge-Kutta method (RK4) with a uniform time step on the same meshes. Furthermore, LTS3 is up to 35% faster in the best cases, showing that LTS schemes are viable for use in MPAS-O with the added benefit of substantially less computational cost. The results of these performance experiments inform us of the requirements for efficient mesh design for LTS schemes. In particular, we see that for LTS to be efficient on a given mesh, it is important to have enough cells using the coarse time-step relative to those using the fine time-step, typically at least 1:5.

physics.flu-dyn

An unstructured CD-grid variational formulation for sea ice dynamics

For the numerical simulation of earth system models, Arakawa grids are largely employed. A quadrilateral mesh is assumed for their classical definition, and different types of grids are identified depending on the location of the discretized quantities. The B-grid has both velocity components at the center of a cell, the C-grid places the velocity components on the edges in a staggered fashion, and the D-grid is a ninety-degree rotation of a C-grid. Historically, B-grid formulations of sea ice dynamics have been dominant because they have matched the grid type used by ocean models. In recent years, as ocean models have increasingly progressed to C-grids, sea ice models have followed suit on quadrilateral meshes, but few if any implementations of unstructured C-grid sea ice models have been developed. In this work, we present an unstructured CD-grid type formulation of the elastic-viscous-plastic rheology, where the velocity unknowns are located at the edges, rather than at the vertices, as in the B-grid. The notion of a CD-grid has been recently introduced and assumes that the velocity components are co-located at the edges. The mesh cells in our analysis have $n$ sides, with $n$ greater than or equal to four. Numerical results are included to investigate the features of the proposed method. Our framework of choice is the Model for Prediction Across Scales (MPAS) within E3SM, the climate model of the U.S. Department of Energy, although our approach is general and could be applied to other models as well. While MPAS-Seaice is currently defined on a B-grid, MPAS-Ocean runs on a C-grid, hence interpolation operators are heavily used when coupled simulations are performed. The discretization introduced here aims at transitioning the dynamics of MPAS-Seaice to a CD-grid, to ultimately facilitate improved coupling with MPAS-Ocean and reduce numerical errors associated with this communication.

physics.comp-ph

Efficient form of the LANS-alpha turbulence model in a primitive-equation ocean model

The Lagrangian-Averaged Navier-Stokes alpha (LANS-alpha) model is a turbulence parameterization that has been shown to capture some of the most important features of high resolution ocean modeling at lower resolution. Simulations using LANS-alpha in the POP primitive-equation ocean model resemble doubled-resolution simulations of standard POP in statistics like kinetic energy, eddy kinetic energy, and potential temperature fields. The computational cost of adding LANS-alpha is only 27% for our most efficient implementation, as compared to a factor of 8-10 for a doubling of resolution. The LANS-alpha model improves turbulence statistics with an additional nonlinear term and a smoothed advecting velocity. In this work we investigate different kinds of smoothing techniques and their effect on the LANS-alpha model's results and efficiency. We show that we can substitute convolution filters for full Hemlholtz inversions and produce similar results at a significantly lower expense. When constructing filters for LANS-alpha in a primitive-equation ocean model the filter weights must be chosen carefully, otherwise a pressure-velocity instability will be excited. We show analytically that certain ranges of filter weights are unstable, and confirm this with numerical experiments. Our stability criterion also guarantees that the kinetic energy is well defined, and that the filtered velocity is smoother that the original velocity.

physics.ao-ph

Implementation of the LANS-alpha turbulence model in a primitive equation ocean model

This paper presents the first numerical implementation and tests of the Lagrangian-averaged Navier-Stokes-alpha (LANS-alpha) turbulence model in a primitive equation ocean model. The ocean model in which we work is the Los Alamos Parallel Ocean Program (POP); we refer to POP and our implementation of LANS-alpha as POP-alpha. Two versions of POP-alpha are presented: the full POP-alpha algorithm is derived from the LANS-alpha primitive equations, but requires a nested iteration that makes it too slow for practical simulations; a reduced POP-alpha algorithm is proposed, which lacks the nested iteration and is two to three times faster than the full algorithm. The reduced algorithm does not follow from a formal derivation of the LANS-alpha model equations. Despite this, simulations of the reduced algorithm are nearly identical to the full algorithm, as judged by globally averaged temperature and kinetic energy, and snapshots of temperature and velocity fields. Both POP-alpha algorithms can run stably with longer timesteps than standard POP. Comparison of implementations of full and reduced POP-alpha algorithms are made within an idealized test problem that captures some aspects of the Antarctic Circumpolar Current, a problem in which baroclinic instability is prominent. Both POP-alpha algorithms produce statistics that resemble higher-resolution simulations of standard POP. A linear stability analysis shows that both the full and reduced POP-alpha algorithms benefit from the way the LANS-alpha equations take into account the effects of the small scales on the large. Both algorithms (1) are stable; (2) make the Rossby Radius effectively larger; and (3) slow down Rossby and gravity waves.

physics.ao-ph

Baroclinic Vorticity Production in Protoplanetary Disks; Part II: Vortex Growth and Longevity

The factors affecting vortex growth in convectively stable protoplanetary disks are explored using numerical simulations of a two-dimensional anelastic-gas model which includes baroclinic vorticity production and radiative cooling. The baroclinic feedback, where anomalous temperature gradients produce vorticity through the baroclinic term and vortices then reinforce these temperature gradients, is found to be an important process in the rate of growth of vortices in the disk. Factors which strengthen the baroclinic feedback include fast radiative cooling, high thermal diffusion, and large radial temperature gradients in the background temperature. When the baroclinic feedback is sufficiently strong, anticyclonic vortices form from initial random perturbations and maintain their strength for the duration of the simulation, for over 600 orbital periods. Based on both simulations and a simple vortex model, we find that the local angular momentum transport due to a single vortex may be inward or outward, depending its orientation. The global angular momentum transport is highly variable in time, and is sometimes negative and sometimes positive. This result is for an anelastic gas model, and does not include shocks that could affect angular momentum transport in a compressible-gas disk.

astro-ph

Baroclinic Vorticity Production in Protoplanetary Disks; Part I: Vortex Formation

The formation of vortices in protoplanetary disks is explored via pseudo-spectral numerical simulations of an anelastic-gas model. This model is a coupled set of equations for vorticity and temperature in two dimensions which includes baroclinic vorticity production and radiative cooling. Vortex formation is unambiguously shown to be caused by baroclinicity because (1) these simulations have zero initial perturbation vorticity and a nonzero initial temperature distribution; and (2) turning off the baroclinic term halts vortex formation, as shown by an immediate drop in kinetic energy and vorticity. Vortex strength increases with: larger background temperature gradients; warmer background temperatures; larger initial temperature perturbations; higher Reynolds number; and higher resolution. In the simulations presented here vortices form when the background temperatures are $\sim 200K$ and vary radially as $r^{-0.25}$, the initial vorticity perturbations are zero, the initial temperature perturbations are 5% of the background, and the Reynolds number is $10^9$. A sensitivity study consisting of 74 simulations showed that as resolution and Reynolds number increase, vortices can form with smaller initial temperature perturbations, lower background temperatures, and smaller background temperature gradients. For the parameter ranges of these simulations, the disk is shown to be convectively stable by the Solberg-Høiland criteria.

astro-ph

On the Role of Global Warming on the Statistics of Record-Breaking Temperatures

We theoretically study long-term trends in the statistics of record-breaking daily temperatures and validate these predictions using Monte Carlo simulations and data from the city of Philadelphia, for which 126 years of daily temperature data is available. Using extreme statistics, we derive the number and the magnitude of record temperature events, based on the observed Gaussian daily temperatures distribution in Philadelphia, as a function of the number of elapsed years from the start of the data. We further consider the case of global warming, where the mean temperature systematically increases with time. We argue that the current warming rate is insufficient to measurably influence the frequency of record temperature events over the time range of the observations, a conclusion that is supported by numerical simulations and the Philadelphia temperature data.

physics.data-an