pith. sign in

arxiv: 2604.15539 · v1 · submitted 2026-04-16 · 🧮 math.NA · cs.NA

Efficient and well-conditioned ghost-point discretization of boundary operators on unfitted domains

Armando Coco , Alessandro Coclite , St\'ephane Clain , Rui Miguel Pereira This is my paper

Pith reviewed 2026-05-10 09:42 UTC · model grok-4.3

classification 🧮 math.NA cs.NA
keywords ghost-point methodsunfitted domainsleast-squares reconstructioncompact stencilsfinite differencesboundary conditionsconvection-diffusion
0
0 comments X

The pith

Least-squares reconstruction on compact stencils produces stable high-order ghost-point boundary conditions for unfitted Cartesian grids.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that a boundary operator built via least-squares reconstruction near each ghost node can enforce boundary conditions through compact linear relations between interior and ghost points. This replaces the wide stencils that high-order accuracy normally demands in unfitted finite-difference schemes on structured grids. A sympathetic reader would care because Cartesian unfitted methods avoid costly body-fitted mesh generation while offering straightforward parallelization, yet traditional ghost-point approaches lose those advantages at high order. The authors supply stencil-selection rules based on conditioning and iterative refinement to keep the resulting global systems stable and accurate. Experiments across geometries and convection-diffusion regimes, including boundary-layer cases, show that the target order is retained without new instabilities.

Core claim

The authors introduce a boundary operator that locally approximates the boundary condition near each ghost node via least-squares reconstruction on compact stencils. This operator replaces the usual wide-stencil ghost relations with linear equations coupling interior and ghost points, while preserving the desired order of accuracy. Stencil selection guided by conditioning criteria together with iterative refinement procedures are shown to deliver globally stable discretizations, as verified by numerical tests on multiple geometries and convection-diffusion regimes.

What carries the argument

A least-squares reconstructed boundary operator that converts the boundary condition into compact linear relations between interior and ghost points.

If this is right

  • High-order accuracy is retained even when thin boundary layers are present in convection-dominated problems.
  • The linear systems exhibit improved conditioning relative to wide-stencil ghost methods.
  • Stencil compactness increases, reducing communication overhead in parallel implementations.
  • The same accuracy level is achieved on a range of geometries without boundary-conforming grids.
  • Iterative refinement prevents order reduction across the tested convection-diffusion regimes.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The local reconstruction could be adapted to enforce interface conditions in multiphase or embedded-boundary problems.
  • Compact stencils may simplify coupling with fast iterative solvers or multigrid techniques for very large systems.
  • Extension to three-dimensional or time-dependent problems would test whether the conditioning criteria remain sufficient.
  • The approach might integrate naturally with adaptive Cartesian refinement because small stencils localize the boundary treatment.

Load-bearing premise

That least-squares reconstruction on compact stencils, combined with the proposed selection and iterative refinement, produces globally stable and accurate discretizations without instabilities or order reduction in all convection-diffusion regimes.

What would settle it

A sequence of refined grids on an irregular domain with thin boundary layers where the observed convergence order falls below the design order or the linear systems become singular under the proposed stencil rules.

Figures

Figures reproduced from arXiv: 2604.15539 by Alessandro Coclite, Armando Coco, Rui Miguel Pereira, St\'ephane Clain.

Figure 1
Figure 1. Figure 1: Stencils S1 (a), S2 (b) and S3 (c). The red point denotes the ghost node [PITH_FULL_IMAGE:figures/full_fig_p010_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Stencils S4.1, with ϑ = 45◦ (a), 60◦ (b), 90◦ (c), 120◦ (d), 360◦ (e). The red point denotes the ghost node xk from which the stencil is constructed, while the blue circle indicates the corresponding boundary point pk. L1 error and the L∞ errors defined on the discrete mesh Mh as: ∥e∥L1 = P xi∈Mh [PITH_FULL_IMAGE:figures/full_fig_p011_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Stencils S4.2 (left) and S4.3 (right) with [PITH_FULL_IMAGE:figures/full_fig_p012_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Numerical errors for the test described in Section 4.2. We use the stencils S1 (a), S2 (b), and S3 (c). [PITH_FULL_IMAGE:figures/full_fig_p013_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Numerical errors for the test described in Section 4.2. We use the stencils S4.1 (left), S4.2 (middle), and [PITH_FULL_IMAGE:figures/full_fig_p014_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Boxplots comparing local and global conditioning for the three stencils S4.1, S4.2, and S4.3, with [PITH_FULL_IMAGE:figures/full_fig_p015_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Numerical errors for the test described in Section 4.2. We use the stencil S4.3 and a vertex angle of [PITH_FULL_IMAGE:figures/full_fig_p016_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Boxplots of stencil diameters (maximum pairwise distance between stencil points divided by the spatial [PITH_FULL_IMAGE:figures/full_fig_p017_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Complex geometries that do not fit with the Cartesian grid: the rotated leaf (left), the flower (middle), [PITH_FULL_IMAGE:figures/full_fig_p018_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Numerical errors for the test described in Section 4.3: the rotated leaf (left), the flower (middle), and the [PITH_FULL_IMAGE:figures/full_fig_p018_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Numerical errors for the tests described in Section 4.4: Case 1 with [PITH_FULL_IMAGE:figures/full_fig_p019_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Numerical errors for the tests described in Section 4.4: Benchmark 1: [PITH_FULL_IMAGE:figures/full_fig_p020_12.png] view at source ↗
read the original abstract

Unfitted boundary methods are widely used to numerically solve partial differential equations (PDEs) on irregular domains, avoiding the computational burden of generating boundary-conforming grids. In the finite-difference framework, structured Cartesian grids offer advantages such as ease of implementation and efficient parallelization, while geometry is represented implicitly, for instance, through level-set functions. In this setting, ghost point methods are commonly employed to enforce boundary conditions by introducing additional relations between interior and ghost nodes. However, constructing these relations becomes challenging for high-order accurate discretizations, which often rely on wide stencils that can reduce computational efficiency and degrade performance in large-scale parallel simulations. In this work, we investigate alternative ghost-point discretizations based on compact stencils. We introduce a formulation based on a boundary operator that locally approximates the boundary condition near each ghost node, replacing it with linear relations involving both interior and ghost points. The operator is constructed via least-squares reconstruction, allowing flexible stencil configurations while preserving the desired order of accuracy. Several strategies for selecting and adapting compact stencils are proposed, guided by conditioning criteria and iterative refinement procedures to improve global stability. Numerical experiments on various geometries and convection-diffusion regimes demonstrate the effectiveness of the proposed approach, showing that it maintains high accuracy even in the presence of boundary layers and improves stencil compactness and conditioning of the resulting linear systems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 2 minor

Summary. The manuscript proposes a ghost-point method for enforcing boundary conditions in finite-difference discretizations of convection-diffusion PDEs on unfitted domains. It replaces the boundary condition at each ghost node with a linear relation obtained from a least-squares reconstruction of a local boundary operator on compact stencils, and introduces conditioning-guided selection and iterative refinement procedures to preserve the target order of accuracy while improving global stability and matrix conditioning. Numerical experiments on multiple geometries and flow regimes are reported to support maintained high-order accuracy and better-conditioned linear systems compared with wide-stencil alternatives.

Significance. If the stability and accuracy claims are substantiated, the work provides a practical route to high-order unfitted schemes that avoid the efficiency and parallelization penalties of wide stencils, which is valuable for large-scale simulations on complex domains. The construction rests on standard least-squares approximation theory rather than data-driven fitting, and the emphasis on explicit conditioning control is a constructive addition to the ghost-point literature.

major comments (1)
  1. Numerical Experiments section: the abstract states that experiments 'demonstrate maintained high accuracy' and 'improved conditioning,' yet the summary provides no tabulated L2 or L∞ errors, observed convergence rates, or side-by-side comparisons against wide-stencil baselines. Without these quantitative anchors it is difficult to verify that the compact-stencil variants retain the design order across the reported convection-dominated regimes.
minor comments (2)
  1. The description of the iterative refinement procedure would benefit from an explicit pseudocode or flowchart showing how the conditioning threshold interacts with the least-squares solve at each ghost point.
  2. Notation for the boundary operator and the reconstructed linear relations should be introduced with a single consistent symbol set in the formulation section to avoid later ambiguity when discussing stencil adaptation.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the positive overall assessment and the constructive comment regarding the presentation of numerical results. We address the major comment in detail below.

read point-by-point responses

  1. Referee: Numerical Experiments section: the abstract states that experiments 'demonstrate maintained high accuracy' and 'improved conditioning,' yet the summary provides no tabulated L2 or L∞ errors, observed convergence rates, or side-by-side comparisons against wide-stencil baselines. Without these quantitative anchors it is difficult to verify that the compact-stencil variants retain the design order across the reported convection-dominated regimes.

    Authors: We agree that tabulated quantitative data strengthens verifiability. Although the Numerical Experiments section presents convergence plots that visually confirm maintained high-order accuracy and improved conditioning across multiple geometries and convection-dominated regimes, we acknowledge the benefit of explicit side-by-side comparisons. In the revised manuscript we will add tables reporting L2 and L∞ error norms, observed convergence rates, and direct comparisons with wide-stencil ghost-point baselines for the convection-dominated test cases. These additions will provide the requested quantitative anchors without altering the existing figures or conclusions. revision: yes

Circularity Check

0 steps flagged

No significant circularity

full rationale

The paper's central construction uses least-squares reconstruction to define a local boundary operator on compact stencils, replacing boundary conditions with linear relations between interior and ghost points. This step follows directly from standard approximation theory and numerical linear algebra without reducing the claimed order of accuracy or stability to any fitted quantity derived from the same data or to a self-citation chain. Stencil selection and refinement are guided by explicit conditioning criteria, and the overall discretization remains consistent with prior ghost-point literature while being validated through independent numerical experiments on multiple geometries and regimes. No load-bearing step equates a prediction to its own input by definition or by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract provides insufficient detail to enumerate specific free parameters or axioms; the approach implicitly relies on standard least-squares minimization and finite-difference consistency assumptions, with stencil-size choices likely acting as tunable parameters.

pith-pipeline@v0.9.0 · 5546 in / 1093 out tokens · 26223 ms · 2026-05-10T09:42:02.859087+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

25 extracted references · 25 canonical work pages

  1. [1]

    Osher and R

    S. Osher and R. Fedkiw.Level Set Methods and Dynamic Implicit Surfaces. Springer-Verlag New York, Applied Mathematical Sciences, 2002

    work page 2002

  2. [2]

    Guidelines for Poisson solvers on irregular domains with Dirichlet boundary conditions using the ghost fluid method.J

    Yen Ting Ng, Han Chen, Chohong Min, and Frédéric Gibou. Guidelines for Poisson solvers on irregular domains with Dirichlet boundary conditions using the ghost fluid method.J. Sci. Comput., 41:300–320, 2009

    work page 2009

  3. [3]

    C. S. Peskin. Numerical analysis of blood flow in the heart.Journal of Computational Physics, 25:220–252, 1977

    work page 1977

  4. [4]

    Coclite, S

    A. Coclite, S. Ranaldo, M.D. de Tullio, P. Decuzzi, and G. Pascazio. Kinematic and dy- namic forcing strategies for predicting the transport of inertial capsules via a combined lattice boltzmann immersed boundary method. Computers & Fluids, 180:41–53, 2019. ISSN 0045-7930. doi: https://doi.org/10.1016/j.compfluid.2018.12.014. URLhttp://www. sciencedirect....

    work page doi:10.1016/j.compfluid.2018.12.014 2019

  5. [5]

    de Tullio

    Alessandro Coclite, Sergio Ranaldo, Giuseppe Pascazio, and Marco D. de Tullio. A lat- tice boltzmann dynamic-immersed boundary scheme for the transport of deformable iner- tial capsules in low-re flows. Computers & Mathematics with Applications, 80(12):2860 – 2876, 2020. ISSN 0898-1221. doi: https://doi.org/10.1016/j.camwa.2020.09.017. URL http://www.scie...

    work page doi:10.1016/j.camwa.2020.09.017 2020

  6. [6]

    LeVeque and Zhilin Li

    Randall J. LeVeque and Zhilin Li. The immersed interface method for elliptic equations with discontinuous coefficients and singular sources.Journal of Numerical Analysis, 31:1019–1044, 1994

    work page 1994

  7. [7]

    A second-order- accurate symmetric discretization of the poisson equation on irregular domains.Journal of computational physics, 176(1):205–227, 2002

    Frederic Gibou, Ronald P Fedkiw, Li-Tien Cheng, and Myungjoo Kang. A second-order- accurate symmetric discretization of the poisson equation on irregular domains.Journal of computational physics, 176(1):205–227, 2002

    work page 2002

  8. [8]

    Fedkiw, T

    R.P. Fedkiw, T. Aslam, B. Merriman, and S. Osher. A non-oscillatory Eulerian approach to interfaces in multimaterial flows (the ghost fluid method).Journal of Computational Physics, 152:457–492, 1999

    work page 1999

  9. [9]

    Gibou and R

    F. Gibou and R. Fedkiw. A fourth order accurate discretization for the Laplace and heat equa- tions on arbitary domains, with applications to the stefan problem.Journal of Computational Physics, 202:577–601, 2005

    work page 2005

  10. [10]

    An efficient fluid-solid coupling algorithm for single-phase flows.J

    Yen Ting Ng, Chohong Min, and Frédéric Gibou. An efficient fluid-solid coupling algorithm for single-phase flows.J. Comput. Phys., 228:8807–8829, 2009

    work page 2009

  11. [11]

    Finite-difference ghost-point multigrid methods on carte- sian grids for elliptic problems in arbitrary domains.Journal of Computational Physics, 241: 464–501, 2013

    Armando Coco and Giovanni Russo. Finite-difference ghost-point multigrid methods on carte- sian grids for elliptic problems in arbitrary domains.Journal of Computational Physics, 241: 464–501, 2013. 22

    work page 2013

  12. [12]

    Armando Coco and Giovanni Russo. Second order finite-difference ghost-point multigrid meth- ods for elliptic problems with discontinuous coefficients on an arbitrary interface.Journal of Computational Physics, 361:299–330, 2018

    work page 2018

  13. [13]

    A multigrid ghost-point level-set method for incompressible navier-stokes equations on moving domains with curved boundaries.Journal of computational physics, 418: 109623, 2020

    Armando Coco. A multigrid ghost-point level-set method for incompressible navier-stokes equations on moving domains with curved boundaries.Journal of computational physics, 418: 109623, 2020

    work page 2020

  14. [14]

    A hydro-geophysical simulator for fluid and mechanical processes in volcanic areas.Journal of Mathematics in Industry, 6(1):6, 2016

    Armando Coco, Gilda Currenti, Joachim Gottsmann, Giovanni Russo, and Ciro Del Negro. A hydro-geophysical simulator for fluid and mechanical processes in volcanic areas.Journal of Mathematics in Industry, 6(1):6, 2016

    work page 2016

  15. [15]

    A level-set multigrid technique for nonlinear diffusion in the numerical simulation of marble degradation under chemical pollu- tants

    Armando Coco, Matteo Semplice, and S Serra Capizzano. A level-set multigrid technique for nonlinear diffusion in the numerical simulation of marble degradation under chemical pollu- tants. Applied Mathematics and Computation, 386:125503, 2020

    work page 2020

  16. [16]

    A finite-difference ghost-point multigrid method for multi-scale modelling of sorption kinetics of a surfactant past an oscillating bubble

    Clarissa Astuto, Armando Coco, and Giovanni Russo. A finite-difference ghost-point multigrid method for multi-scale modelling of sorption kinetics of a surfactant past an oscillating bubble. Journal of Computational Physics, 476:111880, 2023

    work page 2023

  17. [17]

    High order multiscale methods for advection-diffusion equation in highly oscillatory regimes: application to surfactant diffusion and generalization to arbitrary domains

    Clarissa Astuto. High order multiscale methods for advection-diffusion equation in highly oscillatory regimes: application to surfactant diffusion and generalization to arbitrary domains. Communications in Computational Physics, 37(5):1327–1357, 2025

    work page 2025

  18. [18]

    Clarissa Astuto, Daniele Boffi, Giovanni Russo, and Umberto Zerbinati. A nodal ghost method based on variational formulation and regular square grid for elliptic problems on arbitrary domains in two space dimensions.Computer Methods in Applied Mechanics and Engineering, 443:118041, 2025

    work page 2025

  19. [19]

    A Comparison of the Coco-Russo Scheme and ghost-FEM for elliptic Equations in Arbitrary Domains

    Clarissa Astuto, Armando Coco, and Umberto Zerbinati. A Comparison of the Coco-Russo Scheme and ghost-FEM for elliptic Equations in Arbitrary Domains. pages 1–21, 2026

    work page 2026

  20. [20]

    High-order finite-difference ghost-point methods for el- liptic problems in domains with curved boundaries.Open Mathematics, 22(1):20240072, 2024

    Armando Coco and Giovanni Russo. High-order finite-difference ghost-point methods for el- liptic problems in domains with curved boundaries.Open Mathematics, 22(1):20240072, 2024

    work page 2024

  21. [21]

    Appelo and N.A

    D. Appelo and N.A. Peterson. A fourth-order accurate embedded boundary method for the wave equation.SIAM J. Sci. Comput., 50(6):2982–3008, 2012. doi: 10.1137/09077223X

    work page doi:10.1137/09077223x 2012

  22. [22]

    Baeza, P

    A. Baeza, P. Mulet, and D. Zorío. High order boundary extrapolation technique for finite difference methods on complex domains with cartesian meshe.J. Sci. Comput., 66:761–791,

    work page

  23. [23]

    doi: 10.1007/s10915-015-0043-2

    work page doi:10.1007/s10915-015-0043-2

  24. [24]

    Javier Fernández-Fidalgo, Stéphane Clain, Luis Ramírez, Ignasi Colominas, and Xesús Nogueira. Very high-order method on immersed curved domains for finite difference schemes with regular cartesian grids.Computer Methods in Applied Mechanics and Engineering, 360: 112782, 2020

    work page 2020

  25. [25]

    Very high-order finite dif- ference method on arbitrary geometries with cartesian grids for non-linear convection diffusion reaction equations

    Stéphane Clain, Diogo Lopes, Rui MS Pereira, and Paulo A Pereira. Very high-order finite dif- ference method on arbitrary geometries with cartesian grids for non-linear convection diffusion reaction equations. Journal of Computational Physics, 498:112667, 2024. 23 Appendix A. The S4 algorithm. Algorithm 1Construction of the S4.1, S4.2 and S4.3 isosceles t...

    work page 2024