Skip to main content
Log in

Stability of Radial Basis Function Methods for Convection Problems on the Circle and Sphere

  • Published:
Journal of Scientific Computing Aims and scope Submit manuscript

Abstract

This paper investigates the stability of the radial basis function (RBF) collocation method for convective problems on the circle and sphere. We prove that the RBF method is Lax-stable for problems on the circle when the collocation points are equispaced and the transport speed is constant. We also show that the eigenvalues of discretization matrices are purely imaginary in the case of variable coefficients and equispaced nodes. By studying the \(\epsilon \)-pseudospectra of these matrices we argue that approximations are also Lax-stable in the latter case. Based on these results, we conjecture that the discretization of transport operators on the sphere present a similar behavior. We provide strong evidence that the method is Lax-stable on the sphere when the collocation points come from certain polyhedra. In both geometries, we demonstrate that eigenvalues of the differentiation matrix deviate from the imaginary axis linearly with perturbations off the set of ideal collocation points. When the ideal set is impractical or unavailable, we propose a least-squares method and present numerical evidence suggesting that it can substantially improve stability without any increase to computational cost and with only a minor cost to accuracy.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12

Similar content being viewed by others

References

  1. Boyd, J.: Chebyshev and Fourier Spectral Methods. Dover Publications, Mineola (2001)

    MATH  Google Scholar 

  2. Buhmann, M.: Radial Basis Functions. Cambridge University Press, New York (2003)

    Book  MATH  Google Scholar 

  3. Driscoll, T.A., Fornberg, B.: Interpolation in the limit of increasingly flat radial basis functions. Comput. Math. Applic. 43, 413–422 (2002)

    Article  MathSciNet  MATH  Google Scholar 

  4. Fasshauer, G.: Meshfree Approximation Methods with MATLAB. World Scientific Publishing Co. Inc., River Edge (2007)

    Book  MATH  Google Scholar 

  5. Flyer, N., Lehto, E., Blaise, S., Wright, G.B., St-Cyr, A.: A guide to RBF-generated finite differences for nonlinear transport: Shallow water simulations on a sphere. J. Comput. Phys. 231(11), 4078–4095 (2012)

    Article  MathSciNet  MATH  Google Scholar 

  6. Flyer, N., Wright, G.B.: Transport schemes on a sphere using radial basis functions. J. Comput. Phys. 226, 1059–1084 (2007)

    Article  MathSciNet  MATH  Google Scholar 

  7. Flyer, N., Wright, G.B.: A radial basis function method for the shallow water equations on a sphere. Proc. R. Soc. A Math. Phys. Eng. Sci. 465(2106), 1949–1976 (2009)

    Article  MathSciNet  MATH  Google Scholar 

  8. Fornberg, B., Larsson, E., Flyer, N.: Stable computations with Gaussian radial basis functions. SIAM J. Sci. Comput. 33(2), 869–892 (2011)

    Article  MathSciNet  MATH  Google Scholar 

  9. Fornberg, B., Merrill, D.: Comparison of finite difference and pseudospectral methods for convective flow over a sphere. Gheophys. Res. Lett. 24(24), 3245–3248 (1997)

    Article  Google Scholar 

  10. Fornberg, B., Piret, C.: A stable algorithm for flat radial basis functions on a sphere. SIAM J. Sci. Comput. 30, 60–80 (2007)

    Article  MathSciNet  MATH  Google Scholar 

  11. Fornberg, B., Piret, C.: On choosing a radial basis function and a shape parameter when solving a convective pde on a sphere. J. Comput. Phys. 227, 2758–2780 (2008)

    Article  MathSciNet  MATH  Google Scholar 

  12. Fornberg, B., Wright, G.B.: Stable computation of multiquadric interpolants for all values of the shape parameter. Computers Math. Applic. 48, 853–867 (2004)

    Article  MathSciNet  MATH  Google Scholar 

  13. Fornberg, B., Zuev, J.: The Runge phenomenon and spatially variable shape parameters in rbf interpolation. Comput. Math. Appl. 54, 379–398 (2007)

    Article  MathSciNet  MATH  Google Scholar 

  14. Martel, J.M., Platte, R.B.: Matlab code for convection problems on the circle and sphere. https://math.la.asu.edu/~platte/software.html (2015)

  15. Platte, R.B., Driscroll, T.A.: Eigenvalue stability of radial basis function discretizations for time-dependent problems. Comput. Math. Appl. 51, 1251–1268 (2006)

    Article  MathSciNet  MATH  Google Scholar 

  16. Reddy, S., Trefethen, L.N.: Stability of the method of lines. Numer. Math. 62, 235–267 (1992)

    Article  MathSciNet  MATH  Google Scholar 

  17. Trefethen, L.N.: Spectral Methods in MATLAB. SIAM, Philadelphia (2000)

    Book  MATH  Google Scholar 

  18. Trefethen, L.N., Embree, M.: Spectra and Pseudospectra: the Behavior of Nonnormal Matrices and Operators. Princeton University Press, Princeton (2005)

    MATH  Google Scholar 

  19. Wendland, H.: Scattered Data Approximation. Cambridge University Press, Cambridge (2005)

    MATH  Google Scholar 

  20. Wright, T.G.: EigTool. http://www.comlab.ox.ac.uk/pseudospectra/eigtool/ (2002)

Download references

Acknowledgments

The authors would like to thank Anne Gelb, Toby Driscoll, Bengt Fornberg and Natasha Flyer for their insight and suggestions. The first author is particularly grateful for the support of Bengt Fornberg and Natasha Flyer at CU Boulder and NCAR, respectively, who gave invaluable commentary on the present manuscript. We have benefited from using EigTool [20] in our computations.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Rodrigo B. Platte.

Additional information

This work is supported in part by NSF-DMS-CSUMS 0703587, NSF-DMS1522639 and AFOSR FA9550-12-1-0393.

Appendix

Appendix

1.1 Proof that A and B commute

Proof

Fix \(1\le i,j\le N\) and let \(\psi _{ij}:\{1,\ldots ,N\}\rightarrow \{1,\ldots ,N\}\) be given by

$$\begin{aligned} \psi _{ij}(k)=i+j-k+N(\mathbf {1}_{i+j-k\,<\,1}-\mathbf {1}_{N\,<\,i+j-k}) \end{aligned}$$
(72)

where \(\mathbf {1}\) denotes the indicator function. It can be shown that \(\psi _{ij}\) is an involution (i.e. \(\psi _{ij}\)=\(\psi _{ij}^{-1}\)) and hence a bijection. Next, fix \(1\le k\le N\) and observe that

$$\begin{aligned} \cos (2\pi (\psi _{ij}(k)-j)/N)&=\cos (2\pi (i-k)/N-2\pi (\mathbf {1}_{i+j-k\,<\,1}-\mathbf {1}_{N\,<\,i+j-k})) \nonumber \\&=\cos (2\pi (i-k)/N), \end{aligned}$$
(73)
$$\begin{aligned} \cos (2\pi (i-\psi _{ij}(k))/N)&=\cos (2\pi (k-j)/N-2\pi (\mathbf {1}_{i+j-k\,<\,1}-\mathbf {1}_{N\,<\,i+j-k}))\nonumber \\&=\cos (2\pi (k-j)/N) \end{aligned}$$
(74)

and similarly for sine. Comparing (73)–(74) to (21)–(22), we see that

$$\begin{aligned} a_{ik}b_{kj}=b_{i\psi _{ij}(k)}a_{\psi _{ij}(k)j}. \end{aligned}$$
(75)

Since \(\psi _{ij}\) is a bijection,

$$\begin{aligned} AB\equiv X=[x_{ij}]=\left[ \sum _{k=1}^{N}a_{ik}b_{kj}\right] =\left[ \sum _{k=1}^{N}b_{i\psi _{ij}(k)}a_{\psi _{ij}(k)j}\right] =[y_{ij}]=Y\equiv BA \end{aligned}$$
(76)

and so A and B commute.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Martel, J.M., Platte, R.B. Stability of Radial Basis Function Methods for Convection Problems on the Circle and Sphere. J Sci Comput 69, 487–505 (2016). https://doi.org/10.1007/s10915-016-0206-9

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10915-016-0206-9

Keywords

Mathematics Subject Classification

Navigation