Skip to main content
Log in

A Unified View of Space–Time Covariance Functions Through Gelfand Pairs

  • Published:
Journal of Fourier Analysis and Applications Aims and scope Submit manuscript

Abstract

We give a characterization of positive definite integrable functions on a product of two Gelfand pairs as an integral of positive definite functions on one of the Gelfand pairs with respect to the Plancherel measure on the dual of the other Gelfand pair. In the very special case where the Gelfand pairs are Euclidean groups and the compact subgroups are reduced to the identity, the characterization is a much cited result in spatio-temporal statistics due to Cressie, Huang and Gneiting. When one of the Gelfand pairs is compact the characterization leads to results about expansions in spherical functions with positive definite expansion functions, thereby recovering recent results of the author in collaboration with Peron and Porcu. In the special case when the compact Gelfand pair consists of orthogonal groups, the characterization is important in geostatistics and covers a recent result of Porcu and White.

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.

Similar content being viewed by others

References

  1. Andrews, G.E., Askey, R., Roy, R.: Special Functions. Cambridge University Press, Cambridge (1999)

    Book  Google Scholar 

  2. Apanasovich, T.V., Genton, M.G.: Cross-covariance functions for multivariate random fields based on latent dimensions. Biometrika 97, 15–30 (2010)

    Article  MathSciNet  Google Scholar 

  3. Berg, C.: Dirichlet forms on symmetric spaces. Ann. Inst. Fourier Grenoble 23(1), 135–156 (1973)

    Article  MathSciNet  Google Scholar 

  4. Berg, C., Forst, G.: Potential Theory on Locally Compact Abelian Groups. Springer, Berlin (1975)

    Book  Google Scholar 

  5. Berg, C., Porcu, E.: From Schoenberg coefficients to Schoenberg functions. Constr. Approx. 45, 217–241 (2017)

    Article  MathSciNet  Google Scholar 

  6. Berg, C., Peron, A.P., Porcu, E.: Orthogonal expansions related to compact Gelfand pairs. Expo. Math. 36, 259–277 (2018)

    Article  MathSciNet  Google Scholar 

  7. Charalambides, C.A.: Enumerative Combinatorics. Chapman & Hall, Boca Raton (2002)

    MATH  Google Scholar 

  8. Cooper, J.L.B.: Positive definite functions of a real variable. Proc. Lond. Math. Soc. 10, 53–66 (1960)

    Article  MathSciNet  Google Scholar 

  9. Cressie, N., Huang, H.-C.: Classes of nonseparable, spatiotemporal stationary covariance functions. J. Am. Stat. Assoc. 94, 1330–1340 (1999)

    Article  Google Scholar 

  10. Dai, F., Xu, Y.: Approximation Theory and Harmonic Analysis on Spheres and Balls. Springer, Berlin (2013)

    Book  Google Scholar 

  11. Dawson, M., Ólafsson, G., Wolf, J.A.: Direct systems of spherical functions and representations. J. Lie Theory 23, 711–729 (2013)

    MathSciNet  MATH  Google Scholar 

  12. Dieudonné, J.: Éléments D’Analyse, Tome VI. Chapitre XXII. Gauthier-Villars, Paris (1975)

    MATH  Google Scholar 

  13. Dixmier, J.: Les \(C^*\)-algèbres et Leurs Représentations. Gauthier-Villars, Paris (1964)

    MATH  Google Scholar 

  14. Faraut, J.: Analysis on Lie Groups. An Introduction. Cambridge University Press, Cambridge (2008)

    Book  Google Scholar 

  15. Furrer, R., Porcu, E., Nychka, D.: A 30 years of space-time covariance functions. WIREs Comput. Stat. (2020). https://doi.org/10.1002/wics.1512

    Article  Google Scholar 

  16. Gneiting, T.: Nonseparable, stationary covariance functions for space–time data. J. Am. Stat. Assoc. 97, 590–600 (2002)

    Article  MathSciNet  Google Scholar 

  17. Gneiting, T.: Strictly and non-strictly positive definite functions on spheres. Bernoulli 19(4), 1327–1349 (2013)

    Article  MathSciNet  Google Scholar 

  18. Kotz, S., Ostrovskii, I.V., Hayfavi, A.: Analytic and asymptotic properties of Linnik’s probability densities. I. J. Math. Anal. Appl. 193, 353–371 (1995)

    Article  MathSciNet  Google Scholar 

  19. Linnik, Ju.V.: Linear forms and statistical criteria, I, II. Select. Transl. Math. Stat. Probab. 3, 1–90 (1963). [Original paper appeared in Ukrainskii Mat. Zh. 5 (1953), 207–290.]

  20. Mateu, J., Porcu, E.: Positive Definite Functions: From Schoenberg to Space–Time Challenges. University Jaume I, Castellon, Spain, Department of Mathematics (2008)

  21. Phillips, T.R.L., Schmidt, K.M.: On unbounded positive definite functions. Math. Pannon. 26(2), 33–51 (2018)

    Google Scholar 

  22. Porcu, E., White, P.: Towards a complete picture of stationary covariance functions on spheres cross time. Electron. J. Stat. 13, 2566–2594 (2019)

    Article  MathSciNet  Google Scholar 

  23. Porcu, E., Cleanthous, G., Georgiadis, G., White, P.A., Alegria A.: Random fields on the hypertorus: Covariance modeling, regularities and approximations. Technical Report: Submitted for publication

  24. Porcu, E., Gregori, P., Mateu, J.: Nonseparable stationary anisotropic space-time covariance functions. Stoch. Environ. Res. Risk Assess. 21, 113–122 (2006)

    Article  MathSciNet  Google Scholar 

  25. Rudin, W.: Fourier Analysis on Groups. Interscience, New York (1962)

    MATH  Google Scholar 

  26. Sasvári, Z.: Positive Definite and Definitizable Functions. Akademie, Berlin (1994)

    MATH  Google Scholar 

  27. Schoenberg, I.J.: Positive definite functions on spheres. Duke Math. J. 9, 96–108 (1942)

    Article  MathSciNet  Google Scholar 

  28. Stein, E.M., Weiss, G.: Introduction to Fourier Analysis on Euclidean Spaces. Princeton University Press, Princeton (1971)

    MATH  Google Scholar 

  29. van Dijk, G.: Introduction to Harmonic Analysis and Generalized Gelfand Pairs De Gruyter Studies in Mathematics, 36. Walter de Gruyter & Co, Berlin (2009)

    Book  Google Scholar 

  30. Widder, D.V.: The Laplace Transform. Princeton University Press, Princeton (1941)

    MATH  Google Scholar 

  31. Wolf, J.A.: Harmonic Analysis on Commutative Spaces. Mathematical Surveys and Monographs, vol. 142. American Mathematical Society, New York (2007)

    Book  Google Scholar 

Download references

Acknowledgements

The author wants to thank Emilio Porcu, Zoltán Sasvári and Ryszard Szwarc for valuable advice during the preparation of this paper. The author also wants to thank two independent referees for valuable comments.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Christian Berg.

Additional information

Communicated by Marcin Bownik.

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendix

Appendix

We shall give an example showing that the functions \(C({\mathbf {h}};u)\) constructed in Proposition 1.1 need not be integrable.

In the following \({\mathcal { F}} f\) denotes the Fourier transform of a function \(f:{\mathbb {R}}\rightarrow {\mathbb {C}}\) given by

$$\begin{aligned} {\mathcal { F}} f(t)=\int _{-\infty }^\infty e^{-itx}f(x)\,dx,\quad t\in {\mathbb {R}}. \end{aligned}$$

Let \(C_0(]-1,1[)\) denote the set of continuous functions \(f:]-1,1[\rightarrow {\mathbb {C}}\) vanishing at “infinity”, i.e.,

$$\begin{aligned} \lim _{x\rightarrow -1}f(x)=\lim _{x\rightarrow 1} f(x)=0. \end{aligned}$$

It is a Banach space under the uniform norm

$$\begin{aligned} ||f||=\sup \{|f(x)|\mid -1<x<1\}. \end{aligned}$$

We proceed in a number of steps.

1: There exists \(f\in C_0(]-1,1[)\) such that \({\mathcal { F}} f\notin L^1({\mathbb {R}})\).

This is a classical application of the Banach-Steinhaus Theorem to the continuous linear functionals on the Banach space \(C_0(]-1,1[)\):

$$\begin{aligned} L_n(f)=\int _{-n\pi }^{n\pi } {\mathcal { F}} f(t)dt=\int _{-1}^1f(x)\frac{2\sin (n\pi x)}{x}dx,\quad f\in C_0(]-1,1[). \end{aligned}$$

In fact, assuming that \(L_n, n\ge 0\) is pointwise bounded, we get that \(||L_n||\) is bounded, which is a contradiction because

$$\begin{aligned} ||L_n||=\int _{-1}^1\left| \frac{2\sin (n\pi x)}{x}\right| dx=4\int _0^{n\pi }\frac{|\sin (u)|}{u}du \end{aligned}$$

tends to infinity for \(n\rightarrow \infty \). This shows the existence of an \(f\in C_0(]-1,1[)\) such that \((L_n(f))\) is an unbounded sequence, and in particular \({\mathcal { F}} f\notin L^1({\mathbb {R}})\).

2: There exists \(f\in C_0^+(]-1,1[)\) with \(\max f=1\) such that \({\mathcal { F}} f\notin L^1({\mathbb {R}})\).

This is an easy consequence of 1.

3: There exists \(f\in C_c^+({\mathbb {R}})\) with \(\max f=1\) and \(f(x)>0\) for \(x\in [0,1]\) such that \({\mathcal { F}} f\notin L^1({\mathbb {R}})\).

For f as in 2 let \(x_0\in {\mathbb {R}}\) satisfy \(f(x_0)=1\). Then there exists \(\delta >0\) such that \(f(x)>0\) for \(x\in [x_0,x_0+\delta ]\), and \(x\mapsto f(\delta x+x_0)\) satisfies 3.

4: There exists \(f\in C({\mathbb {R}})\cap L^1({\mathbb {R}})\) such that \(0<f(x)<1\) for all \(x\in {\mathbb {R}}\) and \({\mathcal { F}} f\notin L^1({\mathbb {R}})\).

Let h have the properties of 3 and let \(a_n>0, n\in {\mathbb {Z}}\) be such that \(a_0=1/2\), \(\sum _{n\in {\mathbb {Z}},n\ne 0} a_n=1/4\). Then

$$\begin{aligned} f(x)=\sum _{n\in {\mathbb {Z}}} a_n h(x-n),\quad x\in {\mathbb {R}}\end{aligned}$$

is continuous and has the properties \(0<f(x)\le 3/4\) for \(x\in {\mathbb {R}}\) and \(\int f(x)dx=3/4\int h(x)dx<\infty \). Furthermore,

$$\begin{aligned} {\mathcal { F}} f(t)=\mathcal F h(t)\sum _{n\in {\mathbb {Z}}}a_n e^{-itn}, \end{aligned}$$

hence

$$\begin{aligned} |{\mathcal { F}} f(t)|=|\mathcal F h(t)|\left| \sum _{n\in {\mathbb {Z}}}a_n e^{-itn}\right| \ge |\mathcal F h(t)|\left( a_0-\sum _{n\in {\mathbb {Z}},n\ne 0}a_n\right) =1/4|\mathcal F h(t)|, \end{aligned}$$

showing that \({\mathcal { F}} f\notin L^1({\mathbb {R}})\).

Conclusion Define

$$\begin{aligned} g(\omega ;\tau )=\frac{1}{\sqrt{2\pi }}f(\omega )e^{-\tau ^2/2},\quad (\omega ,\tau )\in {\mathbb {R}}\times {\mathbb {R}}, \end{aligned}$$

where f satisfies 4. Then g is strictly positive, continuous and integrable and

$$\begin{aligned} h(\omega ;u):=\int _{-\infty }^\infty g(\omega ;\tau )e^{iu\tau }d\tau =f(\omega )e^{-u^2/2} \end{aligned}$$

satisfies (C1’) and (C2’) of Proposition 1.1 with \(d=1\) and

$$\begin{aligned} C(h;u)=\int _{-\infty }^\infty e^{ih\omega }h(\omega ;u)d\omega =e^{-u^2/2}{\mathcal { F}} f(-h)\notin L^1({\mathbb {R}}^2). \end{aligned}$$

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Berg, C. A Unified View of Space–Time Covariance Functions Through Gelfand Pairs. J Fourier Anal Appl 26, 90 (2020). https://doi.org/10.1007/s00041-020-09793-z

Download citation

  • Received:

  • Revised:

  • Published:

  • DOI: https://doi.org/10.1007/s00041-020-09793-z

Keywords

Mathematics Subject Classification

Navigation