Springer Nature is making SARS-CoV-2 and COVID-19 research free. View research | View latest news | Sign up for updates

The random pseudo-metric on a graph defined via the zero-set of the Gaussian free field on its metric graph


We further investigate properties of the Gaussian free field (GFF) on the metric graph associated to a discrete weighted graph (where the edges of the latter are replaced by continuous line-segments of appropriate length) that has been introduced by the first author. On such a metric graph, the GFF is a random continuous function that generalises one-dimensional Brownian bridges so that one-dimensional techniques can be used. In the present paper, we define and study the pseudo-metric defined on the metric graph (and therefore also on the discrete graph itself), where the length of a path on the metric graph is defined to be the local time at level zero accumulated by the Gaussian free field along this path. We first derive a pathwise transformation that relates the GFF on the metric graph with the reflected GFF on the metric graph via the pseudo-distance defined by the latter. This is a generalisation of Paul Lévy’s result relating the local time at zero of Brownian motion to the supremum of another Brownian motion. We also compute explicitly the distribution of certain functionals of this pseudo-metric and of the GFF. In particular, we point out that when the boundary consists of just two points, the law of the pseudo-distance between them depends solely on the resistance of the network between them. We then discuss questions related to the scaling limit of this pseudo-metric in the two-dimensional case, which should be the conformally invariant way to measure distances between CLE(4) loops introduced and studied by the second author with Wu, and by Sheffield, Watson and Wu. Our explicit laws on metric graphs also lead to new conjectures for related functionals of the continuum GFF on fairly general Riemann surfaces.

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

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8


  1. 1.

    Since this formula will be used throughout the paper, let us recall briefly a possible direct proof: For a Brownian motion W started from \(W_0\), one knows from the reflection principle [28] the joint law of \((W_T, I_T)\), from which one can deduce the law of \((|W_T|, L_T)\) by Lévy’s theorem. We note that by reflection, the laws of \((W_T, L_T) 1_{L_T >0}\) and of \((-W_T, L_T) 1_{L_T >0}\) coincide while \(L_T = 0\) implies that \(W_T\) has the same sign as \(W_0\). From this, one can deduce the joint law of \((W_T, L_T)\), and by conditioning on the value of \(W_T\), one gets the law of \(L_T\) for the Brownian bridge.


  1. 1.

    Ahlfors, L.V.: Conformal Invariants. Topics in Geometric Function Theory. AMS Chelsea Publishing, New York (2010)

  2. 2.

    Aru, J., Sepulveda, A., Werner, W.: On bounded-type thin local sets of the two-dimensional Gaussian free fields. J. Inst. Math. Jussieu (2016). arXiv:1603.03362 (to appear)

  3. 3.

    Bertoin, J., Pitman, J.: Path transformations connecting Brownian bridge, excursion and meander. Bull. Sci. Math. 118, 147–166 (1994)

  4. 4.

    Borodin, A., Salminen, P.: Handbook of Brownian Motion: Facts and Formulae, Volume XIV of Probability and Its Applications. Birkhäuser, Basel (1996)

  5. 5.

    Ding J., Li, L.: Chemical distances for level-set percolation of two-dimensional discrete Gaussian free fields (2016). arXiv:1605.04449

  6. 6.

    Dubédat, J.: SLE and Virasoro representations: localization. Commun. Math. Phys. 336(2), 695–760 (2015)

  7. 7.

    Duffin, R.J.: The extremal length of a network. J. Math. Anal. Appl. 5, 200–215 (1962)

  8. 8.

    Duplantier, B., Sheffield, S.: Liouville quantum gravity and KPZ. Invent. Math. 185, 333–393 (2010)

  9. 9.

    Farkas, H.M., Kra, I.: Riemann Surfaces, Volume 71 of Graduate Text in Mathematics, 2nd edn. Springer, Berlin (1992)

  10. 10.

    Garaberdian, P.R.: Partial Differential Equations. Chelsia Publishing Company, New York (1986)

  11. 11.

    Jurchescu, M.: Bordered Riemann surfaces. Math. Ann. 143, 264–292 (1961)

  12. 12.

    Karatzas, I., Shreve, S.E.: Brownian Motion and Stochastic Calculus, Volume 113 of Grad. Texts in Math, 2nd edn. Springer, Berlin (2010)

  13. 13.

    Kemppainen, A., Werner, W.: The nested simple conformal loop ensembles in the Riemann sphere. Probab. Theory Relat. Fields 165, 835–866 (2016)

  14. 14.

    Lawler, G.F.: Partition functions, loop measure, and versions of SLE. J. Stat. Phys. 134, 813–837 (2009)

  15. 15.

    Lawler, G.F., Limic, V.: Random Walk: A Modern Introduction. Cambridge Stud. Adv. Math, vol. 123. Cambridge University Press, Cambridge (2010)

  16. 16.

    Lawler, G.F., Werner, W.: Universality for conformally invariant intersection exponents. J. Eur. Math. Soc. 2, 291–328 (2000)

  17. 17.

    Lawler, G.F., Werner, W.: The Brownian loop-soup. Probab. Theory Relat. Fields 128, 565–588 (2004)

  18. 18.

    Le Jan, Y.: Markov paths, loops and fields. In: 2008 St-Flour Summer School, L.N. Math., Vol. 2026. Springer, Berlin (2011)

  19. 19.

    Le Jan, Y., Marcus, M.B., Rosen, J.: Permanental fields, loop soups and continuous additive functionals. Ann. Probab. 43(1), 44–84 (2015)

  20. 20.

    Lupu, T.: From loop clusters and random interlacements to the free field. Ann. Probab. 44(3), 2117–2146 (2016)

  21. 21.

    Lupu,T.: Convergence of the two-dimensional random walk loop soup clusters to CLE. J. Eur. Math. Soc. (2015). arXiv:1502.06827 (to appear)

  22. 22.

    Lyons, R., Peres, Y.: Probability on Trees and Networks. Cambridge University Press, Cambridge (2017)

  23. 23.

    Miller, J., Sheffield, S.: Imaginary geometry I: interacting SLEs. Probab. Theory Relat. Fields 164, 553–705 (2016)

  24. 24.

    Miller, J., Sheffield, S.: Quantum Loewner evolution. Duke Math. J. 165(17), 3241–3378 (2016)

  25. 25.

    Miller, J., Sheffield, S.: CLE(4) and the Gaussian free field. In preparation

  26. 26.

    Miller, J., Watson, S.S., Wilson, D.B.: The conformal loop ensemble nesting field. Probab. Theory Relat. Fields 163, 769–801 (2015)

  27. 27.

    Qian, W., Werner, W.: Decomposition of Brownian loop-soup clusters (2015). arXiv:1509.01180

  28. 28.

    Revuz, D., Yor, M.: Continuous Martingales and Brownian Motion, Volume 293 of Grundlehren Math. Wiss, 3rd edn. Springer, Berlin (1999)

  29. 29.

    Rozanov, Y.A.: Markov Random Fields, 1st edn. Springer, Berlin (1982)

  30. 30.

    Schramm, O., Sheffield, S.: Contour lines of the two-dimensional discrete Gaussian free field. Acta Math. 202, 21–137 (2009)

  31. 31.

    Schramm, O., Sheffield, S.: A contour line of the continuum Gaussian free field. Probab. Theory Relat. Fields 157, 47–80 (2013)

  32. 32.

    Sheffield, S.: The Gaussian free field for mathematicians. Probab. Theory Relat. Fields 139, 521–541 (2007)

  33. 33.

    Sheffield, S., Watson, S.S., Wu, H.: A conformally invariant metric on CLE(4). In preparation

  34. 34.

    Sheffield, S., Werner, W.: Conformal loop ensembles: the Markovian characterization and the loop-soup construction. Ann. Math. 176(3), 1827–1917 (2012)

  35. 35.

    Sznitman, A.-S.: Disconnection and level-set percolation for the Gaussian free field. J. Math. Soc. Japan 67(4), 1801–1843 (2015)

  36. 36.

    Wang, M., Wu, H.: Level lines of Gaussian free field I: zero-boundary GFF. Stoch. Process. Appl. 127(4), 1045–1124 (2017)

  37. 37.

    Werner, W.: Topics on the two-dimensional Gaussian free field. Lecture notes (2014)

  38. 38.

    Werner, W., Wu, H.: On conformally invariant CLE explorations. Commun. Math. Phys. 320, 637–661 (2013)

  39. 39.

    Wu, H.: Autour des relations entre SLE, CLE, champ libre gaussien, et conséquences. Ph.D. thesis, Ecole doctorale Mathématiques de la région Paris-Sud (2013)

  40. 40.

    Yor, M.: Some Aspects of Brownian Motion, Part I: Some Special Functionals. Birkhäuser, Basel (1992)

Download references


T.L. acknowledges the support of Dr. Max Rössler, the Walter Haefner Foundation and the ETH Zurich Foundation. W.W. acknowledges the support of the SNF Grant SNF-155922. The authors are also part of the NCCR Swissmap of the SNF. They also thank the referees for their comments.

Author information

Correspondence to Titus Lupu.

Additional information

Dedicated to the memory of Marc Yor.



Here we provide some details about the rather brute-force Itô formula computation of the semi-martingale decomposition of \(\varPsi _{\ell }(t)\) that proves Lemma 8. We use the notation of the beginning of Sect. 3. To simplify the expressions, we will use the following notation:

At each time t, the \(\tilde{\phi }^{*}_{i}(t)\) for different values of i have the same sign. We will denote this common sign by \(\sigma (t)\in \lbrace -1,+1\rbrace \). If all the \(\tilde{\phi }^{*}_{i}(t)\) are zero, the value of \(\sigma (t)\) does not matter. Similarly, we will denote the common sign of h on . If h vanishes on , the value of the sign does not matter.

According to Lemma 2, \(\tilde{\phi }^{*}_{i}(t)\) has the following semi-martingale decomposition:

where \(M^{*}_{i}(t)\) is an \((\widehat{{\mathcal {F}}}_{t})_{t\geqslant 0}\)-martingale. Moreover, \(\langle M_{i}^{*},M_{i}^{*}\rangle _{t}=r^{*}_{i}(t)\) and \( \langle M_{i}^{*},M_{j}^{*}\rangle _{t}=0\) for \(i\ne j\). Further, \(d S_{i\check{x}}(t)= d \tilde{\phi }^{*}_{i}(t)\), \(d S_{i\check{x}}^{+}(t)= \sigma (t) d\tilde{\phi }^{*}_{i}(t)\), and \(d\langle S_{i\check{x}},S_{i\check{x}}\rangle _{t}= d\langle S^{+}_{i\check{x}},S^{+}_{i\check{x}}\rangle _{t} = d r^{*}_{i}(t).\) According to Lemma 3, Eqs. (8) and (9),

Using Itô’s formula we get that

where \(\theta _{\check{x}}(t)=(\sigma (t)\vert h(\check{x})\vert + h(\check{x})+\sigma (t)(\ell -L^{*}(t)))\). Further


It follows that \(\varPsi _{\ell }(t)\) is a local martingale.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Lupu, T., Werner, W. The random pseudo-metric on a graph defined via the zero-set of the Gaussian free field on its metric graph. Probab. Theory Relat. Fields 171, 775–818 (2018).

Download citation


  • Random distances
  • Gaussian free field
  • Metric graph
  • Local time
  • Conformal loop ensemble

Mathematics Subject Classification

  • 60G60
  • 60G15
  • 60K35
  • 60J55
  • 60J65
  • 60J67