On the Computation of the Extremal Index for Time Series


The extremal index is a quantity introduced in extreme value theory to measure the presence of clusters of exceedances. In the dynamical systems framework, it provides important information about the dynamics of the underlying systems. In this paper we provide a review of the meaning of the extremal index in dynamical systems. Depending on the observables used, this quantity can inform on local properties of attractors such as periodicity, stability and persistence in phase space, or on global properties such as the Lyapunov exponents. We also introduce a new estimator of the extremal index and show its relation with those previously introduced in the statistical literature. We reserve a particular focus to the systems perturbed with noise as they are a good paradigm of many natural phenomena. Different kind of noises are investigated in the annealed and quenched situations. Applications to climate data are also presented.

This is a preview of subscription content, access via your institution.

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


  1. 1.

    Abadi et al. [16] built a dynamically generated stochastic processes with an extremal index for which that equality does not hold. They considered observable functions maximised at least two points of the phase space, where one of them is an indifferent periodic point and another one is either a repelling periodic point or a non periodic point. We will not consider these kind of observables in this paper.

  2. 2.

    This condition holds for instance when the invariant measure \(\mu \) is mixing with decay of correlation fast enough; sometimes a rate of decay as \(n^{-2}\) is sufficient.

  3. 3.

    Notice that this is equivalent to Eq. (22) by defining the map \(\omega \rightarrow f_{(\omega )'}\), where \((\omega )'\) is the \(\omega _1\) coordinate of \(\omega \).

  4. 4.

    These conditions essentially ensure that the transfer operator associated with the map T has a spectral gap and that the density h has finite oscillation in the neighborhood of the diagonal.

  5. 5.

    This time scale is relevant for the underlying dynamics of the atmospheric circulation.


  1. 1.

    Faranda, D., Moreira Freitas, A.C., Milhazes Freitas, J., Holland, M., Kuna, T., Lucarini, V., Nicol, M., Todd, M., Vaienti, S.: Extremes and Recurrence in Dynamical Systems. Wiley, New York (2016)

    MATH  Google Scholar 

  2. 2.

    Faranda, D., Lucarini, V., Manneville, P., Wouters, J.: On using extreme values to detect global stability thresholds in multi-stable systems: the case of transitional plane Couette flow. Chaos Solit. Fractals 64, 26–35 (2014)

    ADS  Article  Google Scholar 

  3. 3.

    Lucarini, V., Faranda, D., Wouters, J., Kuna, T.: Towards a general theory of extremes for observables of chaotic dynamical systems. J. Stat. Phys. 154(3), 723–750 (2014)

    ADS  MathSciNet  Article  Google Scholar 

  4. 4.

    Lucarini, V., Blender, R., Herbert, C., Ragone, F., Pascale, S., Wouters, J.: Mathematical and physical ideas for climate science. Rev. Geophys. 52(4), 809–859 (2014)

    ADS  Article  Google Scholar 

  5. 5.

    Moloney, N., Faranda, D., Sato, Y.: An overview of the extremal index. J. Nonlinear Sci. 29, 022101 (2018)

    MathSciNet  Google Scholar 

  6. 6.

    Faranda, D., Messori, G., Yiou, P.: Dynamical proxies of North Atlantic predictability and extremes. Sci. Rep. 7, 41278 (2017)

    ADS  Article  Google Scholar 

  7. 7.

    Caballero, R., Faranda, D., Messori, G.: A dynamical systems approach to studying midlatitude weather extremes. Geophys. Res. Lett. 44(7), 3346–3354 (2017)

    ADS  Article  Google Scholar 

  8. 8.

    Faranda, D., Sato, Y., Messori, G., Moloney, N.R., Yiou, P.: Minimal dynamical systems model of the northern hemisphere jet stream via embedding of climate data. Earth Syst. Dyn. 10, 555–567 (2019). https://doi.org/10.5194/esd-2018-80

    ADS  Article  Google Scholar 

  9. 9.

    Faranda, D., Alvarez-Castro, M.C., Messori, G., Rodrigues, D., Yiou, P.: The hammam effect or how a warm ocean enhances large scale atmospheric predictability. Nat. Commun. 10(1), 1316 (2019)

    ADS  Article  Google Scholar 

  10. 10.

    Faranda, D., Vaienti, S.: Correlation dimension and phase space contraction via extreme value theory. Chaos 28, 041103 (2018)

    ADS  MathSciNet  Article  Google Scholar 

  11. 11.

    Freitas, A.C.M., Freitas, J.M., Todd, M.: The extremal index, hitting time statistics and periodicity. Adv. Math. 231(5), 2626–2665 (2012)

    MathSciNet  Article  Google Scholar 

  12. 12.

    Caby, T., Faranda, D., Mantica, G., Vaienti, S., Yiou, P.: Generalized dimensions, large deviations and the distribution of rare events. arXiv:1812.00036

  13. 13.

    Keller, G., Liverani, C.: Rare events, escape rates and quasistationarity: some exact formulae. J. Stat. Phys. 135, 519–534 (2009)

    ADS  MathSciNet  Article  Google Scholar 

  14. 14.

    Azevedo, D., Freitas, A.-C.M., Freitas, J.-M., Rodrigues, F.B.: Extreme value laws for dynamical systems with countable extremal sets. J. Stat. Phys. 167(5), 12441261 (2017)

    MathSciNet  Article  Google Scholar 

  15. 15.

    Keller, G.: Rare events, exponential hitting times and extremal indices via spectral perturbation. Dyn. Syst. 27, 11–27 (2012)

    MathSciNet  Article  Google Scholar 

  16. 16.

    Abadi, M., Freitas, A.C.M., Freitas, J.M.: Dynamical counterexamples regarding the Extremal Index and the mean of the limiting cluster size distribution. arXiv:1808.02970

  17. 17.

    Haydn, N., Vaienti, S.: The compound Poisson distribution and return times in dynamical systems. Probab. Theory Related Fields 144(3–4), 517–542 (2009)

    MathSciNet  Article  Google Scholar 

  18. 18.

    Aytac, H., Freitas, J.M., Vaienti, S.: Laws of rare events for deterministic and random dynamical systems. Trans. Am. Math. Soc. 367(11), 8229–8278 (2015)

    MathSciNet  Article  Google Scholar 

  19. 19.

    Leadbetter, M.R.: On extreme values in stationary sequences. Z. Wahrscheinlichkeitstheorie und Verw. Gebiete, 28, 289–303 (1973/1974)

  20. 20.

    O’Brien, G.L.: Extreme values for stationary and Markov sequences. Ann. Probab. 15, 281–291 (1987)

    MathSciNet  Article  Google Scholar 

  21. 21.

    Bradley, R.C.: Basic properties of strong mixing conditions. A survey and some open questions. Probab. Surv. 2, 107–144 (2005)

    MathSciNet  Article  Google Scholar 

  22. 22.

    Chernick, M.R., Hsing, T., McCormick, W.P.: Calculating the extremal index for a class of stationary sequences. Adv. Appl. Probab. 23(4), 835–850 (1981)

    MathSciNet  Article  Google Scholar 

  23. 23.

    Carvalho, M., Freitas, A.C.M., Milhazes Freitas, J., Holland, M., Nicol, M.: Extremal dichotomy for uniformly hyperbolic systems. Dyn. Syst. 30(4), 383–403 (2015)

    MathSciNet  Article  Google Scholar 

  24. 24.

    Holland, M., Nicol, M., Torok, A.: Extreme value theory for non-uniformly expanding dynamical systems. Trans. Am. Math. Soc. 364(2), 661–688 (2012)

    MathSciNet  Article  Google Scholar 

  25. 25.

    Azevedo, D., Freitas, A.C.M., Freitas, J.M., Rodrigues, F.B.: Clustering of extreme events created by multiple correlated maxima. Phys. D 315, 3348 (2016)

    MathSciNet  Article  Google Scholar 

  26. 26.

    Mantica, G., Perotti, L.: Extreme value laws for fractal intensity functions in dynamical systems: Minkowski analysis. J. Phys. A 49(37), 374001 (2016)

    MathSciNet  Article  Google Scholar 

  27. 27.

    Freitas, A.C.M., Freitas, J.M., Rodrigues, F.B., Soares, J.V.: Rare events for cantor target sets. arXiv:1903.07200 (2019)

  28. 28.

    Chazottes, J.-R., Coelho, Z., Collet, P.: Poisson processes for subsystems of nite type in symbolic dynamics. Stoch. Dyn. 9(3), 393422 (2009)

    Article  Google Scholar 

  29. 29.

    Keller, G., Liverani, C.: Rare events, escape rates and quasistationarity: some exact formulae. J. Stat. Phys. 135(3), 519–534 (2009)

    ADS  MathSciNet  Article  Google Scholar 

  30. 30.

    Faranda, D., Ghoudi, H., Guiraud, P., Vaienti, S.: Extreme value theory for synchronization of coupled map lattices. Nonlinearity 31(7), 332–3358 (2018)

    MathSciNet  Article  Google Scholar 

  31. 31.

    Beirlant, J., Goegebeur, Y., Segers, J., Teugels, J.L., De Waal, D., (Contributions by), Ferro, C. (Contributions by): Statistics of Extremes: Theory and Applications

  32. 32.

    Süveges, M.: Likelihood estimation of the extremal index. Extremes 10, 41–55 (2017)

    MathSciNet  Article  Google Scholar 

  33. 33.

    Ferreira, M.: Heuristic tools for the estimation of the extremal index: a comparison of methods. REVSTAT Stat. J. 16, 115–136 (2018)

    MathSciNet  MATH  Google Scholar 

  34. 34.

    Arnold, V.I., Avez, A.: Ergodic Problems of Classical Mechanics. The Mathematical Physics Monograph Series. W. A. Benjamin Inc., New York (1968)

    MATH  Google Scholar 

  35. 35.

    Freitas, A.C.M., Freitas, J.M., Vaienti, S.: Extreme value laws for nonstationary processes generated by sequential and random dynamical systems. Annales de l’Institut Henri Poincaré 53, 1341–1370 (2017)

    MathSciNet  Article  Google Scholar 

  36. 36.

    Conze, J.P., Raugi, A.: Limit theorems for sequential expanding dynamical systems on \([0, 1]\). Ergod. Theory Relat. Fields. Contemp. Math. 430, 89–121 (2007)

    MathSciNet  Article  Google Scholar 

  37. 37.

    Haydn, N., Nicol, M., Torok, A., Vaienti, S.: Almost sure invariance principle for sequential and non-stationary dynamical systems. Trans. Am. Math. Soc. 36, 5293–5316 (2017)

    MathSciNet  Article  Google Scholar 

  38. 38.

    Arnold, L.: Random Dynamics. Springer Monographs in Mathematics. Springer, Berlin (1998)

    Google Scholar 

  39. 39.

    Gonzalez-Tokman, C.: Multiplicative ergodic theorems for transfer operators: towards the identification and analysis of coherent structures in non-autonomous dynamical systems, Contributions of Mexican mathematicians abroad in pure and applied mathematics Guanajuato, Mexico American Mathematical Society. Contemp. Math. 709, 31–52 (2018)

    MathSciNet  Article  Google Scholar 

  40. 40.

    Freitas, A.C.M., Freitas, J.M., Magalhaes, M., Vaienti, S.: Point processes of non stationary sequences generated by sequential and random fynamical systems. arXiv:1904.05761 (2019)

  41. 41.

    Faranda, D., Freitas, J.-M., Lucarini, V., Turchetti, G., Vaienti, S.: Extreme value statistics for dynamical systems with noise. Nonlinearity 26, 2597–2622 (2013)

    ADS  MathSciNet  Article  Google Scholar 

  42. 42.

    Faranda, D., Ghoudi, H., Guiraud, P., Vaienti, S.: Extreme value theory for synchronization of coupled map lattices. Nonlinearity 31, 3326–3358 (2018)

    MathSciNet  Article  Google Scholar 

  43. 43.

    Faranda, D., Vaienti, S.: Extreme value laws for dynamical systems under observational noise. Phys. D 280, 86–94 (2014)

    MathSciNet  Article  Google Scholar 

  44. 44.

    Moreira Freitas, A.C., Milhazes Freitas, J., Todd, M.: The compound Poisson limit ruling periodic extreme behaviour of non-uniformly hyperbolic dynamics. Commun. Math. Phys. 321(2), 483527 (2013)

    MathSciNet  MATH  Google Scholar 

  45. 45.

    Haydn, N., Vaienti, S.: Limiting entry times distribution for arbitrary sets. arXiv:1904.08733 (2019)

  46. 46.

    Coelho, Z., Collet, P.: Asymptotic limit law for the close approach of two trajectories in expanding maps of the circle. Prob. Theory Relat. Fields 99, 237–250 (1994)

    MathSciNet  Article  Google Scholar 

  47. 47.

    Gora, P., Boyarsky, A.: Laws of Chaos. Birkhäuser, Basel (1997)

    MATH  Google Scholar 

  48. 48.

    Gallo, S., Haydn, N., Vaienti, S.: in preparation

Download references


We would like to thank Jorge Freitas, Michele Gianfelice, Nicolai Haydn and Giorgio Mantica for several discussions related to different parts of this work. PY was supported by ERC Grant No. 338965-A2C2. The authors thank the anonymous referees who suggested to formulate Proposition 1 in a better form and indicated us the appropriate reference [14].

Author information



Corresponding author

Correspondence to S. Vaienti.

Additional information

Publisher's Note

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

Communicated by Valerio Lucarni.



As mentioned in Sect. 3, results proving the existence of extreme value laws for non-stationary sequential systems are lacking. Nevertheless, the methods of estimation described in the paper can still be used to evaluate a quantity \(\theta \) that would correspond to the extremal index in the eventuality that an extreme value law holds for this kind of systems, which by the way was the case in [35]. We could also alternatively define the extremal index by computing the statistics of the number of visit and evaluating the expectation in Eq. (46). We choose this second option and consider the motion given by the concatenation

$$\begin{aligned} f_{\bar{\xi }}^n(x)=f_{\xi _n} \circ \cdots \circ f_{\xi _1} (x), \end{aligned}$$

where the probability law of the \(\xi _i\) changes over time. In particular, we consider the 10 maps

$$\begin{aligned} f_i(x)=2x+b_i \mod \ 1, \end{aligned}$$

where \(b_i\) is the ith component of a vector \({\bar{b}}\) of size 10, with entries equally spaced between 0 and 1 / 2. We consider sequences of time intervals \([\tau k+1,(k+1)\tau ]\), with \(\tau =10\), for \(k=0,1,2,\dots \) in which the weights associated to \(\xi _i\) are equal to \(p_i^k\). For every \(\tau \) iterations, the weight associated to each \(\xi _i\) changes randomly, with the only constraint that they sum to 1. \({\hat{\theta }}_5\) is computed considering a trajectory of \(5\times 10^7\) points, and a threshold u corresponding to the 0.995-quantile of the observable distribution. Using the same trajectory, we computed the empirical distribution of the number of visits in a ball centered at the origin and of radius \(r=e^{-u}\) in intervals of time of length \(\lfloor 2tr \rfloor \), with \(t=50\). This is indeed t times the Lebesgue measure (which is the invariant measure associated to all the maps \(f_i\)) of a ball of radius r centered in the origin. Choosing the same trajectory for the computation of \(\theta \) and for the statistics of visits is of crucial importance here, because different probability laws imply large variations of \(\theta \) depending on the trajectory considered. In Fig. 12, we observe again a perfect agreement between the empirical distribution of the number of visits and the Polyà-Aeppli distribution of parameters \(t=50\) and \({\hat{\theta }}_5\), which is equal to 0.92 for the trajectory presented in the Fig. 12. Although different trajectories give variations for \(\theta \), this agreement is stable against different trajectories and different values of t and \(\tau \), suggesting that a geometric Poisson distribution for the number of visits given by \(\theta \) is a universal feature, even for non-stationary scenarii.

Fig. 12

Comparison between the empirical distribution \(\mu (N_n(50)=k)\) and the Polyà-Aeppli distribution of parameters \(t=50\) and \({\hat{\theta }}_5\approx 0.92\), computed with a trajectory of length \(5\times 10^7\) of the sequential system described in the text. The procedure used to compute the empirical distribution is described in the text

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Caby, T., Faranda, D., Vaienti, S. et al. On the Computation of the Extremal Index for Time Series. J Stat Phys 179, 1666–1697 (2020). https://doi.org/10.1007/s10955-019-02423-z

Download citation


  • Extreme value theory
  • Extremal index
  • Random dynamical systems
  • Poisson statistics