1: \documentstyle[aps,prl,epsfig,amssymb,balanced,times]{revtex}
2: \begin{document}
3: \title{Extensive Chaos in the Nikolaevskii Model}
4: \author{Hao-wen Xi$^1$, Ra\'ul Toral$^2$, J. D. Gunton$^3$ and
5: Michael I. Tribelsky$^4$}
6: \address{$^1$Department of Physics and Astronomy,
7: Bowling Green State University, Bowling Green OH 43403\\
8: $^2$Instituto Mediterr\'aneo de Estudios Avanzados (IMEDEA),
9: CSIC-UIB, E-07071 Palma de Mallorca, Spain\\ $^3$Department of
10: Physics, Lehigh University, Bethlehem, PA 18015\\ $^4$Department of
11: Applied Physics, Faculty of Engineering, Fukui University, Bunkyo
12: 3-9-1, Fukui 910-8507, Japan\\}
13: \maketitle
14: \pacs{PACS numbers: 05.45.+b, 47.20.Ky, 47.27.Eq, 47.52.+j}
15: \begin{abstract}
16: We carry out a systematic study of a novel type of chaos at onset
17: (``soft-mode turbulence") based on numerical integration of the
18: simplest one dimensional model. The chaos is characterized by a
19: smooth interplay of different spatial scales, with defect
20: generation being unimportant. The Lyapunov exponents are calculated
21: for several system sizes for fixed values of the control parameter
22: $\epsilon$. The Lyapunov dimension and the Kolmogorov-Sinai entropy
23: are calculated and both shown to exhibit extensive and
24: microextensive scaling. The distribution functional is shown to
25: satisfy Gaussian statistics at small wavenumbers and small
26: frequency.\\
27: \end{abstract}
28:
29: \begin{twocolumns}
30: Spatiotemporal chaos (STC) is a subject of considerable
31: experimental and theoretical importance and occurs in a wide
32: variety of driven, dissipative systems\cite{CH93,PM90,HG98}. Such
33: chaotic behavior in spatially extended systems is extremely
34: difficult to characterize quantitatively, as the dynamics involves
35: a large number of degrees of freedom. The most common and useful
36: tool for the characterization of chaos is given by the Lyapunov
37: exponents $\{\lambda_i\}$. Knowledge of this Lyapunov spectrum
38: permits one to estimate the number of effective degrees of freedom
39: of the system (i.e., the dimension of the attractor), using for
40: example the Kaplan-Yorke\cite{KY79} formula for the Lyapunov
41: dimension $D(L)$, where $L$ is the linear system size. It also
42: permits one to test the important concept of extensivity of chaos,
43: defined as the case in which $\lim_{L\rightarrow\infty} D(L)\sim
44: L^{d}$, where $d$ is the spatial dimension of the
45: system\cite{CH93,HS89}. An interpretation of extensive chaos is
46: that the whole system can then be thought of in some sense as the
47: union of almost independent subsystems. This was originally
48: proposed by Ruelle\cite{DR82}, who argued that widely separated
49: subsystems of a turbulent system should be weakly correlated, so
50: that the spectrum of Lyapunov exponents would be the union of
51: exponents associated with each of the subsystems. The question is
52: closely related to the fundamental problem of ergodicity of
53: nonequilibrium systems. If the chaos is extensive and each
54: subsystem evolves in time practically independently of the others,
55: then in a steady (non-transient) chaotic state the time average is
56: equivalent to the ensemble average and the system should be
57: ergodic. Much work has focused on attempting to characterize
58: spatiotemporal dynamics in these terms (see, e.g.,
59: \cite{VY89,LLPP93,CH95,WH99,COB99}). However, in spite of the
60: fundamental importance of the question practically all the results
61: are related just to a few discrete coupled map
62: lattices\cite{LPR92,HC93,PP98} and two continuous systems, namely
63: the complex Ginzburg-Landau (CGL) and Kuramoto-Sivashinsky (KS)
64: equations (see, e.g., refs.\cite{HC93,PM85,EG95}). This is partly
65: due to the computational complexity of the problem but primarily to
66: the lack of simple models exhibiting the requisite chaotic
67: behavior. It would therefore be of considerable interest to
68: characterize quantitatively other types of STC.
69:
70: Recently attention was drawn to the existence of a new wide class
71: of systems displaying such a behavior\cite{RBR95,KHH96,TT96,MIT97}.
72: Their properties are qualitatively different from those of the CGL
73: and KS models. In contrast to both these models the chaos is
74: associated with smooth, random long-wavelength modulations of a
75: short-wavelength pattern, with defect generation being unimportant.
76: The short-wavelength pattern arises due to a single supercritical
77: bifurcation of the Turing type such as occurs in
78: Rayleigh-B\'{e}nard convection. The long-wavelength modes belong to
79: a Goldstone branch of the spectrum originated in a broken
80: continuous symmetry. The symmetry makes the system degenerate to
81: the extent that instead of a single, unique spatially uniform
82: state, it has a {\it continuous family} of equivalent spatially
83: uniform states, which may be obtained from each other by the
84: symmetry transformation. This symmetry, which is additional to the
85: trivial groups of translations and rotations, can be one of many
86: different types. For this reason the STC in question is quite a
87: common phenomenon and occurs, for example, in electroconvection in
88: liquid crystals\cite{RBR95,KHH96}, in convection in a fluid with
89: stress-free boundary conditions\cite{SZ82,BB84,XLG97,N}, etc., see
90: ref.\cite{MIT97} for further discussion. The chaos observed in such
91: cases may be interpreted as a macroscopic dynamical analog of
92: second order phase transitions, where the order parameter is
93: related to the amplitudes of turbulent modes. Due to this analogy
94: it has been called {\it soft-mode turbulence} (SMT)\cite{KHH96}.
95: The simplest model exhibiting SMT was introduced by
96: Nikolaevskii\cite{VN89,BN93} to describe longitudinal seismic waves
97: in viscoelastic media. In what follows we exploit the simplicity of
98: this model to shed light on general features of this new type of
99: STC.
100:
101: We present a detailed systematic study of the Nikolaevskii model,
102: including calculation of the Lyapunov exponents, fractal dimension,
103: etc. We show that the Lyapunov dimension and Kolmogorov-Sinai
104: entropy are extensive quantities, which supports the validity of
105: the Ruelle's ideas for SMT. We also show that the power spectrum
106: can be described by Gaussian statistics for small wavenumbers and
107: frequencies, in agreement with a general argument of Hohenberg and
108: Shraiman\cite{HS89}.
109:
110: The model is defined by the following partial differential equation
111: for the real scalar field $v(x,t)$ (longitudinal mode of the
112: displacement velocity in the original formulation)\cite{VN89,BN93}:
113: \begin{equation}
114: \frac{\partial v}{\partial t} + \frac{\partial^{2}}{\partial x^{2}}
115: [\epsilon - (1 + \frac{\partial^{2}}{\partial x^{2}})^{2}]v +
116: v\frac{\partial v}{\partial x}= 0
117: \end{equation}
118: with $0\leq x \leq L$ and periodic boundary conditions. This model
119: has two control parameters, $\epsilon$ (the distance from onset)
120: and $L$, in contrast to, e.g., the KS model, where the only
121: non-trivial control parameter is the system size $L$. An essential
122: feature of the model is that even at small $\epsilon$ it cannot be
123: reduced to any $\epsilon$-free form\cite{MIT97,BM92}, which makes
124: the hierarchy of characteristic scales much more complex than those
125: in the KS and CGL models\cite{TV96,MT97}. Eq. (1) may be regarded
126: as a generalized Burgers equation and shares with it the same group
127: of symmetry, namely trivial symmetries under shifts of the
128: spatiotemporal coordinate system, and nontrivial invariance with
129: respect to the Galilean transformation $v(x,t)\rightarrow
130: v(x-v_{o}t,t)+v_{o}$, where $v_{o}$ is an arbitrary constant. The
131: Galilean invariance plays the role of the above specified
132: additional symmetry, generating for Eq. (1) the continuous family
133: of solutions $v=v_{o}$. The Nikolaevskii equation admits a
134: continuous set of spatially periodic, stationary solutions, whose
135: instability has been proved analytically\cite{TV96}. Computer
136: simulations\cite{TT96} of an equivalent version of this model for
137: an order parameter $u(x,t)\equiv\int v(x,t)dx$ showed that even at
138: extremely small $\epsilon=10^{-4}$, the system exhibits STC.
139: However, this simulation did not provide any quantitative results
140: about the STC. The only result of this kind is in ref.\cite{KM97},
141: in which just a single quantitative characteristic, namely the
142: dependence of the mean amplitude of chaotic patterns on the control
143: parameter, was studied. In this Letter we span the gap in our
144: knowledge of this new type of STC, providing a detailed
145: quantitative description of its most important properties based on
146: numerical integration of Eq. (1). The simulations were carried out
147: using the pseudo-spectral method combined with a fourth-order
148: predictor-corrector integrator, for several different values of $L$
149: and two values of $\epsilon$ (0.2 and 0.5, respectively)\cite{NID}.
150: The Lyapunov exponents were calculated by linearizing the equation
151: along the trajectory, performing a re-orthonormalization after a
152: few integration steps to prevent the largest Lyapunov exponent from
153: swamping all the others\cite{PC89}. A typical pattern $v(x,t)$ as a
154: function of space and time in the steady chaotic regime is shown in
155: Fig.~1. The time averaged power spectrum $\langle |\hat
156: v(k)|^{2}\rangle $ obtained over a time period of $T=10^{4}$ for
157: system size $L=78$ with $\epsilon=0.2$ is shown in Fig.~2. As can
158: be seen, the dominant modes occur in the vicinity of $k=\pm 1$.
159: Note, however, the smaller peak near zero wavenumber, which arises
160: from coupling between unstable short-wavelength modes centered
161: about $k=\pm 1$ and the slowly decaying modes from the Goldstone
162: branch of the spectrum centered at $k = 0$ (using terminology based
163: upon the linear stability analysis of the spatially uniform
164: solution). The power spectrum for $\hat u(k)=\hat v(k)/ik$ (shown
165: in the insert in Fig.~2) has the dominant peak near $k=0$, as was
166: originally demonstrated in\cite{TT96}. One would also expect on
167: general grounds\cite{HS89} each Fourier transform variable $\hat
168: v(k,\omega)$ to be governed by a Gaussian probability distribution
169: functional, $\exp(-D(|\hat {v}(k,\omega)|^{2})$, for small $k$ and
170: $\omega$. We have verified this for several values of $k$ and
171: $\omega$. In particular, for $\omega=0.6$ the Gaussian distribution
172: holds for $0 < k < 2$ (see Fig. 3).
173:
174: \begin{figure}
175: \epsfxsize = 0.45\textwidth
176: \epsfysize = 0.4\textwidth
177: \makebox{\epsfbox{figure1.eps}}
178: \caption{Typical space-time configuration for $v(x,t)$
179: in the steady chaotic regime is shown here. The configuration shown has
180: evolved from random initial condition within a system size $L=78$
181: and $\epsilon=0.5$.\label{fig1}}
182: \end{figure}
183:
184: \begin{figure}
185: \epsfxsize = 0.45\textwidth
186: \epsfysize = 0.35\textwidth
187: \makebox{\epsfbox{figure2.eps}}
188: \caption{Time average power spectra $\langle|\hat {v}(k)|^{2}\rangle$
189: and $\langle|\hat {u}(k)|^{2}\rangle$ in $k$-space are plotted. The
190: system size is $L=78$ and control parameter $\epsilon=0.2$. Note
191: that the power spectrum has symmetry with respect to $ k\rightarrow
192: -k$.\label{fig2}}
193: \end{figure}
194:
195: As has already been emphasized, the key question for any system
196: exhibiting STC is whether it can be represented as a union of
197: weakly correlated subsystems. If this is the case, then the
198: spectrum of Lyapunov exponents for the entire system should be {\it
199: intensive} in the sense that $\lambda_{i}$ is a function only of
200: the intensive index $i/V$, i.e., $\lambda_{i}=f(i/V)$, where V
201: stands for the volume of the system\cite{HG98}. The question is not
202: trivial for the type of STC considered here, because of the
203: importance of the long-wavelength modes and the divergence of the
204: two point correlation length $\xi_{2}$\cite{MT99} as $\epsilon
205: \rightarrow 0$\cite{MIT97}. To answer this question, a detailed
206: study of the Lyapunov spectrum for the Nikolaevskii model was
207: conducted. The results are shown in Fig. 4, where the number of
208: Lyapunov exponents greater than a particular value $\lambda_{i}$,
209: scaled by the system size $L$, is plotted versus $\lambda_{i}$ for
210: $\epsilon=0.5$. A similar curve is found at $\epsilon=0.2$, but the
211: maximum positive eigenvalue is now smaller (one expects this
212: eigenvalue to vanish as $\epsilon \rightarrow 0$.) The intensive
213: nature of the Lyapunov density is evident. In this case, one also
214: expects the fractal dimension $D(L)$ of an attractor to be
215: extensive for large enough $L$ (extensive chaos), as was first
216: shown by Manneville\cite{PM85} for chaotic solutions of the
217: Kuramoto-Sivashinsky equation. We have checked this for the
218: Nikolaevskii model using the Kaplan-Yorke formula\cite{KY79} for
219: the Lyapunov dimension,
220: \begin{equation}
221: D(L) = K + \sum_{i=1}^{K} \lambda_{i}/|\lambda_{k+1}|
222: \end{equation}
223: where the integer $K$ is the largest integer such that the sum of
224: the first $K$ Lyapunov exponents is nonnegative. We have also
225: calculated the Kolmogorov-Sinai entropy $H(L)$ from the definition
226: \begin{equation}
227: H(L) = \sum_{i=1}^{i_{+}} \lambda_{i}
228: \end{equation}
229: where the sum is over the positive Lyapunov exponents. The
230: Kolmogorov-Sinai entropy\cite{PC89} is a measure of the mean rate
231: of information production in a system, or the mean rate of growth
232: of uncertainty in a system subjected to small perturbations. We
233: find that for large enough $L$ both $D(L)$ and $H(L)$ are
234: extensive. Our results for the Lyapunov dimension $D(L)$ are shown
235: in Figure 5 for $\epsilon=0.5$. The same behavior is found at
236: $\epsilon=0.2$, but with a different slope (naturally the slope of
237: the straight line D(L) at $\epsilon=0.2$ is smaller than that at
238: $\epsilon=0.5$). In addition, we find that the upper index $i_{+}$
239: in (3), which corresponds to the smallest positive Lyapunov
240: exponent, is also proportional to $L$.
241:
242: \begin{figure}
243: \epsfxsize = 0.45\textwidth
244: \epsfysize = 0.35\textwidth
245: \makebox{\epsfbox{figure3.eps}}
246: \caption{Probability density distribution of
247: power spectrum $\hat {v}(k)$ for $k=1.5$ with system size $L=78$
248: and $\epsilon=0.5$. Notice that the validity of the Gaussian
249: distribution for $\hat v(k)$ (see the main text) implies an
250: exponential distribution $f(z)=D{\rm e}^{-D z}$ for the variable
251: $z\equiv |\hat {v}(k)|^{2}$. In this case we find the value
252: $D\approx 47$.\label{fig3}}
253: \end{figure}
254:
255: The important characteristic of STC is the dimension correlation
256: length $\xi_{\delta}$\cite{CH93,HG98}. This length is defined as
257: $\xi_{\delta} \equiv \delta^{-1/d}$, where $\delta \equiv
258: lim_{L\rightarrow \infty} D(L)/L^{d}$. It can be thought of as the
259: ``radius" of a volume that contains one degree of freedom, or, as
260: the linear size of the subsystem described above. The value of this
261: dimension correlation length for Eq.~(1) is $\xi_{\delta}= 3.0$ for
262: $\epsilon=0.5$ and $\xi_{\delta}= 3.3$ for $\epsilon=0.2$. In
263: contrast to $\xi_{\delta}$ the above-mentioned two point
264: correlation length $\xi_2$ governs the spatial decay of the
265: correlation function. In general these two lengths are
266: different\cite{HG98}. We found that $\xi_{2} \cong 4.9$ for the
267: Nikolaevskii model at $\epsilon = 0.5$ and $\xi_{2}
268: \cong 5.6$ at $\epsilon=0.2$.
269:
270: Finally, Tajima and Greenside\cite{TG99} have recently found for
271: the one dimensional Kuramoto-Sivashinsky model that $D(L)$ is also
272: "microextensive." Namely, they found that if one increases L by a
273: small amount $\delta L$, with $\delta L\ll \xi_{\delta}$ ($\delta
274: L=0.8$ in our simulation), one finds that $D(L)$ satisfies the same
275: linear relationship as that characterizing extensive chaos. We have
276: examined this for two different domains of $L$ for our model and
277: found that microextensivity holds for both $D(L)$ and $H(L)$. Our
278: results for microextensivity for $D(L)$ are shown in the insert in
279: Fig. 5.
280:
281: \begin{figure}
282: \epsfxsize = 0.45\textwidth
283: \epsfysize = 0.35\textwidth
284: \makebox{\epsfbox{figure4.eps}}
285: \caption{Here $N_{>i}$ is the number of Lyapunov exponents greater than
286: a particular value $\lambda_{i}$. We plot $N_{>i}/L$ (scaled by the
287: system size $L$) vs $\lambda_{i}$ in the case $\epsilon=0.5$.\label{fig4}}
288: \end{figure}
289:
290: In conclusion, our detailed study of this new type of STC based on
291: the numerical integration of the Nikolaevskii model shows that for
292: sufficiently large system size the chaos is both extensive and
293: microextensive. We also found that the system satisfies Gaussian
294: statistics at sufficiently small wavenumbers and frequencies. We
295: believe these results are quite general and reflect intrinsic
296: features of this type of STC, rather than specific peculiarities of
297: the model.
298:
299: There are several interesting questions to investigate in the limit
300: $\epsilon \rightarrow 0$, including the dependence of quantities
301: such as the correlation lengths and Lyapunov exponents on
302: $\epsilon$ as well as the possible scaling of the power spectrum.
303: This study is in progress and will be reported elsewhere.
304:
305: \begin{figure}
306: \epsfxsize = 0.45\textwidth
307: \epsfysize = 0.35\textwidth
308: \makebox{\epsfbox{figure5.eps}}
309: \caption{Lyapunov dimension $D(L)$ vs system size $L$ for $\epsilon=0.5$.
310: $\delta L=15.5$ for extensive, and $\delta L=0.8$ for
311: microextensive.\label{fig5}}
312: \end{figure}
313:
314: We would like to thank Henry Greenside for a helpful discussion.
315: This work was supported by a grant from NATO CRG.CRG972822, by NSF
316: Grant DMR9810409, projects PB94-1167 and PB97-0141-C02-01 (Spain)
317: and the Grant-in-Aid for Scientific Research (No. 11837006) from
318: the Ministry of Education, Science, Sports and Culture (Japan). We
319: also wish to acknowledge an allocation of time on the Pittsburgh
320: Supercomputer Center, where some of this work was carried out.
321:
322: \begin{references}
323: \bibitem{CH93} M. C. Cross and P. C. Hohenberg, Rev. Mod. Phys. {\bf 65}
324: 851 (1993).
325:
326: \bibitem{PM90} {\it Dissipative Structures and Weak Turbulence}, P. Manneville,
327: Academic Press (1990).
328:
329: \bibitem{HG98} H. Greenside, "Spatiotemporal Chaos in Large Systems:
330: The Scaling of Complexity with Size", in Proceedings of the CRM
331: Workshop, {\it Semi-Analytic Methods for the Navier Stokes
332: Equations}, ed. K. Coughlin (1998), and references therein.
333:
334: \bibitem{KY79} J. L. Kaplan, and J. A. Yorke, in {\it Lecture Notes in Mathematics},
335: 228, Springer-Verlag (1979).
336:
337: \bibitem{HS89} P. C. Hohenberg, and B. I. Shraiman, Physica D {\bf 37}, 109
338: (1989).
339:
340: \bibitem{DR82} D. Ruelle, Commun. Math. Phys. {\bf 87}, 287
341: (1982).
342:
343: \bibitem{VY89} V. Yakhot, Phys. Rev. A {\bf 24}, 642 (1989).
344:
345: \bibitem{LLPP93} V. S. L'vov, V. V. Lebedev, M. Paton, and I. Procaccia,
346: Nonlinearity {\bf 6}, 25 (1993).
347:
348: \bibitem{CH95} C. C. Chow, and T. Hwa, Physica D {\bf 84}, 494 (1995).
349:
350: \bibitem{WH99} R. W. Wittenberg, and P. Holmes, Chaos {\bf 9}, 452 (1999).
351:
352: \bibitem{COB99} R. Carretero-Gonzalez, S. Orstavik, J. Huke, D. S.
353: Broomhead, and J. Stark, Chaos {\bf 9}, 466 (1999).
354:
355: \bibitem{LPR92} R. Livi, A. Politi, and S. Ruffo, J. Phys. A {\bf 25}, 4813 (1992).
356:
357: \bibitem{HC93} H. Chate, Europhys. Lett. {\bf 21}, 419 (1993).
358:
359: \bibitem{PP98}A. Pikovsky, and A. Politi, Nonlinearity {\bf 11}, 1049 (1998).
360:
361: \bibitem{PM85} P. Manneville, "Liapounov exponents for the
362: Kuramoto-Sivashinsky Model", in {\it Macroscopic Modeling of
363: Turbulent Flows}, ed. O. Pironneau, Lecture Notes in Physics, No.
364: 230 Springer-Verlag (1985), p. 319.
365:
366: \bibitem{EG95} D. A. Egolf, and H. S. Greenside, Phys. Rev. Lett.
367: {\bf 74}, 1751 (1995).
368:
369: \bibitem{RBR95} H. Richter, N. Kloepper, A. Hertrich,
370: and A. Buka, Europhys. Lett. {\bf 30}, 37 (1995).
371:
372: \bibitem{KHH96} S. Kai, K. Hayashi, and Y. Hidaka, J. Phys.
373: Chem. {\bf 100}, 19007 (1996).
374:
375: \bibitem{TT96} M. I. Tribelsky, and K. Tsuboi, Phys. Rev.
376: Lett. {\bf 76}, 1631 (1996).
377:
378: \bibitem{MIT97} M. I. Tribelskii, Usp. Fiz. Nauk. {\bf 167},
379: 167 (1997) [Phys. Usp. {\bf 40}, 159 (1997)].
380:
381: \bibitem{SZ82} E.D. Siggia, and A. Zippelius, Phys. Rev. Lett. {\bf 47}, 835 (1981).
382:
383: \bibitem{BB84} F. H. Busse and E.W. Bolton, J. Fluid. Mech. {\bf 146}, 115
384: (1984).
385:
386: \bibitem{XLG97} H. W. Xi, X. J. Li and J. D. Gunton, Phys.
387: Rev. Lett. {\bf 78}, 1046 (1997).
388:
389: \bibitem{N} It is relevant to mention that the
390: consideration of convection with the stress-free boundary
391: conditions\cite{SZ82,BB84} was the very first analysis of pattern
392: formation in the degenerate systems. However, at that time the
393: dramatic destabilization of steady patterns obtained in this case
394: was regarded as a peculiarity of the particular problem,
395: rather than a generic attribute of the degeneracy.
396:
397: \bibitem{VN89} V.N. Nikolaevskii, in {\it Recent Advances in
398: Engineering Science}, ed. S. L. Koh and C. G. Speciale, Lecture
399: Notes in Engineering, No. 39, Springer-Verlag (1989), p. 210.
400:
401: \bibitem{BN93}I. A. Beresnev and V. N. Nikolaevskii, Physica D {\bf 66},
402: 1 (1993).
403:
404: \bibitem{BM92} B. A. Malomed, Phys. Rev. A {\bf 45}, 1009 (1992).
405:
406: \bibitem{TV96} M. I. Tribelsky and M. G. Velarde, Phys. Rev. E {\bf
407: 54}, 4973 (1996).
408:
409: \bibitem{MT97} M. I. Tribelsky, Inter. Jour. Bifur. Chaos
410: {\bf 7}, 997 (1997).
411:
412: \bibitem{KM97} I. L. Kliakhandler and B. A. Malomed, Phys. Lett.
413: A {\bf 231}, 191 (1997).
414:
415: \bibitem{NID} Although the original numerical study\cite{TT96}
416: was carried out in terms of the order parameter $u(x,t)$, we have
417: found that Eq.~(1) is more convenient for numerical study as it
418: exhibits greater numerical stability. The reason is that the
419: conservation law $d\bar v/dt=0$ holding for this equation (where
420: the bar denotes a spatial average) helps to stabilize the code. In
421: our simulations the value $\bar v=0$ was imposed by the initial,
422: random, conditions. For the space discretization we have found that
423: $\Delta x=0.31$ provides a good enough resolution.
424:
425: \bibitem{PC89}T. S. Parker and L. O. Chua, {\it Practical Numerical
426: Algorithms for Chaotic Systems} (Springer-Verlag, 1989).
427:
428: \bibitem{MT99}At small $\epsilon$ in this model the two point
429: correlation function ($\langle v(x,t)v(x+x',t)\rangle$, where the
430: $\langle\ldots\rangle$ denotes the time average) is generally
431: oscillatory with a spatial period of the oscillations close to
432: $2\pi$ and a slow decay of its amplitude over a scale $\xi_{2}$.
433: Thus, the model has two correlation length scales, with $\xi_2$
434: being the larger of the two.
435:
436: \bibitem{TG99} S. Tajima and H. S. Greenside, "Microextensive
437: Chaos of a Spatially Extended System", March Meeting of APS,
438: Atlanta, Georgia (1999).
439: \end{references}
440:
441: \end{twocolumns}
442:
443: \end{document}
444: