1: \documentclass{elsart}
2: \usepackage{epsfig}
3:
4: \newcommand{\be}{\begin{equation}}
5: \newcommand{\ee}{\end{equation}}
6: \newcommand{\bc}{\begin{center}}
7: \newcommand{\ec}{\end{center}}
8: \newcommand{\bi}{\begin{itemize}}
9: \newcommand{\ei}{\end{itemize}}
10: \newcommand{\ba}{\begin{eqnarray}}
11: \newcommand{\ea}{\end{eqnarray}}
12:
13:
14: \begin{document}
15:
16: \begin{frontmatter}
17:
18: \title{Boundary effects in extended dynamical systems}
19:
20: \author{V\'{\i}ctor M. Egu\'{\i}luz,
21: Emilio Hern\'andez-Garc\'{\i}a and Oreste Piro}
22: \address{Instituto Mediterr\'aneo de Estudios Avanzados, IMEDEA (CSIC-UIB)\\
23: E-07071 Palma de Mallorca, Spain}
24: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%ABSTRACT%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
25: \begin{abstract}
26: In the framework of spatially extended dynamical systems, we
27: present three examples in which the presence of walls lead to
28: dynamic behavior qualitatively different from the one obtained in
29: an infinite domain or under periodic boundary conditions. For a
30: nonlinear reaction-diffusion model we obtain boundary-induced
31: spatially chaotic configurations. Nontrivial average patterns
32: arising from boundaries are shown to appear in spatiotemporally
33: chaotic states of the Kuramoto-Sivashinsky model. Finally, walls
34: organize novel states in simulations of the complex
35: Ginzburg-Landau equation.
36: \end{abstract}
37: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
38: \date{7 November 1999}
39: \end{frontmatter}
40: \smallskip
41:
42: \noindent PACS: 05.45.-a, 47.54.+r \\
43: Keywords: Spatiotemporal chaos, pattern formation, boundary
44: conditions, Kuramoto-Sivashinsky, Ginzburg-Landau.
45:
46: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
47: \section{Introduction.}
48:
49: Most of our current knowledge of spatiotemporal chaos comes from
50: its analysis in the infinite size limit or from simulations in
51: finite domains with periodic boundary conditions\cite{CrossHoh}.
52: However, different geometries or boundary conditions may lead to
53: substantially different dynamical behavior. We will exemplify this
54: assertion by showing results from three different extended
55: dynamical systems in which the dynamics is strongly influenced by
56: the boundaries.
57:
58: \begin{figure}[t]
59: \epsfig{figure=froz.eps, width=1.\textwidth}
60: \put(-190,130){(b)}
61: \put(-400,130){(a)}
62: \caption{
63: {\small (a) Top: two-dimensional configuration of our reaction-diffusion
64: equation starting
65: from random initial conditions. Bottom: Amplitude at the
66: center of the domain.
67: (b) Stroboscopic Poincar\'e map of the phase space of an
68: approximation to our model equation.
69: }}
70: \end{figure}
71:
72: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
73: \section{Frozen spatial chaos induced by boundaries}
74: \label{frozen}
75:
76: We first show how the presence of nontrivial boundaries can induce
77: the appearance of {\sl spatial chaos} in a system for which no
78: chaotic behavior is found neither in the infinite size limit nor
79: with purely periodic boundary conditions\cite{Eguiluz99b}. The
80: model we consider is a nonlinear diffusion equation of the
81: Fisher-Kolmogorov type: $\partial_t A = \nabla^2 A +A-A^3$. The
82: real quantity $A=A(x,y,t)$ is a twodimensionally extended field.
83: When solved in doubly periodic integration domains, regions in
84: which $A\approx
85: \pm 1$ form, grow, and compete until one of the two phases
86: takes over the whole system. When solved in regions such that
87: Dirichlet (that is $A=0$) conditions are applied in lateral
88: boundaries which are not straight but undulating (see Fig. 1) the
89: result is different: Frozen states in which the $A=+1$ and $A=-1$
90: phases alternate in space become stable and attract most of the
91: initial conditions. The alternation of the two phases is random
92: and produces static but spatially chaotic configurations. The
93: justification of the 'chaotic' adjective can be done with
94: different dynamical systems tools. For example Fig. 1b shows a
95: Poincar\'{e} map of some of the spatial configurations obtained
96: from an approximation to our model equation. KAM tori and other
97: fractal structures are evident, in direct analogy with the
98: classical picture of Hamiltonian systems with chaotic time
99: trajectories. Theoretical arguments can be developed to show that
100: the effect of the spatially undulated boundaries on the spatial
101: pattern is similar to time-periodic parametric forcing in common
102: temporal dynamical systems, from which the above chaotic
103: phenomenology can be understood. Further details are given in
104: \cite{Eguiluz99b}
105:
106: \begin{figure}[t]
107: \epsfig{figure=aver.ps, height=1.\textwidth,angle=90,clip=true}
108: \caption{
109: {\small (a) A
110: characteristic configuration of the one-dimensional
111: Kuramoto-Sivashinsky equation with stress-free boundary conditions.
112: (b) The time averaged field.
113: }}
114: \end{figure}
115: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
116: \section{Average patterns of spatiotemporal chaos.}
117: \label{averages}
118:
119: Chaotic pattern dynamics in many experimental systems
120: \cite{Gluckman95,Ahlers} show structured
121: time averages. In this second Section, we suggest that simple
122: universal boundary effects underlie this phenomenon and exemplify
123: them with the Kuramoto-Sivashinsky equation in a finite domain.
124: Figure~2b shows a structured average pattern for the
125: Kuramoto-Sivashinsky equation in one dimension
126: ($\dot\varphi=-\partial^2_x
127: \varphi-\partial^4_x \varphi + (\partial_x \varphi)^2$) with
128: stress-free boundary conditions (null first and third spatial
129: derivatives at the boundaries). In contrast, the strong
130: fluctuations of the instantaneous field are also shown in Fig.~2a
131: for comparison. As in the experiments, the average pattern
132: recovers the symmetries which are respected by both the equation
133: and the boundary conditions (in this case left-right symmetry)
134: locally broken in the chaotic field. The amplitude is strongest at
135: the boundaries and decays through the center of the average
136: pattern. The strength of the oscillations in the average pattern
137: follows a $L^{-1/2}$ dependence on system size. Plateaus in the
138: average-pattern wavenumber as a function of the system size are
139: observed\cite{Eguiluz99}. Most of these observations are also
140: found in experimental systems\cite{Gluckman95,Ahlers} for which
141: the Kuramoto-Sivashinsky equation does not apply, thus indicating
142: its generic, mainly geometrical, origin: what is relevant for
143: these phenomena to appear is the occurrence of strong enough
144: chaotic fluctuations in the presence of non-trivial boundaries.
145:
146: \begin{figure}[t]
147: \begin{center}
148: \epsfig{figure=cgl.ps, width=.8\textwidth,angle=0}
149: \caption{{\small
150: Frozen structures for the complex Ginzburg-Landau equation
151: in the Benjamin-Feir stable regime under null Dirichlet boundary conditions in
152: different domains. Top raw: modulus; bottom raw: phase
153: }}
154: \end{center}
155: \end{figure}
156: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
157: \section{The complex Ginzburg-Landau equation in bounded domains.}
158: \label{scgle}
159:
160: In our third example, the effect of a finite geometry on the
161: two-dimensional complex Ginzburg-Landau equation (in the
162: Benjamin-Feir stable regime\cite{CrossHoh}) is addressed. Boundary
163: conditions induce the formation of novel states. For example
164: target like-solutions (Fig.~3a-b) appear as robust solutions under
165: Dirichlet boundary conditions, whereas they are not observed under
166: periodic boudary conditions. Dirichlet boundary conditions play a
167: double r\^ole as sources (or sinks) of defects and as emitters of
168: plane waves. The interplay between these two properties of the
169: boundaries gives rise to interesting
170: behavior\cite{Eguiluz99a,Eguiluz99c}. In a square, walls emit
171: waves that develop shock lines when they cross. Spiral defects
172: form chains anchored by these shock lines. In circular domains,
173: however, the emission is definitively dominated by the internal
174: spiral defects. In a stadium geometry (Figs.~3c-d), typically a
175: chain of defects links the centers of the circular regions.
176: Synchronization of boundary wave emission is also
177: found\cite{Eguiluz99a,Eguiluz99c}. Most of these phenomena can be
178: understood from the emission properties of Dirichlet
179: walls\cite{Eguiluz99c}.
180:
181: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
182: \section{Conclusions.}
183:
184: We have shown three examples where the boundary conditions play a
185: key r\^ole in extended dynamical systems. Further work is needed
186: to clarify the degree of universality of these results and to find
187: general properties of different kinds of boundaries.
188:
189: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
190: Financial support from DGICyT projects PB94-1167 and
191: PB97-0141-C02-01 is greatly acknowledged.
192:
193: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
194: \begin{thebibliography}{10}
195:
196: \bibitem{CrossHoh}
197: M.C. Cross, P.C. Hohenberg, Rev. Mod. Phys. {\bf 65}, 851 (1993).
198:
199: \bibitem{Eguiluz99b}
200: {V.M. Egu\'{\i}luz, E. Hern\'{a}ndez-Garc\'{\i}a, O. Piro, S.
201: Balle}, Phys. Rev. E {\bf 60} 6571 (1999). e-print: {\tt
202: chao-dyn/9906014}.
203:
204: \bibitem{Gluckman95}
205: B.J. Gluckman, C.B. Arnold and J.P. Gollub, Phys. Rev. E {\bf 51},
206: 1128 (1995).
207:
208: \bibitem{Ahlers}
209: {L. Ning, Y. Hu, R.E. Ecke and G. Ahlers}, Phys. Rev. Lett. {\bf
210: 71}, 2216 (1993).
211:
212: \bibitem{Eguiluz99}
213: V.M. Egu\'{\i}luz, P.Alstr{\o}m, E. Hern\'andez-Garc\'{\i}a, O.
214: Piro, Phys. Rev. E. {\bf 59}, 2822 (1999). e-print: {\tt
215: chao-dyn/9805003}.
216:
217: \bibitem{Eguiluz99a}
218: V.M. Egu\'{\i}luz, E. Hern\'andez-Garc\'{\i}a, O. Piro, Int. J.
219: Bif. Chaos {\bf 9}, 2209 (1999). e-print: {\tt chao-dyn/9812010}.
220:
221: \bibitem{Eguiluz99c}
222: V.M. Egu\'{\i}luz, E. Hern\'andez-Garc\'{\i}a, O. Piro,
223: in preparation.
224:
225:
226: \end{thebibliography}
227: \newpage
228:
229: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%FIGURES%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
230:
231: \end{document}
232: