Lecture notes for 12.006J/18.353J/2.

050J, Nonlinear Dynamics: Chaos

D. H. Rothman, MIT
September 8, 2022

Contents
1 Introduction 1
1.1 Who am I? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 What is this course? . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.1 Nonlinear systems . . . . . . . . . . . . . . . . . . . . . 3
1.2.2 Dissipative systems . . . . . . . . . . . . . . . . . . . . 5
1.3 Course goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.4 Administrative details . . . . . . . . . . . . . . . . . . . . . . 6
1.5 Syllabus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.6 Course material . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.7 Students . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.8 Handouts and further reading . . . . . . . . . . . . . . . . . . 9

1 Introduction

This course is
12.006J/18.353J/2.050J Nonlinear Dynamics: Chaos

Prof. Daniel Rothman

TA: Constantin Arnscheidt

1.1 Who am I?

A professor of geophysics. My current interests focus on

• How the carbon cycle works, including its relation to abrupt climate

1
 change (via “tipping points”)
• Dynamical mechanisms underlying the coevolution of life and the envi-
ronment, and catastrophes such as mass extinctions.
• Complex systems in general.

I created this course long ago, but have not taught it since 2006. I’m delighted
to teach it again, and am exploring many ways of reinvigorating its content.

1.2 What is this course?

An undergraduate introduction to the theory and phenomenology of dissipa-

tive nonlinear dynamical systems.

Let’s parse that out:

• Dynamical system: anything (physical, chemical, biological) that evolves

with time. Here we consider systems parameterized by only a few vari-
ables (e.g., position and momentum. . .).
• Dissipative: system has some friction (e.g., viscosity). As t → ∞, sys-
tem approaches an attractor which does not depend (usually) on initial
conditions (e.g., rest state of, say, a pendulum; terminal velocity of a
falling object).
Almost all systems in Nature are dissipative. Counter-examples: So-
lar system dynamics are conservative (“Hamiltonian”). Also molecular
dynamics of an ideal gas (elastic collisions).
• Nonlinear. Nonlinear science is literally the study of systems (theoretical
or real) that are not linear.

Let’s look at the last two points more closely.

2
1.2.1 Nonlinear systems

The Polish-American scientist Stanislas Ulam once famously remarked that

defining nonlinear science as above is “like defining the bulk of zoology by
calling it the study of non-elephant animals [1].”

Consider, for example, the usual assumptions that

• stress ∝ strain;
• flux ∝ force; or
• current ∝ voltage

We often think that, e.g. pushing something twice as hard yields twice the
velocity.

But consider these examples:

• Push a block with a weak force. If the force is too weak, the block
sticks to surface. If the force exceeds a threshold, the block slips.
terminal
velocity
simple friction
pinned

force
This is the basis of “stick-slip” models for the dynamics of, e.g., earth-
quake faults. Or violin bows on a string.
• Make a pile of sand by adding one grain at a time. Most of the time
the grains are at rest. But occasionally there are avalanches. Most are
small, but some are quite big.
• Heat a fluid from below. If the thermal gradient is weak, heat diffuses
upward but the fluid does not move
Stronger thermal gradients: convection (fluid motion) carries warm, less
dense fluid upward, and cold, more dense fluid downward.

3
 If we place a probe somewhere in the fluid, the joint possibility of an
upward or downward velocity x leads to a picture that looks like this
(where r measures the thermal gradient):

© Informa UK Limited. All rights reserved.

This content is excluded from our Creative
Commons license. For more information,
see https://ocw.mit.edu/help/faq-fair-use.

Strogatz [2], Fig. 3.4.2

Here, when r is less than a critical value rc , the zero-velocity state is sta-
ble; above rc it is unstable and the typical velocity grows like (r − rc )1/2 .
We’ll look at such situations more generally in the next lecture.
But for now, we note that this is an example of a general characteristic of
nonlinear systems: small changes in parameters can lead to qualitatively
different behavior.
• Fluid dynamics more generally:
u(x)
u(x+dx)

The fluid velocity ~u(~x) changes in part because the fluid flows; i.e.,

~u(~x) → ~u(~x + d~x)

But ~u also governs how fast this change occurs. Therefore d~u/dt depends
nonlinearly on u, and includes a change like

(~u · r)~u ∼ u2

i.e., a particle moves at velocity ~u along a velocity gradient r~u to a place

where the velocity is different.
When this nonlinearity is weak (because u is small), flow is smooth and
laminar. When it is stronger, flow becomes turbulent.
• Climate. The climate “system’ involves fluids, convection, and perhaps
the most nonlinear system of all: life. So it is unquestionably nonlinear.

4
 But is it unstable? There are many known—and unknown—positive
feedbacks. And plenty of examples of abrupt climate change in the past.
Understanding the methods and concepts in this course is a necessary
first step toward determining whether climate “tipping points” exist.
• Social systems (e.g., political or economic). The same general remarks
hold.

1.2.2 Dissipative systems

In dissipative systems, energy input to a system is eventually balanced by

friction. The resulting “steady state,” a kind of “attractor” is sometimes
quite simple.

But we shall see that the combination of nonlinearity and dissipation can lead
to strange attractors, on which there is sensitivity to initial conditions.

The overall idea is that small changes in initial conditions lead to large
changes in the long term.

The classic example is the weather, explained first, in 1963, by MIT professor
Edward Lorenz.

Lorenz’s discovery was eventually termed the “butterfly effect”: a butterfly

that flaps its wings in, say, Brazil, can affect, at a later time (in principle)
the weather in New York.

This deterministic unpredictability we call chaos.

The idea has now entered the cultural mainstream. But this course shows
that the notion of chaos in dissipative systems is really quite non-intuitive:
we’ll understrand why it is possible for a system to be attracted to a statisti-

5
cally steady state—its attractor—regardless of initial conditions, while being
sensitive to to initial conditions on the attractor.

1.3 Course goals

We teach:

• Elementary aspects of the theory of nonlinear dynamics and chaos.

• Phenomenology (e.g., aspects of fluid turbulence, scaling laws, experi-
mental phenomena).
• Computer experimentation.
• Analysis of experimental data.

Our computational experiments are exploratory. Rather than focusing on

the computation of a specific quantity (e.g., some integral), we construct
simple models and compute their evolution to determine qualitative aspects
of dynamics.

These qualitative dynamics are often quite general and apply to a wide array
of problems in science and engineering.

Thus a major goal of the course is for students to learn why such wide-ranging
applicability exists in problems that may superficially appear quite different.

1.4 Administrative details

• TA: Constantin Arnscheidt.

• All course materials will be available on Canvas.
• Prereqs: Must know o.d.e.’s (18.03). Some linear algebra (e.g., eigenval-
ues and eigenvectors).
• Problem sets: Some analytic, some require numerical simulation.

6
 • We usually provide Matlab and Python codes, but modifications are
often necessary. Only rudimentary coding skills are required, and they
can be learned in this course.
• The objective of the numerical experiments is to impart a sense of dis-
covery in the exploration of dynamical systems and comparison with
theoretical predictions (in the spirit of Lorenz and Feigenbaum).
• Students with no experience in numerical computation may wish to con-
sult the TA for assistance.
• Requirements
– No exams.
– There will be about 9 problem sets.
– A final project: either a review of a topic in the literature, your own
attempt to apply or extend what you’ve learned to a problem that
interests you, or a combination of both. A written report will be
due in the last class (Dec. 13), at which time students will also give
brief presentations. You should choose your topic by Nov. 8, and
submit it for approval. Further guidelines will be given.
– Problems sets count for about 80% of the grade and the final project
about 20%.
• The first pset is due next Thursday. It is easy, but we want to be sure
that everyone is comfortable with the (modest) numerical computation.

1.5 Syllabus

1. Elementary nonlinear dynamics and its empirical analysis

(a) Flows and bifurcations in 1D.
(b) Oscillators, phase space, stability, conservation/contraction of areas
in phase space,
(c) Limit cycles, Hopf bifurcations, excitability
(d) Power spectra, autocorrelations, Poincaré sections, maps (e.g., xk+1 =
f (xk )). Phase space reconstruction.

7
 (e) Fluid dynamics and Rayleigh-Bénard (thermal) convection.
2. Deterministic chaos in low-dimensional systems.
(a) Strange attractors. Sensitivity to initial conditions. Lorenz attrac-
tor, Hénon attractor, etc.
(b) Quantifying chaos (“measuring the strangeness of strange attrac-
tors”).
i. Fractal dimension (how many “degrees of freedom”?; dynamics
becomes geometry).
ii. Lyaponov exponents (How sensitive to initial conditions?).
(c) Transitions to chaos, scaling and universality.
i. Period doubling. Oscillations of successively longer periods 2n
occur when the control parameter has value µn , with

(µ∞ − µn ) ∝ δ −n

where the system is chaotoc at µ∞ and δ = 4.669... is universal.

ii. Intermittency.
iii. Quasiperiodicity.
The second half of the course stresses the relations between pde’s, ode’s,
and discrete mapping. We shall see that much of the complexity of non-
linear pde’s is contained in simple 1-D maps!
3. Remaining time (if any): physical models of scale invariance (fractals)
in nature.

1.6 Course material

Strongly recommended: Strogatz [2]. Beautifully written. Only a few lectures

will follow it in detail, but nearly all the subjects we cover are addressed in
the book. PDF is downloadable from the MIT Library.

All of our lectures will be accompanied by detailed lecture notes that will be
posted to the Canvas website before or shortly after the lecture.

8
1.7 Students

Who should take this course?

• Scientists and engineers who desire to learn how and why nonlinearity
manifests itself in natural systems.
• Mathematicians who seek a scientific, physical, and phenomenological
inspiration for the further development of mathematical theory.
• Anyone interested in how one “does science” with computers.

1.8 Handouts and further reading

You should have

• Guidelines
• Syllabus
• Problem Set 1. Much easier than usual, just to get started, and an op-
portunity to resolve now any technical problems with Python or Matlab.

Also: For general background, read Chapter 1 of Strogatz.

References
1. Campbell, D. K. Nonlinear science. Los Alamos Science 15, 218–262
(1987).
2. Strogatz, S. Nonlinear Dynamics and Chaos: With Applications to Physics,
Biology, Chemistry, and Engineering (CRC Press, 2018).

9
MIT OpenCourseWare
https://ocw.mit.edu

12.006J/18.353J/2.050J Nonlinear Dynamics: Chaos

Fall 2022

For information about citing these materials or our Terms of Use, visit: https://ocw.mit.edu/terms.