1

Crackling noise

James P. Sethna*, Karin A. Dahmen†, and Christopher R. Myers¶

* Laboratory of Atomic and Solid State Physics, Clark Hall, Cornell University, Ithaca,
NY 14853-2501, [email protected]. † Department of Physics, 1110 West Green
Street, University of Illinois at Urbana-Champaign, IL 61801-3080,

[email protected]. ¶ Cornell Theory Center, Frank H.T. Rhodes Hall, Cornell

University, Ithaca, NY 14853-3801, [email protected].

Crackling noise arises when a system responds to changing external conditions

through discrete, impulsive events spanning a broad range of sizes. A wide variety

of physical systems exhibiting crackling noise have been studied, from earthquakes

on faults to paper crumpling. Because these systems exhibit regular behavior over

many decades of sizes, their behavior is likely independent of microscopic and

macroscopic details, and progress can be made by the use of very simple models.

The fact that simple models and real systems can share the same behavior on a
wide range of scales is called universality. We illustrate these ideas using results for

our model of crackling noise in magnets, explaining the use of the renormalization
group and scaling collapses. This field is still developing: we describe a number of

continuing challenges.
 2

Crackling noise: a new realm for science

In the past decade or so, science has broadened its purview to include a new range of
phenomena. Based on advances in the 1970's on second-order phase transitions0.5-4 and

stochastic theories of turbulence5 and in the 1980's on disordered systems,6-8 we now

claim that we should be able to explain how and why things crackle.

Figure 1: The Earth Crackles. (a) Time history of radiated energy from earthquakes throughout all of
1995.10-12 The earth responds to the slow strains imposed by continental drift through a series of
earthquakes: impulsive events well separated in space and time. This time series, when sped up, sounds
remarkably like the crackling noise of paper, magnets, and Rice Krispies© (hear it in Ref. 12). (b)
Histogram of number of earthquakes in 1995 as function of their magnitude (or, alternatively, their energy
release). Earthquakes come in a wide range of sizes, from unnoticeable trembles to catastrophic events.
The smaller earthquakes are much more common: the number of events of a given size forms a power
law9 called the Gutenberg-Richter law. (Earthquake magnitude scales with the logarithm of the strength
of the earthquake, e.g. its radiated energy. On a log-log plot of number vs. radiated energy, a power law is
a straight line, as we observe in the plotted histogram.) One would hope that such a simple law should
have an elegant explanation.

Many systems crackle; when pushed slowly, they respond with discrete events
of a variety of sizes. The earth responds9 with violent and intermittent earthquakes as

two tectonic plates rub past one another (see figure 1). A piece of paper14 (or a candy

wrapper at the movies15,16) emits intermittent, sharp noises as it is slowly crumpled or

rumpled. (Try it: preferably not with this page.) A magnetic material in a changing
external field magnetizes in a series of jumps.13,17 These individual events span many
orders of magnitude in size – indeed, the distribution of sizes forms a power law with no

characteristic size scale. In the past few years, scientists have been making rapid
 3

progress in developing models and theories for understanding this sort of scale-invariant
behavior in driven, nonlinear, dynamical systems.

Interest in these sorts of phenomena goes back several decades. The work of

Gutenberg and Richter9 in the 1940's and 1950's established the well-known frequency-

magnitude relationship for earthquakes that bears their names (figure 1). A variety of

many-degree-of-freedom dynamical models,18-30 with and without disorder, have been

introduced in the years since to investigate the nature of slip complexity in earthquakes.

More recent impetus for work in this field came from the study of the depinning

transition in sliding charge-density wave (CDW) conductors in the 1980's and early

1990's.31-37 Interpretation of the CDW depinning transition as a dynamic critical

phenomenon sprung from Fisher's early work,31,32 and several theoretical and numerical

studies followed. This activity culminated in the RG solution by Narayan and Fisher34

and the numerical studies by Middleton35 and Myers,36 which combined to provide a

clear picture of depinning in CDWs and open the doors to the study of other disordered,

nonequilibrium systems.

Bak, Tang, and Wiesenfeld inspired much of the succeeding work on crackling

noise.38,39 They introduced the connection between dynamical critical phenomena and
crackling noise, and they emphasized how systems may naturally end up at the critical

point through a process of self-organized criticality. (Their original model was that of
avalanches in growing sandpiles. Sand has long been used as an example of crackling

noise.40,41 However, it turns out that real sandpiles don't crackle at the longest
scales.42,43)

Researchers have studied many systems that crackle. Simple models have been
developed to study bubbles rearranging in foams as they are sheared,44 biological

extinctions45 (where the models are controversial:46,47 of course we personally believe

that the asteroid did in the dinosaurs), fluids invading porous materials and other
 4

problems involving invading fronts48-53 (where the model we describe was

invented48,49), the dynamics of superconductors54-54.4 and superfluids,55,56 sound emitted

during martensitic phase transitions,57 fluctuations in the stock market,58,59 solar flares,60

cascading failures in power grids,61,62 failures in systems designed for optimal

performance,63-65 group decision making,65.1 and fracture in disordered materials.65.2-65.7

These models are driven systems with many degrees of freedom, which respond to the

driving in a series of discrete avalanches spanning a broad range of scales – what we are
calling crackling noise.

There has been healthy skepticism by some established professionals in these

fields to the sometimes grandiose claims by newcomers proselytizing for an overarching

paradigm. But often confusion arises because of the unusual kind of predictions the new

methods provide. If our models apply at all to a physical system, they should be able to
predict all behavior on long length and time scales, independent of many microscopic

details of the real world. This predictive capacity comes, however, at a price: our

models typically don’t make clear predictions of how the real-world microscopic

parameters affect the long-length-scale behavior.

In this paper, we will provide an overview of the renormalization-group0.5-4

many researchers use to understand crackling noise. Briefly, the renormalization group

discusses how the effective evolution laws of our system change as we measure on
longer and longer length scales. (It works by generating a coarse-graining mapping in

system space, the abstract space of all possible evolution laws.) The broad range of

event sizes will be attributed to a self-similarity, where the evolution laws look the same
under different length scales. Using this self-similarity, we are led to a method for
scaling experimental data. In the simplest case this yields power laws and fractal

structures, but more generally it leads to universal scaling functions – where we argue
the real predictive power lies. We will only touch upon the dauntingly complex
 5

analytical methods used in this field, but we believe we can explain faithfully and fully
both what our tools are useful for, and how to apply them in practice. The

renormalization group is perhaps the most impressive use of abstraction in science.

Why should crackling noise be comprehensible?

Not all systems crackle. Some respond to external forces with lots of similar-sized,

small events (popcorn popping as it is heated). Others give way in one single event

(chalk snapping as it is stressed). Roughly speaking, crackling noise is in between these

limits: when the connections between parts of the system are stronger than in popcorn
but weaker than in the grains making up chalk, the yielding events can span many

decades of sizes. Crackling forms the transition between snapping and popping.

Figure 1b presents a pretty simple relationship. We expect that there ought to be

a simple, underlying reason that earthquakes occur on all different sizes. The very small

earthquake properties probably depend a lot on the kind of dirt (fault gouge) in the

crack. The very largest earthquakes will depend on the geography of the continental

plates. But the smooth power-law behavior suggests that something simpler is

happening in between, independent of either the microscopic or the macroscopic details.

There is a nice analogy with the behavior of a fluid. A fluid is very complicated
on the microscopic scale, where molecules are bumping into one another: the
trajectories of the molecules are chaotic, and depend both on exactly what direction they

are moving and what they are made of. However, a simple law describes most fluids on

long length and time scales. This law, the Navier-Stokes equation, depends on the
constituent molecules only through a few parameters (the density and viscosity).
Physics works because simple laws emerge on large scales. In fluids, these microscopic

fluctuations and complexities go away on large scales: for crackling noise, they become
scale invariant and self-similar.
 6

How do we derive the laws for crackling noise? There are two approaches. First,
one can analytically calculate the behavior on long length and time scales by formally

coarse-graining over the microscopic fluctuations. This leads us to renormalization-

group methods,0.5-4 which we discuss in the next section. The analytic approach can get

pretty hairy, but it can give useful results and (more importantly) is the only explanation

for why events on all scales should occur. Second, one can make use of universality. If
the microscopic details don’t matter for the long length scale behavior, why not make

up a simple model with the same behavior (in the same universality class) and solve it?

Figure 2: Magnets Crackle.66-69 Magnets respond to a slowly varying external field by changing their
magnetization in a series of bursts, or avalanches. These bursts, called Barkhausen noise, are very similar
(albeit on very different time and size scales) to those shown in figure 1 for the earthquakes. The
avalanches in our model have a power law distribution only at a special value of the disorder, Rc=2.16.
Shown is the histogram giving the number of avalanches Dint(S,R) of a given size S at various disorders R
ranging from 4 to 2.25; the thin lines are theoretical predictions from our model. (Dint gives all the
avalanches during our simulation, integrated over the external field -∞ < H(t) < +∞). The straight dashed
line shows the power-law distribution at the critical point. Notice that fitting power laws to the data would
work only very near to Rc: even with six decades of avalanche sizes, the slope hasn’t converged to the
asymptotic value. On the other hand, the scaling function predictions (theoretical curves) work well quite
far from the critical point. The inset shows a scaling collapse of the avalanche size distribution (scaled
probability versus scaled size), which is used to provide the theoretical curves as described in the text.
 7

The model we’ll focus on in this paper is a caricature of a magnetic

material.17,48,49,66-70 A piece of iron will “crackle” as it enters a strong magnetic field,

giving what is called Barkhausen noise. We model the iron as a cubic grid of magnetic

domains Si, whose north pole is either pointing upwards (Si = +1) or downward (Si = -1).
The external field pushes on our domain with a force H(t), which will increase with

time. Iron can be magnetized because neighboring domains prefer to point in the same
direction: if the six neighbors of our cubic domain are Sj, then in our model we let their

force on our domain be ∑j J Sj. Finally, we model dirt, randomness in the domain

shapes, and other kinds of disorder by introducing a random field hi, different for each
domain and chosen at random from a normal distribution with standard deviation R,

which we call the disorder. The net force on our domain is thus

Force on domain i = H(t) + ∑j J Sj + hi. [1]

The domains in our model all start pointing down (-1), and flip up as soon as the net

force on them becomes positive. This can occur either because H(t) increases

sufficiently (spawning a new avalanche), or because one of their neighbors flipped up,

kicking them over (propagating an existing avalanche). (Thermal fluctuations are

ignored: a good approximation in many experiments because the domains are large.) If

the disorder R is large, so the hi are typically big compared to J, then most domains flip
independently: all the avalanches are small, and one gets popping noise. If the disorder
is small compared to J, then typically most of the domains will be triggered by one of

their neighbors: one large avalanche will snap up most of our system. In between, we

get crackling noise. When the disorder R is just large enough so that each domain flip
on average triggers one of its neighbors (at the critical disorder Rc), then we find
avalanches on all scales (figures 2 and 3).
 8

Figure 3: Self-similarity. These are cross-sections of the avalanches during the magnetization of our
model.17,66-69 Here each avalanche is drawn in a separate color. (a) shows a 1003 simulation and (b) shows
a 10003 simulation (a billion domains21b); both are run at the critical point Rc=2.16 J where avalanches
just barely continue. The black background represents a large avalanche that spans the system: the critical
point occurs when avalanches would first span an infinite system.

What do these avalanches represent? In nonlinear systems with many degrees of

freedom, there are often large numbers of metastable states. Local regions in the system

can have multiple stable configurations, and many combinations of these local

configurations are possible. (A state is metastable when it cannot lower its energy by

small rearrangements. It’s distinguished from the globally stable state, which is the

absolute lowest energy possible for the system.) Avalanches are precisely the

rearrangements as our system shifts from one metastable state to another. Our specific
interest is in systems with a broad distribution of avalanche sizes, where shifting

between metastable states can rearrange anything between a few domains and millions
of domains.

There are lots of choices we made in our model that don’t matter. Because of
universality, we can argue72-73 that the behavior would be the same if we chose a

different grid of domains, or if we changed the distribution of random fields, or if we

introduced more realistic random anisotropies and random coupling constants. Were

this not the case, we could hardly expect our simple model to explain real experiments.
 9

The Renormalization Group and Scaling

To study crackling noise, we use renormalization-group0.5-4,71,72 tools developed in the

study of second-order phase transitions. The word renormalization has roots in the study

of quantum electrodynamics, where the effective charge changes in size (norm) as a

function of length scale. The word group refers to the family of coarse-graining

operations basic to the method: the group product is composition (coarsening

repeatedly). The name is unfortunate, however, as the basic coarse-graining operation

does not have an inverse, and thus the renormalization group does not have the
mathematical structure of a group.

The renormalization group studies the way the space of all physical systems

maps into itself under coarse-graining (see figure 4). The coarse-graining operation

shrinks the system, and removes degrees of freedom on short length scales. Under

coarse-graining, we often find a fixed point S*: many different models flow into the

fixed point and hence share long-wavelength properties. To get a schematic view of

coarse-graining, look at figure 3: the 10003 cross section looks (statistically) like the
1003 section if you blur your eyes by a factor of 10. Much of the mathematical

complexity of this field involves finding analytical tools for computing the flow

diagram in figure 4. Using methods developed to study thermodynamical phase

transitions2 and the depinning of charge-density waves,34 we can calculate for our model
the flows for systems in dimensions close to six (the so-called ε expansion,71-73 where ε

=6-d, d being the dimension of the system). Interpolating between dimensions may

seem a surprising thing to do. In our system it gives rather good predictions even in
three dimensions (i.e., ε=3), but it’s hard work, and we won’t discuss it here. Nor will
we discuss real-space renormalization-group methods0.5 or series expansion methods.

We focus on the relatively simple task of using the renormalization group to justify and

explain the universality, self-similarity, and scaling observed in nature.

 10

Figure 4: Renormalization-group flows. The renormalization-group is a theory of how coarse-graining to

longer length scales introduces a mapping from the space of physical systems to itself. Consider the space
of all possible models of magnetic hysteresis (what an abstraction!). Each model can be coarse-grained,
removing some fraction of the microscopic degrees of freedom and introducing more complicated rules
so that the remaining ones still flip at the same external fields. This defines a mapping from our space into
itself. A fixed point S* in this space will be self-similar: since it maps to itself upon coarse-graining, it
must have the same behavior on different length scales. Points that flow into S* under coarse-graining
share this self-similar behavior on sufficiently long length scales: they all share the same universality
class.

Consider the "system space" for disordered magnets. There is a separate

dimension in system space for each possible parameter in a theoretical model (disorder,
coupling, next-neighbor coupling, dipolar fields ...) or in an experiment (temperature,

annealing time, chemical composition ...). Coarse-graining, however one implements it,
gives a mapping from system space into itself: shrinking the system and ignoring the

shortest length scales yields a new physical system with identical long-distance physics,
 11

but with different (renormalized) values of the parameters. We’ve abstracted the
problem of understanding crackling noise in magnets into understanding a dynamical

system acting on the space of all dynamical systems.

Figure 4 represents a two-dimensional cross section of this infinite-dimensional

system space. We've chosen the cross section to include our model (equation [1]): as we

vary the disorder R, our model sweeps out a straight line (red) in system space. The
cross section also includes a fixed point S*, which maps into itself under coarse-

graining. The system S* looks the same on all length and time scales, because it coarse-
grains into itself. We can picture the cross section of figure 4 either as a plane in system

space (in which case the arrows and flows depict projections, since in general the real

flows will point somewhat out of the plane), or as the curved manifold swept out by our

one-parameter model as we coarse grain (in which case the flows above our model and

below the green curved line should be ignored).

The flow near S* has one unstable direction, leading outward along the green

curve (the unstable manifold). In system space, there is a surface of points C which flow

into S* under coarse-graining. Because S* has only one unstable direction, C divides

system space into two phases. To the left of C, the systems will have one large, system-

spanning avalanche (a snapping noise). To the right of C, all avalanches are finite and

under coarse-graining they all become small (popping noise). Our model, as it crosses C

at the value Rc, goes through a phase transition.

Our model at Rc is not self-similar on the shortest length scales (where the

square lattice of domains still is important), but because it flows into S* as we coarse-
grain we deduce that it is self-similar on long length scales. Some phase transitions, like

ice melting into water, are abrupt and don't exhibit self-similarity. Continuous phase
transitions like ours almost always have self-similar fluctuations on long length scales.

Also, we must note that our model at Rc will have the same self-similar structure as S*
 12

does. Indeed, any experimental or theoretical model lying on the critical surface C will
share the same long-wavelength critical behavior. This is the fundamental explanation

for universality.

Figure 5: Flows in the space of earthquake models. A model for earthquakes will have a force F applied
across the front. In models ignoring inertia and velocity-dependent friction30, there is a critical force Fc
that just allows the fault to slip forward. (a) Coarse-graining defines a flow on the space of earthquake
models. The fixed point S*eq will have a different local flow field from other renormalization-group fixed
points, yielding its own universality class of critical exponents and scaling functions. The critical
manifold C, consisting of models which flow into S*eq, separates the stuck faults from those which slide
forward with an average velocity v(F). (b) The velocity varies with the external force as a power law
v(F)~Fβ. The motion of the continental plates, however, does not fix the force F across the fault: rather, it
sets the average relative velocity to a small value vs (centimeters per year). This automatically sets the
force across the fault very close to its critical force Fc. This is one example of self-organized
criticality.38,39
 13

The flows in system space can vary from one class of problems to another: the
system space for some earthquake models (figure 5a) will have a different flow, and its

fixed point will have different scaling behavior (yielding a different universality class).
In some cases, a fixed point will attract all the systems in its vicinity (no unstable

directions, figure 6). Usually at such attracting fixed points the fluctuations become
unimportant at long length scales: the Navier-Stokes equation for fluids described

earlier can be viewed as a stable fixed point.74,75 The coarse-graining process, averaging

over many degrees of freedom, naturally smoothens out fluctuations, if they aren’t

amplified near a critical point by the unstable direction. Fluctuations can remain

important when a system has random noise in a conserved property, so that fluctuations

can only die away by diffusion: in these cases, the whole phase will have self-similar

fluctuations, leading to generic scale invariance.76,77

Figure 6: Attracting fixed point. Often there will be fixed points that attract in all directions. These fixed
points describe phases rather than phase transitions. Most phases are rather simple, with fluctuations that
die away on long length scales.74 When fluctuations remain important, they will exhibit self-similarity
and power laws called generic scale invariance.76,77
 14

Sometimes, even when the system space has an unstable direction like in figure
4, the observed behavior always has avalanches of all scales. This can occur simply

because the physical system averages over a range of model parameters (i.e., averaging
over a range of R including Rc in figure 4). For example, this can occur by the sweeping

of a parameter78 slowly in time, or varying it gradually in space – either deliberately or

through large-scale inhomogeneities.

One can also have self-organized criticality,38,39 where the system is controlled

so that it naturally sits on the critical surface. Self-organization to the critical point can

occur via many mechanisms. In some models of earthquake faults (figure 5b), the

external force naturally stays near the rupture point because the plates move at a fixed,

but very small22 velocity with respect to one another (figure 5b). (This probably does

not occur during large earthquakes, where inertial effects lead to temporary strain relief

.30,79) Sandpile models self-organized (if sand is added to the system at an infinitesimal

rate) when open boundary conditions80 are used (which allows sand to leave until the
sandpile slope falls to the critical value). Long-range interactions81-83 between domains

can act as a negative feedback in some models, yielding a net external field that remains

at the critical point. For each of these cases, once the critical point is understood adding

the mechanism for self-organization is relatively easy.

The case shown in figure 4 of plain old criticality is what’s seen in some17,66-69
but not all81-85 models of magnetic materials, in foams,44 and in some models of

earthquakes.30

Beyond Power Laws

The renormalization group is the theoretical underpinning for understanding

why universality and self-similarity occur. Once we grant that different systems should
 15

sometimes share long-distance properties, though, we can quite easily derive some
powerful predictions.

To take a tangible example, let’s consider the relation between the duration of an
avalanche and its size. In paper crumpling, this isn’t interesting: all the avalanches seem

to be without internal temporal structure.14 But in magnets large events take longer to
finish, and have an interesting internal statistical self-similarity (figure 7a). If we look at

all avalanches of a certain duration T in an experiment, they will have a distribution of

sizes S around some average <S>experiment(T). If we look at a theoretical model, it will

have a corresponding average size <S>theory(T). If our model describes the experiment,

these functions must be essentially the same at large S and large T. We must allow for

the fact that the experimental units of time and size will be different from the ones in

our model: the best we can hope for is that <S>experiment(T) = A <S>theory(T/B), for some

rescaling factors A and B.

Now, instead of comparing to experiment, we can compare our model to itself

on a slightly larger time scale.88 If the time scale is expanded by a small factor B=1/(1-
δ), then the rescaling of the size will also be small, say 1+a δ.

<S>(T) = (1 + a δ) <S>( (1-δ) T)

Making δ very small yields the simple relation a <S>=T d<S>/dT, which can be solved
to give the power law relation <S>(T) = S0 Ta. The exponent a is called a critical
exponent, and is a universal prediction of a given theory. (That means that if the theory
correctly describes an experiment, the critical exponents will agree.) In our work, we

write the exponent a relating time to size in terms of three other critical exponents,

a=1/σνz.
 16

Figure 7: Scaling of avalanche

shapes. (a) Voltage (number of
domains flipped) pulse during a
single large avalanche
(arbitrary units). Notice how
the avalanche almost stops
several times: if the forcing
were slightly smaller, this large
avalanche would have broken
up into two or three smaller
ones. The fact that the forcing
is just large enough to on
average keep the avalanche
growing is the cause of the self-
similarity: on average a partial
avalanche of size S will trigger
one other on size S. (b)
Average avalanche shapes86 for
avalanches of different
durations for our model (A.
Mehta and K. A. Dahmen). Our
theories don’t only predict
power laws: they should
describe all behavior on long
length and time scales (at least
in a statistical sense). In
particular, by fixing parameters
one can predict what are called
scaling functions. If we average
the voltage as a function of
time over all avalanches of a
fixed duration, we get an
average shape. In our
simulation, this shape is the
same for different durations. (c)
Experimental data of
Spasojević et al.,87 showing all
large avalanches averaged after
scaling to fixed duration and
area. The experimental average
shape is very asymmetric and is
not described correctly by our
model.
 17

To take a tangible example, let’s consider the relation between the duration of an
avalanche and its size. In paper crumpling, this isn’t interesting: all the avalanches seem

to be without internal temporal structure.14 But in magnets large events take longer to

finish, and have an interesting internal statistical self-similarity (figure 7a). If we look at

all avalanches of a certain duration T in an experiment, they will have a distribution of

sizes S around some average <S>experiment(T). If we look at a theoretical model, it will
have a corresponding average size <S>theory(T). If our model describes the experiment,

these functions must be essentially the same at large S and large T. We must allow for

the fact that the experimental units of time and size will be different from the ones in

our model: the best we can hope for is that <S>experiment(T) = A <S>theory(T/B), for some
rescaling factors A and B.

Now, instead of comparing to experiment, we can compare our model to itself

on a slightly larger time scale.88 If the time scale is expanded by a small factor B=1/(1-

δ), then the rescaling of the size will also be small, say 1+a δ.

<S>(T) = (1 + a δ) <S>( (1-δ) T)

Making δ very small yields the simple relation a <S>=T d<S>/dT, which can be solved

to give the power law relation <S>(T) = S0 Ta. The exponent a is called a critical

exponent, and is a universal prediction of a given theory. (That means that if the theory

correctly describes an experiment, the critical exponents will agree.) In our work, we
write the exponent a relating time to size in terms of three other critical exponents,
a=1/σνz.

There are several basic critical exponents, which arise in various combinations
depending on the physical property being studied. The details of the naming and

relationships between these exponents aren’t a focus of our paper. Briefly, the cutoff in
the avalanche size distribution in figure 2 gets larger as one approaches the critical
 18

disorder as (R-Rc)−1/σ (figure 2). The typical length of the largest avalanche goes as (R-
Rc)−ν. At Rc, the probability of having an avalanche of size S goes as S−(τ+σβδ) (figure 2);

if one sits just at the critical field, it goes as S−τ. (Don’t confuse the small change in

scale δ with the critical exponent δ.) The fractal dimension of the avalanches is 1/σν,
meaning the spatial extent L of an avalanche is proportional to the size Sσν. The duration

T of an avalanche of spatial extent L goes as Lz…

Figure 8: Critical exponents in various dimensions. We test our ε-expansion predictions72 by measuring69
the various critical exponents numerically in up to five spatial dimensions. The various exponents are
described in the text. All of the exponents are calculated only to linear order in ε, except for the
correlation length exponent ν, where we use results from other models.71,72 The agreement even in three
dimensions is remarkably good, considering that we’re expanding in ε where ε=3! We should note that
perturbing in dimension for our system is not only complicated, but also controversial89 (see also section
VI.C of Ref. 72 and section V of Ref. 69).

To specialists in critical phenomena, these exponents are central; whole

conversations will seem to rotate around various combinations of Greek letters. Critical
exponents are one of the relatively easy things to calculate from the various analytic
 19

approaches, and so have attracted the most attention. They are derived from the
eigenvalues of the linearized flows about the fixed point S* in figure 4. Figure 8 shows

our numerical estimates69 for several critical exponents in our model in various spatial

dimensions, together with our 6-ε expansions71,72 for them. Of course the key challenge
is not to get analytical work to agree with numerics: it’s to get theory to agree with

experiment. Figure 9 shows that our model does rather well in describing a wide variety
of experiments, but that two rival models (with different flows around their fixed

points) also fit.

Figure 9: Experiments vs. Theory: Critical Exponents. Different experiments on crackling noise in
magnets measure different combinations of the universal critical exponents. Here we compare
experimental measurements81,87,90-101 (see table I of Ref. 66) to the theoretical predictions for three
models: our model,17,66-69 the front-propagation model48-53,100 and mean-field theory. (In mean-field theory
our coupling J in equation [1] couples all pairs of spins: such long-range interactions occur because of the
boundaries in models with magnetic dipolar forces.83 Mean-field theory is equivalent to a rather
successful single-degree of freedom model.102,103) Power laws giving the probability of getting an
avalanche of a given size, duration, or energy at the critical point are shown; also shown is the critical
exponent giving the power as a function of frequency86 (due to the internal structure of the avalanches,
Figure 7a). In each pair of columns, the first column includes only avalanches at external fields H in
equation [1] where the largest avalanches occur, and the second column (when it exists) includes all
avalanches. The various combinations of the basic critical exponents can be derived from exponent
equality calculations similar to the one discussed in the text.66,71,86 Many of the experiments were done
years before the theories were developed: many did not report error bars. All three theories do well
(especially considering the possible systematic errors in fitting power laws to the experimental
measurements: see figure 2). Recent work suggests a clumping of experimental values around the mean-
field and front-propagation predictions.100
 20

Critical exponents are not the be-all and end-all: many other scaling predictions,
explaining wide varieties of behavior, are quite easy to extract from numerical

simulations. Universality extends even to those long length scale properties for which
one cannot write formulas. Perhaps the most important of these other predictions are the

universal scaling functions. For example, let’s consider the time history of the
avalanches, V(t), denoting the number of domains flipping per unit time. (We call it V

because it’s usually measured as a voltage in a pickup coil.) Each avalanche has large
fluctuations, but one can average over many avalanches to get a typical shape. Figure

7b shows the average over all avalanches of fixed duration T. Let’s call this <V>(T,t).
Universality again suggests that this average should be the same for experiment and a

successful theory, apart from an overall shift in time and voltage scales:
<V>experiment(T,t) = A <V>theory(T/B, t/B). Comparing our model to itself with a shifted

time scale becomes simple if we change variables: let v(T,t/T) = <V>(T,t), so v(T,t/T) =

A v(T/B,t/T). Here t/T is a particularly simple example of a scaling variable. Now, if we

rescale time by a small factor B=1/(1-δ), we have v(T, t/T) = (1 + b δ) v(t/T, (1-δ) T).

Again, making δ small we find b v = T ∂v/∂T, with solution v = v0 Tb. However, the

integration constant v0 will now depend on t/T, v0 =V(t/T), so we arrive at the scaling

form

<V> (t,T) = TbV(t/T), [2]

where the entire scaling function V is a universal prediction of the theory.

Figures 7b and 7c show the universal scaling functions V for our model86 and an

experiment.87 For our model, we’ve drawn what are called scaling collapses, a simple
but powerful way to both check that we’re in the scaling regime, and to measure the

universal scaling function. Using the form of the scaling equation Eq.[2], we simply
plot T–b <V>(t,T) versus t/T, for a series of long times T. All the plots fall onto the same

curve. This tells us that our avalanches are large enough to be self-similar. (If in your
 21

scaling collapse the corresponding plots do not all look alike, then any power laws you
have measured are probably accidental.) It also provides us with a numerical evaluation

of the scaling function V. Note that we use 1/σνz-1 for the critical exponent b. This is an

example of an exponent equality: easily derived from the fact that <S>(T) = ∫ <V>(t,T)

dt = ∫ Tb V(t/T) dt ~ Tb+1, and the scaling relation <S>(T) ~T1/σνz.

Notice that our model and the experiment have quite different shapes for V. The
other two models from Figure 9 also give much more symmetrical forms for V than the

experiment does.86 How do we react to this? Our models are falsified if any of the

predictions are shown to be wrong asymptotically on long length and time scales. If

duplication of this measurement by other groups continues to show this asymmetry,

then our theory is obviously incomplete. Even if later experiments in other systems

agree with our predictions, it would seem that this particular system is described by an

undiscovered universality class. Incorporating insights from careful experiments to

refine the theoretical models has historically been crucial in the broad field of critical
phenomena. The message we emphasize here is that scaling functions can provide a

sharper tool for discriminating between different universality classes than critical
exponents.

Broadly speaking, most common properties that involve large length and time
scales have scaling forms: using self-similarity, one can write functions of N variables

in terms of scaling functions of N-1 variables: F(x,y,z) = z−α F(x/zβ, y/zγ). In the inset to
figure 2, we show the scaling collapse for the avalanche size distribution: D(S,R) =

S−(τ+σβδ) D( (R-Rc) / S−σ). (This example illustrates that scaling works not only at Rc but
also near Rc; the green unstable manifold in figure 4 governs the behavior for systems
near the critical manifold C.)
 22

Figure 10: Fractal spatial structure of an avalanche.67 Fractal structures, as well as power laws, are
characteristic of systems at their critical point. This moderate-sized avalanche involved the flipping of
282,785 domains in our simulation. The colors represent time: the first domains to flip are colored blue,
and the last pink. So far, there have not been many experiments showing the spatial structure of
avalanches.104 When experiments become available, there are a wealth of predictions of the scaling
theories that we could test. Other systems105,55,30 have seen a qualitatively different kind of avalanche
spatial structure, where the avalanche is made up of many small disconnected pieces, which trigger one
another through the waves emitted as they flip.

Many other kinds of properties beyond critical exponents and scaling functions
can be predicted from these theories. Figure 10 shows the spatial structure of a large
avalanche in our model: notice not only that it is fractal (rugged on all scales), but that it

is longer than it is wide,106 and that it is topologically interesting.107 (It has tunnels, and
sometimes during the avalanche it forms a tunnel and later winds itself through it,

forming a knot. It’s interesting that the topology of the interfaces in the three-
dimensional Ising model have applications in quantum gravity.107) The statistics of all

of these properties, in other systems, have been shown to be universal on long length
and time scales.
 23

Continuing challenges

Do not be fooled: our understanding of crackling noise is far from complete. There are
only a few systems30,34,50,51,54.2,54.4,71-73 where the renormalization-group framework has

substantially explained even the behavior of the numerical models. There are several

other approaches63,64,80,108-110 that have been developed to study crackling noise, many
which share our view of developing effective descriptions on long length and time

scales. But the successes remain dwarfed by the bewildering variety of systems that
crackle. Getting a global perspective on the universality classes for crackling noise

remains an open challenge.

An even more important challenge is to make quantitative comparison between

the theoretical models and experimental systems. We believe that going beyond power

laws will be crucial in this endeavor. The past focus on critical exponents has

sometimes been frustrating: it’s too easy to find power laws over limited scaling

ranges,111 and too easy to find models which roughly fit them. It also seems unfulfilling,
summarizing a complex morphology into a single critical exponent. We believe that

measuring a power law is almost never definitive by itself: a power law in conjunction

with evidence that the morphology is scale invariant (e.g., a scaling collapse) is crucial.

By aggressively pursuing quantitative comparisons of other, richer measures of

morphology such as the universal scaling functions, we will be better able both to
discriminate among theories and to ensure that a measured power-law corresponds to a

genuine scaling behavior.

Another challenge is to start thinking about the key ways that these complex
spatiotemporal systems differ from the phase transitions we understand from
equilibrium critical behavior. (The renormalization-group tools developed by our

predecessors are seductively illuminating: it’s always easy to focus where the light is

good.) For example, in several of these systems there are collective, dynamical
 24

“memory” effects17,112-115 that even may have practical applications.116 The quest for a
scaling theory of crackling phenomena needs to be viewed as part of the larger process

of understanding the dynamics of these nonlinear, nonequilibrium systems.

A final challenge is to make the study of crackling noise profitable. Less noise
from candy wrappers14-16 in movie theaters is not the most pressing of global concerns.

Making money from fluctuations in stock prices is already big business.58,59 Predicting

earthquakes over the short term probably will not be feasible using these approaches,117
but longer term prediction of impending large earthquakes may be both possible79 and

useful, say, for guiding local building codes. Understanding that the large-scale

behavior relies only on a few emergent material parameters (disorder and external field

for our model of magnetism) leads one to study how these parameters depend on the

microphysics: one has dreams, for example, of learning how to shift an active

earthquake fault into a harmless, continuously sliding regime by adding lubricants to the

fault gouge. In the meantime, crackling noise is basic research at its elegant,
fundamental best.

References
0.5. Kadanoff, L. P., Scaling Laws for Ising Models near Tc, Physics (Long
Island City, NY) 2, 263-272 (1966).
1. Wilson, K.G., Problems in physics with many scales of length, Scientific American 241, 140-57 (1979).
2. Pfeuty, P. & Toulouse, G., Introduction to the Renormalization Group and to Critical Phenomena,
Wiley and Sons, London, 1977.
3. Yeomans, J. M., Statistical Mechanics of Phase Transitions, Oxford, 1992.
4. Fisher, M. E., Renormalization group theory: Its basis and formulation in statistical physics, Rev. Mod.
Phys. 70, 653-681 (1998).
5. Martin, P. C., Siggia, E. D., & Rose, H. A., Statistical Dynamics of Classical Systems, Phys. Rev. A 8,
423-437 (1973).
6. De Dominicis, C., Dynamics as a substitute for replicas in systems with quenched random impurities,
Phys. Rev. B 18, 4913-4919 (1978).
7. Sompolinsky, H. & Zippelius, A., Relaxational dynamics of the Edwards-Anderson model and the
mean-field theory of spin-glasses, Phys. Rev. B 25, 6860-6875 (1982).
8. Zippelius, A., Critical-dynamics of spin-glasses, Phys. Rev. B 29, 2717-2723 (1984).
9. Gutenberg, B. & Richter, C. F., Seismicity of the earth and associated phenomena (Princeton
University, Princeton, NJ, 1954).
10. Composite Earthquake Catalog Archive of the Council of the National Seismic System,
www.cnss.org.
11. USGS National Earthquake Information Center: Earthquake Information for the World,
wwwneic.cr.usgs.gov.
12. Sethna, J. P., Kuntz, M. C., & Houle, P. A., Crackling Noise, http://simscience.org/crackling.
 25

13. Sethna, J. P., Hysteresis and Avalanches,

http://www.lassp.cornell.edu/sethna/hysteresis/hysteresis.html.
14. Houle, P. A. & Sethna, J. P., Acoustic Emission from Crumpling Paper, Phys. Rev. E 54, 278-283
(1996).
15. Kramer, E. M. & Lobkovsky, A. E., Universal Power Law in the Noise from a Crumpled Elastic
Sheet, Phys. Rev. E 53, 1465-1469 (1996).
16. Glanz, J., No Hope of Silencing the Phantom Crinklers of the Opera, New York Times National Desk,
June 1, 2000.
17. Sethna, J. P., Dahmen, K. A., Kartha, S., Krumhansl, J. A., Roberts, B. W., & Shore, J. D., Hysteresis
and Hierarchies: Dynamics of Disorder-Driven First-Order Phase Transformations, Phys. Rev. Lett.. 70,
3347–3351 (1993).
18. Burridge, R., & Knopoff, L., Model and theoretical seismicity, Bull. Seismol. Soc. Am. 57, 3411-3471
(1967).
19. Rice, J. R. & Ruina, A. L., Stability of steady frictional slipping, J. Appl. Mech. 50, 343 (1983).
20. Carlson, J. M. & Langer, J. S., Mechanical model of an earthquake fault, Phys. Rev. A 40, 6470-6484
(1989).
21. Bak, P. & Tang, C., Earthquakes as a self-organized critical phenomenon, J. Geophysical Res. 94,
15635-7 (1989).
22. Chen, K., Bak, P., & Obukhov, S. P., Self-organized criticality in a crack-propagation model of
earthquakes, Phys. Rev. A 43, 625-630 (1991).
23. Olami, Z., Feder, H. J. S., & Christensen, K., Self-organized criticality in a continuous,
nonconservative cellular automaton modeling earthquakes, Phys. Rev. Lett. 68, 1244-1247 (1992).
24. Miltenberger, P., Sornette, D., & Vanette, C., Fault Self-Organization and Optimal Random Paths
Selected by Critical Spatiotemporal Dynamics of Earthquakes, Phys. Rev. Lett. 71, 3604-3607 (1993).
25. Crowie, P.A., Vanette, C., & Sornette, D., Statistical physics model for the spatiotemporal evolution
of faults, J. Geophysical Res.-Solid Earth 98, 21809-21821 (1993).
26. Carlson, J. M., Langer, J. S., & Shaw, B.E., Dynamics of earthquake faults, Rev. Mod. Phys. 66, 657-
70 (1994).
27. Myers, C. R., Shaw, B. E., & Langer, J. S., Slip complexity in a crustal-plane model of an earthquake
fault, Phys. Rev. Lett. 77, 972-975 (1996).
28. Shaw, B. E. & Rice, J. R., Existence of continuum complexity in the elastodynamics of repeated fault
ruptures, J. Geophysical Res.,105, 23791-23810 (2000).
29. Ben-Zion, Y. & Rice, J. R., Slip patterns and earthquake populations along different classes of faults
in elastic solids, J. Geophysical Res., 100, 12959-12983 (1995).
30. Fisher, D. S., Dahmen, K., Ramanathan, S., & Ben-Zion, Y., Statistics of Earthquakes in Simple
Models of Heterogeneous Faults, Phys. Rev. Lett. 78, 4885-4888 (1997).
31. Fisher, D. S., Threshold behavior of charge-density waves pinned by impurities, Phys. Rev. Lett. 50,
1486-1489 (1983).
32. Fisher, D. S., Sliding charge-density waves as a dynamic critical phenomenon, Phys. Rev. B 31, 1396-
1427 (1985).
33. Littlewood, P. B., Sliding charge-density waves: a numerical study, Phys. Rev. B 33, 6694-6708
(1986).
34. Narayan, O. & Fisher, D. S., Critical behavior of sliding charge-density waves in 4-ε dimensions,
Phys. Rev. B 46, 11520-11549.
35. Middleton, A. A. & Fisher, D. S., Critical behavior of charge-density waves below threshold:
numerical and scaling analysis, Phys. Rev. B 47, 3530-52 (1993).
36. Myers, C. R. & Sethna, J. P., Collective dynamics in a model of sliding charge-density waves. I.
Critical behavior, Phys. Rev. B 47, 11171-11192 (1993).
37. Thorne, R. E., Charge-density-wave conductors, Physics Today 49, 42-47 (1996).
38. Bak, P., Tang, C., & Wiesenfeld K., Self-organized Criticality: An explanation for 1/f noise, Phys.
Rev. Lett. 59, 381-384 (1987).
39. Bak, P., Tang, C., & Wiesenfeld K., Self-organized Criticality, Phys. Rev. A 38, 364-374 (1988)
40. deGennes, P. G., Superconductivity of Metals and Alloys, Benjamin, New York, 1966, p. 83.
41. Feynman, R. P., Leighton, R. B., & Sands, M., The Feynman Lectures on Physics, Addison Wesley,
1963-5, Volume II, section 37-3.
42. Jaeger, H. M., Liu, C., & Nagel, S. R., Relaxation at the Angle of Repose, Phys. Rev. Lett. 62, 40-43
(1989).
43. Nagel, S. R., Instabilities in a sandpile, Rev. Mod. Phys. 64, 321-5 (1992)).
 26

44. Tewari, S., Schiemann, D., Durian, D. J., Knobler, C. M., Langer, S. A., & Liu, A. J., Statistics of
shear-induced rearrangements in a two-dimensional model foam, Phys. Rev. E 60, 4385-4396, (1999).
45. Solé, R. V. & Manrubia, S. C., Extinction and self-organized criticality in a model of large-scale
evolution, Phys. Rev. E 54, R42-R45 (1996).
46. Newman, M. E. J., Self-organized criticality, evolution, and the fossil extinction record, Proc. R. Soc.
London B 263, 1605-1610 (1996).
47. Newman, M. E. J. & Palmer, R. G., Models of Extinction: A Review, http://xxx.lanl.gov/abs/adap-
org/9908002.
48. Cieplak, M. & Robbins, M. O., Dynamical Transition in Quasistatic Fluid Invasion in Porous Media,
Phys. Rev. Lett. 60, 2042-2045 (1988).
49. Koiller, B. & Robbins, M. O., Morphology transitions in three-dimensional domain growth with
Gaussian random fields, Phys. Rev. B 62, 5771-5778 (2000), and references therein.
50. Nattermann, T., Stepanow, S., Tang, L. H., & Leschhorn N., Dynamics of Interface Depinning in a
Disordered Medium, J. de Physique II 2, 1483-1488 (1992).
51. Narayan, O. & Fisher, D. S., Threshold Critical Dynamics of Driven Interfaces in Random Media,
Phys. Rev. B 48, 7030-7042 (1993).
52. Leschhorn, H., Nattermann, T., Stepanow, S., & Tang, L.-H., Driven interface depinning in a
disordered medium, Ann. Phys. (Leipzig) 6, 1-34, (1997).
53. Roters, L., Hucht, A., Lubeck, S., Nowak, U., & Usadel, K.D., Depinning transition and thermal
fluctuations in the random-field Ising model, Phys. Rev. E 60, 5202-7 (1999).
54. Field, S., Witt, J., Nori, F., & Ling, X., Superconducting Vortex Avalanches, Phys. Rev. Lett. 74,
1206-1209, 1995.
54.2. Ertaş, D. & Kardar, M., Anisotropic Scaling in Depinning of a Flux Line, Phys. Rev. Lett. 73, 1703-
1706 (1994).
54.4. Ertaş, D. & Kardar, M., Anisotropic scaling in threshold critical dynamics of driven directed lines,
Phys. Rev. B 53, 3520-3542 (1996).
55. Lilly, M. P., Wootters, A. H., & Hallock, R. B., Spatially extended avalanches in a hysteretic capillary
condensation system: Superfluid He-4 in nuclepore, Phys. Rev. Lett. 77, 4222-4225 (1996).
56. Guyer, R. A., McCall, K. R., Capillary condensation, invasion percolation, hysteresis, and discrete
memory, Phys. Rev. B 54, 18-21 (1996).
57. Ortín, J., Ràfols, I., Carrillo, L., Goicoechea, J., Vives, E., Mañosa, L., & Planes, A., Experiments and
models of avalanches in martensites, J. de Physique IV 5, 209-214 (1995), and many papers thereafter.
58. Bouchaud, J.P., Power-laws in economy and finance: some ideas from physics, cond-mat/0008103,
proceedings of the Santa Fe Conference `Beyond Efficiency’, to be published in J. Quant. Finance, and
references therein.
59. Bak, P., Paczuski, M., & Shubik, M., Price Variations in a Stock Market with Many Agents, Physica
A 246, 430-453 (1997).
60. Lu, E. T., Hamilton, R. J., McTiernan, J. M., & Bromond, K. R., Solar Flares and Avalanches in
Driven Dissipative Systems, Astrophysical Journal 412, 841-852 (1993).
61. Carreras, B. A., Newman, D. E., Dobson, I., & Poole, A. B., Initial Evidence for Self-Organized
Criticality in Electric Power System Blackouts, Proceedings of the 33rd Hawaii International Conference
on System Sciences - 2000, January 2000.
62. Sachtjen, M. L., Carreras, B. A., & Lynch, V. E., Disturbances in a power transmission system, Phys.
Rev. E 61, 4877-4882 (2000).
63. Carlson, J. M. & Doyle, J., Highly optimized tolerance: A mechanism for power laws in designed
systems, Phys. Rev. E 60, 1412-1427 (1999).
64. Carlson, J. M. & Doyle, J., Highly optimized tolerance: Robustness and design in complex systems,
Phys. Rev. Lett. 84, 2529-2532 (2000).
65. Newman, M., The power of design, Nature 405, 412-413 (2000).
65.1. Galam, S., Rational group decision making: a random field Ising model at T=0, Physica A 238, 66-
80 (1997).
65.2. Petri, A., Paparo, G., Vespignani, A., Alippi, A., & Costantini, M., Experimental Evidence for
Critical Dynamics in Microfracturing Processes, Phys. Rev. Lett. 73, 3423-3426 (1994).
65.3. Garcimartín, A., Guarino, A., Bellon, L., & Ciliberto, S., Statistical Properties of Fracture
Precursors, Phys. Rev. Lett. 79, 3202-3205 (1997).
65.4. Curtin, W. A., & Scher, H., Analytic model for scaling of breakdown, Phys. Rev. Lett. 67, 2457-
2460 (1991).
 27

65.5. Herrman, H. J. & Roux, S., eds., Statistical Models for the Fracture of Disordered Media, (North
Holland, Amsterdam, 1990), and references therein.
65.6. Chakrabarti, B. K. & Benguigui, L. G., Statistical Physics of Fracture and Breakdown in
Disordered Systems, (Clarendon Press, Oxford, 1997) and references therein.
65.7. Zapperi, S., Ray, P., Stanley, H. E., & Vespignani, A., First-Order Transition in the Breakdown of
Disordered Media, Phys. Rev. Lett. 78, 1408-1411 (1997).
66. Perković, O., Dahmen, K. A., & Sethna, J. P., Avalanches, Barkhausen Noise, and Plain Old
Criticality, Phys. Rev. Lett. 75, 4528-4531 (1995).
67. Kuntz, M. C., Perković, O., Dahmen, K. A., Roberts, B. W., & Sethna, J. P., Hysteresis, Avalanches,
and Noise: Numerical Methods, Computing in Science and Engineering 1, 73-81 (1999).
68. M. C. Kuntz & J. P. Sethna, Hysteresis, Avalanches, and Noise: Numerical Methods,
http://www.lassp.cornell.edu/sethna/hysteresis/code/.
69. Perković, O., Dahmen, K. A., & Sethna, J. P., Disorder-induced Critical Phenomena in Hysteresis:
Numerical Scaling in Three and Higher Dimensions, Phys. Rev. B 59, 6106-6119 (1999).
70. Berger, A., Inomata, A., Jiang, J. S., Pearson, J. E., & Bader, S. D., Experimental Observation of
Disorder-Driven Hysteresis-Loop Criticality, Phys. Rev. Lett. 85, 4176-4179 (2000).
71. Dahmen, K. A. & Sethna, J. P., Hysteresis Loop Critical Exponents in 6-ε Dimensions, Phys. Rev.
Lett. 71, 3222-3225 (1993)
72. Dahmen, K. A. & Sethna, J. P., Hysteresis, Avalanches, and Disorder Induced Critical Scaling: A
Renormalization Group Approach, Phys. Rev. B 53, 14872-14905 (1996).
73. da Silveira, R. & Kardar, M., Critical Hysteresis for N-Component Magnets, Phys. Rev. E 59, 1355-
1367 (1999).
74. Visscher, P. B., Renormalization-Group Derivation of Navier-Stokes Equation, J. Stat. Phys. 38, 989-
1013 (1985).
75. Kadanoff, L. P., McNamara, G. R., & Zanetti, G., From automata to fluid flow: comparisons of
simulation and theory, Phys. Rev. A 40, 4527-4541 (1989).
76. Hwa, T. & Kardar, M., Dissipative Transport in Open Systems: An Investigation of Self-Organized
Criticality, Phys. Rev. Lett. 62, 1813-1816 (1989).
77. Grinstein, G., Lee, D.-H., & Sachdev, S., Conservation Laws, Anisotropy, and "Self-Organized
Criticality" in Noisy Nonequilibrium Systems, Phys. Rev. Lett. 64, 1927-1930 (1990).
78. Sornette, D., Sweeping of an instability – An alternative to self-organized criticality to get power laws
without parameter tuning. Journal de Physique I, 4, 209-221 (1994).
79. Sykes, L. R., Shaw, B. E., & Scholz, C. H., Rethinking earthquake prediction, Pure and Applied
Geophysics 155, 207 (1999).
80. Carlson, J. M., Chayes, J. T., Grannan, E. R., & Swindle, G. H., Self-organized criticality and singular
diffusion, Phys. Rev. Lett. 65, 2547-2550 (1990).
81. Urbach, J. S., Madison, R. C., Markert, J. T., Interface Depinning, Self-Organized Criticality, and the
Barkhausen Effect, Phys. Rev. Lett. 75, 276-279 (1995)
82. Narayan, O., Self-Similar Barkhausen Noise in Magnetic Domain Wall Motion, Phys. Rev. Lett. 77,
3855-3857 (1996).
83. Zapperi, P., Cizeau, P., Durin, G., & Stanley, H. E., Dynamics of a ferromagnetic domain wall:
Avalanches, depinning transition, and the Barkhausen effect, Phys. Rev. B 58, 6353-6366 (1998)
84. Pazmandi F., Zarand G., & Zimanyi G. T., Self-organized criticality in the hysteresis of the
Sherrington-Kirkpatrick model, Phys. Rev. Lett. 83, 1034-1037 (1999).
85. Pazmandi F., Zarand G., & Zimanyi G. T., Self-organized criticality in the hysteresis of the
Sherrington-Kirkpatrick model, Physica B 275, 207-211 (2000).
86. Kuntz, M. C. & Sethna, J. P., Noise in disordered systems: The power spectrum and dynamic
exponents in avalanche models, Phys. Rev. B 62, 11699-11708 (2000).
87. Spasojević, D., Bukvić, S., Milošević, S., & Stanley, H. E., Barkhausen noise: Elementary signals,
power laws, and scaling relations, Phys. Rev. E 54, 2531-2546 (1996).
88. Perković, O., Dahmen, K. A., & Sethna, J. P., Disorder-Induced Critical Phenomena in Hysteresis: A
Numerical Scaling Analysis, cond-mat/9609072, appendix A.
89. Brézin E. & De Dominicis C., Dynamics versus replicas in the random field Ising model, Comptes
rendus d l’academie des sciences II fascicule b-mechanique physique astronimie 327, 383-390 (1999).
90. Cote, P.J. & Meisel, L.V., Self-organized criticality and the Barkhausen effect, Phys. Rev. Lett. 67,
1334-7 (1991).
91. Meisel, L.V. & Cote, P.J., Power laws, flicker noise, and the Barkhausen effect. Phys. Rev. B 46,
10822-8 (1992).
 28

92. Stierstadt, K. & Boeckh, W., Die Temperaturabhangigkeit des Magnetischen Barkhauseneffekts .3.
Die Sprunggrössenverteilung längs der Magnetisierungskurve, Z. Physik 186, 154 (1965).
93. deleted.
94. Bertotti, G., Durin, G., Magni, A., Scaling aspects of domain wall dynamics and Barkhausen effect in
ferromagnetic materials, J. Appl. Phys. 75, 5490-2, (1994).
95. Bertotti, G., Fiorillo, F., & Montorsi, A., The role of grain size in the magnetization process of soft
magnetic materials, J. Appl. Phys. 67, 5574-6, (1990).
96. Lieneweg, U., Barkhausen noise of 3% Si-Fe strips after plastic deformation, IEEE Trans. Magn. 10,
118-20, (1974).
97. Lieneweg, U. & Grosse-Nobis, W., Distribution of size and duration of Barkhausen pulses and energy
spectrum of Barkhausen noise investigated on 81% nickel-iron after heat treatment. Intern. J. Magnetism
3, 11-16, (1972).
98. Bittel, H., Noise of ferromagnetic materials, IEEE Transactions on Magnetics 5, 359-65, (1969).
99. Montalenti, G., Barkhausen noise in ferromagnetic materials, Zeitschrift fur Angewandte Physik 28,
295-300, (1970).
100. Durin, G. & Zapperi, S., Scaling Exponents for Barkhausen Avalanches in Polycrystalline and
Amorphous Ferromagnets, Phys. Rev. Lett. 84, 4705-4708 (2000).
101. Petta, J. R. & Weissmann, M. B., Barkhausen Pulse structure in an amorphous ferromagnet:
Characterization by high-order spectra, Phys. Rev. E 57, 6363-6369 (1998).
102. Alessandro, B., Beatrice, C., Bertotti, G., & Montorsi, A., Domain-wall dynamics and Barkhausen
effect in metallic ferromagnetic materials. 1. Theory, J. Appl. Phys. 68, 2901-2907.
103. Alessandro, B., Beatrice, C., Bertotti, G., & Montorsi, A., Domain-wall dynamics and Barkhausen
effect in metallic ferromagnetic materials. 2. Experiment 2908-2915 (1990).
104. Walsh, B., Austvold, S., & Proksch, R., Magnetic force microscopy of avalanche dynamics in
magnetic media, J. Appl. Phys. 84, 5709-5714 (1998).
105. Krysac, L. C. & Maynard, J. D., Evidence for the role of propagating stress waves during fracture,
Phys. Rev. Lett. 81, 4428-4431 (1998).
106. Family, F., Vicsek, T., & Meakin, P., Are Random Fractal Clusters Isotropic?, Phys. Rev. Lett. 55,
641-644 (1985).
107. Dotsenko, V. S., Windey, P., Harris, G., Marinari E., Martinec E., & Picco, M., Critical and
Topological Properties of Cluster Boundaries in the 3D Ising Model, Phys. Rev. Lett. 71, 811-814 (1993).
108. Kadanoff, L. P., Nagel, S. R., Wu, L., & Zhou, S.-M., Scaling and Universality in Avalanches, Phys.
Rev. A 39, 6524-6537 (1989).
109. Dhar, D., The Abelian Sandpile and Related Models, Physica A 263, 4-25 (1999).
110. Paczuski, M., Maslov, S., & Bak, P., Avalanche Dynamics in Evolution, Growth, and Depinning
Models, Phys. Rev. E 414-443 (1996).
111. Malcai, O., Lidar, D. A., Biham, O., & Avnir, D., Scaling Range and Cutoffs in Empirical Fractals,
Phys. Rev. E 56, 2817-2828 (1997).
112. Fleming, R. M. & Schneemeyer, L. F., Observation of a Pulse-Duration Memory Effect in
K0.30MoO3, Phys. Rev. Lett. 33, 2930-29321(1986).
113. Coppersmith, S. N. & Littlewood, P. B., Pulse-Duration Memory Effect and Deformable Charge-
Density Waves, Phys. Rev. B 36, 311-317 (1987).
114. Middleton, A. A., Asymptotic Uniqueness of the Sliding State for Charge-Density Waves, Phys.
Rev. Lett. 68, 670-673 (1992).
115. Amengual, A., Mañosa, Ll., Marco, F., Picornell C., Segui, C., & Torra, V., Systematic Study of the
Martensitic Transformation in a Cu-Zn-Al Alloy – Reversibility versus Irreversibility Via Acoustic
Emission, Thermochimica Acta 116, 195-308 (1987).
116. Perković, O. & Sethna, J. P., Improved magnetic information storage using return-point memory, J.
Appl. Phys. 81, 1590-1597 (1997).
117. Pepke, S. L., Carlson, J. M., & Shaw, B. E., Prediction of Large Events on a Dynamical Model of a
Fault, J. Geophysical Research 99, 6769 (1994).

Acknowledgements

Much of the work reviewed in this paper grew out of a collaboration with Matt Kuntz2. We thank Amit
Mehta for supplying the data for Figure 7b. We’d like to thank Drew Dolgert, Mark Newman, Jean-
Phillipe Bouchaud, L. Cindy Krysac, Daniel Fisher, and Jim Thorpe for helpful comments and references.
 29

This work was supported by NSF grants DMR 9805422, 9873214, 00-72783, 99-76550, ASC-9523481,
the Cornell Theory Center (through NSF #9972853) and IBM.