BULLETIN (New Series) OF THE

AMERICAN MATHEMATICAL SOCIETY

Volume 44, Number 4, October 2007, Pages 603–621
S 0273-0979(07)01184-6
Article electronically published on July 5, 2007

ON THE EULER EQUATIONS OF INCOMPRESSIBLE FLUIDS

PETER CONSTANTIN

Abstract. Euler equations of incompressible fluids use and enrich many

branches of mathematics, from integrable systems to geometric analysis. They
present important open physical and mathematical problems. Examples in-
clude the stable statistical behavior of ill-posed free surface problems such as
the Rayleigh-Taylor and Kelvin-Helmholtz instabilities. The paper describes
some of the open problems related to the incompressible Euler equations, with
emphasis on the blowup problem, the inviscid limit and anomalous dissipation.
Some of the recent results on the quasigeostrophic model are also mentioned.

1. Introduction
Euler’s equations for incompressible fluids, like number theory, are the well-
spring of many mathematical streams. Linear partial differential equations (PDEs,
henceforth), spectral theory, dynamical systems, nonlinear PDEs, geometric PDEs,
harmonic analysis, completely integrable systems find in the Euler equations source,
challenge and inspiration.
Euler had been involved in acoustics, hydrostatics and hydraulics research for
many years by the time he wrote his treatises on fluids ([77]). After the efforts of
the Bernoullis and D’Alembert, Euler’s work represented a crowning manifestation
of the eighteenth century’s confidence in the mathematical foundations of the laws
of nature. The equations are concise and capture in an idealized fashion the essence
of fluid behavior.
No full account of the mathematical activity surrounding the Euler equations
since their inception to this day can be attempted in a few pages. A complete
description of their impact outside mathematics, from weather prediction to ex-
ploding supernovae, would fill volumes. There are books ([6], [28], [33], [105], [107])
and expository articles ([10], [42], [80]) on the subject, too numerous to all be listed
here. My own biases notwithstanding, I would like to be able to give a glimpse
of some of the current activities surrounding the Euler equations and of the major
directions needing further progress.
I start by casting a very wide net and describing concisely some of the areas of
current research. The Euler equations cannot be separated from the surrounding
scientific topics. Exposure to the broader physical context provides the mathe-
matical researcher with both vision and power. Vision because it directs research

Received by the editors May 31, 2007.

2000 Mathematics Subject Classification. Primary 76B47; Secondary 35Q30.
Key words and phrases. Euler, anomalous dissipation, blowup, free surface, quasigeostrophic
equation, vanishing viscosity, blowup, ill-posed, turbulence.

2007
c American Mathematical Society

603
604 PETER CONSTANTIN

towards fertile ground, and power because it gives the opportunity of invention
of new technical tools. Therefore, in the next few pages I will present the Euler
equations in connection with physical problems surrounding them.
Waves and jets and drops come to mind when thinking of fluids. These objects
turn out to be mathematically the most challenging, because they involve one or two
different fluids separated by an unknown (“free”) surface. In the PDE community
these problems go by the name of “free boundary problems”. Examples include wa-
ter waves, Hele-Shaw cells (oil/water interfaces in nearly two-dimensional settings),
drop formation, thin films, flame fronts and more. The subject of hydrodynamic
free boundary problems is vast: it includes systems that are completely integrable,
nonlinear dispersive equations and stochastic models of front propagation. Imag-
ine for instance that a very idealized description of the surface of a body of water
is given by one function, h, the height of the water. No matter how simple, this
function will have to depend at least on two variables: a position variable x and
a time variable t. The evolution in time of the function h will be dictated by the
Euler equations for the fluids coupled with boundary conditions for the interface.
It is sometimes possible to reduce all this to an equation for h, expressed in terms
of h alone. Insisting upon such economy of the unknown comes at a price, and this
equation is complicated: it is nonlinear and integro-differential (it involves simul-
taneously derivatives and integrals of h). Simplifications (neglecting some terms,
approximating some others) lead to famous equations of water waves, such as the
Korteweg de Vries equation. Murky modeling yields a pure mathematical treasure:
completely integrable PDEs. Integrable PDEs are a source of significant mathe-
matical developments that are not limited to fluid mechanics nor to PDE theory
([131]), but are outside the scope of this presentation.
Integrability in PDEs requires the ability to produce solution formulas depending
on infinitely many parameters, providing the “general” solution. The well-posedness
of a PDE requires the ability to solve the equation and prove the continuity of
the solution with respect to initial conditions in an appropriate functional setting.
It might seem that integrability would imply well-posedness, but it does not in
general, and important examples of integrable free surface equations (Hele-Shaw)
are ill-posed. Integrability of two-dimensional Euler equations with free surface is
a current subject of research ([132]). The modern well-posedness theory for free
surfaces uses harmonic analysis and geometric PDEs ([3], [4], [35], [63], [65], [66],
[102], [126], [127]) and is a subject in full swing.
Problems of hydrodynamic stability ([9], [83], [124], [129]) are of classical im-
portance. The classical work concerns the study of the evolution of small pertur-
bations of special, time independent solutions, usually by linearization. The linear
equations obtained produce examples of nonnormal operators. The more modern
theory studies the growth in time of the distance between solutions of the nonlin-
ear equations, i.e. nonlinear instability. Instability of the solution is different from
ill-posedness; actually it does not make much sense to speak of nonlinear instability
of a solution without having a well-posed evolution. Ill-posedness usually means
that the small scale features of the perturbation grow faster than the large scale
features, and there is no cutoff in the time scales as we go to finer and finer spatial
scales. This catastrophic growth prevents the continuity of the solution map in all
spaces except spaces of analytic functions. Unfortunately, the distinction between
 EULER EQUATIONS OF INCOMPRESSIBLE FLUIDS 605

well-posed instability and ill-posed catastrophic growth is not made in the clas-
sical physics literature, and the meanings of “instability” are multiple, including
ill-posedness (in the Rayleigh-Taylor and Kelvin-Helmholtz instabilities).
Solutions of Euler equations might seem more unstable than they really are, or
to be more precise, the notion of stability appropriate for them is a more generous
one, that of orbital stability. An example of this nuance is the case of Kirchhoff
ellipses, which are special solutions of two-dimensional Euler equations. These are
ellipses that rotate at constant angular velocity, proportional to the area of the
ellipse. If two ellipses have close but unequal areas, they will become, at some
point in time, very different from each other. This should not imply automatically
that the evolution is unstable; it just means that one should “mod-out” by the
constant rotation and stay on the “leaf” defined by a fixed value of the area (which
happens to be an invariant of the equation).
In the context of free surface flows, the main puzzle is a sort of “stability of
instability”: characteristic, predictable features arise in the physical realization of
ill-posed classical instabilities such as the Rayleigh-Taylor and Kelvin-Helmholtz
instabilities. For instance, characteristic plumes appear and grow in the Rayleigh-
Taylor instability (heavy fluid on top of light fluid) and generate a mixing layer that
grows at a predictable rate. Also, the eye easily discerns the typical roll-up that
occurs in wind-driven interfaces in the Kelvin-Helmholtz instability in clouds and
waves (think of the famous Hokusai wave). Modern research established recently
the elliptic nature of these ill-posed problems ([90], [97]): if solutions have a certain
minimal regularity, then they are actually analytic. The “stability of instability”
has yet to find a good mathematical explanation. Scaling ([11]) is the key aspect in
this problem. Perhaps generalized relative entropies ([49], [110]) might be relevant
for this.
Mixing and transport are subjects of great practical significance, as they relate to
the diffusion of passive and active tracers, for instance pollutants in the atmosphere
or plankton in the oceans. The mathematical issue is to understand the relation-
ship between the underlying particle trajectory dynamics and macroscopic mixing
properties of the flow. Classical methods of dynamical systems are relevant for the
long-time effects ([118]). Strong rapid effects are studied using PDE methods ([17],
[50], [55]).
When mixing and transport are studied for small scale particles, it is necessary to
consider the dissipation of kinetic energy due to internal molecular processes. The
Newtonian stress balance gives rise to the Navier-Stokes equations in which the
dissipation of kinetic energy is represented macroscopically by the addition to the
Euler equation of a Laplacian term multiplied by a positive coefficient, the kinematic
viscosity. Complex fluids are fluids in which microscopic particles are suspended,
altering the Newtonian stress balance and conferring new physical properties to the
fluid.
The applications of complex fluids range from biology to materials science. PDE
models include non-Newtonian viscoelastic models like the Oldroyd-B equations,
tensor models, and kinetic models, in which Navier-Stokes equations are coupled to
linear or nonlinear Fokker-Planck equations. The well-posedness theory is difficult
even in two space dimensions, and consequently the mathematical theory of complex
fluids is in its developing stages ([43], [48], [52], [76], [89], [98], [100], [101], [103],
[115]). Some of the models of complex fluids involve stochastic PDEs, or hybrid
606 PETER CONSTANTIN

systems, in which PDEs are coupled to stochastic differential equations. These are
nonlinear systems with many degrees of freedom subjected to thermal noise. The
mathematical theory here is also developing. The numerical analysis of these models
leads naturally to questions of dimension reduction and effective algorithms that
are very much in the forefront of current applied mathematical research. Stochastic
models also have applications to modeling based on information theoretical ideas
in geophysics and atmospheric science ([106]).
In this paper I will concentrate on a few specific problems related to the incom-
pressible Euler equations,
(1) ∂t u + u · ∇u + ∇p = 0,
with
(2) ∇ · u = 0.
A few words

about notation: u = (uj ) is the velocity vector with n components;
u · ∇ = nj=1 uj ∂j is a first order differential operator with coefficients uj . The
velocity components are functions uj = uj (x, t) where x ∈ Rn and t ∈ R. The
∂ ∂
notations ∂t = ∂t , ∂j = ∂x represent partial derivatives. The equation (2) is the
j
n
condition of incompressibility: the divergence ∇ · u = j=1 ∂j uj vanishes. The
scalar function p = p(x, t) whose gradient ∇p appears in (1) is called the pressure.
The pressure is determined up to the addition of a term that is constant in space
but not necessarily in time. The equations are nonlinear, which is to say, they
look nonlinear to the naked eye, with quadratic nonlinearity. (Some completely
integrable equations look the same and, even more, are derived from them.) The
Euler equations are also nonlocal. That means that one cannot compute the time
derivative of the solution u at (x, t) only from knowledge of the function u in a
neighborhood of x at time t. Taking the divergence of (1) we see that
(3) −∆p = ∇ · (u · ∇u) .
This is the reason for the nonlocality: the Laplacian ∆p is determined by local
information about u, and ∇p is not.
The equations are covariant under Galilean transformations: if the coordinates
move at constant speed y = x + vt, then the velocity in the y reference frame is
u(x, t) + v. If the coordinates are rotated by a fixed angle, then the velocities are
rotated by the same angle. The equation (1) is the Eulerian description of the flow.
The equations are in a sense hyperbolic: information is carried by the flow, to some
extent. This is seen in the Lagrangian description (due also to Euler). We consider
time dependent maps X : Rn → Rn
a → X(a, t), X(a, 0) = a.
These maps represent marked fluid particle trajectories; a is the label of the particle.
It is customary to take this label to be the position of the particle at the beginning
of the observation (t = 0). The fact that particles travel with velocity u is expressed
in the system of ordinary differential equations
(4) ∂t X = u(X, t).
The Lagrangian formulation of (1) is Newton’s second law
(5) ∂t2 X + (∇x p)(X, t) = 0,
 EULER EQUATIONS OF INCOMPRESSIBLE FLUIDS 607

and the Lagrangian form of (2) is

(6) det (∇a X) = 1.
Incompressibility refers to the fact that the maps a → X(a, t) are volume preserv-
ing (in view of (6)). The divergence-free condition (2) is the Eulerian equivalent
counterpart. (As is easily seen, if everything is smooth, by differentiating (4)). The
Lagrangian formulation involves the Eulerian gradient of the pressure ∇x p which
could be computed using (3).
The system describes with great simplification the time evolution of an incom-
pressible fluid of constant density in the absence of any forces other than those
arising from the incompressibility constraint. Considering internal processes that
lead to energy dissipation leads to the Navier-Stokes equations,
(7) ∂t u + u · ∇u − ν∆u + ∇p = 0,
with the same incompressibility constraint (2). Here ν > 0 represents the kinematic
viscosity, which is a constant (at constant temperature). The introduction of a
second order operator requires (and permits) boundary conditions; these are the
essential physical ingredient introduced by the presence of the Laplacian. The
Navier-Stokes equations, in the presence of forces and boundary conditions, are a
good description of Newtonian fluids and, when additional stresses are introduced,
of non-Newtonian fluids.

2. The blowup problem

2.1. Dimension and geometry. The blowup problem is: Can singularities arise
in finite time from smooth initial velocities with finite kinetic energy?
The blowup problem for the Euler equations is a major open problem of PDE
theory, of far greater physical importance than the blowup problem for the Navier-
Stokes equation, which of course is known to nonspecialists because it is a Clay
Millennium Problem. There is a deep connection between the two problems, but
there are major differences as well. The blowup problem for the Euler equation
concerns growth of velocity gradients of solutions and is not tamed by putting an
upper bound on the velocity itself: an Euler equation singularity does not require
infinite momentum, whereas a Navier-Stokes blowup does. (Interesting criteria for
absence of inviscid hydrodynamic blowup based on momentum do exist, but they
are for limited configurations ([64]).) The occurrence of an infinite momentum from
finite smooth initial data, without external forcing, with zero or periodic bound-
ary conditions and with positive kinematic viscosity has no physical justification.
Obviously, continuum mechanics breaks down at very small scales, so there is no
doubt that the equations do not hold at all scales. While this might be the case, it
is not the question. The question, not to put too fine a point on it, is: do we need
Schrödinger’s equations to calculate the flow around a moving car? Or to predict
tomorrow’s weather? So far, no calculation of the purely decaying Navier-Stokes
equation suggested the need for a cutoff at fixed viscosity. There are many ways one
can regularize the Euler equations, and the kinematic viscosity might not be the
only regularizing mechanism. On the other hand, if regularity for Navier-Stokes
equations is to be proven, then the proof cannot be perturbative and new ideas
are needed. Indeed, writing the Navier-Stokes equations as a perturbed heat equa-
tion is useful for small data arguments, but when the nonlinear term dominates,
608 PETER CONSTANTIN

this approach ceases to be sufficient. New methods that are nonlinear, solution-
dependent, need to be invented. They are to be sought in the study of the Euler
equations. In fact, growth of gradients in Euler equation fashion might lead to
vortex reconnection and regularity in Navier-Stokes equations ([37]). Contrary to
infinite momentum singularities, production of very large gradients is physically im-
portant as it is related to the anomalous dissipation of energy, a well documented
experimental fact. It is known that if there are no singularities in the solution of the
Euler equations with initial data u0 on the time interval [0, T ], then there can be
no singularities in the Navier-Stokes solution with the same initial data and small
enough viscosity ([36]). The regularity for large enough viscosities is also known.
Unfortunately, there is a gap between the two ranges of viscosities, and it is not
clear how to close it.
The solutions of Euler equations do not blow up from smooth, finite kinetic
energy initial data when the dimension of space n = 2 ([8]). When n = 3, the
problem is open. The easiest way to see the difference between the two situations
is to look at the vorticity, curl of velocity ω = ∇ × u in three dimensions. In two
dimensions, ω = ∇⊥ · u. Taking the curl of (1) we obtain the vorticity evolution
equation

(8) ∂t ω + u · ∇ω − ω · ∇u = 0.

In n = 2 term ω · ∇u vanishes identically and the vorticity equation is a transport

equation

(9) ∂t ω + u · ∇ω = 0.

In two dimensions it follows then that

(10) ω(x, t) = ω0 (A(x, t))

where ω0 is the curl of the initial data and A(x, t) = X −1 (x, t) is the “back-to-
labels” map: the inverse of the Lagrangian path trajectory a → X(a, t). Be-
cause the back-to-labels map preserves volume, it follows that ω(x, t)Lp (dx) =
ω0 Lp (dx) holds for any 1 ≤ p ≤ ∞ in n = 2. (Lp (dx) is the usual Lebesgue
space.) This quantitative information was extracted from the system because of a
cancellation, without using quantitative information about A(x, t). It is a free gift.
Once the magnitude of the vorticity is controlled, persistence of smoothness follows
with a little analysis.
The analogue of (10) in three dimensions is

(11) ω(x, t) = (∇a X) (A(x, t), t)ω0 (A(x, t))

which expresses the fact that the integral curves of the vorticity, the vortex lines,
are carried by the flow. The two-dimensional vorticity is carried along by particle
paths, its magnitude unchanged. The three-dimensional vorticity is carried as well,
but its magnitude is amplified or diminished by the gradient of the flow map. If
one allows for infinite kinetic energy solutions, then one can find blowup ([32], [39],
[85], [86], [113], [122]). If one considers complex solutions, then again one can find
blowup ([22]). Unfortunately, the infinite energy blowup also occurs in 2D, where
in fact finite energy solutions do not blow up.
 EULER EQUATIONS OF INCOMPRESSIBLE FLUIDS 609

The Beale-Kato-Majda ([15]) criterion (see also ([94]) for an extension) says that
the time integral of the maximum magnitude of the vorticity
 T
ω(t)L∞ (dx) dt
0

controls blowup or its absence. If the integral is finite and if the initial velocity is
in a Sobolev space H s with large enough exponent (s > 5/2) or in a C s space with
s > 1 (and some decay in physical space, for instance ω0 ∈ Lp with p > 1), then
the solution remains smooth on the time interval [0, T ]. Of course, if the integral
is infinite, then there is finite time blowup.
The Euler equations possess local-in-time (i.e. for small T ) unique smooth solu-
tions if the initial velocity is in the spaces described above. These function spaces
are Banach algebras for the gradients.
Because the magnitude of vorticity controls the blowup, it is useful to look more
closely at its evolution:
(12) (∂t + u · ∇) |ω| = α|ω|.
The scalar quantity α(x, t) is the vortex stretching factor, and obviously, by Gron-
T
wall’s lemma, when 0 α(t)L∞ dt is finite then the Beale-Kato-Majda criterion
implies that no singularities are generated spontaneously in the time interval [0, T ].
If α ≥ c|ω| holds for a long enough period of time, with c > 0, then blowup ensues.
It turns out ([37]) that α has a representation

3 dy
(13) α(x, t) = P.V. D(
y , ξ(x + y, t), ξ(x, t))|ω(x + y, t)| 3
4π R 3 |y|
where
D(e1 , e2 , e3 ) = (e1 · e3 ) det(e1 , e2 , e3 )
ω(x,t)
and ξ(x, t) = |ω(x,t)| . The integral representing α is a principal value type integral,
and consequently the dimensional analysis of the equation would predict blowup
along characteristics. But the representation leads to criteria for absence of blowup
([46], [47]) that are based on the observation that if the direction field ξ is regular,
then the integral in (13) is not of principal value type any more, and in the case
of the Navier-Stokes equations that is enough to rule out blowup. The regularity
of ξ in two dimensions is clear: ξ = (0, 0, 1) in some system of coordinates. The
fact ξ has to have some roughness in order for blowup to be possible indicates some
necessary complexity of the underlying geometric support of blowup. Moreover,
single-scale self-similar behavior is impossible ([25]).
The criteria based on ξ have been extended and refined ([19], [69]) and validated
numerically ([87]). There exists an extensive numerical literature on the subject of
finite time singularities in the Euler equations, but the problem remains undecided
numerically as well.

2.2. Conserved quantities. From past experience with nonlinear PDE of physical
origin, it is imperative to take advantage of conserved quantities, the “free gifts”.
The smooth solutions of the Euler equations conserve kinetic energy,
 
|u(x, t)|2 dx = |u0 (x)|2 dx,
R3
610 PETER CONSTANTIN

helicity ([111])
 
u(x, t) · ω(x, t)dx = u0 (x) · ω0 (x)dx,
R3 R3
and circulation  
u(x, t) · dx = u0 (a) · da.
X(γ,t) γ

In two space dimensions, additional conservation laws are integrals of arbitrary

functions of the vorticity. The circulation is the integral of the form u · dx on
arbitrary closed loops transported by the flow. The helicity and circulation conser-
vation laws have not been sufficiently exploited for analysis. The helicity encodes
topological information about the average linking number between vortex lines.
The circulation provides a connection between particle path length elements and
momentum.
The conservation of energy follows easily from the form (1) of the equations.
The conservation of circulation follows from the Weber formula (see ([40], [42]) and
references therein)
(14) u(x, t) = P [(∇A)∗ (u0 ◦ A)] .
Here P is the Leray-Hodge projector on divergence-free functions and (∇A)∗ is the
transposed matrix. Conservation of helicity follows also from the Weber formula.
The Euler equations are expressed succinctly as a system of active scalars
(15) ∂t A + u · ∇A = 0
with u given from A by (14). In this formulation, the conservation of circulation
and of helicity are transparent. The Navier-Stokes equations have a deterministic
parabolic formulation ([41]) that resembles (14), (15). A stochastic formulation
based on (14), (15) and on seminal ideas of ([34]) was obtained as well ([88]).

2.3. Weak solutions. Because singularities cannot be ruled out, and because
sometimes it makes physical sense to admit solutions with singularities in them,
the notion of solution can be extended to that of weak solution. These solve a
weak form of the equation in the distribution sense, placing the equations in large
spaces using duality. The Euler and Navier-Stokes equations are in divergence-
form, and integration by parts moves derivatives on test functions. The notions of
weak solution depend on the equation and need to be handled with care.
The Navier-Stokes equations possess global weak solutions in 3D ([99]), and their
singularities, if any, are confined to a space-time set of dimension less than 1 ([20]).
Their uniqueness is not known.
There is no similar notion of weak solution of Euler equations in 3D. Nonunique-
ness of putative weak solutions of the 3D Euler equations has been demonstrated
([67], [119], [121]). For 2D Euler there exist unique weak solutions ([130]) if the
vorticity is bounded. The existence (but not the uniqueness) of weak solutions is
established for ω ∈ Lp (R2 ) for 1 < p < ∞, and a framework for measure-valued
solutions exists ([71], [72]). The existence of weak solutions in the limit case when
the initial vorticity is a positive Radon measure with velocity locally uniformly in
L2 (positive vortex sheet data) was established in ([68]). The vortex sheet problem
itself is the problem of keeping track of the evolution of the support of the mea-
sure if initially this was a smooth curve. Although there is a simple equation for
 EULER EQUATIONS OF INCOMPRESSIBLE FLUIDS 611

the evolution of such a curve, the Birkhoff-Rott equation, the problem is ill-posed
([128]).
Among the two-dimensional problems that attracted recently renewed interest
is the surface quasi-geostrophic active scalar (QG) ([37], [53]). This is a model of
geophysical origin, but it has been studied mathematically for more than a decade,
mainly because of its similarity to the 3D Euler equations and because, as a scalar
model in 2D, it is more amenable to numerical simulations then the full 3D Euler
equation. The equation is deceptively simple,

(16) ∂t θ + u · ∇θ = 0

with

(17) u = R⊥ θ

where R = ∇Λ−1 are Riesz transforms, Λ = (−∆) 2 the Zygmund operator. The
1

analogy to the Euler equations is in the fact that the level sets of θ are like vortex
lines. The analogue of the 3D vortex equation is the equation for ∇⊥ θ, a vector
field that is tangent to the level sets. This vector field obeys (8) with ω replaced
by ∇⊥ θ. Because θ is carried by the flow it creates, it is expressed in terms of the
back-to-labels map as θ(x, t) = θ0 (A(x, t)). The analogue of (11) for ∇⊥ θ holds in
the same form,

(∇⊥ θ)(x, t) = (∇a X) (A(x, t), t)(∇⊥ θ0 )(A(x, t)).

There is a BKM theorem that establishes that finiteness of

 T
∇⊥ θ(t)L∞ (dx) dt
0

controls absence of blowup. The magnitude |∇⊥ θ| obeys the same equation as
(12), and the stretching factor α has a principal value representation like (13) that
reveals the same geometric depletion if the direction field ξ = ∇⊥ θ/|∇⊥ θ| is locally
regular. Remarkably, there exist weak solutions in L2 ([116]). There is numerical
([114]) and theoretical ([61]) evidence that no blowup occurs, but the problem is
still open.
Recent new developments have been achieved in dissipative forms of the equation

(18) ∂t θ + u · ∇θ + κΛs θ = 0,

with κ > 0, 0 < s. If s > 1 we say that the equation is subcritical, if s < 1 the
equation is supercritical, and if s = 1, critical. When s = 1 the model is physical:
the dissipation represents friction with boundaries. The subcritical equations have
smooth solutions ([27], [58]). In the critical case, it was known for some time that
if the initial data are small in L∞ they remain small and the solution is regular
([44], [62]). Recently, two remarkable and quite different proofs of global existence
in the critical case have been obtained. In one ([92]) the result is one of global
persistence of regularity, based on a new and promising idea, a maximum modulus
of continuity principle. The other proof ([21]) uses harmonic extension to prove a
gain of regularity of weak solutions, in the spirit of the De Giorgi, from L2 to L∞ ,
from L∞ to Hölder continuous, and beyond. Work on the supercritical case along
similar lines is in progress ([59], [60]).
612 PETER CONSTANTIN

3. The vanishing viscosity limit and anomalous dissipation

3.1. Singular limits and boundary layers. The vanishing viscosity limit of solu-
tions of Navier-Stokes equations is a singular limit (that means that the type of the
equation threatens to change in the limit: the highest order derivative is multiplied
by a vanishing coefficient). Singular limits are ubiquitous in applied mathematics
and correspond to physical reality. In bounded domains Ω, the fact that the vanish-
ing viscosity limit is a singular limit is most vividly manifested in boundary layers.
The Navier-Stokes equations are second–order differential equations and require
boundary conditions, for instance homogeneous Dirichlet boundary conditions,
u| ∂Ω = 0.
For the Euler equation the only requirement is that the particle paths be tangent
to the boundary
u · N| ∂Ω = 0
where N is the normal to the boundary. In the limit of zero viscosity there are thin
layers near the boundary where the change from a viscosity-dominated behavior to a
nonviscous behavior takes place, or at least that is the physical picture in quiescent
situations near flat boundaries. Boundary layers are very real but difficult to study
rigorously, especially if the domains are curved and the flows are time-dependent
([23], [24], [74], [75]). At present the problem of the vanishing viscosity limit is
open even in two space dimensions, in bounded domains. The physical problem of
boundary layers, for instance in wall bounded flow, has a long history ([120]), but
the classical Prandtl boundary layer theory has been challenged ([12], [13]).
More progress has been made in the study of the limit when boundaries are
absent (flow in Rn or Tn ). Even in this restricted situation, there are two distinct
concepts of vanishing viscosity limit. The finite–time, zero–viscosity limit is the
limit limν→0 S ν (t)(ω0 ) of solutions S ν (t)(ω0 ) of the Navier-Stokes equations with
a fixed initial datum ω0 and with time t in some finite interval [0, T ]. By contrast,
in the infinite–time, zero–viscosity limit, long–time averages of functionals of the
solutions  
1 t
lim Φ(S (s)ω0 )ds = Φ(ω)dµν (ω)
ν
t→∞ t 0

are considered first, at fixed ν. These are represented by measures µν in function

space. The long–time, zero–viscosity limit is then limν→0 µν ,
  
1 t ν
lim lim Φ(S (s)ω0 )ds .
ν→0 t→∞ t 0

The two kinds of limits are not the same. This is most clearly seen in the situation
of two-dimensional, unforced Navier-Stokes equations. In this case, any smooth
solution of the Euler equations is a finite–time inviscid limit, but the infinite–time
inviscid limit is unique: it is the function identically equal to zero. This simple
example points out the fact that the infinite–time, zero–viscosity limit is more
selective. In less simple situations, when the Navier-Stokes equations are forced,
the long–time inviscid limit is not well understood. The finite–time, zero–viscosity
limit is the limit that has been most studied. For smooth solutions in R3 , the
zero–viscosity limit is given by solutions of the Euler equations, for short time, in
classical ([123]) and Sobolev ([91]) spaces; the limit holds for as long as the Euler
solution is smooth ([36]). The convergence occurs in the Sobolev space H s as long
 EULER EQUATIONS OF INCOMPRESSIBLE FLUIDS 613

as the solution remains in the same space ([108]). The rates of convergence are
optimal in the smooth regime, O(ν). In some nonsmooth regimes (smooth vortex
patches), the finite time inviscid limit exists and optimal rates of convergence for
the velocity can be obtained ([1], [108]), but the rates (for the vorticity) deteriorate
when the smoothness of the initial data deteriorates – for nonsmooth vortex patches
([56], [57]).
3.2. Anomalous dissipation. One of the most fundamental questions concerning
the inviscid limit is: what happens to ideally conserved quantities?
We start by discussing this problem for two-dimensional flows where the anal-
ysis is simpler. Two-dimensional flows are relevant for atmospheric flow, which is
nearly incompressible and nearly two-dimensional. In n = 2 one of the important
conserved quantities for smooth solutions of the Euler equations (one of the “free
gifts”) is the enstrophy, 
|ω(x, t)|2 dx.
R2
When we solve the unforced Navier-Stokes equations and the initial vorticity is in
L2 , then  
d 2 2
|ω(x, t)| dx + ν |∇ω(x, t)| dx = 0
2dt R2 R 2

and the quantity 

ν |∇ω(x, t)|2 dx
R2
is the instantaneous rate of dissipation of enstrophy. If the gradients are bounded
in L2 , then this quantity vanishes as ν → 0. But of course, as ω depends on ν and
as the gradients of vorticity might be large, the limit might be nonzero. The term
“anomalous dissipation”, favored by physicists with field-theoretical background,
refers to the case when the rate of dissipation has a nonzero limit, in the zero
viscosity limit. The existence of anomalous dissipation of enstrophy is essential
in Kraichnan’s theory of two-dimensional turbulence ([84], [95]). Bounds on the
dissipation of enstrophy in physical terms were given in ([2]). Anomalous dissipation
of enstrophy was studied in the framework of finite–time inviscid limits with rough
initial data ([79], [104]). It was established that if the initial vorticity belongs to
L2 (R2 ), then weak solutions of the Euler equations conserve enstrophy, and that
implies absence of anomalous dissipation of enstrophy for finite time.
The physical context for anomalous dissipation is, however, turbulence theory,
and that refers to the long–time limit, in the presence of forces.
For long–time behavior, in the case of forced two-dimensional Navier-Stokes
equations, there is a difficulty due to the inverse cascade of kinetic energy: the
energy piles up at the largest scale and grows without bound. In numerical sim-
ulations and in experimental studies, energy injected at small scales is transferred
to large scales (from high wave numbers to low wave numbers). Physically, this
corresponds to the coalescence of small vortices into large ones. There cannot exist
a steady vanishing viscosity state when the forcing does not vanish with viscosity
([96]).
Wave-number independent damping is used to model friction with boundaries
and a variety of other physical mechanisms responsible for attenuating large scale
turbulent energy. For example, a well-known model, the Charney-Stommel model
of the Gulf Stream ([7]), is a two-dimensional, damped and driven Euler system.
614 PETER CONSTANTIN

Numerical studies of two-dimensional turbulence employ devices to remove the

energy that piles up at the large scales, and damping is the most common such
device. D. Bernard ([18]) suggested, on the basis of field–theoretical computations,
that there is no anomalous dissipation of enstrophy in damped and driven two-
dimensional Navier-Stokes equations.
This was recently proved rigorously in ([54]). A first result concerns sequences
of time–independent individual solutions. The sequences have enough compactness
to pass to convergent subsequences. The resulting solution is a weak solution of
the damped and driven Euler equations. The existence of weak solutions of such
equations in the case of the Charney-Stommel model was first obtained in ([7]). The
weak solution of the damped and driven Euler equation is a renormalized solution
in the sense of ([70]). This implies that the weak solution obeys an enstrophy
balance, and that is used to show that there is no anomalous dissipation. The
time–dependent situation is more complicated. Stationary statistical solutions of
the damped and driven Navier-Stokes equations in the spirit of ([81], [82], [125])
need to be considered. In the case of finite–dimensional dynamical systems
dω
= N (ω),
dt
invariant measures µ are Borel measures in the ω phase space such that
∇ω Ψ(ω)N (ω)dµ(ω) = 0 for any test function Ψ. In infinite dimensions, we need
to restrict the test functions to a limited class T of admissible functions, but the
notion of a stationary statistical solution of the damped and driven Navier-Stokes
equation is a natural extension of the notion of invariant measure for deterministic
finite-dimensional dynamical systems.
Weak limits of stationary statistical solutions of the damped and driven Navier-
Stokes equations are renormalized stationary statistical solutions of the damped and
driven Euler equations, in the spirit of ([70]). Stationary statistical solutions µν
of the damped and driven Navier-Stokes equations can be obtained by the Krylov-
Bogoliubov procedure of taking long–time averages. These solutions have good
enough properties so that their weak limits are renormalized stationary statistical
solutions µ0 of the damped and driven Euler equations that obey the enstrophy
balance law (expected formally from the damped and driven Euler equation by
taking long–time averages). This is the central fact that allows one to prove that
the zero viscosity limit of the long–time average enstrophy dissipation rate vanishes:
  
1 t
lim ν lim sup ∇ω(s + t0 )L2 (R2 ) ds = 0
2
ν→0 t→∞ t 0

holds for all solutions ω(t) = S N S,γ (t)(ω0 ) of the damped and driven Navier-Stokes
equation with fixed damping coefficient γ, all t0 > 0, and all ω0 ∈ Lp (R2 )∩L∞ (R2 ).
The convergence in this class of statistical solutions is such that
 
lim ω2L2 (R2 ) dµν (ω) = ω2L2 (R2 ) dµ0 (ω).
ν→0 L2 (R2 ) L2 (R2 )

The problems of anomalous dissipation in 3D are substantially more difficult.

The ideally conserved quantity under investigation is the kinetic energy,

|u(x, t)|2 dx.
R3
 EULER EQUATIONS OF INCOMPRESSIBLE FLUIDS 615

The kinetic energy is conserved by smooth solutions of the Euler equations. For
smooth solutions of unforced Navier-Stokes equations, the kinetic energy decays,
 
d 2 2
|u(x, t)| dx + ν |∇u(x, t)| dx = 0,
2dt R3 R 3

and 
ν |∇u(x, t)|2 dx
R3
is the instantaneous rate of dissipation of kinetic energy. There is experimental and
numerical evidence that the rate of dissipation of kinetic energy in Navier-Stokes
equations is bounded away from zero, even at very high Reynolds numbers (small
viscosities). Turbulence theory and practice is rather solidly anchored in this fact
([80], [84], [117]). The famous Kolmogorov-Obukhov ([93]) power law for the energy
spectrum of turbulent fluctuations is
2/3 −5/3
E(k) = C k
where > 0 is the nonzero rate of dissipation of kinetic energy,
2
= ν |∇u|
where . . . is the expected value with respect to an invariant measure supported on
solutions of the Navier-Stokes equations at high Reynolds number. The fact that
is bounded below, independently of viscosity, is a fundamental part of the theory.
The wave number magnitude k belongs to a range of values called the inertial
range that extends to positive infinity in the limit of infinite Reynolds number (the
vanishing viscosity limit). Independently of Kolmogorov, Onsager ([112]) suggested
that there exists (in modern language) anomalous dissipation of energy and that it
is supported by weak solutions of the Euler equations. Onsager conjectured that
weak solutions of the Euler equation with Hölder continuity exponent h > 1/3 do
conserve energy and that turbulent or anomalous dissipation occurs when h ≤ 1/3.
The Hölder exponent 1/3 corresponds precisely to the Kolmogorov-Obukhov energy
spectrum exponent −5/3. More specifically
p
p
((u(x + re) − u(x)) · e) ∼ ( r) 3
for p = 2 is the Kolomogorov 2/3 law ([93]); the Kolmogorov-Obukhov energy
spectrum was derived from it by dimensional analysis. In the expression above e is
a unit vector. Corrections to the Kolmogorov scaling for high p are known in the
turbulence community as “anomalous scaling”. The 2/3 law was challenged ([16])
and defended ([14]). Anomalous scaling has been extensively examined numerically
and theoretically (but not rigorously) ([84]), but there are competing explanations
of the data ([5]).
One direction of the Onsager conjecture was addressed mathematically ([45],
[73], [78]), and it was established that energy conservation occurs for u in Besov
s s
spaces B3,∞ , s > 1/3. Besov spaces Bp,q are interpolation spaces, the index s
measures the number of derivatives, the index p the Lp (dx) space and the index q
is an interpolation index. Recently, the conservation of kinetic energy was proved
1/3 1/3
([29]) for velocities in B3,c(N) , a space that contains B3,q for all 1 ≤ q < ∞, and
s
it contains also B3,∞ for s > 1/3. This is fairly close to optimal. There exist
616 PETER CONSTANTIN

1/3
divergence-free vector fields in B3,∞ for which the energy flux is bounded away
from zero. This implies that the instantaneous time derivative of kinetic energy of
the weak solution of the Euler equations is nonzero at these fields. The existence
1/3
of weak solutions of the Euler equations in B3,∞ has not been proved, so the
counterexample does not prove the second direction of the Onsager conjecture. The
proofs and examples employ the Littlewood-Paley decomposition and the flux of
the Littlewood-Paley spectrum ([38]), which is a mathematically convenient variant
of the physical concept of flux from the turbulence literature. For the helicity
conservation, similar results apply but require more regularity (see [26] for previous
2/3
sufficient conditions). If the velocity is in B3,c(N) , then the helicity is conserved.
2/3
There exist divergence-free velocity fields in B3,∞ for which the helicity flux is
bounded away from zero.
Shell models are sequences of ODEs, resembling the form the Euler equation
takes when written in terms of the Littlewood-Paley decomposition, but greatly
simplified and truncated. A proof of one direction of the Onsager conjecture near
the exponent 1/3 was done in ([51]); there are models that have anomalous dissi-
pation ([30], [31]). This can occur in linear systems as well ([109]).
It is not clear however how one would go about constructing weak solutions of
1/3
the initial value problem for 3D Euler equations in B3,∞ that dissipate energy. It
is possible however that the long–time, zero–viscosity limit selects very particular
stationary measures supported in this space.

4. Conclusion
It is no exaggeration to say that the Euler equations are the very core of fluid
dynamics. They use and enrich several branches of mathematics and provide im-
portant open problems. The “stability of instability” of mixing layers in ill–posed
problems such as the Rayleigh-Taylor and Kelvin-Helmholtz instabilities is one such
problem.
The mathematical question of blowup in the Euler equations is still one of the
most challenging and meaningful problems in nonlinear PDE. The study of the
vanishing viscosity limit (more properly, the infinite Reynolds number limit) in
domains with boundaries is very far from completion. The problem in bounded
domains is extremely challenging. The existence of weak solutions that dissipate
energy by constant flux of energy, in the whole space, in the correct functional space
is not known. The characterization of the zero–viscosity long–time limit statistics,
including Kolmogorov-Obukhov spectrum, is open.

Acknowledgment
Work partially supported by NSF DMS grant No. 0504213 and by the ASC Flash
Center at the University of Chicago.

About the author

Peter Constantin is Louis Block Professor at the University of Chicago. He was
an invited speaker at the ICM in Zurich, the ICMP in Paris, and the ICIAM in
Edinburgh. He is a Fellow of the Institute of Physics, London.
 EULER EQUATIONS OF INCOMPRESSIBLE FLUIDS 617

References
[1] H. Abidi, R. Danchin, Optimal bounds for the inviscid limit of Navier-Stokes equations.
Asymptot. Anal. 38 (2004), 35-46. MR2060619 (2005c:35227)
[2] A. Alexakis, C. Doering, Energy and enstrophy dissipation in steady state 2D turbulence.
Physics Lett. A 359 (2006), 652-657.
[3] D.M. Ambrose, Well-posedness of vortex sheets with surface tension. SIAM J. Math. Anal.
35 (2003), 211-244. MR2001473 (2005g:76006)
[4] D. M. Ambrose, N. Masmoudi, The zero surface tension limit of two dimensional water
waves. Commun. Pure Appl. Math. 58 (2005), 1287-1315. MR2162781 (2006d:35189)
[5] A. Arenas, A. J. Chorin, On the existence of scaling functions in turbulence according to
the data. Proc. Natl. Acad. Sc. USA 103 (2006), 4352-4355. MR2213976 (2006j:76060)
[6] V. Arnold, B. Khesin, Topological methods in hydrodynamics. Springer, 1998. MR1612569
(99b:58002)
[7] V. Barcilon, P. Constantin, E. Titi, Existence of solutions to the Stommel-Charney model
of the Gulf Stream. SIAM J. Math. Anal. 19 (1988), 1355-1364. MR965256 (89k:35241)
[8] C. Bardos, Existence et unicité de la solution de l’équation d’Euler en dimension deux. J.
Math. Anal. Appl. 40 (1972), 769-790. MR0333488 (48:11813)
[9] C. Bardos, Y. Guo, W. Strauss, Stable and unstable ideal plane flows. Chinese Annals of
Math. 23 B (2002), 149-164. MR1924132 (2003k:35192)
[10] C. Bardos, E. Titi, Euler equations of incompressible ideal fluids. Preprint 2007.
[11] G. I. Barenblatt, Scaling. CUP, Cambridge, 2003. MR2034052 (2005e:00011)
[12] G.I. Barenblatt, A.J. Chorin, Scaling laws and zero viscosity limits for wall-bounded shear
flows and for local structure in developed turbulence. Comm. Pure Appl. Math. 50 (1997),
381-391. MR1438152 (98a:76041)
[13] G.I. Barenblatt, A. J. Chorin, Scaling laws and vanishing viscosity limits in turbulence
theory. Proc. Symp. Appl. Math. AMS 54 (1998), 1-25. MR1492690 (99a:76065)
[14] G.I. Barenblatt, A. J. Chorin, V.M. Prostokishin, The Kolmogorov-Obukhov exponent in the
inertial range of turbulence: A reexamination of experimental data. Physica D 127 (1999),
105-110. MR1677445 (2000c:76034)
[15] J.T. Beale, T. Kato, A. Majda, Remarks on the breakdown of smooth solutions for the 3-D
Euler equations. Commun. Math. Phys. 94 (1984), 61-66. MR763762 (85j:35154)
[16] R. Benzi, S. Ciliberto, C. Baudet, G. Ruiz Chavarria, On the scaling of three dimensional ho-
mogeneous and isotropic turbulence. Physica D 80, 385-398 (1995). MR1312600 (95i:76046)
[17] H. Berestycki, F. Hamel, N. Nadirashvili, Elliptic eigenvalue problems with large drift and
applications to nonlinear propagation phenomena. Comm. Math. Phys. 253 (2005), 451-480.
MR2140256 (2006b:35057)
[18] D. Bernard, Influence of friction on the direct cascade of 2D forced turbulence. Europhys.
Lett. 50 (2000), 333–339.
[19] H. Beirão da Veiga, L. Berselli, On the regularizing effect of the vorticity direction in in-
compressible viscous flows. Diff. Int. Eqns. 15 (2002), 345-356. MR1870646 (2002k:35248)
[20] L. Caffarelli, R. Kohn, L. Nirenberg, Partial regularity of suitable weak solutions of
the Navier-Stokes equations. Comm. Pure Appl. Math. 35 (1982), 771-831. MR673830
(84m:35097)
[21] L. Caffarelli, A. Vasseur, Drift diffusion equations with fractional diffusion and the quasi-
geostrophic equation. ArXiv: Math.AP/0608447 (2006).
[22] R. Caflisch, Singularity formation for complex solutions of the 3D incompressible Euler
equations. Physica D 67 (1993), 1-18. MR1234435 (94h:76013)
[23] R. Caflisch, M. Sammartino, Zero viscosity limit for analytic solutions of the Navier-Stokes
equations on a half space, I. Existence for Euler and Prandtl equations. Comm. Math. Phys.
192 (1998), 433-461. MR1617542 (99d:35129a)
[24] R. Caflisch, M. Sammartino, Zero viscosity limit for analytic solutions of the Navier-Stokes
equations on a half space, II. Construction of the Navier-Stokes solution. Comm. Math.
Phys. 192 (1998), 463-491. MR1617538 (99d:35129b)
[25] D. Chae, Nonexistence of self-similar singularities for the 3D incompressible Euler equa-
tions. Commun. Math. Phys. (2007) (to appear).
[26] D. Chae, On the conserved quantities for the weak solutions of the Euler equations and the
Quasi-Geostrophic equation. Commun. Math. Phys. 266 (2006), 197-210. MR2231970
618 PETER CONSTANTIN

[27] D. Chae, On the regularity conditions for the dissipative quasi-geostrophic equations. SIAM
J. Math. Anal. 37 (2006), 1649-1656. MR2215601 (2007c:76008)
[28] J-Y. Chemin, Perfect Incompressible Fluids. Clarendon Press, Oxford Univ., 1998.
MR1688875 (2000a:76030)
[29] A. Cheskidov, P. Constantin, S. Friedlander, R. Shvydkoy, Energy conservation and On-
sager’s conjecture for the Euler equations. ArXiv; Math.AP/0704.0759 (2007).
[30] A. Cheskidov, S. Friedlander, N. Pavlović, An inviscid dyadic model of turbulence: the fixed
point and Onsager’s conjecture. Journal of Mathematical Physics, to appear.
[31] A. Cheskidov, S. Friedlander, N. Pavlović, An inviscid dyadic model of turbulence: the global
attractor. Preprint (2007).
[32] S. Childress, G. R. Ierly, E. A. Spiegel, W. R. Young, Blow up on unsteady two-dimensional
Euler and Navier-Stokes solutions having stagnation point form. J. Fluid. Mechanics 203,
(1989), 1-22. MR1002875 (90e:76054)
[33] A. Chorin, J. Marsden, A mathematical introduction to fluid mechanics. 3rd edition,
Springer, 1993. MR1218879 (94c:76002)
[34] A. J. Chorin, Numerical study of slightly viscous flow. J. Fluid Mech. 57 (1973), 785-796.
MR0395483 (52:16280)
[35] D. Christodoulou, H. Lindblad, On the motion of the free surface of a liquid. Commun.
Pure Appl. Math. 53 (2000), 1536-1602. MR1780703 (2002c:76025b)
[36] P. Constantin, Note on loss of regularity for solutions of the 3D incompressible Euler and
related equations. Commun. Math. Phys. 104 (1986), 311-326. MR836008 (87f:35200)
[37] P. Constantin, Geometric statistics in turbulence. SIAM Review 36 (1994), 73-98.
MR1267050 (95d:76057)
[38] P. Constantin, The Littlewood-Paley spectrum in 2D turbulence. Theor. Comp. Fluid Dyn.
9 (1997), 183-189.
[39] P. Constantin, The Euler equations and nonlocal conservative Riccati equations. Intern.
Math. Res. Notes 9 (2000), 455-465. MR1756944 (2001b:76007)
[40] P. Constantin, An Eulerian-Lagrangian approach for incompressible fluids: local theory.
JAMS 14 (2001), 263-278. MR1815212 (2002e:76008)
[41] P. Constantin, An Eulerian-Lagrangian approach to the Navier-Stokes equations. Commun.
Math. Phys. 216 (2001), 663-686. MR1815721 (2002m:76023)
[42] P. Constantin, Euler equations, Navier-Stokes equations and turbulence. In Mathemati-
cal foundation of turbulent viscous flows: Lectures given at the C.I.M.E. Summer School,
Martina Franca, Italy. Editors: M. Cannone and T. Miyakawa, Springer Lecture Notes in
Mathematics 1871 (2005), 1-43. MR2196360 (2007c:76001)
[43] P. Constantin, Nonlinear Fokker-Planck Navier-Stokes systems. Commun. Math. Sci. 3 (4)
(2005), 531-544. MR2188682 (2007b:35249)
[44] P. Constantin, D. Cordoba, J. Wu, On the critical dissipative quasigeostrophic equation.
Indiana U. Math. Journal 50 (2001) 97-107. MR1855665 (2002h:35246)
[45] P. Constantin, W. E, E. Titi, Onsager’s conjecture on the energy conservation for solutions
of Euler’s equation. Commun. Math. Phys. 165 (1994), 207–209. MR1298949 (96e:76025)
[46] P. Constantin, C. Fefferman, Direction of vorticity and the problem of global regularity for
the Navier-Stokes equations. Indiana Univ. Math. J. 42 (1993), 775. MR1254117 (95j:35169)
[47] P. Constantin, C. Fefferman, A. Majda, Geometric constraints on potentially singular so-
lutions for the 3-D Euler equations. Commun. in PDE 21 (1996), 559-571. MR1387460
(97c:35154)
[48] P. Constantin, C. Fefferman, E. Titi, A. Zarnescu, Regularity for coupled two-dimensional
nonlinear Fokker-Planck and Navier-Stokes systems. Comm. Math. Phys. (to appear, 2007-
08). MR2276466
[49] P. Constantin, G. Iyer, Stochastic Lagrangian transport and generalized relative entropies.
Commun. Math. Sci. 4 (2006), 767-777. MR2264819
[50] P. Constantin, A. Kiselev, L. Ryzhik, A. Zlatos, Diffusion and mixing in fluid flow. Annals
of Math., to appear (2007).
[51] P. Constantin, B. Levant, E. Titi, Regularity of inviscid shell models of turbulence. Physical
Review E 75 1 (2007), 016305.
[52] P. Constantin, N. Masmoudi, Global well-posedness for a Smoluchowski equation coupled
with Navier-Stokes equations in 2D. Commun. Math. Phys., to appear (07-08).
 EULER EQUATIONS OF INCOMPRESSIBLE FLUIDS 619

[53] P. Constantin, A. Majda, E. Tabak, Formation of strong fronts in the 2D quasi-geostrophic

thermal active scalar. Nonlinearity 7 (1994), 1495-1533. MR1304437 (95i:76107)
[54] P. Constantin, F. Ramos, Inviscid limit for damped and driven incompressible Navier-Stokes
equations in R2 . Commun. Math. Phys., to appear (2007).
[55] P. Constantin, L. Ryzhik, A. Novikov, Relaxation in reactive flows. GAFA (2007) (to ap-
pear).
[56] P. Constantin, J. Wu, Inviscid limit for vortex patches. Nonlinearity 8 (1995), 735-742.
MR1355040 (96g:76012)
[57] P. Constantin, J. Wu, The inviscid limit for non-smooth vorticity. Indiana U. Math. J. 45
(1996), 67-81. MR1406684 (97g:35129)
[58] P. Constantin, J. Wu, Behavior of solutions to 2D quasigeostrophic equations. SIAM J.
Math. Anal. 30 (1999), 937-948. MR1709781 (2000j:76019)
[59] P. Constantin, J. Wu, Regularity of Hölder continuous solutions of the supercritical quasi-
geostrophic equation. ArXiv: Math.AP/0701592 (2007).
[60] P. Constantin, J. Wu, Hölder continuity of solutions of supercritical dissipative hydrody-
namic transport equations. ArXiv: Math.AP/0701594 (2007).
[61] D. Córdoba, Nonexistence of simple hyperbolic blow-up for the quasigeostrophic equation.
Annals of Math. (2) 148 (1998), 1135-1152. MR1670077 (2000j:76020)
[62] A. Córdoba, D. Córdoba, A maximum principle applied to quasi-geostrophic equations.
Commun. Math. Phys. 249 (2004), 511-528. MR2084005 (2005f:76011)
[63] D. Cordoba, F. Gancedo, Contour dynamics of incompressible 3D fluids in a porous medium
with different densities. Commun. Math. Phys., to appear (2007).
[64] D. Cordoba, C. Fefferman, R. de la Llave, On squirt singularities in hydrodynamics. SIAM
J. Math. Analysis 36 (2004), 204-213. MR2083858 (2005i:76032)
[65] D. Coutaud, S. Shkoller, Well-posedness of the free surface incompressible Euler equations
with or without surface tension. JAMS (2007).
[66] W. Craig, An existence theory for water waves in the Boussinesq and the Korteweg-de-Vries
scaling limits. Commun. PDE 10 (1985), 787-1003. MR795808 (87f:35210)
[67] C. De Lellis, L. Szekelyhidi, The Euler equations as differential inclusions. Preprint (2007).
[68] J-M. Delort, Existence de nappes de tourbillon en dimension deux. JAMS 4 (1991), 553-586.
MR1102579 (92f:76019)
[69] J. Deng, T. Y. Hou, X. Yu, Geometric properties and non-blowup for 3-D incompressible
Euler flow. Commun. PDE 30 (2005), 225-243. MR2131052 (2006f:35215)
[70] R. DiPerna, P-L. Lions, Ordinary differential equations, transport theory and Sobolev spaces.
Invent. Math. 98 (1989) 511-547. MR1022305 (90j:34004)
[71] R. DiPerna, A. Majda, Oscillations and concentrations in weak solutions of the incompress-
ible fluid equations. Commun. Math. Phys. 108 (1987), 667-689. MR877643 (88a:35187)
[72] R. DiPerna, A. Majda, Reduced Hausdorff dimension and concentration-cancellation for
two dimensional incompressible flow. JAMS 1 (1988), 59-95. MR924702 (89e:35126)
[73] J. Duchon, R. Robert, Inertial energy dissipation for weak solutions of incompressible Euler
and Navier-Stokes equations. Nonlinearity 13 (2000), 249–255. MR1734632 (2001c:76032)
[74] W. E, Boundary layer theory and the zero viscosity limit of the Navier-Stokes equation.
Acta Math. Sinica (Engl. ser.) 16 (2000), 207-218. MR1778702 (2001d:76037)
[75] W. E, B. Enquist, Blow up of solutions to the unsteady Prandtl equation. Comm. Pure Appl.
Math. 50 (1997), 1287-1293. MR1476316 (99c:35196)
[76] W. E, T. Li, P. Zhang. Well-posedness for the dumbbell model of polymeric fluids. Comm.
Math. Phys. 248(2) (2004), 409–427. MR2073140 (2005d:35210)
[77] L. Euler, Principes généraux du mouvement des fluides. Mémoires de L’Académie Royale
des Sciences et des Belles-Lettres de Berlin 11 (4 September 1755, printed 1757), 217-273;
reprinted in Opera Omnia ser. 2 12, 219-250.
[78] G. L. Eyink, Energy dissipation without viscosity in ideal hydrodynamics. I. Fourier analysis
and local energy transfer. Phys. D 78 (1994), 222–240. MR1302409 (95m:76020)
[79] G. Eyink, Dissipation in turbulent solutions of 2D Euler equations. Nonlinearity 14 (2001),
787-802. MR1837638 (2002g:76070)
[80] G. L. Eyink, K. R. Sreenivasan, Onsager and the theory of hydrodynamic turbulence. Rev.
Mod. Phys. 78 (2006), 87–135. MR2214822 (2007g:76108)
[81] C. Foiaş, Statistical study of the Navier-Stokes equations I. Rend. Sem. Mat. Univ. Padova
48 (1972), 219–348. MR0352733 (50:5220)
620 PETER CONSTANTIN

[82] C. Foiaş, Statistical study of the Navier-Stokes equations II. Rend. Sem. Mat. Univ. Padova
49 (1973), 9–123. MR0352733 (50:5220)
[83] S. Friedlander, A. Lipton-Lifschitz, Localized instabilities in fluids. Handbook of mathe-
matical fluid dynamics, Vol. II, North-Holland, Amsterdam (2003), 289-354. MR1984155
(2004g:76072)
[84] U. Frisch, Turbulence. The legacy of A. N. Kolmogorov. Cambridge University Press, Cam-
bridge, 1995. MR1428905 (98e:76002)
[85] J.D. Gibbon, D.R. Moore, J.T. Stuart, Exact, infinite energy blow-up solutions of the three-
dimensional Euler equations. Nonlinearity 16 (2003), 1823-1831. MR1999581 (2004j:35232)
[86] J. D. Gibbon, A. Fokas, C.R. Doering, Dynamically stretched vortices as solutions of the
3D Navier-Stokes equations. Physics D 132 (1999), 497-510. MR1704825 (2000h:76047)
[87] T. Y. Hou, R. Li, Dynamic depletion of vortex stretching and non-blow up of the 3-D incom-
pressible Euler equations. Nonlinear Science 16 (2006), 639-664. MR2271429 (2007f:76014)
[88] G. Iyer, A stochastic Lagrangian formulation of the incompressible Navier-Stokes and re-
lated transport equations. PhD Thesis, The University of Chicago (2006).
[89] B. Jourdain, T. Lelièvre, C. Le Bris. Existence of solution for a micro-macro model of
polymeric fluid: the FENE model. J. Funct. Anal. 209(1) (2004), 162–193. MR2039220
(2005a:76006)
[90] V. Kamotski, G. Lebeau, On the 2D Rayleigh-Taylor instabilities. Asymptot. Anal. 42
(2005), 1-27. MR2133872 (2006b:76038)
[91] T. Kato, Nonstationary flows of viscous and ideal fluids in R3 . J. Funct. Anal. 9 (1972),
296-305. MR0481652 (58:1753)
[92] A. Kiselev, F. Nazarov, A. Volberg, Global well-posedness for the critical 2D dissipative
quasi-geostrophic equation. Invent. Math. 167 (2007), 445–453. MR2276260
[93] A. N. Kolmogorov, The local structure of turbulence in incompressible viscous fluids at very
large Reynolds numbers. Dokl. Akad. Nauk. SSSR 30 (1941), 301–305. MR0004146 (2:327d)
[94] H. Kozono, Y. Taniuchi, Limiting case of the Sobolev inequality in BMO, with application to
the Euler equations. Commun. Math. Phys. 214 (2000), 191-200. MR1794270 (2002k:46081)
[95] R. H. Kraichnan, Inertial ranges in two-dimensional turbulence. Phys. Fluids 10 (1967),
1417-1423.
[96] S. Kuksin, A. Shirikyan, Some limiting properties of randomly forced 2D Navier-Stokes equa-
tions. Proc. Roy. Soc. Edinburgh sect. A 133 (2003), 875-891. MR2006207 (2005c:60079)
[97] G. Lebeau, Régularité du problème de Kelvin-Helmholtz pour l’équation d’Euler 2D.
ESAIM:COCV 8 (2002), 801-825. MR1932974 (2004a:76011)
[98] C. Le Bris, P-L. Lions, Existence and uniqueness of solutions to Fokker-Planck type equa-
tions with irregular coefficients. Rapport de receherche du CEREMICS 349, April 2007.
[99] J. Leray, Essai sur le mouvement d’un liquide visqueux emplissant l’espace. Acta Mathe-
matica 63 (1934), 193-248. MR1555394
[100] F.-H. Lin, C. Liu, P. Zhang. On hydrodynamics of viscoelastic fluids. Comm. Pure Appl.
Math. 58(11) (2005), 1437–1471. MR2165379 (2006d:76005)
[101] F.-H. Lin, P. Zhang, Z. Zhang. On the global existence of smooth solution to the 2-d FENE
dumbbell model. Preprint, 2007.
[102] H. Linblad, Well-posedness for the motion of an incompressible liquid with free surface
boundary. Annals of Math. (2) 162 (2005), 109-194. MR2178961 (2006g:35293)
[103] P.-L. Lions, N. Masmoudi. Global existence of weak solutions to micro-macro models. C. R.
Math. Acad. Sci. Paris, 2007.
[104] M. Lopes Filho, A. Mazzucato, H. Nussenzveig-Lopes, Weak solutions, renormalized solu-
tions and enstrophy defects in 2D turbulence. ARMA 179 (2006), 353-387. MR2208320
(2006k:35234)
[105] A. Majda, A. Bertozzi, Vorticity and incompressible flow. CUP, Cambridge, 2002.
MR1867882 (2003a:76002)
[106] A. Majda, X. Wang, Nonlinear dynamics and statistical theories for basic geophysical flows.
CUP, Cambridge (2006).
[107] C. Marchioro, M. Pulvirenti, Mathematical theory of incompressible nonviscous fluids. Ap-
plied Mathematical Sciences, 96, Springer, 1994. MR1245492 (94k:76001)
[108] N. Masmoudi, Remarks about the inviscid limit of the Navier-Stokes system. Comm. Math.
Phys. 270 (2007), 777–788. MR2276465
 EULER EQUATIONS OF INCOMPRESSIBLE FLUIDS 621

[109] J. Mattingly, T. Suidan, E. Vanden-Eijnden, Simple systems with anomalous dissipation

and energy cascade. Commun. Math. Phys., to appear 2007-08.
[110] P. Michel, S. Mischler, B. Perthame, General entropy equations for structured population
models and scattering. C.R. Acad. Sci. Paris Ser. I 338 (2004), 697-702. MR2065377
(2005b:35019)
[111] H. K. Moffatt, A. Tsinober, Helicity in laminar and turbulent flow. Ann. Rev. Fluid Mech.
24 (1992), 281–312. MR1145012 (93b:76018)
[112] L. Onsager, Statistical hydrodynamics. Nuovo Cimento (Supplemento) 6 (1949), 279–287.
MR0036116 (12:60f)
[113] K. Ohkitani, J.D. Gibbon, Numerical study of singularity formation in a class of Euler and
Navier-Stokes flows. Physics of Fluids 12 (2000), 3181-3194. MR1796376 (2001i:76008)
[114] K. Ohkitani, M. Yamada, Inviscid and inviscid-limit behavior of a surface quasi-geostrophic
flow. Phys. Fluids 9 (1997), 876-882. MR1437554 (97m:76032)
[115] F. Otto, A.E. Tzavaras, Continuity of velocity gradients in suspensions of rod-like molecules.
SFB preprint Nr. 141 (2004).
[116] S. Resnick, Dynamical problems in nonlinear advective partial differential equations. Ph.D.
Thesis, University of Chicago, 1995.
[117] R. Robert, Statistical Hydrodynamics (Onsager revisited). Handbook of Mathematical
Fluid Dynamics, vol. 2 (2003), 1–55. Eds. Friedlander and Serre. Elsevier. MR1983588
(2005a:76004)
[118] V. Rom-Kedar, A. Leonard, S. Wiggins, An analytical study of transport, mixing and chaos
in unsteady vortical flow. J. Fluid Mech. 214 (1990), 347-394. MR1054106 (91d:76017)
[119] V. Scheffer, An inviscid flow with compact support in space-time. J. Geom. Anal. 3(4)
(1993), 343–401. MR1231007 (94h:35215)
[120] H. Schlichting, K. Gersten, E. Krause, H. Oertel, C. Mayes, Boundary-Layer Theory. 8th
edition, Springer, 2000. MR1765242 (2001c:76039)
[121] A. Shnirelman, On the nonuniqueness of weak solution of the Euler equation. Comm. Pure
Appl. Math. 50 (1997), 1261–1286. MR1476315 (98j:35149)
[122] J. T. Stuart, Nonlinear Euler partial differential equations: singularities in their solution.
In Applied Mathematics, Fluid Mechanics, Astrophysics (Cambridge, MA, 1987), World.
Scien., Singapore (1988), 81-95. MR973917
[123] H.S.G. Swann, The convergence with vanishing viscosity of nonstationary Navier-Stokes
flow to ideal flow in R3 . Trans. Amer. Math. Soc. 157 (1971), 698-726. MR0277929
(43:3662)
[124] M. Vishik, Spectrum of small oscillations of an ideal fluid and Lyapunov exponents. J.
Math. Pures Appl. (9) 75 (1996), 531-557. MR1423046 (97k:35203)
[125] M. Vishik, A. Fursikov, Mathematical problems of statistical hydrodynamics. Kluwer AP,
Dordrecht (1988).
[126] S. Wu, Well-posedness in Sobolev spaces of the full water wave problem in 2-D. Invent.
Math. 130 (1997), 39-72. MR1471885 (98m:35167)
[127] S. Wu, Well-posedness in Sobolev spaces of the full water wave problem in 3-D. JAMS 12
(1999), 445-495. MR1641609 (2001m:76019)
[128] S. Wu, Mathematical analysis of vortex sheets. Commun. Pure Appl. Math. 59 (2006),
1065-1206. MR2230845
[129] V. I. Yudovich, The linearization method in hydrodynamical stability theory. Translations
of Mathematical Monographs, 74, AMS, Providence, RI (1989) (Translated from Russian
by J.R. Schulenberger). MR1003607 (90h:76001)
[130] V. Yudovich, Non-stationary flow of an ideal incompressible liquid. Zh. Vych. Mat. Mat.
Fiz. 3 (1963), 1032-1066. MR0158189 (28:1415)
[131] V. Zakharov (editor), What Is Integrability? Springer Series in Nonlinear Dynamics,
Springer (1991). MR1098334 (91k:58005)
[132] V. Zakharov, talk at the University of Chicago, May 2007.

Department of Mathematics, University of Chicago, Chicago, Illinois 60637

E-mail address: const@math.uchicago.edu