Fluid kinematics — Eulerian vs Lagrangian description: Question bank

Before you start, keep these two pictures in your head (the whole topic hangs on them):

The fish — you tag one lump of fluid and follow its journey. This is the Lagrangian view; the derivative that goes with it is $\frac{D}{D t}$ (holds the particle's name tag fixed).
The bridge — you stand at one fixed spot and watch whatever fluid passes. This is the Eulerian view; the derivative that goes with it is $\frac{\partial}{\partial t}$ (holds the position fixed).

The number that links them is the convective term $(v \cdot \nabla) ϕ$ , which measures "how much the field value changes just because the particle moved to a new place."

True or false — justify

A "steady" flow means every fluid particle moves at constant speed.

False. Steady means the field at each fixed point is unchanging in time (

\partial / \partial t = 0

), not that a particle's speed is constant. A particle can still speed up by moving into a faster region — that is the convective term

(v \cdot \nabla) v

In steady flow the acceleration of a fluid particle is always zero.

False. Only the local part

\partial v / \partial t

vanishes. The convective part

(v \cdot \nabla) v

can be large — e.g. fluid accelerating into a nozzle throat while the flow pattern itself never changes.

\frac{D ϕ}{D t}

and

\frac{\partial ϕ}{\partial t}

are just two notations for the same derivative.

False.

\partial / \partial t

holds position fixed (the bridge reading);

D / D t

holds the particle label fixed (the fish reading). They differ by exactly

(v \cdot \nabla) ϕ

, and agree only when that term is zero.

If a flow is unsteady (

\partial v / \partial t \neq = 0

), then every particle must be accelerating.

False in general. The local and convective terms can cancel:

\frac{\partial v}{\partial t} + (v \cdot \nabla) v

could sum to zero at a point even though each piece is nonzero, giving zero particle acceleration there.

A Lagrangian velocity and an Eulerian velocity of the same flow are numerically different quantities.

False — they are the same velocity, just labelled differently. Lagrangian writes it as a function of the particle tag

a

and

t

; Eulerian writes the identical value as a function of the current position

x

and

t

. Evaluated at the same particle/place/time they agree.

The material derivative can be applied to a scalar (like temperature) but not to a vector.

False. It applies to any property carried by the particle. For a scalar you get one equation; for a vector like velocity you apply it component-by-component, giving the acceleration

\frac{D v}{D t}

\nabla ϕ = 0

everywhere, the material derivative reduces to the local derivative.

True. The convective term

(v \cdot \nabla) ϕ

vanishes when

ϕ

has no spatial variation, so a particle only feels the field's change in time — moving doesn't take it anywhere "different."

The convective term is what makes fluid dynamics mathematically hard.

True (for velocity).

(v \cdot \nabla) v

multiplies velocity by its own gradient, making the equations nonlinear — this nonlinearity is the origin of turbulence and why closed-form solutions are rare.

Spot the error

"The flow is steady, so

D T / D t = 0

for a drifting parcel."

Wrong. Steady kills only

\partial T / \partial t

. If temperature varies in space, the parcel still feels

(v \cdot \nabla) T \neq = 0

as it drifts into warmer or cooler fluid. Steady

\neq =

zero material rate.

"A student derived

u = a e^{t}

from the path

x = a e^{t}

and reported this as the Eulerian velocity field."

Wrong — it still contains the particle label

a

, which an Eulerian field cannot reference. Invert the path (

a = x e^{- t}

) and substitute to get

u (x, t) = x

, a function of current position and time only.

"Because instruments sit at fixed points, Newton's second law is naturally written in the Eulerian frame."

Wrong.

F = ma

is a statement about a specific particle's acceleration, so it is naturally Lagrangian. We must convert the Lagrangian acceleration into Eulerian language using

\frac{D v}{D t} = \frac{\partial v}{\partial t} + (v \cdot \nabla) v

before applying it at fixed instrument points.

(v \cdot \nabla) ϕ

means take the dot product of

v

with

ϕ

Wrong.

ϕ

(a scalar) has no direction to dot with. The operator

v \cdot \nabla = u \partial_{x} + v \partial_{y} + w \partial_{z}

acts on

ϕ

; you first form the operator

v \cdot \nabla

, then let it differentiate

ϕ

"In the Lagrangian description

x

is an independent variable I can vary freely."

Wrong. In Lagrangian bookkeeping the label

a

is the independent variable;

x = x (a, t)

is a dependent output telling you where that tagged particle has gone. It is the Eulerian description where

x

is independent.

"A fixed thermometer and a balloon drifting with the wind must read the same rate of temperature change."

Wrong. The thermometer reads

\partial T / \partial t

(local only); the balloon reads

D T / D t = \partial T / \partial t + (v \cdot \nabla) T

, which adds the change from riding into new temperatures. They match only if the flow speed or the spatial temperature gradient is zero.

Why questions

Why do we bother keeping two descriptions instead of picking one?

Because the physics (Newton's law) is naturally Lagrangian — about particles — while our measurements are Eulerian — at fixed points. The material derivative is the bridge that lets us write particle laws in field variables we can actually measure.

Why does the convective term appear at all — why isn't the particle's rate just the field's time rate?

Because the particle moves. Between two instants it sits at a new location where the field may already have a different value. The convective term accounts for that "change due to relocation," which

\partial / \partial t

(fixed spot) completely misses.

Why is the acceleration term

(v \cdot \nabla) v

called nonlinear?

Because velocity appears multiplied by a derivative of velocity — the unknown

v

times its own spatial gradient. Doubling

v

more than doubles this term, so the equation is not linear in

v

Why must the particle label be eliminated when converting to an Eulerian field?

Because an Eulerian field answers "what is the value here, now?" — it cannot depend on which particle happens to be passing. The label

a

is Lagrangian-only information, so it must be traded for

(x, t)

via the path equation.

Why does the chain rule produce exactly two kinds of terms in the material derivative?

Because

ϕ (x (t), y (t), z (t), t)

has two sources of time dependence: the explicit

t

slot (gives the local term) and the moving coordinates

x (t), y (t), z (t)

(give the three convective pieces). The chain rule separates them cleanly.

Why can a pressure gauge on a pipe wall never directly report a particle's material rate?

Because it is glued to one location and measures

\partial p / \partial t

there. It has no way to sense the

(v \cdot \nabla) p

contribution a moving parcel feels — that requires following the parcel, which a fixed gauge cannot do.

Edge cases

If the fluid is at rest (

v = 0

) everywhere, what does the material derivative become?

It collapses to the local derivative

\partial ϕ / \partial t

, because the convective term

(v \cdot \nabla) ϕ

is multiplied by zero velocity. With no motion, Lagrangian and Eulerian rates coincide.

For a uniform field (same value everywhere) that varies in time, what does a moving particle feel?

Exactly the local rate

\partial ϕ / \partial t

. The spatial gradient is zero so the convective term vanishes — moving doesn't matter when every location has the same value.

Consider a particle momentarily at rest (

v = 0

at one instant) in an unsteady flow. Is its material acceleration zero?

Not necessarily. At that instant the convective term is zero, but the local term

\partial v / \partial t

can be nonzero, so the particle can still be accelerating from rest.

In perfectly steady flow along a streamline where speed also happens to be constant everywhere, what is the particle's acceleration?

Zero along the flow direction — both terms vanish (local from steadiness, convective because there's no spatial change in speed). But if the streamline is curved, the convective term still supplies centripetal acceleration even at constant speed.

What happens to the distinction between the two descriptions in the limit of vanishingly slow flow with strong time-dependence (e.g. a slowly seeping fluid under a fast-changing tap)?

The convective term (small

v

) becomes negligible next to the local term, so

D / D t \approx \partial / \partial t

. The Eulerian and Lagrangian rates nearly coincide because relocation contributes little.

If two different particles occupy the same point

x

at two different times, do they carry the same velocity there?

In an unsteady flow, no — the Eulerian value

v (x, t)

depends on

t

, so each arrives to a different local velocity. In steady flow, yes — the field at

x

is time-independent, so every particle passing that point has the same velocity there.

Recall Self-check before you leave

Which term of $\frac{D ϕ}{D t}$ survives in steady flow? ::: The convective term $(v \cdot \nabla) ϕ$ ; the local term $\partial ϕ / \partial t$ is zero. What single quantity converts a Eulerian time rate into a Lagrangian one? ::: The convective term $(v \cdot \nabla) ϕ$ , added on to $\partial ϕ / \partial t$ . What must you remove to turn a Lagrangian velocity into an Eulerian field? ::: The particle label $a$ , eliminated via the path equation $x = x (a, t)$ .