← Spatial Mathematical Biology

Advection

Advection is directed transport through space. A population, chemical or pathogen is carried systematically in a particular direction by a velocity field.

Examples include plankton carried by an ocean current, seeds transported by prevailing wind, cells carried by fluid flow, or organisms moving with a persistent migration velocity.

Core idea. Diffusion spreads a distribution because of undirected local movement. Advection carries the distribution because there is a preferred transport direction.

Imagine a population patch moving downstream

Suppose a group of organisms is concentrated around one location in a river. If the current flows to the right at constant velocity, the whole concentration profile is transported downstream.

With pure advection and constant velocity, the profile moves without being intrinsically widened or flattened.

earlierlatertransport directionposition xdensity u
Pure constant advection translates the density profile. Its location changes, but its shape is preserved.

Density and velocity

Let

\[u(x,t)\]

denote density at position \(x\) and time \(t\), and let \(v\) be the transport velocity.

If \(v>0\), transport is toward increasing \(x\). If \(v<0\), transport is toward decreasing \(x\).

The units of \(v\) are distance per unit time.

The advection equation

For constant velocity \(v\), pure advection is described by

\[\boxed{\frac{\partial u}{\partial t}+v\frac{\partial u}{\partial x}=0.}\]

Equivalently,

\[\frac{\partial u}{\partial t}=-v\frac{\partial u}{\partial x}.\]

What does the equation mean?

The term \(u_x\) describes how density changes across space. The velocity \(v\) converts that spatial change into a temporal change as the profile passes a fixed location.

The equation does not say the population disappears. It says that changes observed at one fixed point arise because the spatial profile is being transported past that point.

The travelling solution

If the initial density profile is

\[u(x,0)=u_0(x),\]

then for constant \(v\),

\[\boxed{u(x,t)=u_0(x-vt).}\]

This formula is extremely useful for understanding advection.

At time \(t\), the original profile has shifted by the distance

\[vt.\]

Why is it x - vt for motion to the right?

Suppose a particular feature of the initial profile was at position \(x_0\). At time \(t\), that feature is at

\[x=x_0+vt.\]

Therefore

\[x-vt=x_0,\]

which explains why the translated profile is written as \(u_0(x-vt)\).

For \(v>0\), \(u_0(x-vt)\) moves to the right. The minus sign inside the function does not mean leftward motion.

Following the moving population

Along a path moving with the flow,

\[x(t)=x_0+vt.\]

The chain rule gives

\[\frac{d}{dt}u(x(t),t)=\frac{\partial u}{\partial t}+\frac{dx}{dt}\frac{\partial u}{\partial x}.\]

Since \(dx/dt=v\),

\[\frac{d}{dt}u(x(t),t)=u_t+vu_x.\]

For pure advection, the right-hand side is zero. Thus

\[\frac{d}{dt}u(x(t),t)=0.\]

The density value carried along each characteristic path remains constant.

Characteristics

The lines

\[x=x_0+vt\]

are called characteristics. They show how information is transported through space and time.

position xtime tcharacteristicsx = x₀ + vt
For positive constant velocity, characteristics move toward larger positions as time increases. Density values are carried along these paths.

Advection as a conservation law

Transport can also be written using a flux. If material with density \(u\) moves at velocity \(v\), the advective flux is

\[J=vu.\]

Conservation gives

\[\frac{\partial u}{\partial t}=-\frac{\partial J}{\partial x}.\]

Therefore

\[\boxed{u_t+\frac{\partial(vu)}{\partial x}=0.}\]

If \(v\) is constant, this reduces to

\[u_t+vu_x=0.\]

Why the conservative form matters

If velocity changes with position, we should not generally replace \((vu)_x\) by \(vu_x\).

Using the product rule,

\[\frac{\partial(vu)}{\partial x}=v\frac{\partial u}{\partial x}+u\frac{\partial v}{\partial x}.\]

The second term accounts for compression or expansion caused by spatial variation in velocity.

The simple equation \(u_t+vu_x=0\) assumes constant velocity, or a setting where the non-conservative form is specifically appropriate. For conservation of transported density with variable velocity, use \(u_t+(vu)_x=0\).

Compression and expansion

If trajectories converge, transported density can become concentrated. If they diverge, density can become diluted.

This is fundamentally different from diffusion: the change can arise from the velocity field itself rather than random spreading.

Advection versus diffusion

The two mechanisms have very different effects on a density profile.

initial profileadvection: shifteddiffusion: widened and smoothedposition x
Advection primarily transports the profile; diffusion spreads and smooths it.
AdvectionDiffusion
directed transportundirected spreading
controlled by velocity \(v\)controlled by diffusion coefficient \(D\)
constant advection preserves profile shapediffusion generally broadens and smooths profile
first spatial derivative for constant velocitysecond spatial derivative
velocity units: distance/time\(D\) units: distance²/time

Advection and diffusion together

Many biological systems contain both directed transport and random spreading:

\[\boxed{\frac{\partial u}{\partial t}=D\frac{\partial^2u}{\partial x^2}-v\frac{\partial u}{\partial x}.}\]

For constant \(D\) and \(v\), the advection term moves the profile while the diffusion term broadens it.

Advection–diffusion with local biology

Birth, death, infection or reaction can also be included:

\[\frac{\partial u}{\partial t}=D u_{xx}-v u_x+f(u).\]

Here the three mechanisms are easy to distinguish:

TermRole
\(D u_{xx}\)random spatial spreading
\(-v u_x\)directed transport
\(f(u)\)local biological change

A river example

Consider microorganisms living in flowing water. Reproduction changes abundance locally. Small random movements and mixing cause diffusion. The downstream current causes advection.

A model may therefore combine growth, diffusion and downstream transport. Whether the population persists at a location can depend on the balance between reproduction and washout.

Wind-dispersed organisms

Seeds, spores, pollen and small insects may experience a prevailing wind. The systematic component of wind transport can be represented by advection, while turbulent or irregular movement can contribute a diffusive component.

Spatial epidemics

Advection can be relevant when infectious material or hosts are transported systematically, for example through water flow or directional movement.

Different compartments may have different velocities or movement mechanisms, so spatial epidemic models should reflect the biology of each population rather than automatically assigning the same transport term to all compartments.

Two-dimensional advection

In two dimensions, let the velocity field be

\[\mathbf{v}(x,y,t)=(v_x,v_y).\]

For a transported scalar in a prescribed incompressible flow, the advective term can be written

\[\mathbf{v}\cdot\nabla u.\]

The advection equation is then

\[u_t+\mathbf{v}\cdot\nabla u=0.\]

For conservation of density in a general velocity field, the conservative form is

\[\boxed{u_t+\nabla\cdot(u\mathbf{v})=0.}\]

Boundary conditions

Advection makes boundary direction especially important. Where flow enters the domain, an inflow condition may be needed to specify what is entering. At an outflow boundary, material is transported out of the domain.

Boundary conditions should therefore reflect the physical or biological meaning of the edges.

Does advection conserve total population?

Advection redistributes population internally. Total population changes only if material crosses the boundary or if birth, death or other reaction terms are present.

Thus an open downstream boundary can cause loss from the modelled region even though advection itself does not destroy individuals.

Advection is not taxis

Advection usually means transport by a prescribed velocity field or systematic bulk motion. Taxis describes directed biological movement in response to a stimulus, such as chemotaxis toward a chemical signal.

Both are directional, but the mechanism and mathematical dependence can be different.

Advection is not migration between discrete patches

Advection is usually formulated in continuous space. Migration between distinct habitats can instead be represented with movement rates in a patch model.

The appropriate representation depends on the spatial scale and biological question.

Numerical difficulty

Advection equations can be more delicate to simulate than their simple appearance suggests. Some naive numerical methods create artificial oscillations or excessive numerical smoothing.

This is why spatial discretisation methods designed for transport, such as upwind ideas, are often used.

A practical interpretation workflow

Identify what is being transported and what determines the velocity. Establish the positive spatial direction and interpret the sign of \(v\). Decide whether velocity is constant or spatially varying. Use the conservative form when modelling conservation of density. Add diffusion only if random spreading is also present, and add reaction terms only for genuine local biological processes.

Key idea. Advection transports a biological distribution systematically through space. For constant velocity, the profile moves by the distance \(vt\) without changing shape. Conservation gives the flux \(J=vu\), while combining advection with diffusion and local reactions produces richer spatial biological models.