Introduction to Vector Differential Calculus

The Fluid Model - A Field of Arrows

Picture a river seen from above. At every point, the water is moving - some direction, some speed. If you could freeze time and plant a tiny arrow at each location showing the velocity there, you'd get a forest of arrows covering the whole plane.

That forest is a vector field.

Let's build some intuition. Below we have three very different flows. Click the presets to switch between them, and notice how the arrows change character completely.

The key idea: a vector field assigns a vector to every point in space. Not just one vector - a whole field of them. The field might be steady (the same pattern forever) or it might change with time. We'll start with steady fields.

We can write this as $\mathbf{v} = \mathbf{v}(x, y)$, or in 3D, $\mathbf{v} = \mathbf{v}(x, y, z)$. Each component is a function of position:

And these aren't just fluid velocities. Gravitational fields, electric fields, magnetic fields - think of iron filings lining up around a magnet. Any situation where a vector is defined at every point gives you a vector field.

Stream Lines - Following the Flow

Imagine you drop a leaf into a stream. It gets carried along by the current, tracing out a path. That path is called a stream line (or flow line).

Here's the crucial connection: at every instant, the leaf's velocity equals the field's velocity at that point. So the velocity vector is always tangent to the stream line:

Click anywhere in the field below to drop a "leaf" and watch it trace its stream line. Try different starting points!

Click to drop a particle. It will trace the stream line in real-time.

Notice something: stream lines never cross each other (as long as the field is well-behaved). Why? Because at any given point, there's exactly one velocity vector. The particle can only go one way. If two stream lines crossed, you'd have two different velocities at the same point - a contradiction.

There are two very different ways to think about $\mathbf{v}$:

• Follow one particle through time: $\mathbf{v}(t)$ gives velocity along a path.
• Snapshot the whole field at one instant: $\mathbf{v}(x,y)$ gives velocity at every point.

Both viewpoints are useful, and we'll need both as we go deeper.

A Taste of Divergence - Expansion vs. Compression

Before we define anything precisely, let's build the intuition.

Imagine injecting a small blob of dye into the fluid - a little circle of colored dots. Now watch what happens as the flow carries them along. In some fields the blob spreads out. In others it compresses. And in some it just moves without changing size.

What would you predict happens to the blob in a source field, where everything radiates outward? And what about pure rotation?

Let's check. Hit "Play" below and watch the dots evolve.

As you can see,

• div v > 0 - stuff is spreading out. The fluid is expanding.
• div v < 0 - stuff is piling up. The fluid is compressing.
• div v = 0 - the fluid is incompressible. The blob moves and deforms, but its area stays the same.

The divergence $\operatorname{div} \mathbf{v}$ measures the net rate at which matter is transported away from each point. It's a scalar - just a number at each point, not a vector.

We're just previewing this idea here. The precise formula and derivation come in §3.4.

A Taste of Curl - Rotation in the Field

Here's a different question. Drop a tiny paddlewheel into the flow. Does it spin?

If the fluid is just pushing the paddlewheel forward in a straight line, it won't rotate. But if one side of the paddlewheel gets pushed harder than the other, it'll start spinning. The curl measures exactly this local rotation.

Try the different fields below. Does the paddlewheel spin?

Here's a beautiful fact: for a rigid body rotating at angular velocity $\omega$ about the $z$-axis, the velocity field is $\mathbf{v} = \omega(-y\,\mathbf{i} + x\,\mathbf{j})$, and:

Twice the angular velocity! The curl captures the local spinning of the fluid.

And here's the trap: the shear field $\mathbf{v} = y\,\mathbf{i}$ doesn't look like it's rotating - the arrows all point to the right. But the paddlewheel spins anyway, because the arrows are stronger at the top than the bottom. A field can have curl even when it doesn't "look" rotational at first glance.

We'll make this precise in §3.5. For now, the physical picture is what matters.

It turns out that just two numbers - the divergence and the curl - capture almost everything about how a vector field behaves locally. That's the central story of vector calculus, and we've just had our first glimpse.