Directional Derivatives and Gradients

students, imagine standing on a hillside 🎒. If you walk straight east, the ground may go up fast. If you walk north, it may go down. If you walk diagonally, the slope could be different again. In multivariable calculus, directional derivatives measure how a function changes when you move in a chosen direction, and the gradient helps you find the direction of steepest increase. These ideas connect directly to tangent planes, linear approximation, and the chain rule, because they describe how a function changes near a point in space.

What You Will Learn

By the end of this lesson, students, you should be able to:

• Explain what a directional derivative means in words and symbols.
• Compute a directional derivative using partial derivatives and a unit vector.
• Describe what the gradient vector tells us about a surface or function.
• Use the gradient to find the direction of steepest increase and the rate of change in a chosen direction.
• Connect directional derivatives and gradients to tangent planes, linear approximation, and the chain rule.

Directional Derivatives: Change in a Chosen Direction

In single-variable calculus, the derivative of $f(x)$ tells how fast $f$ changes as $x$ changes. In multivariable calculus, a function might depend on two or more variables, such as $f(x,y)$. Now the rate of change depends on the direction you move.

A directional derivative measures the rate of change of $f(x,y)$ at a point in a specific direction. If $u$ is a unit vector showing the direction you want to travel, then the directional derivative of $f$ at $(x_0,y_0)$ in the direction of $u$ is written as $D_u f(x_0,y_0)$.

A common formula is

$$D_u f(x_0,y_0)=\nabla f(x_0,y_0)\cdot u$$

when $f$ is differentiable. Here $\nabla f$ is the gradient, and $\cdot$ means the dot product.

This tells us that to find the rate of change in a direction, we combine the partial derivatives with the direction vector. The direction must be a unit vector because we want the rate of change per unit distance, not per arbitrary vector length.

Real-World Meaning

Suppose $f(x,y)$ represents the temperature on a metal plate. If students moves across the plate in a direction $u$, the directional derivative tells how quickly the temperature changes along that path. A positive value means the temperature increases, a negative value means it decreases, and $0$ means no change in that direction.

Computing Directional Derivatives

To compute a directional derivative, you often follow these steps:

1. Find the partial derivatives $f_x$ and $f_y$.
2. Evaluate them at the point of interest.
3. Find the unit vector $u$ for the direction.
4. Use the gradient formula $D_u f=\nabla f\cdot u$.

Let’s look at an example.

Suppose

$$f(x,y)=x^2y+y^3$$

and we want the directional derivative at $(1,2)$ in the direction of the vector $v=\langle 3,4\rangle$.

First, find the gradient:

$$\nabla f(x,y)=\langle 2xy, x^2+3y^2\rangle$$

At $(1,2)$, this becomes

$$\nabla f(1,2)=\langle 4, 13\rangle$$

Next, turn $v$ into a unit vector. Its length is

$$\|v\|=\sqrt{3^2+4^2}=5$$

so

$$u=\left\langle \frac{3}{5},\frac{4}{5}\right\rangle$$

Now compute the dot product:

$$D_u f(1,2)=\langle 4,13\rangle\cdot\left\langle \frac{3}{5},\frac{4}{5}\right\rangle$$

$$D_u f(1,2)=\frac{12}{5}+\frac{52}{5}=\frac{64}{5}$$

So the function increases at a rate of $\frac{64}{5}$ units of $f$ per unit distance in that direction.

The Gradient: A Vector of Partial Derivatives

The gradient of a function $f(x,y)$ is the vector of its partial derivatives:

$$\nabla f(x,y)=\langle f_x(x,y), f_y(x,y)\rangle$$

For a function of three variables $f(x,y,z)$, the gradient is

$$\nabla f(x,y,z)=\langle f_x, f_y, f_z\rangle$$

The gradient gives important information about the function:

• It points in the direction of steepest increase.
• Its length tells how steep that increase is.
• It is perpendicular to level curves in two dimensions and level surfaces in three dimensions.

This last fact is very important for tangent planes and surfaces. A level curve is where $f(x,y)=c$, and a level surface is where $f(x,y,z)=c$. The gradient is normal, meaning perpendicular, to those shapes.

Why the Gradient Points Uphill

The directional derivative formula

$$D_u f=\nabla f\cdot u$$

shows that the rate of change depends on the angle between $\nabla f$ and $u$. The dot product is largest when $u$ points in the same direction as $\nabla f$. Therefore, the gradient gives the direction where $f$ increases most rapidly.

If $u$ points opposite the gradient, then the directional derivative is as negative as possible, meaning the function decreases fastest.

Example: Finding the Steepest Direction

Let

$$f(x,y)=x^2+2y^2$$

Find the gradient at $(1,-1)$.

We compute

$$f_x=2x, \qquad f_y=4y$$

so

$$\nabla f(x,y)=\langle 2x,4y\rangle$$

and

$$\nabla f(1,-1)=\langle 2,-4\rangle$$

This vector points in the direction of fastest increase. Its magnitude is

$$\|\nabla f(1,-1)\|=\sqrt{2^2+(-4)^2}=\sqrt{20}=2\sqrt{5}$$

That value is the maximum directional derivative at the point. So the steepest upward rate is $2\sqrt{5}$, and it happens in the direction of $\langle 2,-4\rangle$.

If students wanted the direction of steepest decrease, the direction would be the opposite vector $\langle -2,4\rangle$.

Connection to Tangent Planes and Linear Approximation

Directional derivatives and gradients are closely tied to tangent planes. A differentiable function near a point behaves almost like a plane when you zoom in enough. This is the idea of linear approximation.

For $z=f(x,y)$ near $(a,b)$, the tangent plane is

$$z=f(a,b)+f_x(a,b)(x-a)+f_y(a,b)(y-b)$$

The terms $f_x(a,b)$ and $f_y(a,b)$ are the components of the gradient. They tell how the surface changes if you move along the $x$-direction or $y$-direction.

In fact, the total differential is

$$dz=f_x\,dx+f_y\,dy$$

This is a small-change version of the tangent plane idea. If you choose a direction $u=\langle a,b\rangle$, then the directional derivative tells the rate of change per unit distance in that direction, while the differential tells the approximate change for small moves $dx$ and $dy$.

So, students, directional derivatives are not isolated facts. They are part of the same framework that gives tangent planes and linear approximation.

Connection to the Chain Rule

The chain rule in multivariable calculus explains how a function changes when its inputs depend on another variable, such as time $t$.

If $z=f(x,y)$, where $x=x(t)$ and $y=y(t)$, then

$$\frac{dz}{dt}=f_x\frac{dx}{dt}+f_y\frac{dy}{dt}$$

This can be written using the gradient:

$$\frac{dz}{dt}=\nabla f\cdot\left\langle \frac{dx}{dt},\frac{dy}{dt}\right\rangle$$

This is another directional derivative idea. The vector

$$\left\langle \frac{dx}{dt},\frac{dy}{dt}\right\rangle$$

shows the direction and speed of motion in the $xy$-plane, and the gradient tells how the function changes in that direction.

For example, if a hiker walks across a landscape with elevation $f(x,y)$, the hiker’s position changes with time, so the elevation changes with time too. The chain rule gives the rate of elevation change, and that rate is based on the gradient and the direction of motion.

Example: Chain Rule and Directional Rate

Suppose $f(x,y)=x^2y$, with $x(t)=t$ and $y(t)=t^2$.

First compute the partial derivatives:

$$f_x=2xy, \qquad f_y=x^2$$

Then use the chain rule:

$$\frac{dz}{dt}=f_x\frac{dx}{dt}+f_y\frac{dy}{dt}$$

Since

$$\frac{dx}{dt}=1 \qquad \text{and} \qquad \frac{dy}{dt}=2t$$

we get

$$\frac{dz}{dt}=2xy(1)+x^2(2t)$$

Substitute $x=t$ and $y=t^2$:

$$\frac{dz}{dt}=2(t)(t^2)+t^2(2t)=2t^3+2t^3=4t^3$$

This result shows how the surface changes as a point moves along the curve $(x(t),y(t))$.

Conclusion

Directional derivatives and gradients help students measure and understand change in many directions, not just one axis. The directional derivative tells the rate of change of a function in a chosen direction, while the gradient gathers all partial derivatives into one vector that points uphill fastest. Together, they explain tangent planes, linear approximation, and the multivariable chain rule. These ideas are central in studying how surfaces behave near a point and how quantities change along paths through space. Whether you are modeling temperature, elevation, or motion, gradients and directional derivatives provide the language for describing local change clearly and accurately.

Study Notes

• The directional derivative $D_u f$ measures the rate of change of $f$ in the direction of a unit vector $u$.
• For differentiable functions, $D_u f=\nabla f\cdot u$.
• The gradient is $\nabla f=\langle f_x,f_y\rangle$ for $f(x,y)$ and $\nabla f=\langle f_x,f_y,f_z\rangle$ for $f(x,y,z)$.
• The gradient points in the direction of steepest increase.
• The magnitude $\|\nabla f\|$ gives the maximum directional derivative.
• The gradient is perpendicular to level curves and level surfaces.
• The tangent plane for $z=f(x,y)$ at $(a,b)$ is $z=f(a,b)+f_x(a,b)(x-a)+f_y(a,b)(y-b)$.
• The total differential is $dz=f_x\,dx+f_y\,dy$.
• The multivariable chain rule can be written as $\frac{dz}{dt}=f_x\frac{dx}{dt}+f_y\frac{dy}{dt}=\nabla f\cdot\left\langle \frac{dx}{dt},\frac{dy}{dt}\right\rangle$.
• Directional derivatives and gradients are key tools for connecting local change, tangent planes, and motion along paths.