Calculus III: Parametric Surfaces

Moving from curves to surfaces is a pivotal leap in multivariable calculus, unlocking the ability to model and analyze complex 3D shapes essential in engineering fields like fluid dynamics, computer-aided design, and heat transfer. Parametric surfaces provide the most flexible framework for this, representing a surface not as a single static equation, but as a dynamic mapping from a flat parameter region into three-dimensional space. Mastering this representation is the key to performing calculus on surfaces, including finding tangent planes, computing area, and integrating functions over complex geometries.

The Foundation: Surface Parametrization

A parametric surface is defined by a vector-valued function of two independent parameters, typically $u$ and $v$. The general form is: $$\vec{r}(u,v)=⟨x(u,v),y(u,v),z(u,v)⟩,(u,v)\in D$$ where $D$ is a region in the $uv$-plane. This function $\vec{r}$ maps each point $(u,v)$ in the parameter domain $D$ to a unique point $(x,y,z)$ on the surface $S$ in ${\mathbb{R}}^3$. The core idea is that two degrees of freedom (the parameters) are needed to trace out a two-dimensional object. Choosing an effective parametrization is the first critical skill, as it simplifies all subsequent calculations.

Essential Parametrization Examples

Engineers must be fluent in parametrizing standard surfaces. Each example highlights a different geometric relationship between the parameters and the coordinates.

Planes: A plane through a point $P_0$ with position vector ${\vec{r}}_0$ and spanned by two non-parallel vectors $\vec{a}$ and $\vec{b}$ is parametrized linearly: $$\vec{r}(u,v)={\vec{r}}_0+u\vec{a}+v\vec{b},u,v\in \mathbb{R}.$$ Here, $u$ and $v$ act as scalar multipliers moving you along the two direction vectors that lie in the plane.

Cylinders: Consider a right circular cylinder of radius $R$ along the $z$-axis. The $x$ and $y$ coordinates trace a circle, while $z$ is free. A natural parametrization uses an angle $\theta$ and height $z$: $$\vec{r}(\theta ,z)=⟨R\cos \theta ,R\sin \theta ,z⟩,\theta \in [0,2\pi ],z\in \mathbb{R}.$$ The parameter domain $D$ is an infinite strip. This showcases how one parameter often handles a "curled" dimension while the other handles a "linear" one.

Spheres: For a sphere of radius $\rho$ centered at the origin, spherical coordinates provide the perfect parametrization. Let $\varphi$ be the polar angle from the positive $z$-axis ($0\le \varphi \le \pi$) and $\theta$ be the azimuthal angle in the $xy$-plane ($0\le \theta <2\pi$): $$\vec{r}(\varphi ,\theta )=⟨\rho \sin \varphi \cos \theta ,\rho \sin \varphi \sin \theta ,\rho \cos \varphi ⟩.$$ The parameters $\varphi$ and $\theta$ correspond to lines of latitude and longitude on the globe.

Cones: A right circular cone with height $H$ and base radius $R$, aligned along the $z$-axis, can be parametrized by relating the radial distance to the height. Using a "radius fraction" parameter $u$ (where $u=0$ at the apex and $u=1$ at the base) and an angle $v$: $$\vec{r}(u,v)=⟨uR\cos v,uR\sin v,H(1-u)⟩,u\in [0,1],v\in [0,2\pi ].$$ This linear relationship between the $xy$-radius and the $z$-coordinate is the defining feature of the cone.

Tangent Planes and Surface Normal Vectors

To find a tangent plane at a point on a parametric surface, we need two vectors that lie in that plane. These are provided by the partial derivatives of the parametrization. The partial derivative with respect to $u$, ${\vec{r}}_u=\frac{\partial \vec{r}}{\partial u}=⟨x_u,y_u,z_u⟩$, gives a tangent vector in the direction of increasing $u$ with $v$ held constant. Similarly, ${\vec{r}}_v$ gives a tangent vector in the direction of increasing $v$.

Provided ${\vec{r}}_u$ and ${\vec{r}}_v$ are continuous and non-parallel at a point, they span the tangent plane at that point. A vector normal to this plane—and hence normal to the surface—is given by their cross product: $$\vec{n}={\vec{r}}_u\times {\vec{r}}_v.$$ This surface normal vector is crucial. For example, the equation of the tangent plane at a point $\vec{r}(u_0,v_0)$ is $\vec{n}\cdot ⟨x-x_0,y-y_0,z-z_0⟩=0$, where $\vec{n}$ is evaluated at $(u_0,v_0)$. In fluid flow problems, this normal vector is essential for calculating flux across the surface.

Computing Surface Area

We cannot use the formula for the area of a parallelogram to find the area of a curved surface directly. However, we can approximate a small piece of the surface $dS$ by the area of the parallelogram spanned by the small tangent vectors ${\vec{r}}_{u}du$ and ${\vec{r}}_{v}dv$. The area of this infinitesimal parallelogram is: $$dS=\Vert {\vec{r}}_u\times {\vec{r}}_v\Vert dudv.$$ To find the total surface area of $S$, we integrate this quantity over the parameter domain $D$: $$\text{Area}(S)={\iint }_D\Vert {\vec{r}}_u\times {\vec{r}}_v\Vert dA.$$ The term $\Vert {\vec{r}}_u\times {\vec{r}}_v\Vert$ acts as a "scaling factor" that converts an area element $dA=dudv$ in the parameter plane into the corresponding area element $dS$ on the curved surface. For a cylinder $\vec{r}(\theta ,z)=⟨R\cos \theta ,R\sin \theta ,z⟩$, we find ${\vec{r}}_{\theta }\times {\vec{r}}_z=⟨R\cos \theta ,R\sin \theta ,0⟩$, which has magnitude $R$. The area formula then integrates $Rd\theta dz$, correctly yielding the familiar formula $2\pi R\cdot \text{height}$.

Choosing Parametrizations for Integration

The final and most applied skill is selecting a parametrization to evaluate a surface integral, either of a scalar function $f(x,y,z)$ or a vector field $\vec{F}(x,y,z)$. The choice is dictated by two factors: the geometry of the surface and the function being integrated. A good parametrization should simplify the integrand and/or have a simple parameter domain $D$.

For a surface integral of a scalar function, the formula is: $${\iint }_{S}f(x,y,z)dS={\iint }_{D}f(\vec{r}(u,v))\Vert {\vec{r}}_u\times {\vec{r}}_v\Vert dA.$$ Your goal is to choose $\vec{r}(u,v)$ so that $f(\vec{r}(u,v))$ and $\Vert {\vec{r}}_u\times {\vec{r}}_v\Vert$ combine into an integrable function over a rectangular or simple polar-like region $D$. For instance, integrating a function symmetric about the $z$-axis over a sphere strongly suggests using the spherical coordinate parametrization. The parameter domain $D$ becomes a simple rectangle in $(\varphi ,\theta )$-space: $0\le \varphi \le \pi$, $0\le \theta \le 2\pi$.

Common Pitfalls

1. Ignoring the Parameter Domain: A parametrization is incomplete without specifying the region $D$ in the $uv$-plane. Forgetting this leads to incorrect bounds of integration when computing area or evaluating a surface integral. Always sketch or explicitly define $D$.
2. Incorrect Normal Vector Orientation: The cross product ${\vec{r}}_u\times {\vec{r}}_v$ gives a normal, but its direction (inward/outward) matters for flux integrals. If the problem requires an outward normal, you must check the sign of your computed $\vec{n}$ against the geometry. For a closed surface, the outward normal is conventionally used.
3. Miscomputing the Area Element: The area element is $\Vert {\vec{r}}_u\times {\vec{r}}_v\Vert dudv$, not simply $dudv$. A frequent error is to forget to take the magnitude of the cross product, or to miscalculate the cross product itself. Always compute ${\vec{r}}_u\times {\vec{r}}_v$ explicitly.
4. Overcomplicating the Parametrization: Use the symmetry of the surface. Don't force a parametrization in $x$ and $y$ if the surface is naturally described in cylindrical or spherical coordinates. The simplest parametrization that respects the geometry will lead to the most manageable integrals.

Summary

• Parametric surfaces are defined by a vector function $\vec{r}(u,v)$ that maps a 2D parameter domain $D$ into 3D space, providing the most versatile way to describe complex geometries.
• The tangent plane at a point is spanned by the partial derivative vectors ${\vec{r}}_u$ and ${\vec{r}}_v$, and the surface normal vector is given by their cross product $\vec{n}={\vec{r}}_u\times {\vec{r}}_v$.
• Surface area is computed by integrating the magnitude of this cross product over the parameter domain: $\text{Area}(S)={\iint }_D\Vert {\vec{r}}_u\times {\vec{r}}_v\Vert dA$.
• The key to evaluating surface integrals is choosing a parametrization that simplifies the integrand $f(\vec{r}(u,v))\Vert {\vec{r}}_u\times {\vec{r}}_v\Vert$ and has a simple parameter domain $D$ for integration.
• Always pay close attention to the parameter domain and the orientation of the computed normal vector, as these are common sources of error in applied engineering calculations.