To parameterize a surface, you express its x, y, and z coordinates as functions of two parameters, typically called u and v. This converts a surface in three-dimensional space into a mapping from a flat two-dimensional region, making it possible to compute surface integrals, find normal vectors, and measure surface area. The core idea is the vector-valued function r(u,v) = x(u,v) i + y(u,v) j + z(u,v) k, where each point (u,v) in some domain D produces a point on the surface S.
What Parameterization Actually Does
A parametric curve uses one parameter to trace out a one-dimensional path. A parametric surface extends this idea by using two parameters to sweep out a two-dimensional sheet. You pick a region D in the uv-plane, and for every point (u,v) in that region, your parametric equations output a specific point (x, y, z) on the surface. The collection of all those output points is the surface S.
The three component functions, written out explicitly, are called the parametric equations of the surface:
- x = x(u, v)
- y = y(u, v)
- z = z(u, v)
Choosing the right parameters and the right domain D is the actual skill. The rest of this article walks through the standard approaches for the surface types you’ll encounter most often.
Graphs of Functions: The Simplest Case
If your surface is already given as z = f(x, y), parameterization is nearly automatic. Let u = x and v = y, and write:
- x = u
- y = v
- z = f(u, v)
The domain D is whatever region in the xy-plane the function is defined over. For example, if z = x² + y² over the unit disk, the parameterization is r(u,v) = u i + v j + (u² + v²) k, with u² + v² ≤ 1. This works for any surface that passes the “vertical line test” in the z-direction, meaning each (x,y) pair gives exactly one z value.
The same logic applies if the surface is given as x = g(y,z) or y = h(x,z). You just let the two independent variables be your parameters and the third coordinate becomes the function.
Cylinders
A circular cylinder of radius r and height h, centered on the z-axis, has a natural parameterization using an angle and a height variable. Let u be the angle around the circle and v be the height:
- x = r cos(u)
- y = r sin(u)
- z = v
The domain is 0 ≤ u ≤ 2π and 0 ≤ v ≤ h. As u sweeps from 0 to 2π, you trace a full circle at each height. As v goes from 0 to h, you stack those circles to build the cylinder. If you only need half the cylinder, restrict u to [0, π]. If the cylinder is centered on a different axis, rearrange which coordinates get the cos and sin terms.
Spheres
A sphere of radius a centered at the origin uses two angles from spherical coordinates. Let θ be the angle in the xy-plane (longitude) and φ be the angle measured down from the positive z-axis (colatitude):
- x = a cos(θ) sin(φ)
- y = a sin(θ) sin(φ)
- z = a cos(φ)
The domain is 0 ≤ θ ≤ 2π and 0 ≤ φ ≤ π. The angle φ only goes to π (not 2π) because sweeping from the north pole to the south pole is half a turn. If you need the upper hemisphere only, restrict φ to [0, π/2].
Be careful with conventions here. Some textbooks swap the roles of θ and φ, or use φ for the azimuthal angle instead. The formulas above follow the physics convention where φ is the polar angle. Check which convention your course uses before plugging in.
Planes
To parameterize a plane, you need a point on the plane and two direction vectors that lie in it. If the point is P₀ = (x₀, y₀, z₀) and the two non-parallel direction vectors are a and b, the parameterization is:
r(u, v) = P₀ + u·a + v·b
This sweeps out the entire infinite plane as u and v range over all real numbers. If you need only a finite piece (a parallelogram, for instance), restrict u and v to bounded intervals. For a triangular region, you might use 0 ≤ u ≤ 1 and 0 ≤ v ≤ 1 − u.
If you’re given three points P, Q, R on the plane, you can form the two direction vectors as a = Q − P and b = R − P, then use P as the base point.
Surfaces of Revolution
When a curve in the xz-plane is revolved around the z-axis, the resulting surface has a clean parameterization. Suppose the generating curve is given by x = g(t), z = h(t) for some parameter t. Revolving it introduces an angle θ around the z-axis:
- x = g(t) cos(θ)
- y = g(t) sin(θ)
- z = h(t)
Here θ ranges from 0 to 2π for a full revolution, and t covers whatever interval defines the original curve. The function g(t) acts as the radius at each height, so it should be non-negative. This approach works for cones (where g is linear), paraboloids (where g involves a square root), and many other shapes.
If the curve is instead revolved around the x-axis, swap the roles of the coordinates: y and z get the cos and sin factors, while x stays as the curve parameter.
Torus
A torus (donut shape) is built by revolving a circle around an axis that doesn’t pass through the circle. If the circle has radius r and its center sits at distance R from the z-axis, the parameterization uses two angles, θ and ψ, both running from 0 to 2π:
- x = (R + r cos θ) cos ψ
- y = (R + r cos θ) sin ψ
- z = r sin θ
The angle θ traces around the small circular cross-section (the tube), while ψ sweeps that tube around the z-axis to form the full donut. R is called the major radius and r is the minor radius.
Choosing Parameter Bounds
Getting the equations right is only half the problem. You also need the correct domain D, meaning the set of (u,v) values that trace out exactly the portion of the surface you want. A few guidelines help:
For angle parameters representing a full rotation, the range is 0 to 2π. For a polar angle measured from one pole to the other (as in a sphere), it’s 0 to π. For height or length parameters, read the bounds directly from the geometry of the problem: a cylinder from z = 0 to z = 5 means v goes from 0 to 5.
When the surface is a graph z = f(x,y) over some region, the domain D in the uv-plane is that region itself. If the region is described by inequalities like x² + y² ≤ 4, the domain is u² + v² ≤ 4. Sometimes switching to polar coordinates for u and v (setting u = s cos t, v = s sin t) makes the domain simpler to describe.
A good check: if you plug in the boundary values of your parameters, you should get the boundary curves of your surface. If the surface is a hemisphere, plugging in the maximum polar angle should give you the equator.
Why Parameterization Matters for Calculus
The main payoff of parameterization is that it lets you compute integrals over surfaces. Once you have r(u,v), you can take its partial derivatives with respect to u and v. These two vectors, r_u and r_v, are tangent to the surface at each point. Their cross product r_u × r_v gives a vector perpendicular to the surface, and the magnitude of that cross product, |r_u × r_v|, measures how much a tiny rectangle in the uv-plane gets stretched when mapped onto the surface.
This stretching factor is sometimes called the Jacobian of the parameterization. The surface area element is dS = |r_u × r_v| du dv, and a surface integral of a function f over S becomes:
∬_S f dS = ∬_D f(r(u,v)) |r_u × r_v| du dv
This formula converts a surface integral into a standard double integral over the flat domain D, which you can evaluate with the techniques from earlier in your calculus course. The quality of your parameterization directly affects how difficult that double integral is. A well-chosen parameterization produces clean partial derivatives and a manageable cross product; a poorly chosen one can make the integral nearly impossible to compute by hand.