Active Calculus · 11: Multiple Integrals

11.6: Surfaces Defined Parametrically and Surface Area

Motivating Questions
  • What is a parameterization of a surface?
  • How do we find the surface area of a parametrically defined surface?

We have now studied at length how curves in space can be defined parametrically by functions of the form r(t)=⟨x(t),y(t),z(t)⟩,\mathbf{r}(t) = \langle x(t), y(t), z(t) \rangle\text{,} and surfaces can be represented by functions z=f(x,y).z = f(x,y)\text{.} In what follows, we will see how we can also define surfaces parametrically. A one-dimensional curve in space results from a vector function that relies upon one parameter, so a two-dimensional surface naturally involves the use of two parameters. If x=x(s,t),x = x(s, t)\text{,} y=y(s,t),y = y(s, t)\text{,} and z=z(s,t)z = z(s, t) are functions of independent parameters ss and t,t\text{,} then the terminal points of all vectors of the form

r(s,t)=x(s,t)i+y(s,t)j+z(s,t)k \mathbf{r}(s, t) = x(s, t) \mathbf{i} + y(s, t) \mathbf{j} + z(s, t) \mathbf{k}

form a surface in space. The equations x=x(s,t),x=x(s,t)\text{,} y=y(s,t),y=y(s,t)\text{,} and z=z(s,t)z=z(s,t) are the parametric equations for the surface, or a parametrization of the surface. In Preview Activity 11.6.111.6.1 we investigate how to parameterize a cylinder and a cone.

Preview Activity 11.6.111.6.1

Recall the standard parameterization of the unit circle that is given by

x(t)=cos⁡(t)     and     y(t)=sin⁡(t), x(t) = \cos(t) \ \ \ \ \text{ and } \ \ \ \ y(t) = \sin(t),

where 0≤t≤2π.0 \le t \le 2\pi\text{.}

  1. Determine a parameterization of the circle of radius 1 in R3\mathbb{R}^3 that has its center at (0,0,1)(0,0,1) and lies in the plane z=1.z=1\text{.}
  2. Determine a parameterization of the circle of radius 1 in 3-space that has its center at (0,0,−1)(0,0,-1) and lies in the plane z=−1.z=-1\text{.}
  3. Determine a parameterization of the circle of radius 1 in 3-space that has its center at (0,0,5)(0,0,5) and lies in the plane z=5.z=5\text{.}
  4. Taking into account your responses in (a), (b), and (c), describe the graph that results from the set of parametric equations
    x(s,t)=cos⁡(t),    y(s,t)=sin⁡(t),     and     z(s,t)=s, x(s,t) = \cos(t), \ \ \ \ y(s,t) = \sin(t), \ \ \ \ \text{ and } \ \ \ \ z(s,t) = s,

    where 0≤t≤2π0 \le t \le 2\pi and −5≤s≤5.-5 \le s \le 5\text{.} Explain your thinking.

  5. Just as a cylinder can be viewed as a “stack” of circles of constant radius, a cone can be viewed as a stack of circles with varying radius. Modify the parametrizations of the circles above in order to construct the parameterization of a cone whose vertex lies at the origin, whose base radius is 4, and whose height is 3, where the base of the cone lies in the plane z=3.z = 3\text{.} Use appropriate technology to plot the parametric equations you develop. (Hint: The cross sections parallel to the xyxy-plane are circles, with the radii varying linearly as zz increases.)

Parametric Surfaces

In a single-variable setting, any function may have its graph expressed parametrically. For instance, given y=g(x),y = g(x)\text{,} by considering the parameterization ⟨t,g(t)⟩\langle t, g(t) \rangle (where tt belongs to the domain of gg), we generate the same curve. What is more important is that certain curves that are not functions may be represented parametrically; for instance, the circle (which cannot be represented by a single function) can be parameterized by ⟨cos⁡(t),sin⁡(t)⟩,\langle \cos(t), \sin(t) \rangle\text{,} where 0≤t≤2π.0 \le t \le 2\pi\text{.}

In the same way, in a two-variable setting, the surface z=f(x,y)z = f(x,y) may be expressed parametrically by considering

⟨x(s,t),y(s,t),z(s,t)⟩=⟨s,t,f(s,t)⟩, \langle x(s,t), y(s,t), z(s,t) \rangle = \langle s, t, f(s,t) \rangle,

where (s,t)(s,t) varies over the entire domain of f.f\text{.} Therefore, any familiar surface that we have studied so far can be generated as a parametric surface. But what is more powerful is that there are surfaces that cannot be generated by a single function z=f(x,y)z = f(x,y) (such as the unit sphere), but that can be represented parametrically. We now consider an important example.

Example 11.6.111.6.1

Consider the torus (or doughnut) shown in Figure 11.6.111.6.1.

fig_11_6_torus.svg

Figure 11.6.111.6.1. A torus.

To find a parametrization of this torus, we recall our work in Preview Activity 11.6.111.6.1. There, we saw that a circle of radius rr that has its center at the point (0,0,z0)(0,0,z_0) and is contained in the horizontal plane z=z0,z = z_0\text{,} as shown in Figure 11.6.211.6.2, can be parametrized using the vector-valued function r\mathbf{r} defined by

r(t)=rcos⁡(t)i+rsin⁡(t)j+z0k \mathbf{r}(t) = r\cos(t)\mathbf{i} + r\sin(t)\mathbf{j} + z_0\mathbf{k}

where 0≤t≤2π.0\leq t\leq 2\pi\text{.}

fig_11_6_revolve.svg

Figure 11.6.211.6.2. A circle in a horizontal plane centered at (0,0,z0).(0,0,z_0)\text{.}

To obtain the torus in Figure 11.6.111.6.1, we begin with a circle of radius aa in the xzxz-plane centered at (b,0),(b,0)\text{,} as shown on the left of Figure 11.6.411.6.4. We may parametrize the points on this circle, using the parameter s,s\text{,} by using the equations

x(s)=b+acos⁡(s) and z(s)=asin⁡(s), x(s) = b + a\cos(s)\ \mbox{and} \ z(s) = a\sin(s),

where 0≤s≤2π.0\leq s \leq 2\pi\text{.}

fig_11_6_torus_cross.svgfig_11_6_torus_revolve.svg 

Figure 11.6.311.6.3. Revolving a circle to obtain a torus.

Let's focus our attention on one point on this circle, such as the indicated point, which has coordinates (x(s),0,z(s))(x(s), 0, z(s)) for a fixed value of the parameter s.s\text{.} When this point is revolved about the zz-axis, we obtain a circle contained in a horizontal plane centered at (0,0,z(s))(0,0,z(s)) and having radius x(s),x(s)\text{,} as shown on the right of Figure 11.6.311.6.3. If we let tt be the new parameter that generates the circle for the rotation about the zz-axis, this circle may be parametrized by

r(s,t)=x(s)cos⁡(t)i+x(s)sin⁡(t)j+z(s)k. \mathbf{r}(s,t) = x(s)\cos(t)\mathbf{i} + x(s)\sin(t)\mathbf{j} + z(s)\mathbf{k}.

Now using our earlier parametric equations for x(s)x(s) and z(s)z(s) for the original smaller circle, we have an overall parameterization of the torus given by

r(s,t)=(b+acos⁡(s))cos⁡(t)i+(b+acos⁡(s))sin⁡(t)j+asin⁡(s)k. \mathbf{r}(s,t) =(b+a\cos(s))\cos(t)\mathbf{i} + (b+a\cos(s))\sin(t)\mathbf{j} + a\sin(s)\mathbf{k}.

To trace out the entire torus, we require that the parameters vary through the values 0≤s≤2π0\leq s\leq 2\pi and 0≤t≤2π.0\leq t\leq 2\pi\text{.}

Activity 11.6.211.6.2

In this activity, we seek a parametrization of the sphere of radius RR centered at the origin, as shown on the left in Figure 11.6.411.6.4. Notice that this sphere may be obtained by revolving a half-circle contained in the xzxz-plane about the zz-axis, as shown on the right.

fig_11_6_sphere.svgfig_11_6_sphere_half.svg 

Figure 11.6.411.6.4. A sphere obtained by revolving a half-circle.
  1. Begin by writing a parametrization of this half-circle using the parameter s:s\text{:}
    x(s)=…         z(s)=…. x(s) = \ldots \ \ \ \ \ \ \ \ \ z(s) = \ldots.

    Be sure to state the domain of the parameter s.s\text{.}

  2. By revolving the points on this half-circle about the zz-axis, obtain a parametrization r(s,t)\mathbf{r}(s,t) of the points on the sphere of radius R.R\text{.} Be sure to include the domain of both parameters ss and t.t\text{.} (Hint: What is the radius of the circle obtained when revolving a point on the half-circle around the zz axis?)
  3. Draw the surface defined by your parameterization with appropriate technology.

The Surface Area of Parametrically Defined Surfaces

Recall that a differentiable function is locally linear — that is, if we zoom in on the surface around a point, the surface looks like its tangent plane. We now exploit this idea in order to determine the surface area generated by a parametrization ⟨x(s,t),y(s,t),z(s,t)⟩.\langle x(s,t), y(s,t), z(s,t) \rangle\text{.} The basic idea is a familiar one: we will subdivide the surface into small pieces, in the approximate shape of small parallelograms, and thus estimate the entire the surface area by adding the areas of these approximation parallelograms. Ultimately, we use an integral to sum these approximations and determine the exact surface area.

Let

r(s,t)=x(s,t)i+y(s,t)j+z(s,t)k \mathbf{r}(s,t) = x(s,t) \mathbf{i} + y(s,t) \mathbf{j} + z(s,t) \mathbf{k}

define a surface over a rectangular domain a≤s≤ba \leq s \leq b and c≤t≤d.c \leq t \leq d\text{.} As a function of two variables, ss and t,t\text{,} it is natural to consider the two partial derivatives of the vector-valued function r,\mathbf{r}\text{,} which we define by

rs(s,t)=xs(s,t)i+ys(s,t)j+zs(s,t)krt(s,t)=xt(s,t)i+yt(s,t)j+zt(s,t)k.\begin{align*} \mathbf{r}_s(s,t) & = x_s(s,t) \mathbf{i} + y_s(s,t) \mathbf{j} + z_s(s,t) \mathbf{k}\\[4pt] \mathbf{r}_t(s,t) & = x_t(s,t) \mathbf{i} + y_t(s,t) \mathbf{j} + z_t(s,t) \mathbf{k}. \end{align*}

In the usual way, we slice the domain into small rectangles. In particular, we partition the interval [a,b][a,b] into mm subintervals of length Δs=b−an\Delta s = \frac{b-a}{n} and let s0,s_0\text{,} s1,s_1\text{,} …,\ldots\text{,} sms_m be the endpoints of these subintervals, where a=s0<s1<s2<⋯<sm=b.a = s_0\lt s_1\lt s_2 \lt \cdots \lt s_m = b\text{.} Also partition the interval [c,d][c,d] into nn subintervals of equal length Δt=d−cn\Delta t = \frac{d-c}{n} and let t0,t_0\text{,} t1,t_1\text{,} …,\ldots\text{,} tnt_n be the endpoints of these subintervals, where c=t0<t1<t2<⋯<tn=d.c = t_0\lt t_1\lt t_2 \lt \cdots \lt t_n = d\text{.} These two partitions create a partition of the rectangle R=[a,b]×[c,d]R = [a,b] \times [c,d] in stst-coordinates into mnmn sub-rectangles RijR_{ij} with opposite vertices (si−1,tj−1)(s_{i-1},t_{j-1}) and (si,tj)(s_i, t_j) for ii between 11 and mm and jj between 11 and n.n\text{.} These rectangles all have equal area ΔA=Δs⋅Δt.\Delta A = \Delta s \cdot \Delta t\text{.}

Now we want to think about the small piece of area on the surface itself that lies above one of these small rectangles in the domain. Observe that if we increase ss by a small amount Δs\Delta s from the point (si−1,tj−1)(s_{i-1},t_{j-1}) in the domain, then r\mathbf{r} changes by approximately rs(si−1,tj−1)Δs.\mathbf{r}_s(s_{i-1},t_{j-1}) \Delta s\text{.} Similarly, if we increase tt by a small amount Δt\Delta t from the point (si−1,tj−1),(s_{i-1},t_{j-1})\text{,} then r\mathbf{r} changes by approximately rt(si−1,tj−1)Δt.\mathbf{r}_t(s_{i-1},t_{j-1}) \Delta t\text{.} So we can approximate the surface defined by r\mathbf{r} on the stst-rectangle [si−1,si]×[tj−1,tj][s_{i-1},s_i] \times [t_{j-1}, t_{j}] with the parallelogram determined by the vectors rs(si−1,tj−1)Δs\mathbf{r}_s(s_{i-1},t_{j-1}) \Delta s and rt(si−1,tj−1)Δt,\mathbf{r}_t(s_{i-1},t_{j-1}) \Delta t\text{,} as seen in Figure 11.6.511.6.5.

fig_11_6_approx.svg

Figure 11.6.511.6.5. Approximation surface area with a parallelogram.

Say that the small parallelogram has area Sij.S_{ij}\text{.} If we can find its area, then all that remains is to sum the areas of all of the generated parallelograms and take a limit. Recall from our earlier work in the course that given two vectors u\mathbf{u} and v,\mathbf{v}\text{,} the area of the parallelogram spanned by u\mathbf{u} and v\mathbf{v} is given by the magnitude of their cross product, ∣u×v∣.| \mathbf{u} \times \mathbf{v}|\text{.} In the present context, it follows that the area, Sij,S_{ij}\text{,} of the parallelogram determined by the vectors rs(si−1,tj−1)Δs\mathbf{r}_s(s_{i-1},t_{j-1}) \Delta s and rt(si−1,tj−1)Δt\mathbf{r}_t(s_{i-1},t_{j-1}) \Delta t is

Sij=∣(rs(si−1,tj−1)Δs)×(rt(si−1,tj−1)Δt)∣=∣rs(si−1,tj−1)×rt(si−1,tj−1)∣ΔsΔt, \begin{align} S_{ij} & = |(\mathbf{r}_s(s_{i-1},t_{j-1}) \Delta s) \times (\mathbf{r}_t(s_{i-1},t_{j-1}) \Delta t)| \notag\\[4pt] & = |\mathbf{r}_s(s_{i-1},t_{j-1}) \times \mathbf{r}_t(s_{i-1},t_{j-1})| \Delta s \Delta t, \end{align}

where the latter equality holds from standard properties of the cross product and length.

We sum the surface area approximations from Equation \ref{eq_11_6_parallelogram_area} over all sub-rectangles to obtain an estimate for the total surface area, S,S\text{,} given by

S≈∑i=1m∑j=1n∣rs(si−1,tj−1)×rt(si−1,tj−1)∣ΔsΔt. S \approx \sum_{i=1}^m \sum_{j=1}^n |\mathbf{r}_s(s_{i-1},t_{j-1}) \times \mathbf{r}_t(s_{i-1},t_{j-1})| \Delta s \Delta t.

Taking the limit as m,n→∞m, n \to \infty shows that the surface area of the surface defined by r\mathbf{r} over the domain DD is given as follows.

Surface area

Let r(s,t)=⟨x(s,t),y(s,t),z(s,t)⟩\mathbf{r}(s,t) = \langle x(s,t), y(s,t), z(s,t) \rangle be a parameterization of a smooth surface over a domain D.D\text{.} The area of the surface defined by r\mathbf{r} on DD is given by

S=∬D∣rs×rt∣ dA. S = \iint_D |\mathbf{r}_s \times \mathbf{r}_t | \ dA.

Activity 11.6.311.6.3

Consider the cylinder with radius aa and height hh defined parametrically by

r(s,t)=acos⁡(s)i+asin⁡(s)j+tk \mathbf{r}(s,t) = a\cos(s) \mathbf{i} + a\sin(s) \mathbf{j} + t \mathbf{k}

for 0≤s≤2π0 \leq s \leq 2\pi and 0≤t≤h,0 \leq t \leq h\text{,} as shown in Figure 11.6.611.6.6.

fig_11_6_cylinder.svg

Figure 11.6.611.6.6. A cylinder.
  1. Set up an iterated integral to determine the surface area of this cylinder.
  2. Evaluate the iterated integral.
  3. Recall that one way to think about the surface area of a cylinder is to cut the cylinder horizontally and find the perimeter of the resulting cross sectional circle, then multiply by the height. Calculate the surface area of the given cylinder using this alternate approach, and compare your work in (b).

As we noted earlier, we can take any surface z=f(x,y)z = f(x,y) and generate a corresponding parameterization for the surface by writing ⟨s,t,f(s,t)⟩.\langle s, t, f(s,t) \rangle\text{.} Hence, we can use our recent work with parametrically defined surfaces to find the surface area that is generated by a function f=f(x,y)f = f(x,y) over a given domain.

Activity 11.6.411.6.4

Let z=f(x,y)z = f(x,y) define a smooth surface, and consider the corresponding parameterization r(s,t)=⟨s,t,f(s,t)⟩.\mathbf{r}(s,t) = \langle s, t, f(s,t) \rangle\text{.}

  1. Let DD be a region in the domain of f.f\text{.} Using Equation \ref{E_surface_area}, show that the area, S,S\text{,} of the surface defined by the graph of ff over DD is
    S=∬D(fx(x,y))2+(fy(x,y))2+1 dA. S = \iint_D \sqrt{\left(f_x(x,y)\right)^2 + \left(f_y(x,y)\right)^2 + 1} \ dA.
  2. Use the formula developed in (a) to calculate the area of the surface defined by f(x,y)=4−x2f(x,y) = \sqrt{4-x^2} over the rectangle D=[−2,2]×[0,3].D = [-2,2] \times [0,3]\text{.}
  3. Observe that the surface of the solid describe in (b) is half of a circular cylinder. Use the standard formula for the surface area of a cylinder to calculate the surface area in a different way, and compare your result from (b).

Summary

  • A parameterization of a curve describes the coordinates of a point on the curve in terms of a single parameter t,t\text{,} while a parameterization of a surface describes the coordinates of points on the surface in terms of two independent parameters.
  • If r(s,t)=⟨x(s,t),y(s,t),z(s,t)⟩\mathbf{r}(s,t) = \langle x(s,t), y(s,t), z(s,t) \rangle describes a smooth surface in 3-space on a domain D,D\text{,} then the area, S,S\text{,} of that surface is given by
    S=∬D∣rs×rt∣ dA. S = \iint_D |\mathbf{r}_s \times \mathbf{r}_t| \ dA.

Credit and licence

This page titled 11.6: Surfaces Defined Parametrically and Surface Area was last modified on Fri, 29 Sep 2023 19:56:12 GMT and is shared under a CC BY-SA 4.0 license and was authored, remixed, and/or curated by Matthew Boelkins, David Austin, and Steven Schlicker ( ScholarWorks @Grand Valley State University ) via source content that was edited to the style and standards of the LibreTexts platform.

Source: https://math.libretexts.org/Bookshelves/Calculus/Active_Calculus_(Boelkins_et_al.)/11%3A_Multiple_Integrals/11.06%3A_Surfaces_Defined_Parametrically_and_Surface_Area

Licence: CC BY-SA 4.0 (creativecommons.org/licenses/by-sa/4.0/). You may share and adapt it under the same licence, with this credit. WorldClass adapted the formatting only: the text is the authors', with interactive parts and links to other sites left out.

Build status · Flash Infinity AI™ is a trademark of AI-Flash Inc.. WorldClass is not affiliated with, approved or accredited by any exam board, curriculum provider or university unless a page says so.