Concentrate on any one the tiny pieces. Here is a greatly magnified sketch of it, looking at it from above.
We wish to compute its area, which we'll call \(\dee{S}\text{.}\) Now this little piece of surface need not be parallel to the \(xy\)-plane, and indeed need not even be flat. But if the piece is really tiny, it's almost flat. We'll now approximate it by something that is flat, and whose area we know. To start, we'll determine the corners of the piece. To do so, we first determine the bounding curves of the piece. Look at the figure above, and recall that, on the surface \(z=f(x,y)\text{.}\)
The upper blue curve was constructed by holding \(x\) fixed at the value \(x_0\text{,}\) and sketching the curve swept out by \(x_0\,\hi+y\,\hj + f(x_0,y)\,\hk\) as \(y\) varied, and
the lower blue curve was constructed by holding \(x\) fixed at the slightly larger value \(x_0+\dee{x}\text{,}\) and sketching the curve swept out by \((x_0+\dee{x})\,\hi+y\,\hj + f(x_0+\dee{x},y)\,\hk\) as \(y\) varied.
The red curves were constructed similarly, by holding \(y\) fixed and varying \(x\text{.}\)
So the four intersection points in the figure are
\begin{alignat*}{1}
P_0&=x_0\,\hi+y_0\,\hj + f(x_0,y_0)\,\hk\\
P_1&=x_0\,\hi+(y_0+\dee{y})\,\hj + f(x_0,y_0+\dee{y})\,\hk\\
P_2&=(x_0+\dee{x})\,\hi+y_0\,\hj + f(x_0+\dee{x},y_0)\,\hk\\
P_3&=(x_0+\dee{x})\,\hi+(y_0+\dee{y})\,\hj + f(x_0+\dee{x},y_0+\dee{y})\,\hk
\end{alignat*}
Now, for any small constants
\(\dee{X}\) and
\(\dee{Y}\text{,}\) we have the linear approximation
1
\begin{align*}
f(x_0+\dee{X},y_0+\dee{Y})
&\approx f(x_0\,,\,y_0)
+\pdiff{f}{x}(x_0\,,\,y_0)\,\dee{X}
+\pdiff{f}{y}(x_0\,,\,y_0)\,\dee{Y}
\end{align*}
Applying this three times, once with \(\dee{X}=0\text{,}\) \(\dee{Y}=\dee{y}\) (to approximate \(P_1\)), once with \(\dee{X}=\dee{x}\text{,}\) \(\dee{Y}=0\) (to approximate \(P_2\)), and once with \(\dee{X}=\dee{x}\text{,}\) \(\dee{Y}=\dee{y}\) (to approximate \(P_3\)),
\begin{alignat*}{1}
P_1&\approx P_0 \phantom{\ +\ \dee{x}\,\hi}
\ +\ \dee{y}\,\hj\ \ +\ \pdiff{f}{y}(x_0\,,\,y_0)\,\dee{y}\,\hk\\
P_2&\approx P_0
\ +\ \dee{x}\,\hi \phantom{\ +\ \dee{y}\,\hj}
\ \ +\ \pdiff{f}{x}(x_0\,,\,y_0)\,\dee{x}\,\hk\\
P_3&\approx P_0
\ +\ \dee{x}\,\hi\ +\ \dee{y}\,\hj \ +\
\Big[ \pdiff{f}{x}(x_0\,,\,y_0)\,\dee{x}
+ \pdiff{f}{y}(x_0\,,\,y_0)\,\dee{y}\Big]\,\hk
\end{alignat*}
Of course we have only approximated the positions of the corners and so have introduced errors. However, with more work, one can bound those errors (like we in the optional §
3.2.4) and show that in the limit
\(\dee{x},\dee{y}\rightarrow 0\text{,}\) all of the error terms that we dropped contribute exactly
\(0\) to the integral.
Denote by
\(\theta\) the angle between the vectors
\(\overrightarrow{P_0P_1}\) and
\(\overrightarrow{P_0P_2}\text{.}\) The base of the parallelogram,
\(\overrightarrow{P_0P_1}\text{,}\) has length
\(\big|\overrightarrow{P_0P_1}\big|\text{,}\) and the height of the parallelogram is
\(\big|\overrightarrow{P_0P_2}\big|\,\sin\theta\text{.}\) So the area of the parallelogram is
2 , by Theorem
1.2.23,
\begin{align*}
\dee{S}
=|\overrightarrow{P_0P_1}|\ |\overrightarrow{P_0P_2}| \ \sin\theta
&= \big|\overrightarrow{P_0P_1}\times\overrightarrow{P_0P_2}\big|\\
&\hskip-0.5in\approx \bigg|\left(\hj\ +\ \pdiff{f}{y}(x_0\,,\,y_0)\,\hk\right)\times
\left(\hi\ +\ \pdiff{f}{x}(x_0\,,\,y_0)\,\hk\right)\bigg|
\dee{x}\dee{y}
\end{align*}
The cross product is easily evaluated:
\begin{align*}
\left(\hj\ +\ \pdiff{f}{y}(x_0\,,\,y_0)\,\hk\right)\times
\left(\hi\ +\ \pdiff{f}{x}(x_0\,,\,y_0)\,\hk\right)
&=\det\left[\begin{matrix}
\hi & \hj & \hk \\
0 & 1 & \frac{\partial f}{\partial y}(x_0,y_0) \\
1 & 0 & \frac{\partial f}{\partial x}(x_0,y_0)
\end{matrix}\right]\\
&\hskip-0.5in= f_x(x_0,y_0)\,\hi + f_y(x_0,y_0)\,\hj - \hk
\end{align*}
as is its length:
\begin{align*}
&\left|\left(\hj\ +\ \pdiff{f}{y}(x_0\,,\,y_0)\,\hk\right)\times
\left(\hi\ +\ \pdiff{f}{x}(x_0\,,\,y_0)\,\hk\right)\right|\\
&\hskip1in= \sqrt{1 + f_x(x_0,y_0)^2 + f_y(x_0,y_0)^2}
\end{align*}
Throughout this computation, \(x_0\) and \(y_0\) were arbitrary. So we have found the area of each tiny piece of the surface \(S\text{.}\)
As a first example, we compute the area of the part of the cone
\begin{equation*}
z=\sqrt{x^2+y^2}
\end{equation*}
with \(0\le z\le a\) or, equivalently, with \(x^2+y^2\le a^2\text{.}\)
Note that \(z=\sqrt{x^2+y^2}\) is the side of the cone. It does not include the top.
To find its area, we will apply
3.4.1 to
\begin{equation*}
z=f(x,y) = \sqrt{x^2+y^2}
\qquad\text{with $(x,y)$ running over } x^2+y^2\le a^2
\end{equation*}
That forces us to compute the first order partial derivatives
\begin{align*}
f_x(x,y) & = \frac{x}{\sqrt{x^2+y^2}}\\
f_y(x,y) & = \frac{y}{\sqrt{x^2+y^2}}
\end{align*}
Substituting them into the first formula in
3.4.1 yields
\begin{align*}
\dee{S}&= \sqrt{1 + f_x(x,y)^2 + f_y(x,y)^2}\ \dee{x}\,\dee{y}\\
&=\sqrt{1+\Big(\frac{x}{\sqrt{x^2+y^2}}\Big)^2
+\Big(\frac{y}{\sqrt{x^2+y^2}}\Big)^2} \ \dee{x}\,\dee{y}\\
&=\sqrt{1+\frac{x^2+y^2}{x^2+y^2}} \ \dee{x}\,\dee{y}\\
&=\sqrt{2} \ \dee{x}\,\dee{y}
\end{align*}
So
\begin{align*}
\text{Area}
&= \dblInt_{x^2+y^2\le a^2} \sqrt{2}\ \dee{x}\,\dee{y}
= \sqrt{2} \dblInt_{x^2+y^2\le a^2} \dee{x}\,\dee{y}
= \sqrt{2} \pi a^2
\end{align*}
because \(\dblInt_{x^2+y^2\le a^2} \dee{x}\,\dee{y}\) is exactly the area of a circular disk of radius \(a\text{.}\)
Find the surface area of the part of the paraboloid \(z=2-x^2-y^2\) lying above the \(xy\)-plane.
Solution.
The equation of the surface is of the form \(z=f(x,y)\) with \(f(x,y)=2-x^2-y^2\text{.}\) So
\begin{gather*}
f_x(x,y) =-2x\qquad
f_y(x,y) =-2y
\end{gather*}
and, by the first part of
3.4.1,
\begin{align*}
\dee{S} &= \sqrt{1 + f_x(x,y)^2 + f_y(x,y)^2}\ \dee{x}\dee{y}\\
&=\sqrt{1+4x^2+4y^2} \ \dee{x}\dee{y}
\end{align*}
The point \((x,y,z)\text{,}\) with \(z=2-x^2-y^2\text{,}\) lies above the \(xy\)-plane if and only if \(z\ge 0\text{,}\) or, equivalently, \(2-x^2-y^2\ge 0\text{.}\) So the domain of integration is \(\Set{(x,y)}{x^2+y^2\le 2}\) and
\begin{equation*}
\text{Surface Area} = \dblInt_{x^2+y^2\le 2}\ \sqrt{1+4x^2+4y^2} \ \dee{x}\dee{y}
\end{equation*}
Switching to polar coordinates,
\begin{align*}
\text{Surface Area}
&=\int_0^{2\pi}\int_0^{\sqrt{2}}\sqrt{1+4r^2}\ r\,\dee{r}\, \dee{\theta}\\
&=2\pi\left[\frac{1}{12}{\big(1+4r^2\big)}^{3/2}\right]_0^{\sqrt{2}}
=\frac{\pi}{6}[27-1]\\
&=\frac{13}{3}\pi
\end{align*}