Change of Variables: Learn It 3

Change of Variables for Double Integrals

We’ve established that under the change of variables [latex]T(u,v) = (x,y)[/latex] where [latex]x = g(u,v)[/latex] and [latex]y = h(u,v)[/latex], a small region [latex]\Delta A[/latex] in the [latex]xy[/latex]-plane is related to the area [latex]\Delta u \Delta v[/latex] in the [latex]uv[/latex]-plane by:

[latex]\Delta A \approx J(u,v) \Delta u \Delta v[/latex]

Let’s revisit the definition of the double integral:

[latex]\underset{R}{\displaystyle\iint}f(x,y) \ dA=\displaystyle\lim_{m,n\to\infty}\displaystyle\sum_{i=1}^m\displaystyle\sum_{j=1}^nf(x_{ij}, y_{ij}) \Delta A[/latex]

Looking at Figure 1, we divide region [latex]S[/latex] in the [latex]uv[/latex]-plane into small subrectangles [latex]S_{ij}[/latex]. The subrectangles [latex]R_{ij}[/latex] in the [latex]xy[/latex]-plane are the images of [latex]S_{ij}[/latex] under transformation [latex]T(u,v) = (x,y)[/latex].

On the left-hand side of this figure, there is a rectangle S with an inscribed red oval and a subrectangle with lower right corner point (u sub ij, v sub ij), height Delta v, and length Delta u given in the Cartesian u v-plane. Then there is an arrow from this graph to the right-hand side of the figure marked with T. On the right-hand side of this figure there is a region R with inscribed (deformed) red oval and a subrectangle R sub ij with corner point (x sub ij, y sub ij) given in the Cartesian x y-plane. The subrectangle is blown up and shown with vectors pointing along the edge from the corner point.
Figure 1. The subrectangles [latex]S_{ij}[/latex] in the [latex]uv[/latex]-plane transform into subrectangles [latex]R_{ij}[/latex] in the [latex]xy[/latex]-plane.

The double integral becomes:

[latex]\underset{R}{\displaystyle\iint}f(x,y) \ dA=\displaystyle\lim_{m,n\to\infty}\displaystyle\sum_{i=1}^m\displaystyle\sum_{j=1}^n{f}(x_{ij},y_{ij})\Delta{A}=\displaystyle\lim_{m,n\to\infty}\displaystyle\sum_{i=1}^m\displaystyle\sum_{j=1}^nf(g(u_{ij},v_{ij}),h(u_{ij},v_{ij}))|J(u_{ij},v_{ij})|\Delta{u}\Delta{v}[/latex].

This is exactly the double Riemann sum for the integral:

[latex]\underset{S}{\displaystyle\iint}f(g(u,v),h(u,v))\Bigg|\dfrac{\partial(x,y)}{\partial(u,v)}\Bigg|du \ dv[/latex].

change of variables for double integrals

Let [latex]T(u,v) = (x,y)[/latex] where [latex]x = g(u,v)[/latex] and [latex]y = h(u,v)[/latex] be a one-to-one [latex]C^1[/latex] transformation with a nonzero Jacobian on the interior of region [latex]S[/latex] in the [latex]uv[/latex]-plane. If [latex]T[/latex] maps [latex]S[/latex] into region [latex]R[/latex] in the [latex]xy[/latex]-plane and [latex]f[/latex] is continuous on [latex]R[/latex], then:

[latex]\underset{R}{\displaystyle\iint}f(x,y) \ dA=\underset{S}{\displaystyle\iint}f(g(u,v),h(u,v))\Bigg|\dfrac{\partial(x,y)}{\partial(u,v)}\Bigg|du \ dv[/latex].

With this theorem, we can change the variables from [latex](x, y)[/latex] to [latex](u, v)[/latex] in a double integral simply by replacing:

[latex]dA=dx \ dy=\Bigg|\dfrac{\partial(x,y)}{\partial(u,v)}\Bigg|du \ dv[/latex]

when we use the substitutions [latex]x=g(u, v)[/latex] and [latex]y=h(u, v)[/latex] and then change the limits of integration accordingly. This change of variables often makes any computations much simpler.

Why the Absolute Value? 

We use [latex]|J(u,v)|[/latex] in the integral because the Jacobian can be negative, but area must be positive. The absolute value ensures we’re measuring area correctly regardless of whether the transformation reverses orientation.

Choosing a Good Transformation

Look for patterns in:

  • The boundary curves of the region (can they be simplified?)
  • The integrand itself (does it suggest a coordinate system?)
  • Symmetry in the problem (circular, elliptical, etc.)

Often the region’s shape will suggest natural substitutions that make the limits constants.

How to: Apply Change of Variables to Double Integrals

  1. Sketch the region in the [latex]xy[/latex]-plane and write the equations of the curves that form the boundary.
  2. Choose the transformation [latex]x = g(u,v)[/latex] and [latex]y = h(u,v)[/latex] based on the region or integrand structure.
  3. Determine the new limits of integration in the [latex]uv[/latex]-plane.
  4. Find the Jacobian [latex]J(u,v) = \dfrac{\partial(x,y)}{\partial(u,v)}[/latex].
  5. Replace the variables in the integrand to obtain the new integrand.
  6. Replace [latex]dA[/latex] (whether [latex]dy , dx[/latex] or [latex]dx , dy[/latex]) by [latex]|J(u,v)| , du , dv[/latex].

Consider the integral

[latex]\displaystyle\int_0^2\displaystyle\int_0^{\sqrt{2x-x^2}}\sqrt{x^2+y^2}dy \ dx[/latex].

Use the change of variables [latex]x=r\cos\theta[/latex] and [latex]y=r\sin\theta[/latex], and find the resulting integral.

Notice in the next example that the region of integration often suggests a suitable transformation. This is a common and important situation.

Consider the integral [latex]\underset{R}{\displaystyle\iint}(x-y)dy \ dx[/latex], where [latex]R[/latex] is the parallelogram joining the points [latex](1, 2)[/latex], [latex](3, 4)[/latex], [latex](4, 3)[/latex], and [latex](6, 5)[/latex] (Figure 3). Make appropriate changes of variables, and write the resulting integral.

A semicircle in the first quadrant of the xy plane with radius 1 and center (1, 0). The equation for this curve is given as y = the square root of (2x minus x squared)
Figure 3. The region of integration for the given integral.

Using the change of variables [latex]u=x-y[/latex] and [latex]v=x+y[/latex], evaluate the integral

[latex]\large{\underset{R}{\displaystyle\iint}(x-y)e^{x^2-y^2}dA}[/latex],

where [latex]R[/latex] is the region bounded by the lines [latex]x+y=1[/latex] and [latex]x+y=3[/latex] and the curves [latex]x^{2}-y^{2}=-1[/latex] and [latex]x^{2}-y^{2}=1[/latex] (see the first region in Figure 5).