Calculus III: Change of Variables in Multiple Integrals

Mastering integration over complex regions is what separates proficient calculus students from true problem-solvers in engineering and physics. The technique of changing variables transforms daunting integrals with convoluted boundaries into manageable calculations over simple rectangles or boxes, providing a powerful systematic approach for analyzing real-world volumes, masses, and field averages.

The Jacobian Determinant: Measuring Distortion

At the heart of every change of variables lies the Jacobian determinant. When you perform a coordinate transformation from $(x,y)$ to $(u,v)$ via functions $x=g(u,v)$ and $y=h(u,v)$, you are mapping a region in the $uv$-plane to a region in the $xy$-plane. The Jacobian measures how this transformation locally stretches or shrinks area. Formally, the Jacobian of the transformation is the determinant of the matrix of all first-order partial derivatives:

$$J(u,v)=\frac{\partial (x,y)}{\partial (u,v)}=\left|\begin{array}{cc}\frac{\partial x}{\partial u}& \frac{\partial x}{\partial v}\\ \frac{\partial y}{\partial u}& \frac{\partial y}{\partial v}\end{array}\right|$$

For a three-dimensional transformation $(x,y,z)=G(u,v,w)$, you compute a $3\times 3$ determinant. The absolute value $|J(u,v)|$ acts as a correction factor. If $|J|=2$, for instance, a tiny unit square in the $uv$-plane maps to a small region with approximately twice the area in the $xy$-plane. A negative Jacobian simply indicates a reversal of orientation, which is why we take the absolute value in the integration formula.

The General Change of Variables Formula

The change of variables theorem provides the complete recipe for transforming a double integral. If $T:S\to R$ is a one-to-one transformation mapping region $S$ in the $uv$-plane onto region $R$ in the $xy$-plane, with a nonzero Jacobian on the interior of $S$, then:

$${\iint }_{R}f(x,y)dA={\iint }_{S}f(g(u,v),h(u,v))\left|\frac{\partial (x,y)}{\partial (u,v)}\right|dudv$$

The process has four clear steps. First, define a transformation $(x,y)=T(u,v)$ that simplifies the description of the original region $R$. Second, find the image $S$ in the $uv$-plane by mapping the boundaries of $R$. Third, compute the Jacobian determinant $J(u,v)$. Finally, rewrite the integrand $f(x,y)$ in terms of $u$ and $v$, assemble the new integrand $f(g(u,v),h(u,v))|J(u,v)|$, and evaluate the integral over the simpler region $S$.

Mapping Regions Between Coordinate Systems

Effective application hinges on choosing a transformation that simplifies the region of integration. For polar coordinates $(x=r\cos \theta ,y=r\sin \theta )$, the region is often a circle, sector, or annulus. The familiar Jacobian is $|J|=r$. Cylindrical coordinates extend this to 3D for regions with circular symmetry, while spherical coordinates are ideal for spheres and spherical sectors.

The key skill is describing the new region $S$. If $R$ is bounded by curves like $y=2x$ and $y=3x$, you might choose $u=y/x$ so these boundaries become $u=2$ and $u=3$. You must then find a complementary variable, like $v=x$, to complete the transformation. The new boundaries in the $uv$-plane will typically be constant values, creating a rectangular region $S$ that is trivial to integrate over. Always sketch both the original region $R$ and the transformed region $S$ to confirm the mapping is one-to-one and to correctly identify the limits of integration.

Advanced Transformations: Beyond the Standard Set

While polar, cylindrical, and spherical coordinates are essential, engineering problems frequently require more tailored transformations. Two common and powerful types are linear transformations and polynomial transformations.

Linear transformations, of the form $x=au+bv,y=cu+dv$, are excellent for integrating over parallelogram-shaped regions. The Jacobian is constant here: $|J|=|ad-bc|$. For example, to integrate over a parallelogram with vertices at $(0,0),(2,1),(1,2),(3,3)$, you could use $x=2u+v,y=u+2v$, mapping the unit square $0\le u,v\le 1$ onto the parallelogram. The constant Jacobian is $|(2)(2)-(1)(1)|=3$.

Polynomial transformations handle curved boundaries. A classic example is integrating over an elliptical region $\frac{x^2}{a^2}+\frac{y^2}{b^2}\le 1$. The transformation $x=au,y=bv$ maps the ellipse to the unit circle $u^2+v^2\le 1$. The Jacobian is a constant $ab$, and the problem reduces to integration over a circle, where you can then apply polar coordinates. This two-step process—first a linear "stretch" to create a circle, then a polar transformation—is a standard technique.

Application: Simplifying Complex Integration Domains

Consider finding the volume under the plane $z=2x+y+1$ over the region $R$ bounded by the lines $y=x-1$, $y=x+1$, $y=2-x$, and $y=4-x$. This region is a parallelogram, but describing it with $x$ and $y$ limits requires splitting into two sub-integrals. Instead, define new variables to straighten the boundaries. Observe the boundaries involve $x+y$ and $x-y$.

Let $u=x+y$ and $v=x-y$. Solving for $x$ and $y$ gives the transformation: $$x=\frac{u+v}{2},y=\frac{u-v}{2}$$ The boundaries become:

• $y=x-1⟹v=1$
• $y=x+1⟹v=-1$
• $y=2-x⟹u=2$
• $y=4-x⟹u=4$

The new region $S$ is the rectangle $2\le u\le 4,-1\le v\le 1$. Compute the Jacobian: $$J(u,v)=\frac{\partial (x,y)}{\partial (u,v)}=\left|\begin{array}{cc}\frac{1}{2}& \frac{1}{2}\\ \frac{1}{2}& -\frac{1}{2}\end{array}\right|=\left(\frac{1}{2}\right)\left(-\frac{1}{2}\right)-\left(\frac{1}{2}\right)\left(\frac{1}{2}\right)=-\frac{1}{2}$$ So, $|J|=\frac{1}{2}$. The integrand becomes $2x+y+1=2(\frac{u+v}{2})+(\frac{u-v}{2})+1=\frac{3u+v}{2}+1$. The volume integral simplifies dramatically to a single iterated integral over a rectangle: $$\text{Volume}={\int }_{-1}^1{\int }_2^4\left(\frac{3u+v}{2}+1\right)\cdot \frac{1}{2}dudv$$ This is straightforward to evaluate, showcasing the power of a well-chosen transformation.

Common Pitfalls

1. Forgetting the Absolute Value of the Jacobian: The most frequent computational error is omitting the Jacobian determinant entirely or dropping the absolute value. Remember, $dA=dxdy$ transforms to $|J(u,v)|dudv$, not just $dudv$. If you compute a negative Jacobian, take its absolute value for the area/volume correction.
2. Incorrectly Describing the Transformed Region $S$: A perfect transformation is useless if you get the new limits wrong. You must transform every boundary curve of $R$ into the $uv$-plane. The best practice is to solve the transformation equations for $u$ and $v$ in terms of $x$ and $y$, then substitute the boundary equations of $R$ to find the corresponding boundaries of $S$.
3. Using a Non-One-to-One Transformation: The standard formula requires the transformation to be one-to-one on the interior of $S$. A classic trap is using a transformation whose Jacobian is zero within the region, which collapses the mapping, or using a transformation that maps multiple $(u,v)$ points to the same $(x,y)$ point inside $R$. Always check that the Jacobian is nonzero on the interior of $S$.
4. Mixing Up Variable Orders in 3D: In triple integrals, the order of variables in the Jacobian determinant matters. For transformation $(x,y,z)=T(u,v,w)$, the columns must be the partial derivatives with respect to $u$, $v$, and $w$ in that order. Swapping columns changes the sign of the determinant, but the absolute value corrects this. Consistency is key to avoiding confusion.

Summary

• The Jacobian determinant $\frac{\partial (x,y)}{\partial (u,v)}$ quantifies the local area-scaling factor of a coordinate transformation, and its absolute value must be included as a multiplier in the new integrand.
• The change of variables formula ${\iint }_{R}f(x,y)dA={\iint }_{S}f(g(u,v),h(u,v))|J|dudv$ provides a four-step framework: choose a transformation, map the region, compute the Jacobian, and set up the new integral.
• The primary goal of changing variables is to map a complex integration region $R$ to a simpler region $S$, such as a rectangle or a standard shape where polar coordinates apply.
• Beyond standard polar/cylindrical/spherical systems, linear transformations are ideal for parallelograms, and targeted polynomial transformations (like scaling for ellipses) can reshape domains into simpler geometries.
• Successful application requires vigilant avoidance of common errors, chiefly forgetting the Jacobian and misidentifying the bounds of the transformed region $S$. Always verify the transformation is one-to-one on the region's interior.