Jacobians and Change of Variables

A change of variables rewrites a multiple integral in new variables u and v to make the region or the integrand simpler. The Jacobian is the determinant that tells how much the substitution stretches area. You replace dA with |∂(x, y)/∂(u, v)| du dv and integrate over the new region.

Before this lesson: Double Integrals over General Regions, Double Integrals in Polar Coordinates, Partial Derivatives and Determinant of a Matrix

Why integrals need a stretch factor

In single-variable u-substitution, dx = g′(u) du. The factor g′(u) says how much a small piece of the u-axis is stretched when it is mapped to the x-axis. In two variables, a transformation x = x(u, v), y = y(u, v) maps a small rectangle in the uv-plane to a small, almost parallelogram-shaped patch in the xy-plane.

The area of that patch is the area of the rectangle, du dv, times a stretch factor. That factor is the absolute value of the Jacobian. You already met two Jacobians: r in polar and cylindrical coordinates, and ρ² sin φ in spherical coordinates.

Two coordinate grids side by side. On the left, the uv-plane shows the rectangle S with corners (0, 0), (2, 0), (2, 3) and (0, 3), shaded light blue. An arrow labeled x = (2u + v)/3, y = (u − v)/3 points to the right grid. On the right, the xy-plane shows the parallelogram R with corners (0, 0), (4/3, 2/3), (7/3, −1/3) and (1, −1), shaded light orange. The sides of R lie on the lines x + y = 0, x + y = 2, x − 2y = 0 and x − 2y = 3.
The transformation in Example 2 maps the 2-by-3 rectangle S onto the slanted parallelogram R. Every area shrinks by the factor |J| = 1/3, so R has area 6 · (1/3) = 2.

Jacobian and change of variables formula

\dfrac{\partial(x, y)}{\partial(u, v)} = \begin{vmatrix} \dfrac{\partial x}{\partial u} & \dfrac{\partial x}{\partial v} \\[2mm] \dfrac{\partial y}{\partial u} & \dfrac{\partial y}{\partial v} \end{vmatrix} = \dfrac{\partial x}{\partial u}\dfrac{\partial y}{\partial v} - \dfrac{\partial x}{\partial v}\dfrac{\partial y}{\partial u}
\iint_R f(x, y)\,dA = \iint_S f\big(x(u, v), y(u, v)\big)\left|\dfrac{\partial(x, y)}{\partial(u, v)}\right|du\,dv

Here S is the region in the uv-plane that the transformation maps onto R. Use the absolute value: the Jacobian can be negative, but area cannot.

  1. Choose u and v from the boundary curves or the integrand. Lines such as x + y = c suggest u = x + y.
  2. Solve for x and y in terms of u and v (or use the shortcut in Related methods).
  3. Compute the Jacobian ∂(x, y)/∂(u, v) and take its absolute value.
  4. Find the new region S by rewriting each boundary of R in u and v.
  5. Rewrite the integrand, multiply by |J| and integrate over S.

Jacobian change of variables examples

Example 1: The Jacobian of polar coordinates

Find the Jacobian of x = r cos θ, y = r sin θ.

\dfrac{\partial x}{\partial r} = \cos\theta, \quad \dfrac{\partial x}{\partial \theta} = -r\sin\theta

Take the partial derivatives of x.

\dfrac{\partial y}{\partial r} = \sin\theta, \quad \dfrac{\partial y}{\partial \theta} = r\cos\theta

Take the partial derivatives of y.

\dfrac{\partial(x, y)}{\partial(r, \theta)} = \cos\theta \cdot r\cos\theta - (-r\sin\theta)\sin\theta

Apply the determinant formula.

= r\cos^2\theta + r\sin^2\theta

Multiply.

= r

Factor out r and use cos² θ + sin² θ = 1 (see fundamental trig identities).

Answer: the Jacobian is r. That is why dA = r dr dθ in double integrals in polar coordinates.

Example 2: A parallelogram region

Evaluate \iint_R (x + y)^2\,dA, where R is the parallelogram bounded by x + y = 0, x + y = 2, x − 2y = 0 and x − 2y = 3.

u = x + y, \quad v = x - 2y

The boundaries suggest these new variables.

0 \le u \le 2, \quad 0 \le v \le 3

The four boundary lines become the sides of a rectangle S.

u - v = 3y

Subtract the two equations: (x + y) − (x − 2y).

y = \dfrac{u - v}{3}

Divide both sides by 3.

x = u - y = \dfrac{2u + v}{3}

Solve u = x + y for x and substitute.

\dfrac{\partial(x, y)}{\partial(u, v)} = \dfrac{2}{3}\cdot\Big(-\dfrac{1}{3}\Big) - \dfrac{1}{3}\cdot\dfrac{1}{3}

Apply the determinant formula: x_u = 2/3, x_v = 1/3, y_u = 1/3, y_v = −1/3.

= -\dfrac{3}{9} = -\dfrac{1}{3}

Combine the fractions.

|J| = \dfrac{1}{3}

Take the absolute value.

\iint_R (x + y)^2\,dA = \int_0^3 \int_0^2 u^2 \cdot \dfrac{1}{3}\,du\,dv

Rewrite the integral: (x + y)² = u².

\int_0^2 \dfrac{u^2}{3}\,du = \Big[\dfrac{u^3}{9}\Big]_0^2 = \dfrac{8}{9}

Integrate in u.

\int_0^3 \dfrac{8}{9}\,dv = \dfrac{8}{3}

Integrate in v: multiply by 3.

Answer: 8/3. In x and y, this region would need three separate integrals.

Example 3: Area of an ellipse

Find the area of the ellipse x²/9 + y²/4 ≤ 1.

x = 3u, \quad y = 2v

Stretch the unit circle: these make x²/9 + y²/4 = u² + v².

u^2 + v^2 \le 1

So S is the unit disk.

\dfrac{\partial(x, y)}{\partial(u, v)} = 3 \cdot 2 - 0 \cdot 0 = 6

Apply the determinant formula: x_u = 3, x_v = 0, y_u = 0, y_v = 2.

\text{Area} = \iint_S 6\,du\,dv

Integrate 1 times |J|.

= 6 \cdot \pi(1)^2

The area of the unit disk is π.

= 6\pi

Multiply.

Answer: 6π ≈ 18.85. In general, an ellipse with semi-axes a and b has area πab.

Example 4: A region between hyperbolas

Evaluate \iint_R xy\,dA, where R is the region in the first quadrant between the hyperbolas xy = 1 and xy = 4 and the lines y = x and y = 4x.

u = xy, \quad v = \dfrac{y}{x}

The boundaries are u = 1, u = 4, v = 1 and v = 4.

\dfrac{\partial(u, v)}{\partial(x, y)} = y \cdot \dfrac{1}{x} - x \cdot \Big(-\dfrac{y}{x^2}\Big)

It is easier to find the Jacobian of u and v first: u_x = y, u_y = x, v_x = −y/x², v_y = 1/x.

= \dfrac{2y}{x} = 2v

Combine like terms and use v = y/x.

\dfrac{\partial(x, y)}{\partial(u, v)} = \dfrac{1}{2v}

Take the reciprocal. It is positive, since v ≥ 1.

\iint_R xy\,dA = \int_1^4 \int_1^4 u \cdot \dfrac{1}{2v}\,du\,dv

Rewrite: xy = u.

\int_1^4 u\,du = \dfrac{16 - 1}{2} = \dfrac{15}{2}

Integrate in u.

\int_1^4 \dfrac{1}{2v}\,dv = \dfrac{1}{2}\ln 4 = \ln 2

Integrate in v (see integrals that give natural logarithms) and use (1/2) ln 4 = ln 2.

\dfrac{15}{2}\ln 2 \approx 5.199

The limits are constants, so multiply the two results.

Answer: (15/2) ln 2 ≈ 5.20.

Example 5: Water in an oval pond

An oval pond is 30 m long and 20 m wide, so its edge is the ellipse x²/225 + y²/100 = 1. Its depth in meters is d = 2(1 − x²/225 − y²/100), so it is 2 m deep in the middle. How much water does it hold?

V = \iint_R 2\Big(1 - \dfrac{x^2}{225} - \dfrac{y^2}{100}\Big)\,dA

Volume = depth integrated over the surface of the pond.

x = 15u, \quad y = 10v

Change variables to turn the ellipse into the unit disk.

\dfrac{\partial(x, y)}{\partial(u, v)} = 15 \cdot 10 = 150

Apply the determinant formula. The off-diagonal entries are 0.

V = \iint_{u^2 + v^2 \le 1} 2(1 - u^2 - v^2) \cdot 150\,du\,dv

Substitute: x²/225 = u² and y²/100 = v².

= 300\int_0^{2\pi} \int_0^1 (1 - r^2)\,r\,dr\,d\theta

Switch the unit disk to polar coordinates.

\int_0^1 (r - r^3)\,dr = \dfrac{1}{2} - \dfrac{1}{4} = \dfrac{1}{4}

Distribute the r and integrate in r.

V = 300 \cdot 2\pi \cdot \dfrac{1}{4}

Integrate in θ: multiply by 2π.

= 150\pi \approx 471.2

Simplify and evaluate.

Answer: about 471 m³ of water. As a check, the surface area is π(15)(10) = 150π m², so the average depth is 1 m, half of the 2 m center depth.

Common mistakes

MistakeExample of the errorFix
Forgetting the JacobianWriting ∫∫ u² du dv in Example 2 and getting 8Always multiply by |∂(x, y)/∂(u, v)|.
Using ∂(u, v)/∂(x, y) as the factorMultiplying by 3 instead of 1/3 in Example 2The factor is ∂(x, y)/∂(u, v), the reciprocal of ∂(u, v)/∂(x, y).
Using a negative JacobianMultiplying by −1/3 and getting −8/3Take the absolute value of the Jacobian.
Keeping the old limitsIntegrating over x and y limits after switching to u and vRewrite every boundary of R in u and v to find S.
Mixing up the determinant entriesComputing x_u y_u − x_v y_vThe determinant is x_u y_v − x_v y_u.

Related methods

The Jacobian of the inverse transformation

If it is hard to solve for x and y, find u and v as functions of x and y and use ∂(x, y)/∂(u, v) = 1 ÷ ∂(u, v)/∂(x, y). Example 4 used this shortcut. In Example 2, ∂(u, v)/∂(x, y) = (1)(−2) − (1)(1) = −3, and its reciprocal is −1/3, as found above.

Jacobians for triple integrals

For x = x(u, v, w), y = y(u, v, w), z = z(u, v, w), the Jacobian is the 3 × 3 determinant of all nine partial derivatives, and dV = |∂(x, y, z)/∂(u, v, w)| du dv dw. For spherical coordinates this determinant is ρ² sin φ.

Practice problems

  1. Find the Jacobian of x = 3u − v, y = u + 2v.
  2. Find the Jacobian of x = uv, y = u.
  3. Use x = 5u, y = 4v to find the area of the ellipse x²/25 + y²/16 ≤ 1.
  4. Evaluate \iint_R (x - y)(x + y)\,dA, where R is the square bounded by x − y = 0, x − y = 2, x + y = 1 and x + y = 3.
  5. Find the area of the region R in Problem 4 with a change of variables.
  6. For u = x + y and v = x − 3y, find ∂(x, y)/∂(u, v) without solving for x and y.
  7. Evaluate \iint_R x^2\,dA, where R is the ellipse x²/4 + y² ≤ 1.
  8. Find the area of the region in the first quadrant between xy = 1 and xy = 3 and between the lines y = x and y = ex.
  9. Use x = 2u, y = 3v, z = 4w to find the volume of the ellipsoid x²/4 + y²/9 + z²/16 ≤ 1.
  10. An oval lawn has semi-axes 40 m and 25 m. One bag of fertilizer covers 400 m². Find the area of the lawn and the number of bags needed.
Show answers

1.

\dfrac{\partial(x, y)}{\partial(u, v)} = 3 \cdot 2 - (-1)(1)

Apply the determinant formula: x_u = 3, x_v = −1, y_u = 1, y_v = 2.

= 7

Simplify: 6 + 1.

Answer: 7

2.

\dfrac{\partial(x, y)}{\partial(u, v)} = v \cdot 0 - u \cdot 1

Apply the determinant formula: x_u = v, x_v = u, y_u = 1, y_v = 0.

= -u

Simplify.

Answer: −u, so the area factor is |u|.

3.

\dfrac{\partial(x, y)}{\partial(u, v)} = 5 \cdot 4 = 20

Apply the determinant formula.

\text{Area} = 20 \cdot \pi = 20\pi

S is the unit disk, with area π.

Answer: 20π ≈ 62.83

4.

u = x - y, \quad v = x + y

Choose new variables: 0 ≤ u ≤ 2 and 1 ≤ v ≤ 3.

x = \dfrac{u + v}{2}, \quad y = \dfrac{v - u}{2}

Add and subtract the two equations, then divide by 2.

\dfrac{\partial(x, y)}{\partial(u, v)} = \dfrac{1}{2} \cdot \dfrac{1}{2} - \dfrac{1}{2} \cdot \Big(-\dfrac{1}{2}\Big) = \dfrac{1}{2}

Apply the determinant formula.

\int_1^3 \int_0^2 uv \cdot \dfrac{1}{2}\,du\,dv

Rewrite: (x − y)(x + y) = uv.

\dfrac{1}{2} \cdot 2 \cdot 4 = 4

Split: the u integral is 2 and the v integral is (9 − 1)/2 = 4.

Answer: 4

5.

\text{Area} = \int_1^3 \int_0^2 \dfrac{1}{2}\,du\,dv

Integrate 1 times |J| over S.

= \dfrac{1}{2} \cdot 2 \cdot 2 = 2

S is a 2-by-2 square.

Answer: 2

6.

\dfrac{\partial(u, v)}{\partial(x, y)} = 1 \cdot (-3) - 1 \cdot 1 = -4

Apply the determinant formula to u and v.

\dfrac{\partial(x, y)}{\partial(u, v)} = -\dfrac{1}{4}

Take the reciprocal.

Answer: −1/4 (area factor 1/4)

7.

x = 2u, \quad y = v, \quad |J| = 2

Turn the ellipse into the unit disk.

\iint_R x^2\,dA = \iint_S 4u^2 \cdot 2\,du\,dv = 8\iint_S u^2\,du\,dv

Substitute x² = 4u².

\iint_S u^2\,du\,dv = \int_0^{2\pi} \cos^2\theta\,d\theta \int_0^1 r^3\,dr

Use polar coordinates: u = r cos θ, with the extra r.

= \pi \cdot \dfrac{1}{4}

Integrate each factor.

8 \cdot \dfrac{\pi}{4} = 2\pi

Multiply.

Answer: 2π ≈ 6.28

8.

u = xy, \quad v = \dfrac{y}{x}, \quad |J| = \dfrac{1}{2v}

Use the result of Example 4: 1 ≤ u ≤ 3 and 1 ≤ v ≤ e.

\text{Area} = \int_1^e \int_1^3 \dfrac{1}{2v}\,du\,dv

Integrate 1 times |J|.

= 2 \cdot \dfrac{1}{2}(\ln e - \ln 1) = 1

The u integral is 2 and the v integral is (1/2) ln v from 1 to e.

Answer: 1

9.

\dfrac{\partial(x, y, z)}{\partial(u, v, w)} = 2 \cdot 3 \cdot 4 = 24

The determinant of a diagonal matrix is the product of its diagonal.

V = 24 \cdot \dfrac{4\pi}{3} = 32\pi

S is the unit ball, with volume 4π/3.

Answer: 32π ≈ 100.53

10.

\text{Area} = \pi(40)(25) = 1000\pi \approx 3141.6

Use x = 40u, y = 25v: |J| = 1000 and the unit disk has area π.

\dfrac{3141.6}{400} \approx 7.85

Divide by the area one bag covers.

Answer: about 3142 m²; 8 bags (7 bags would cover only 2800 m²)

Jacobian and change of variables FAQ

What is a Jacobian in calculus?

It is the determinant of the partial derivatives of a transformation. Its absolute value is the factor by which the transformation stretches small areas (or volumes in 3D).

Why do you take the absolute value of the Jacobian?

A negative Jacobian only means the transformation flips orientation, like a mirror. Area is always positive, so the integral uses |J|.

How do you choose u and v?

Look at the boundary curves. If R is bounded by x + y = 0 and x + y = 2, then u = x + y turns those into u = 0 and u = 2. Sometimes the integrand suggests u and v instead.

What is the Jacobian for polar coordinates?

It is r, so dA = r dr dθ. For cylindrical coordinates it is also r, and for spherical coordinates it is ρ² sin φ.

Summary

  • The Jacobian is ∂(x, y)/∂(u, v) = x_u y_v − x_v y_u.
  • Change of variables: dA = |∂(x, y)/∂(u, v)| du dv, over the new region S.
  • Pick u and v that turn the boundaries of R into constants.
  • If u and v are given in terms of x and y, use the reciprocal of ∂(u, v)/∂(x, y).
  • Polar: J = r. Spherical: J = ρ² sin φ.

Previous lesson: Triple Integrals in Spherical Coordinates

Next lesson: Vector Fields

Back to the Calculus 3 Index