Flux Through Curved Parametric Patches

Flux as a Signed Flow Across a Curved Patch

A vector field describes a quantity with both magnitude and direction at each point. Its flux across a surface measures the net amount of that field passing through the surface, with the sign determined by the chosen orientation. On a curved patch, both the field and the surface direction can vary from point to point, so the calculation must account for them together.

A parametrization provides coordinates for doing this. It describes each point on the surface in terms of parameters, and its two tangent vectors determine an oriented area vector. Substituting the parametrization into the field and taking a dot product with that area vector gives a scalar function on the parameter domain. The flux is then a double integral of that function.

Central formula: For a regular parametrized patch $\mathbf{r}(u,v)$ with the orientation specified by the order $u,v$, the flux of $\mathbf{F}$ is the integral of $\mathbf{F}(\mathbf{r}(u,v))\cdot(\mathbf{r}_u\times\mathbf{r}_v)$ over the parameter domain.

Parametrization and Oriented Area Vector

Let $D$ be a rectangle in the $uv$-plane, and let $\mathbf{r}:D\to\mathbb{R}^3$ be a continuously differentiable parametrization. Write $\mathbf{r}(u,v)=(x(u,v),y(u,v),z(u,v))$. Assume that $\mathbf{r}$ is one-to-one on the interior of $D$, apart from possible boundary identifications, and that $\mathbf{r}_u\times\mathbf{r}_v\ne\mathbf{0}$ there. These conditions describe a regular patch that is covered once, with no overlap of positive area.

The vectors $\mathbf{r}_u$ and $\mathbf{r}_v$ are tangent to the patch. Their cross product is perpendicular to both, and its magnitude is the area scaling factor from parameter area to surface area. Its direction depends on the order: $\mathbf{r}_u\times\mathbf{r}_v$ gives the orientation associated with increasing $u$ followed by increasing $v$. Reversing the order changes its sign.

Definition: If $\mathbf{F}$ is continuous on the patch, the flux through the parametrized surface with the orientation specified by $u,v$ is $$ \iint_D \mathbf{F}(\mathbf{r}(u,v))\cdot \bigl(\mathbf{r}_u(u,v)\times\mathbf{r}_v(u,v)\bigr)\,du\,dv. $$ The result is signed: positive and negative contributions correspond to field flow along and against the chosen normal direction.

If the rectangle is $a\le u\le b,\ c\le v\le d$, this becomes an ordinary iterated integral. The field must first be evaluated at the point $\mathbf{r}(u,v)$. The cross product already includes the surface-area scaling; there is no additional factor of surface area to multiply in.

Why the Formula Works

For a small parameter rectangle with side lengths $\Delta u$ and $\Delta v$, the corresponding piece of a smooth surface is approximated by a tangent parallelogram. Its directed area vector is approximately $(\mathbf{r}_u\times\mathbf{r}_v)\Delta u\,\Delta v$. Over this small piece, continuity makes the field approximately constant, with value $\mathbf{F}(\mathbf{r}(u,v))$. The field’s contribution to signed flux is therefore approximately its dot product with that directed area vector.

\text{flux on a small patch} \ \approx\ \mathbf{F}(\mathbf{r}(u,v))\cdot (\mathbf{r}_u\times\mathbf{r}_v)\,\Delta u\,\Delta v.

Partition $D$ into small rectangles and add these approximations. The resulting sums are Riemann sums for the displayed flux integral. As the partition is refined, differentiability of $\mathbf{r}$ gives the tangent-parallelogram approximation, and continuity of $\mathbf{F}$ makes the field approximation converge. Thus the sum converges to the integral in the definition. This explains why substituting the parametrization, taking the oriented area vector, and integrating over parameter space computes flux.

A Reliable Calculation Sequence

Record the parametrization and domain.
Write $\mathbf{r}(u,v)$, the bounds, and the requested orientation.

Differentiate with respect to each parameter.
Compute $\mathbf{r}_u$ and $\mathbf{r}_v$, keeping their order consistent.

Form the oriented area vector.
Calculate $\mathbf{r}_u\times\mathbf{r}_v$, not its magnitude.

Substitute into the field and take the dot product.
Replace $x,y,z$ in $\mathbf{F}$ by the coordinates of $\mathbf{r}(u,v)$.

Integrate over the parameter rectangle.
Evaluate the resulting scalar double integral using the stated bounds.

Worked Examples

Worked Example: Flux Across a Curved Paraboloid Patch

Let $\mathbf{r}(u,v)=(u\cos v,u\sin v,u^2)$, where $0\le u\le1$ and $0\le v\le\pi/2$. Find the flux of $\mathbf{F}(x,y,z)=(x,y,z)$ with the orientation $\mathbf{r}_u\times\mathbf{r}_v$.

First differentiate: $\mathbf{r}_u=(\cos v,\sin v,2u)$ and $\mathbf{r}_v=(-u\sin v,u\cos v,0)$. Their cross product is \[ \mathbf{r}_u\times\mathbf{r}_v =(-2u^2\cos v,-2u^2\sin v,u). \]

Substitution into the field gives $\mathbf{F}(\mathbf{r}(u,v))=(u\cos v,u\sin v,u^2)$. Thus the scalar integrand is \[ (u\cos v)(-2u^2\cos v) +(u\sin v)(-2u^2\sin v) +(u^2)(u) =-2u^3(\cos^2v+\sin^2v)+u^3=-u^3. \]

Integrating over the rectangle, \[ \int_0^{\pi/2}\int_0^1 -u^3\,du\,dv =\int_0^{\pi/2}\left[-\frac{u^4}{4}\right]_0^1\,dv =\int_0^{\pi/2}-\frac14\,dv =-\frac{\pi}{8}. \] As a check, integrating in the other order gives $\int_0^1\int_0^{\pi/2}-u^3\,dv\,du =\int_0^1-\frac{\pi u^3}{2}\,du=-\frac{\pi}{8}$. The negative sign is the signed flux for the specified orientation.

Worked Example: A Curved Patch Rising Along a Quarter-Cylinder

Use $\mathbf{r}(u,v)=(u\cos v,u\sin v,v)$, with $0\le u\le1$ and $0\le v\le1$. Find the flux of $\mathbf{F}(x,y,z)=(0,0,z)$ using $\mathbf{r}_u\times\mathbf{r}_v$.

The derivatives are $\mathbf{r}_u=(\cos v,\sin v,0)$ and $\mathbf{r}_v=(-u\sin v,u\cos v,1)$. Therefore \[ \mathbf{r}_u\times\mathbf{r}_v =(\sin v,-\cos v,u). \] On the patch, $z=v$, so $\mathbf{F}(\mathbf{r}(u,v))=(0,0,v)$. The dot product is \[ (0,0,v)\cdot(\sin v,-\cos v,u)=uv. \]

Consequently, \[ \int_0^1\int_0^1 uv\,du\,dv =\int_0^1\left[\frac{u^2v}{2}\right]_0^1\,dv =\int_0^1\frac v2\,dv =\left[\frac{v^2}{4}\right]_0^1 =\frac14. \] Checking in the reverse order gives $\int_0^1\int_0^1 uv\,dv\,du =\int_0^1\frac u2\,du=\frac14$. Both orders confirm the flux.

Worked Example: Flux Through a Saddle-Shaped Patch

Let $\mathbf{r}(u,v)=(u,v,uv)$ on $0\le u\le1,\ 0\le v\le1$, and let $\mathbf{F}(x,y,z)=(x,y,1)$. Compute the flux for the orientation $\mathbf{r}_u\times\mathbf{r}_v$.

We have $\mathbf{r}_u=(1,0,v)$ and $\mathbf{r}_v=(0,1,u)$, so \[ \mathbf{r}_u\times\mathbf{r}_v=(-v,-u,1). \] The field along the patch is $\mathbf{F}(\mathbf{r}(u,v))=(u,v,1)$. Taking the dot product gives \[ (u,v,1)\cdot(-v,-u,1)=-uv-uv+1=1-2uv. \]

The flux is \[ \int_0^1\int_0^1(1-2uv)\,du\,dv =\int_0^1\left[u- u^2v\right]_0^1\,dv =\int_0^1(1-v)\,dv =\left[v-\frac{v^2}{2}\right]_0^1 =\frac12. \] For a check, integrating first in $v$ gives $\int_0^1[u-vu^2]_0^1\,du =\int_0^1(1-u^2)\,du =1-\frac13=\frac23$. That is not the same result, so inspect the integration carefully: with $u$ fixed, $\int_0^1(1-2uv)\,dv=[v-uv^2]_0^1=1-u$. Then $\int_0^1(1-u)\,du=\frac12$, confirming the flux. The intermediate expression $1-u^2$ would come from incorrectly integrating the term $2uv$.

Orientation and Common Errors

The parameter order is part of the calculation. If the prescribed orientation agrees with $\mathbf{r}_u\times\mathbf{r}_v$, use that cross product. If it requires the opposite orientation, use $\mathbf{r}_v\times\mathbf{r}_u=-(\mathbf{r}_u\times\mathbf{r}_v)$; the flux changes sign. The parametrization alone does not determine which orientation is intended unless the parameter order or a normal direction is specified.

Common pitfall: Replacing the cross product by its magnitude discards orientation and generally gives a scalar surface integral, not signed flux. For example, in the saddle-patch calculation, the correct area vector is $(-v,-u,1)$. Its magnitude is $\sqrt{v^2+u^2+1}$, which is not the vector to dot with the field. The correct integrand is $(u,v,1)\cdot(-v,-u,1)=1-2uv$, not $\sqrt{v^2+u^2+1}$ or the field magnitude times that quantity.

Another frequent error is evaluating the field at $(u,v)$ as though the parameters were spatial coordinates. In general, the field must be evaluated at $(x(u,v),y(u,v),z(u,v))=\mathbf{r}(u,v)$. A useful check is to write out the three coordinates of $\mathbf{F}(\mathbf{r}(u,v))$ before taking the dot product.

Flux through a curved patch is a signed quantity, so a negative answer is not automatically evidence of an error. It can mean that the field points, on balance, against the chosen normal. To diagnose a calculation, check the parameter bounds, the substitution into the field, the order in the cross product, and the dot product separately.

Check Your Understanding

For each question, keep the parameter order and orientation explicit.

For a regular parametrization $\mathbf{r}(u,v)$, what role does $\mathbf{r}_u\times\mathbf{r}_v$ play in the flux integral?
Why must the field be evaluated at $\mathbf{r}(u,v)$ before integrating over the parameter domain?
What changes in the flux formula if the parameter order is reversed?
For $\mathbf{r}(u,v)=(u,v,uv)$ and $\mathbf{F}(x,y,z)=(x,y,1)$, write the scalar integrand for the orientation $\mathbf{r}_u\times\mathbf{r}_v$.
Why is the magnitude of the cross product not, by itself, the oriented area vector needed for flux?