$$\int_\Omega F(\theta,\phi) \sin(\phi) \, d\Omega$$
where $\Omega$ represents the (outer) surface of a spherical cap on a unit sphere.

Specifically, let the center of the spherical cap be in some direction $(\phi_s,\theta_s)$, and let the apex angle of the spherical sector corresponding to the cap be $2\sigma$.

The problem I'm having is finding the limits of integration. If this was a cap symmetric about the z axis, it'd be easy because of symmetry. It'd just be:
$$\int_{\theta=0}^{2\pi}\int_{\phi=0}^{\sigma} F(\theta,\phi)\sin(\phi) \, d\phi \, d\theta$$.

I can't figure out how to take into consideration the rotation of the spherical cap. I don't see a way to rotate F , and I can't describe the portion of the sphere's surface in terms of just $\theta$ and $\phi$.

I saw this post, which suggests I should convert the function into Cartesian Coordinates, multiply a rotation matrix, and convert it back into Spherical coordinates, but I don't think that is reasonable with my particular F.

(For additional information, F is a second-order polynomial in $\theta$ and $\phi$. I think that if I were to transform,rotate,and transform it back, I would get a ton of nested sinusoidal and inverse sinusoidal functions that would not be able to be integrated properly/handily)

Any insights or pointers would be greatly appreciated!


set up a spherical triangle with one point at the pole ( $\theta =0$) one point at $(\theta_S, \phi_S)$ and the third point being $(\theta, \phi)$ subject to the condition that the length of the arc from $(\theta_S, \phi_S)$ to $(\theta, \phi)$ is $\sigma$ .

You have all three sides of this triangle $\theta, \theta_S \text{ and } \sigma$ , now you can use the cosine law for spherical triangles to solve for the dihedral angle subtended at the pole, which equals $\phi - \phi_S$

$$\cos \sigma = \cos \theta \cos \theta_S + \sin \theta \sin \theta_S \cos(\phi - \phi_S)$$

so that

$$\cos(\phi - \phi_S) = \frac{ \cos \sigma - \cos \theta \cos \theta_S }{ \sin \theta \sin \theta_S} $$

having the solution

$$ \phi = \phi_S \pm g(\theta) \text { where } g(\theta) \equiv \cos^{-1}\left( \frac{ \cos \sigma - \cos \theta \cos \theta_S }{ \sin \theta \sin \theta_S} \right)$$

If $\sigma < \theta_S$ you can set up the limits as ...

$$\int_{\theta_S - \sigma}^{\theta_S + \sigma}\int_{\phi_S-g(\theta)}^{\phi_S+g(\theta)} F(\theta,\phi)\sin(\phi) \, d\phi \, d\theta $$

If $\sigma > \theta_S$ you should split into two integrals ...

$$\int_\Omega F(\theta,\phi) \sin(\phi) \, d\Omega = $$

$$\int_{0}^{ \sigma - \theta_S }\int_{0}^{2 \pi} F(\theta,\phi)\sin(\phi) \, d\phi \, d\theta +\int_{\sigma - \theta_S }^{\sigma + \theta_S }\int_{\phi_S-g(\theta)}^{\phi_S+g(\theta)} F(\theta,\phi)\sin(\phi) \, d\phi \, d\theta $$

  • $\begingroup$ Oh man, that's one ugly looking limit. I didn't know about the Spherical law of cosines! Thanks :) I'll look into this. $\endgroup$ – Ethan Jun 19 '15 at 3:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.