The quantity I am interested in is not the rotation rate $\Omega$, but I will use this quantity nonetheless to make the problem clearer. I am interested in a spherically-symmetric system and in the boundary condition of the rotation rate $\Omega=u/r\sin\theta$. Now, at $\theta=0$, we have an undetermined form $0/0$ and we must use de L'Hopital rule: $$ \Omega (r,0)=\text{lim}_{\theta\rightarrow 0}\frac{du/d\theta}{r\cos\theta}\neq 0 $$

which is not zero as expected. Using central differences (I assume this is still valid):

$$ \frac{du}{d\theta}|_{\theta=0}=\frac{u_1-u-1}{2\Delta\theta}, $$

where the index for radius was ignored (the minus 1 should be as an index but it somehow doesn't work within the fraction). Now, $u_{-1}=-u_1$ and:

$$ \frac{du}{d\theta}|_{\theta=0}=\frac{u_1}{\Delta\theta}=\frac{\Omega_1 r\sin\theta_1}{\Delta\theta}\simeq\Omega_1 r\quad\quad\text{[for $\theta_1=\Delta\theta$ small]}. $$

Finally: $$ \Omega_0=\Omega_1. $$

Now, if I decide to discretize the equations regularly in $\mu=\cos\theta$, so that:

$$ \sin\theta=\sqrt{1-\mu^2}\quad\quad\frac{1}{d\theta}=-\frac{\sin\theta}{d\cos\theta}=-\frac{\sqrt{1-\mu^2}}{d\mu}, $$


$$ \Omega(r,1)=\text{lim}_{\mu\rightarrow 1}\frac{du/d\mu}{rd\sqrt{1-\mu^2}/d\mu}, $$


$$ \Omega(r,1)=\text{lim}_{\mu\rightarrow 1}-\frac{\sqrt{1-\mu^2}}{r\mu}\frac{du}{d\mu}=0 $$

What am I missing here? This should definitely not be zero at the pole (or anywhere).

  • $\begingroup$ What is $u$ in your example? $\endgroup$ Feb 25, 2023 at 20:03
  • $\begingroup$ It's the velocity in the $\phi$-direction $u=r\sin\theta\Omega$ $\endgroup$ Feb 25, 2023 at 20:28


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.