Given the orientation of a great circle as the cartesian components of the normal vector $(a,b,c)$ to its plane, i.e. all points on the circle described in Earth-Centered Earth-Fixed (ECEF) coordinates $(x_p, y_p, z_p)$ satisfy the condition

$$ x_p\cdot a + y_p\cdot b + z_p\cdot c = 0 \;\;\;\;\;\;\;\;\; (1) $$

(I am working on the Earth surface as modeled in the WGS84 standard (oblate spheroid))

Does it exist a (probably not-so-simple) method to compute the latitude of one of such points given the longitude?

I know that if I use a description of a meridian in ECEF coordinates I can "simply" intersect it with the great circle plane by solving a system of equations, but I seem to fail at some point here.

I describe the meridian as an ellipse in its plane, with semi-axes the respective standard radii of the Earth

$$ \frac{x'^2}{WGS84\_A^2} + \frac{z^2}{WGS84\_B^2} = 1 \;\;\;\;\;\;\;\;\; (2) $$

Where $x'$ and $z$ can be described in term of the Latitude (but I fear this is not fully correct): $$ x' = WGS84\_A \cdot cos(Latitude) \;\;\;\;\;\;\;\;\;\;\;\;\;\; \\ z = WGS84\_B \cdot sin(Latitude) \;\;\;\;\;\;\;\;\; (3) $$ and I am looking for the intercept with a positive $x'$.

Obviously the $z$ of Eq.(1) is the same of Eq.(2) and, considering that no point of the meridian will lay outside its plane, thus $y' == 0$, we can say that (and here I fear I might have insterted a mistake)

$$ x' = x \cdot cos(Long) + y \cdot sin(Long) \\ 0 = - x \cdot sin(Long) + y \cdot cos(Long) $$

from the rotation above, by substitution, we obtain $x' = \frac{y}{sin(Long)}$

At this point we can solve the above non-linear sistem for the Latitude:

$$ Latitude = asin(\frac{a\cdot cos(Long) + b\cdot sin(Long)}{WGS84\_B \cdot c \cdot \sqrt(\frac{1}{WGS84\_A^2} + (\frac{a\cdot cos(Long) + b\cdot sin(Long)}{WGS84\_B\cdot cos(Long)} )^2)}) $$

Unfortunately, when I try this out the results are blatantly wrong, I have obviously committed a mistake, but I am unable to spot it.

Is there anyone that would like to help me or that knows the correct solution/approach?


I found the solution by changing approach (thanks to this answer)

Using the notation used there, I have

$$ P = (a,b,c) $$


$$ Q = (cos(Long+\pi/2),sin(Long+\pi/2),0) $$

the (normalized!) product $D = Q\times P$ is pointing in the direction of the point I am interested in, to extract the Latitude all that remains to do is

$$ Latitude = asin(D_z) $$

| cite | improve this answer | |

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.