A Community Devoted to the Preservation and Practice of Celestial Navigation and Other Methods of Traditional Wayfinding
From: Robin Stuart
Date: 2017 Jun 6, 11:18 -0700
The method you describe is the one I used in this post. It is convenient in that, by assuming the altitude around noon is quadratic in time, least-squares regression has a closed form solution. It has a number of disadvantages however
- It works well only for relatively short intervals around local noon.
- The quadratic fit becomes worse as the Sun’s noon altitude increases.
- It does not account for the Sun’s changing declination in shifting the time of peak altitude. This effect then needs to be accounted for separately.
- It is a 3-parameter fit to a problem with 2 free parameters.
In order to address all of these limitations I modified the method to do a least-squares fit to the true formula for the altitude as a function of time and described it here. The 2-parameter minimization is easily done in Excel using “Solver”.