Quite a fascinating little experiement.
Given the lack of resolution on the time scale, the fact that your experimental results were within 3% and 1% of the accepted values is quite amazing.
Some people may question whether it is acceptable to neglect the height of the ionosphere. Taking an average height for the ionosphere of 300 km, and average earth radius 6371 km, assuming the signal is at an average height of 150 km (it may be a little less due to curvature), and that the circumference is proportional to radius, this would increase the distances by 2.35%. This is not significant, given the existing margin of error.
Thanks Chris.
Cheers, Terry