There were enough inaccuracies in the documentation and the navigation methods in use in 1937 to make finding Howland Island visually a very iffy thing.
At the time Earhart disappeared PanAm was flying scheduled passenger service across the Northern Pacific hitting tiny islands (Midway and Wake) with precision week after week. Finding Howland would not have been iffy at all had Earhart used the system pioneered, in part, by Fred Noonan. ...It was an entirely workable system and the basis for what Earhart intended to do. Noonan did his job. She was not able to do hers because she failed to acquire the necessary knowledge and skills.
This is the great tragedy of the whole story to me, although I'm fairly certain Earhart and Noonan might argue that perishing without a trace might have been worse from their perspectives.
One example is that it appears that the maps of the Pacific Ocean in use at that time showed Howland Island in the wrong place. In all the reports the US Navy wrote at the time of the Earhart search they say the island was 1 mile north and 5 miles west of where it actually is. The Clarence Williams strip chart also shows Howland Island at this incorrect location. Even if Noonan could locate his LOP and then navigate along it with GPS like precision, he still would have missed Howland Island by 5 miles.
Probably a myth. Williams' strip charts were done sometime in late 1936 or early 1937 but by that time the correct location of Howland Island was known to the Coast Guard, the Bureau of Air Commerce and the Dept. of the Interior. They were there building runways for Amelia. It seems inconceivable that someone did not pass that information to Earhart.
The very fact that Williams' strip charts are now in the Purdue archive is a pretty good indication that Noonan didn't use them on the second world flight attempt.
Williams' strip charts... Pudue
![Cheesy :D](https://tighar.org/smf/Smileys/no_animation/cheesy.gif)
A five mile discrepancy would not have been even awkward to overcome had Earhart been able to master the DF task. In terms of celestial, I believe it would be immaterial: one cannot reliably navigate that closely by celestial anyway.