GPS Coordinates and Float precision

As I understand it, Arduino Float or Double only has 6-7 decimal digits of precision, but 32 bits of resolution.

Am I right in thinking that I can store a longer decimal, such as 8 or 9 digit GPS coordinate, fine but would have to account for the lower precision if I wanted I do any floating point math.

My main requirement is to accurately store GPS coordinates and transmit them over a radio link. I don't anticipate doing any math on those coordinates on the Arduino.

A GPS system normally sends it's information in the form of a string. If you are just passing them on I would keep it as a string.

You could use 64-bit long long to store/manipulate coordinates down to the meter.

Why risk it?
I'm with Mike - keep it as the original text.