We don't navigate by rhumb lines, we navigate by directions and local shapes, and those are the important things that stay the same in Mercator. If some angles on the map are smaller than they should be while others are larger, that's confusing. Mercator keeps this consistent.
Imagine using a zoomed-in version of Winkel-Tripel to navigate the west coast of the US; that would be pretty terrible.
Of course today with computerised navigation, we don't really have to use a flat projection at all; it would be even more accurate to calculate every view directly from a globe. Have your own custom projection based on where you are.
But if you're going to use a flat map projection for navigation, Mercator is it. Or something Mercatory, at least; centering on Greenwich and having north up is less relevant, although deviating from the latter would probably confuse a lot of people.
And outside of navigation, we really shouldn't ever be using Mercator. I'm not sure if that happens. When I was in school, I remember we had a Goode-Homolosine map on the wall, which is an excellent choice for that. Every map projection has its purposes that it's good at, and those it's not. But especially with the ubiquity of Google Maps, it's not surprising that people get more exposure to Mercator.
> But if you're going to use a flat map projection for navigation, Mercator is it.
For you, perhaps.
As a professional geographer since ~ 1980 (ish) I did enjoy the NZMG (1949) that remained in use until 2001 - it was centred down the mean common spine of the North and South Islands and minimised distortion across the entirety of the New Zealand land and waters [1]. It was one of only three (IIRC) complex-number polynomial expansion projections in common use - but they're ideal for certain land form shapes.
I've seen a lot of local ellipsoid+datum+projection mapping systems globally - I was doing global geophysical mapping during the transition to WGS84 and got into the weeds wrt transforms between that and the 300+ mapping systems commonly in use about the globe prior to todays uniformity.
Pilots, of course, use aeronautical charts based on LCC (Lambert conformal conic projection) as straight lines there approximate Great Circle routes, satellite agencies use variations on the projections of John P. Snyder [2].
I'd humbly suggest that if you honestly believe that the Mercator family is the be all and end all for navigation then perhaps you're in need of further reading [2].
You're right. For local navigation, local projections are of course best. I was stuck on the unspoken assumption of using a global map projection. When dealing with local maps, there's absolutely no reason to do that.
But those local map projections only work for that region. As soon as you want seamless transitions to other regions, these local projections don't work so well anymore.
Imagine using a zoomed-in version of Winkel-Tripel to navigate the west coast of the US; that would be pretty terrible.
Of course today with computerised navigation, we don't really have to use a flat projection at all; it would be even more accurate to calculate every view directly from a globe. Have your own custom projection based on where you are.
But if you're going to use a flat map projection for navigation, Mercator is it. Or something Mercatory, at least; centering on Greenwich and having north up is less relevant, although deviating from the latter would probably confuse a lot of people.
And outside of navigation, we really shouldn't ever be using Mercator. I'm not sure if that happens. When I was in school, I remember we had a Goode-Homolosine map on the wall, which is an excellent choice for that. Every map projection has its purposes that it's good at, and those it's not. But especially with the ubiquity of Google Maps, it's not surprising that people get more exposure to Mercator.