[Tagging] Is waterway=riverbank an 'Old scheme' ?

[Christoph Hormann]

On Saturday 08 September 2018, Graeme Fitzpatrick wrote:
>
> The imagery I was working of had apparently been taken during the dry
> season as the actual river channel with water in it was ~30 - 50 m's
> wide. But the wet season river (or possibly flood?) channel was very
> obviously ~500 - 700 m's wide.
>
> What should we be marking as the "riverbank" - where the water is
> visible "now", or the defined limits of where it spreads out to in
> the wet season?

Note in many cases such situations are mapped based on the assumption of 
armchair mappers that the floodplain of a braided river is fully water 
covered at some time of the year - which is wrong in many cases.  See 
for example

https://mc.bbbike.org/mc/?lon=12.914557&lat=46.081555&zoom=14&num=2&mt0=bing-satellite&mt1=mapnik

The wiki has contained suggestions for varying water levels for quite 
some time:

https://wiki.openstreetmap.org/wiki/Tag%3Awaterway%3Driverbank#Varying_water_level_river

In any case you should never map a water area based on the assumption it 
is actually water covered at some time without evidence it actually is.  
If you are unfamiliar with an area it is usually better to map what you 
can observe on the sources available than to just guess.

-- 
Christoph Hormann
http://www.imagico.de/