I have understood the 2 m value is the height above "mean" sea level. But, natural=coastline is set approx in the high water mark and here tidal range is big. This is my doubt.

Hello, looking satellite imagery it seems Baker Lake is really an enclosed maritime bay instead a freshwater non-maritime body.

Here there is an unconnected coastline segment: http://tools.geofabrik.de/osmi/?view=coastline&lon=-81.50004&lat=52.32722&zoom=16&opacity=1.00&overlays=coastline,coastline_error_lines,line_not_a_ring,line_overlap,line_invalid,line_direction,questionable,coastline_error_points,unconnected,intersections,not_a_ring,double_node,tagged_node

Hola, el cambio que hiciste en el límite internacional es incorrecto. El Tratado de 1982 dice que el límite internacional en la boca oriental del Estrecho de Magallanes es una línea recta que une los hitos de Punta Dungeness con el del Cabo Espíritu Santo. Incluso había una nota que creo no leíste. Lo he vuelto a corregir. Saludos.

Ok, but must be careful to add every island to multipolygon as inner ring to avoid a "flooding". I add recently like twenty islands mapped previously, but was not part of the relation.

changeset/33818750

I didn't change anything. I just revert a changeset mine. Tagging is as before.

Hola, de qué fuente obtuviste los límites del parque? Gracias.

I didn't add the tag name:en of Tierra del Fuego, but thanks for notice it.

Hello, here there is an opened boundary relation named Thondwe EPA: http://overpass-turbo.eu/s/f7x

Thanks for the info. The problem is there are various complex lakes for an multipolygon relation. Great Slave Lake can be a sample.
Regards.

Hello SK53, I'm sorry, apparently I had to read more about it. Before my changeset I didn't know that it was a community decision tag Great Lakes as natural=coastline. But I found strange to see Lake Ontario tagged natural=water and Lake Superior tagged both ways. I thought I was helping with a multipolygon conversion (most of people have difficulty with that). Not that I want to impose my viewpoint. I'm available to help if you need revert part of my changesets. About the rendering it's a problem of Mapnik. Could be requested a natural=water rendering in higher zoom levels like natural=coastline.
Regards.

In this changeset I put source "OSM Inspector". Then you can know that I solve errors in coastline lines (like intersections with itself): http://tools.geofabrik.de/osmi/?view=coastline&lon=-7.28129&lat=57.13217&zoom=14&opacity=1.00&overlays=coastline,coastline_error_lines,line_not_a_ring,line_overlap,line_invalid,line_direction,questionable,coastline_error_points,unconnected,intersections,not_a_ring,double_node,tagged_node

Someone drawing a new coastline, but he didn't complete the work and he didn't remove the old coastline. I only delete the old coastline and fit the boundary to the new coastline (before it was fit to the old coastline) for not to break the boundary relation.

You're welcome. Also if you use JOSM editor you can run the validator, then you can detect any error (like coastline intersections or gaps) before uploading data. Regards.

Do you refer to Treaty of 1908? So you have a source? I haven't found it. But, lakes and rivers are natural features and the lower limit could be the influence of the lake over the river and the upper limit could be the place where the water begins decrease in altitude, I think.

OSM wiki rules say you can use waterway=riverbank or natural=water + water=river. It's an user decision. I prefer the first because with it you get the "desired" rendering.

English is no my mother language. I was just doing an observation. You not consider the comment as an "attack". Must be careful with automated data import. You should think that not all users fill source tag with Bing when it's necessary.
I'm not mapping here, but in your import data you broke the coastline in many parts and I detected that from OSM Inspector.
You can check here the actual coastline (the blue line) and the coastline before this changeset (mapnik rendering): http://tools.geofabrik.de/osmi/?view=coastline&lon=-79.54139&lat=56.80294&zoom=16&opacity=1.00&overlays=coastline,coastline_error_lines,line_not_a_ring,line_overlap,line_invalid,line_direction,questionable,coastline_error_points,unconnected,intersections,not_a_ring,double_node,tagged_node

Regards.

In this import you have replaced some data with resolution worse than they had before (Bing Imagery).
For example here: osm.org/#map=18/56.81269/-79.59474
Press edit button and compare it!

In fact, this rectangle doesn't represented the disputed territory. The claimed boundaries are different. Nevertheless, official cartography show a gap in this part of the boundary, like as Google Maps or Here Maps: http://tools.geofabrik.de/mc/#8/-49.5192/-72.7423&num=4&mt0=mapnik&mt1=google-map&mt2=geofabrik-topo&mt3=nokia-map

Unfortunately OSM require closed boundaries. This is the official map of 1998 Treaty between Argentina and Chile: http://www.difrol.cl/index.php?option=com_docman&task=doc_download&gid=182&Itemid=

The rectangle define the area of cartographic studies for a future demarcation of the boundary. Perhaps is better only keep boundary=administrative tag.

This rectangle isn't a "country", add administrative tags is wrong. This is only used for close admin relations in an area where Argentina-Chile boundary have not delimited.

Creo que es mejor mapear los usos de suelo con polígonos pequeños, un polígono grande no ofrece ninguna ventaja y es difícil de manipular para usuarios con poca experiencia, que pueden toparse con el límite de 2.000 nodos por segmento. Esta creo que es la mejor forma de mapear grandes bosques por ejemplo: http://prnt.sc/abnkm6
Saludos.