Proposal talk:Coastline-River transit placement

Old discussion from 2014

Maybe the intended use of natural=coastline should be explained better.

• It is right or does make sense to have a river (waterway=river + natural=water polygon) outside the coastline?
• Must rivers have their end node in the natural=coastline?
• How can the ocean polygons be obtained if the coastline is moved inside the river? —Preceding unsigned comment added by Muralito (talk • contribs) 01:01, 2 Apr 2014

These are good questions which should probably be explained better on the wiki. Specifically:

• Yes, the river line can and in practical mapping frequently is extended outside the coastline. This makes a lot of sense especially if there is a dug deepwater channel or a marked shipping lane. See for example [1].
• There should always be a shared node but as said they can extend beyond.
• Coastline polygons are available from [2], see also Coastline.

Ocean polygons are the same as coastline polygon? If yes, the Rio de la Plata case is clear and the coastline is where it sould be, in its exterior limit. If no, the question remains, How can the ocean polygons be obtained? —Preceding unsigned comment added by Muralito (talk • contribs) 13:47, 2 Apr 2014

Yes, the only ocean polygon data in OSM are the coastlines so if you need an ocean polygon you use the processed coastline data. Subsections of the Ocean like marginal seas or bays are not currently mapped as polygons. --Imagico (talk) 17:46, 2 April 2014 (UTC)

In addition there is the matter of overlaps between the riverbank polygons and the coastline. These are rare but exist (like [3] and [4]) and there are no rules forbidding those. And overlap would underline the idea that some waterbody is both part of a river and the ocean. I did not specifically cover them in the proposal since that would further complicate things. My opinion is overlaps should be possible within the limits outlined in the draft.

--Imagico (talk) 08:12, 2 April 2014 (UTC)

The case of overlaping between River Elbe and the coastline, maybe is influenced by mapnik not showing the water polygons at low zoom, and probably we wouldn't be talking about this if mapnik rendered the Rio de la Plata water polygon at low zoom.

My opinion is overlaps between riverbank and coastline should be only when there are no clear limit between them. It's not the case of Rio de la Plata (max. wide 219 Km) , which exterior limit is clearly defined by IHO (International Hydrographic Organization [5].

Coastline refers to "marine environment". "marine" is defined as "related to sea or sea transport" [6] and "sea" is defined as "the salty water that covers a large part of the surface of the Earth, or a large area of salty water, smaller than an ocean, that is partly or completely surrounded by land" [7] The water of a river is not salty, so no "marine" and no "coastline"—Preceding unsigned comment added by Muralito (talk • contribs) 13:47, 2 Apr 2014

First of all this is not primarily about the Rio de la Plata but about the consistency of mapping, the proposal is meant to establish guidelines so that rivers around the world are mapped in a consistent way. If the proposal is modified in a way that supports the current Rio de la Plata mapping that would be fine with me, this is however difficult to accomplish without changing a lot of other rivers around the world.

Concerning the limits of rivers in general and the Rio de la Plata in particular - the rules i drafted are oriented at what you describe as marine environment in some of its characteristics. Only if the river current is the dominating current in the water is the ocean water prevented from progressing 'upstream' and extending the marine environment. As can be seen in the literature, for example in [8] the outer part of the Rio de la Plata is clearly a fully marine environment. This can also be seen on satellite images like [9] where the brownish color or the river water ends near Montevideo. The IHO definition does not seem overly relevant here since it only defines the boundary of the Rio de la Plata as a sea and not as a river so it does not deal with the difficulty of separating river and ocean. --Imagico (talk) 17:46, 2 April 2014 (UTC)

The IHO document has the definition of the limits South Atlantic Ocean, (#32 page 18 and 19) and is explicit in that matter, the ocean ends where the river begins, and that limit is establiched in. Is a river, not a sea, "River of Silver" not "Sea of Silver"

Here [10] you have a clear definition of the estuary (from this article [11]

Here [12] you have a study of the plume of the river, including the salinity and some maps of the model.—Preceding unsigned comment added by Muralito (talk • contribs) 23:14, 2 Apr 2014

The IHO document you cited documents the Limits of Oceans and Seas, not those of rivers. You might have noticed that the Rio de la Plata is the only river included there which - if anything - indicates the IHO considers it to be a marginal sea of the Atlantic Ocean. But this part of the discussion is getting very off-topic.

In "Rationale" you talk about different colours rendering to be a problem. In most of the aerial imagery of Rio de la Plata the waters of the river are clearly of another colour, not the same of the colour of the ocean. You talk that "will help avoid extreme cases like above", the Rio de la Plata case is extreme but is also real, so, IMO, there is no need to avoid it.

The more that I read, the most that I'm convinced that all this was to avoid a mapnik limitation, maybe the most old and extended case of Tagging_for_the_renderer—Preceding unsigned comment added by Muralito (talk • contribs) 13:47, 2 Apr 2014

The opposite is the case - as i explained the standard map style is fairly immune against waterbody mapping inconsistencies since all waterbodies are shown in the same color, the only exception are the lowest zoom levels where only the coastline is displayed and therefore you can see this. Most other maps however (like [13], [14], [15], [16]) show ocean and rivers in different color, therefore the transit between river and ocean is always prominently visible and if this is not mapped consistently the map looks bad - but not because some flaw in rendering but because of inconsistencies in the data. Therefore i do not think this proposal is any more tagging for the renderer than any tagging convention in general which always have the primary purpose of ensuring some level of consistency in the data. --Imagico (talk) 17:46, 2 April 2014 (UTC)

The colour of the water in [17] is by the particles in suspension. This is fresh (salinity<0.05%) water. The oceans begins where the salinity is more than 3%, which is near Punta del Este in Uruguay and Punta Rasa in Argentina. This limit is obviously changing by the tidal or climatic conditions, but the scientifically and legally acepted limit is this. The influence of the great volume of fresh water in the ocean is more than most people expect. You should be watching salinity maps or models. Of course, you are also welcome to came here and personally taste the water, survey the zone, and enjoy the beaches. :-)

In wikipedia there are some global maps of salinity http://en.wikipedia.org/wiki/Salinity The detailed maps or models i found are in spanish. I can also link here if you like.—Preceding unsigned comment added by Muralito (talk • contribs) 23:14, 2 Apr 2014

Please see the paper i have linked to at [18] which quite well documents how the water mixes in this case.

But I don't think it is productive to make this all about the Rio de la Plata - the question is if the proposed guidelines for placing the transit would be useful or not. As i have said before a salinity threshold like ocean begins where the salinity is more than 3% would not be a practically useful criterion since many bodies of water clearly belonging to the maritime domain (like Baltic Sea, Hudson Bay, Sea of Azov) feature a very low salinity and large rivers can reduce the salinity of the ocean far in front of the river mouth, especially if you only consider the surface layer.

If there are ideas for a better usable general criterion for placing the coastline i would be eager to hear.

--Imagico (talk) 23:37, 2 April 2014 (UTC)

This may be another interesting example to look at (Amazon and Anapu Rivers). To my non-expert eyes, the coastline extends pretty far upstream. http://overpass-turbo.eu/s/3M0

As far as i can see the coastline location at the Amazon is currently well within the limits of this proposal. Also in global comparison this is a fairly moderate placement. The Amazon is somewhat special due to its very high water throughput which would in theory allow moving the coastline quite a bit out into the ocean under the rules described here, further than it would make sense from a cartographic viewpoint, but this is a singular case and not enough reason to be more restrictive in general IMO. --Imagico (talk) 13:53, 19 June 2014 (UTC)

Bilder

Ein paar illustrierende Bilder wären hilfreich. Gruss, --Markus (talk) 20:04, 12 September 2014 (UTC)

New discussion

I totally agree with this proposal. This is not tagging for the renderer, this is a matter of defining precisely where the coastline closes in the mouth of a river. I think the word "marine" in coastline should not be there, as there are plenty of places around the world where large rivers make water not salty at all, the beach is just like a river beach, and you don't move the coastline offshore just because of that. May be we should ask to remove that word in coastline definition. I also agree on that this proposal should be considered in general, not just for río de la Plata. And I don't see why we have to attend a definition of IHO at all, we could perfectly have our own definitions and draw a different line with different tags if we want to state the limit definition by this organization.--Pertile (talk) 13:00, 10 May 2017 (UTC)

The coastline is a natural feature. Therefore the placement in Rio de la Plata is independent of any classifications defined by an external "authority". That means the coastline is mapped based on the principle of "map what's on the ground", i.e. where is the mean high water spring (for natural=coastline). Unfortunately the coastline was incorrectly moved to external limit of Rio de la Plata based on a boundary treaty (Treaty of the Rio de la Plata). This treaty defines Rio de la Plata as internal waters causing confusion in some people (note that "internal waters" is only a juridical concept, tagged as boundary=maritime+border_type=baseline in the external limit).--19jp87 (talk) 08:40, 17 February 2019 (UTC)

I think that the main issue is to differentiate inland rivers from estuaries like the Rio de la Plata. So it's not a sea/ocean, but it's still rendered similarly. Not much else to say, I'm not too much into geography.Nicfer (talk) 14:58, 10 May 2017 (UTC)

Note this proposal is not about how to map and tag the Rio de la Plata as whatever you want to classify it as, this is about where to place the coastline at a river mouth. It suggests fairly loose rules for that based on physical criteria in the hope that this can ensure some consistency internationally with mapping the coastline. If the Rio de la Plata is fully or partly inside or outside the coastline is not really of concern for this.--Imagico (talk) 15:35, 10 May 2017 (UTC)

Rationale

The section "Rationale" contains a number of factual inaccuracies, and statements made with no justification at all.

There is no justification for the statement "The distinction in mapping between the world ocean as the main reservoir of the global water cycle and the inland waterbodies can and should be considered one of the most fundamental aspects of mapping in OpenStreetMap." In fact when OSM was first started there were many who felt that there was no need to include coatsline data at all. There is no evidence why the distinction between ocean and water bodies is so fumdemental.

We are drifting into the philosophical here - do i need justification for stating a scientific fact as fact? If you are subscribing to a philosophy of radical relativism or social constructivism then probably yes. But this is not compatible to the very foundation of OSM to document knowledge of the intersubjectively verifiable geographic reality.

Anyway - i am not sure if you are questioning the statement that the ocean and rivers (as physically defined concepts, not as cultural concepts) have fundamentally different roles in the global water cycle, that the global water cycle is one of the most fundamental processes in the Earth ecosystem or that these facts make the representation of this distinction fundamental for OSM.

--Imagico (talk) 08:20, 26 November 2020 (UTC)

The statement "While the exact drawing of the line in this classification can be considered arbitrary it is no different in that regard than other classifications we perform in OpenStreetMap like the distinction between natural=wood and natural=scrub or water=lake and water=river." is interesting since, for instance, there is no formal classification of how many trees per hectare there needs to be before an area moves from the definition of "scrub" into "woodland". It is currently left to individual mappers interpretation. This proposal seems to put a complex distinction between ocean and water , a distinction which many will find hard to interpret and implement.

The distinction between natural=scrub and natural=woodland is one of plant height, not of plant density. Your case would be that of when the density of scattered trees turn natural=grassland or natural=bare_rock into natural=wood. In any case we are dealing with the same fundamental problem of applying a discrete classification system to a continuous geographic reality. The need to do that is fundamental to making meaningful statements about the geography of the world and hence for OpenStreetMap. That tags in OSM are usually not precisely delineated upfront when a tag is invented is obvious. But usually a consensus develops over time about this. One of the core elements of the success of OSM as a social project is that we have in many cases managed to develop such a consensus even across language and culture barriers. The purpose of tag documentation on the wiki and the proposal processes (and hence: also of this proposal) is to document this consensus. --Imagico (talk) 08:20, 26 November 2020 (UTC)

The statemnet "There has always been consensus that the coastline tag is meant to document exclusively physical geography and is meant to be used independently of the the culture specific differences in understanding of terms like coast, ocean, sea, river or inland water and similar terms in different languages." is untrue. It is the reason for instance that much of the tdtal section of the River Thames in London is mapped as river not coastline, because no one thinks of London as a coastal city.

I see no indication in the data or in the documentation on the wiki for the coastline tag being systematically used by local communities for anything other than documenting physical geography, not even in the UK (where this due to it being a native English speaking country would be much more likely). If you can show me such evidence i would be very interested. The Thames seems to be - though not e really extreme case on a global level - quite an outlier within the UK in terms of placement of the coastline-River transit. I would probably attribute that to tagging for the renderer - since the riverbank polygons, in contrast to established and documented convention, carry name tags duplicated from the waterway line (https://www.openstreetmap.org/way/66997114, https://www.openstreetmap.org/way/515944056) to label the river more prominently in maps. --Imagico (talk) 08:20, 26 November 2020 (UTC)

Is there a need for such a complex proposal ?

The statement "As of late 2020 for the most part the vast majority of coastline-river transits around the world are mapped within the bounds suggested by this proposal....and very few river mouths will need to be changed to comply with it." indcates that there is not much of a problem with current mapping interpretation and practice. It is unclear why such a complex and difficult to intepret proposal is therefore needed.

--Dmgroom (talk) 00:28, 26 November 2020 (UTC)

I'd question the claim that this is a complex proposal. The proposal outlines four fairly simple and clear rules (for upper and lower limit in the tidal and non-tidal case) that are practically verifiable for the local mapper. As said on the mailing list if anyone has an idea for a simpler (or just otherwise better - even if more complicated) set of rules that:

• is practically verifiable locally for mappers,
• is based on physical geography criteria (in case this is unclear: That means if you'd magically remove every human being from the face of the planet it would still be practically verifiable),
• is globally uniform (i.e. the same rules apply everywhere on the planet),
• matches the majority of existing practice of coastline-river transit placement in the database,
• that provides meaningful practical guidance to mappers where to place the coastline-river transit,

i would be very happy to see that. The need to have some documented guidelines stems from the conflicts linked to from the proposal.

--Imagico (talk) 08:35, 26 November 2020 (UTC)

The most recent activity on the tagging mailing list and the intervention of the DWG in the Chesapeake Bay coastline dispute, as well as the Rio de la Plata edit war, suggest that there is a very real need to document and achieve consensus on this topic, even given that most of the world does map these a certain way. I would encourage this proposal to move forward and attempt to achieve consensus on this topic. --ZeLonewolf (talk) 14:45, 26 November 2020 (UTC)

Tidal bore

This proposal should at least mention the  tidal bore phenomenon and where it fits into the coastline determination. --ZeLonewolf (talk) 14:47, 26 November 2020 (UTC)

In what way do you think that is relevant for mapping the coastline? --Imagico (talk) 15:23, 26 November 2020 (UTC)

Tidal bores occur on tidal portions of a river. If a river experiences this phenomenon, it would be useful to understand under your proposed definition if tidal bores are automatically on one side of the coastline or the other, and could serve as a clear marker for rivers which have this phenomenon. --ZeLonewolf (talk) 15:50, 26 November 2020 (UTC)

I don't know. My understanding is that tidal bores occur under specific circumstances and where they do they might be observable only along a small part of a river. At the same time as a wave they might propagate beyond the tidal range. So i don't think there is any relevancy for mapping of the coastline-river transit. Also note that their occurrence is not even tied to river mouths - they can form at bays as well apparently and do not depend on the inflow of river water. --Imagico (talk) 16:12, 26 November 2020 (UTC)