ne_10m_geography_marine_polys: clipped and missing features

[MLWorley]

Instead of appearing between the Machangelo Peninsula (the peninsula on the southern end) and the Punto da Macaneta to the north, Baia de Maputo is spread out along the coastline in that area.

The ring seems to have some extra points spliced into it.

[nvkelso]

This smells like #422 where somehow the v4 series lost features, here marine areas instead of lakes.

Looking back in time:

• https://github.com/nvkelso/natural-earth-vector/commits/master/10m_physical/ne_10m_geography_marine_polys.shp

So the "min_zoom" commit just before those other v4 changes in 2017 has full shapes for all these features.

Below Bahia de Maputo.

There are 311 features in the older file, and 306 in the newer file. The older file lacks ne_id and other newer columns. The problematic commits were reprocessing the shapes with mapshaper and with buffer(0), which indicates some geometry errors in the original shapes.

[nvkelso]

The 5 missing features are:

• Bering Strait
• ~La Perouse Strait~ Spelling issues instead
• Strait of Juan de Fuca
• Puget Sound
• Strait of Georgia

[nvkelso]

The following are too small:

• Denmark Strait (1159119235)
• Why did Straight of Florida change shape (1159116625)
• Korea Strait should get bigger (1159116907) - it was partially addressed in v5.1.0
• Sea of Japan needs to exclude Korea Straight? (1159115305)

And in the other direction too small:

• Makassar Strait (1159116881)
• Internal Philippines Waters (1159120309)
• Internal Canada Arctic Waters (1159119777)
• Ross Sea (1159115615)
• Internal Canada (B.C.) Waters (1159120473)
• Baia de Maputo (1159118905)
• Karskiye Strait
• St. Helena Bay
• Strait of Gibraltar
• Bab el Mandeb
• Scotia Sea
• Yucatan Channel
• Great Barrier Reef - partially addressed in v5?
• Drake Passage
• Luzon Strait

These were deliberately combined:

• Mediterranean Sea (1159115321)
• Ross Sea

[nvkelso]

From 204b50046f908df927a31aa52b6d3e32f79b9709:

        - Over 175 row changes to polygon geometry and attributes.
        - Restored or remastered missing features, including: Bab el Mandeb, Baia d
          Maputo, Bering Strait, Denmark Strait, Drake Passage, Great Barrier
          Reef (again), Karskiye Strait, Korea Strait, Luzon Strait, Makassar Strait,
          Puget Sound (again), Ross Sea (again), Scotia Sea, Sea of Japan (again),
          St. Helena Bay, Strait of Dover, Strait of Florida (again), Strait of
          Georgia (again), Strait of Gibraltar, Strait of Hormuz, Strait of Juan
          de Fuca (again), and Yucatan Channel.
        - Added new features primarily in Europe, South East Asia, and Southern Ocean,
          including Aru Sea, Bay of Kiel, Bothnia Bay, Bothnian Sea, Celtic Sea,
          Ceram Sea, Cooperation Sea, Cosmonauts Sea, D'Urville Sea, Fehmarnbelt,
          Iceland Sea, Lazarev Sea, Liaodong Wan, Malacca Strait, Mawson Sea, Natuna
          Sea, Riiser-Larsen Sea, Selat Sumba, Selat Sunda, Somov Sea, Sound Sea,
          Storebælt, Strait of Sicily, Teluk Berau, and Teluk Bone.
        - Mediterranean Sea was reworked back to two separate polygons, one each
          for east and west basins, with separate ne_id and min_label values.
        - Added oceanbasin and subbasin columns.
        - See also: https://legacy.iho.int/mtg_docs/com_wg/S-23WG/S-23WG_Misc/Draft_2002/Draft_2002.htm

[MLWorley]

Is 5.2.0 supposed to be downloadable yet for ne_10m_geography_marine_polys? The version linked to on the download page is still 5.1.0, which of course still has the pre-5.2.0 issues.

[MLWorley]

Finally I was able to download and look at the revised hydro stuff (although not through the downloads page but more directly to the repository).

I'm currently stepping through the various bodies of water in the data set; once I updated my copies of the shp, shx, dbf files I could see the changes (the VERSION file still says 5.1.something rather than 5.2.0).

All Most of the bays look fine with the possible exception of ne_id = 1779565055 name = Liaodong Wan. Liaodong Wan seems to have OK geometry, but it overlaps the northern part of Bo Hai; I had to remove Bo Hai to be able to see Liaodong Wan in my viewing / editing tool. Was Bo Hai supposed to be adjusted so that they don't overlap, or are they supposed to overlap?

Further remark: St Helena Bay (ne_id = 1159118925) still seems to have distorted geometry - it's spread out along the shoreline of what I think should probably be the intended bay, rather than forming a proper bay.

[MLWorley]

The Drake Passage in the revised version appears to be partially overlapping the South Atlantic Ocean, so that at first it's difficult to tell that it's present (although the geometry of the Drake Passage appears to have been much improved when the overlapping South Atlantic Ocean geometry is temporarily removed to make the Drake Passage visible).

Sounds like a similar issue to Liaodong Wan above, possibly?

Further remark: the Bransfield Strait looks as though it is in sync with the Drake Passage; the main problem in that area now seems to be that the South Atlantic Ocean isn't clipped to have a common boundary with the Drake Passage and therefore overlaps it heavily.

[MLWorley]

La Perouse Strait now has a problem similar to Liaodong Wan and the Drake Passage - while its boundary with the Sea of Ohotosk seems to be all right, it now overlaps the Sea of Japan, so depending on the rendering order it can't be seen.

[MLWorley]

Kattegat overlaps Samso Baelt (ne_id=1159120631) so that the latter cannot be seen, similar to the Sea of Japan's relationship with La Perouse Strait in my previous comment.

[MLWorley]

The Strait of Dover lies mostly under the English Channel, but also partially overlaps the North Sea. (It's supposed to link the two - it looks as though neither the English Channel nor the North Sea quite respects its boundaries, although the North Sea is better about it.)

[MLWorley]

The Strait of Georgia, the Strait of Juan de Luca and the Salish Sea appear to overlap quite heavily, to the point that if either Strait is selected, it's not visible, although the geometry seems to be present and correct once some of the overlapping geometry is removed.

[almccon]

The Salish Sea is supposed to fully include the Strait of Georgia and the Strait of Juan de Fuca:

From https://en.wikipedia.org/wiki/Salish_Sea:

it includes the Strait of Georgia, Strait of Juan de Fuca, Puget Sound, and an intricate network of connecting channels and adjoining waterways.

[MLWorley]

If the bodies of water are supposed to include each other, then that's the way it is, of course.

In this case I mention it because the overlap is so great that depending on the rendering order, some of the bodies of water cannot be seen.

From my own point of view, I'd take care (once I knew that the bodies of water might overlap one another) to try to do some priority ordering so that I could consistently predict which body of water would be identified as the containing body at a given position. That's on my end, though.

[MLWorley]

Similarly, it appears that the Gulf of Bothnia contains (consists of) the union of Bothnia Bay and Bothnian Sea. If that's the intent, that's fine, of course.

[nvkelso]

Thanks for your feedback!

Most of the overlaps are intentional -- these are meant as labeling areas not anything else. The min_zoom value indicates which label should be sized larger or smaller and win in a collision. The overlap is defined by the international hydro organization.

I'm interested in learning more about your project. Email me at nvkelso at the gmail?

[MLWorley]

Feedback on bays

(Sorry about the delayed feedback; I hadn't had time to get back to this for a while.)

The following images are taken from the repository's version as of this morning, the one that is noted as 5.2.0 updated at the beginning of June.

St. Helena Bay (ne_id = 1159118925) is still distorted (at least, I think so). (I noted this in the above comments but I think I may not have been clear earlier that this was in the repository's pending update.) St Helena Bay did have problems before the pending 5.2.0 fixes but to me it still looks as though it has a problem.

I think Shark Bay (ne_id = 1159117955) is a bit questionable, although not the way St. Helena Bay is - note that it comes to a long sharp point on the northern end, which looks as though it might need a bit of looking at.

Sulzberger Bay (ne_id = 1159117531) similarly has the westernmost corner of the geometry coming to an elongated sharp point.

[nvkelso]

:wave: @MLWorley thanks for your feedback.

As @almccon points out, some overlap is intended and is indeed present in the source IHO. Please turn transparency on, and also review the min_zoom values for the different features which show the larger unit will "label" first, and then the smaller units. For features that do overlap, the smaller min_zoom determines the "parent", or look in the basin and subbasin columns at the far right of the table.

Further remark: St Helena Bay (ne_id = 1159118925) still seems to have distorted geometry - it's spread out along the shoreline of what I think should probably be the intended bay, rather than forming a proper bay.

This sounds like an error for me to research.

I'll also 2x check these two:

Kattegat overlaps Samso Baelt (ne_id=1159120631) so that the latter cannot be seen, similar to the Sea of Japan's relationship with La Perouse Strait in my previous comment.

Probably intended?

Further remark: the Bransfield Strait looks as though it is in sync with the Drake Passage; the main problem in that area now seems to be that the South Atlantic Ocean isn't clipped to have a common boundary with the Drake Passage and therefore overlaps it heavily.

This one sounds fishier.

Agree the 3 bays could use some cleanup, though the one of most concern is the slivery St. Helena Bay.