I'm not really sure how on Earth I managed to get the second half of Greenland done in a week and a half. I thought this would be a three-week job at minimum. I guess I just had a really good week last week.
All that progress came in spite of both Greenland's gnarly geography (I'm beginning to sound like a broken record I know, but damn the ice ages) and some incredibly spotty, inconsistent and often downright contradictory primary sources covering northern and northeastern Greenland. Shout-out to
@Altaic for pointing out just how inaccurate some seemingly reputable sources actually are a couple of months ago, as when I hit the problem in earnest early last week I was somewhat prepared for it. You'll be pleased to know that after quite a lot of cross-referencing and double-checking, I'm pretty sure I have the correct geography nailed down, and even better I suspect I know why such a discrepancy exists in the first place.
The missing piece of the puzzle was a series of maps of the high Arctic produced by the US Department of Defense back in the 60's,
digitised here.
This map in particular plotting the far northeast of Greenland was particularly useful, not only by providing a visualisation of the area that isn't horrendously destroyed by projection warping, but because it's also annotated. At multiple locations on the map, some form of this disclaimer appears; "
Arctic Institute of North America Project Nord (Control Data Corp.) indicates position discrepancies in excess of 11 nautical miles (Nov/68)". There is some variation in punctuation, and a few areas apparently aren't as off from where they should be as others, but this is an otherwise pretty clear warning that though the shapes of the land were about right, the actual positions were dead wrong. When you're as far north as the top of Greenland, a discrepancy of 11 nautical miles can be as much as two degrees of longitude away from where a feature should be, a problem confounded by map distortion once a dataset is reprojected.
If I had to guess, I'd say these problems are due to the harsh and difficult nature of the terrain making accurate field mapping difficult, in addition to many methods for triangulating a position on the ground not working so well so close to the pole, leading to an awful lot of misalignment. I think this inherent inaccuracy in even the best hisotrical maps is why so many sources are wrong - at some point in the 80's or 90's, somebody digitised a physical map to produce a new coastline shapefile without realising that the map they were using as a source was itself wrong, resulting in a tonne of datasets where the land is off from where it should be. Perhaps the best example is google maps (
see the example linked here), where stretches of ocean are included as if they were land and tracts of land are omitted as if they were ocean.
This dataset covering ice sheets of the world is another good example - just zoom in on northern Greenland and notice how the coasts and ice layer doesn't line up with the basemap.
The earliest versions of the 8K-BAM suffer from this rather badly, which is a shame as I've been using a reprojected 8K-BAM for years as a primary source when drawing on coasts. All the way back at the beginning of the thread I asked one of the creators of the 8K-BAM
@Drex for his sources, and got
this map based on 2015 ESA data back as a response. That map is, alas, just flat out wrong. Basically the entire coastline of northern Greenland is shifted west by a degree of latitude or more, and the shift isn't even consistent, with some areas being more off than others. Several fjords and straits are filled in as land (e.g.
Victoria fjord), while large bites have been taken out of the land (e.g., from the
Mylius-Erichsen Land peninsula). It's a complete mess, and it was apparently copied verbatim for the 8K-BAM. I think those problems have been patched in more recent versions, but the versions I have, including a lot of the hisorical and geographical resources, are all rather badly flawed.
But here's the thing. As well as pointing out its own inaccuracies,
the DoD map tries to rectify the problem somewhat by listing the real coordinates of a few key locations, places where genuinely accurate measurements were able to be taken. For example, for the weather station that would become
Station Nord (the furthest north permanently manned station on Greenland and the second furthers north permanently occupied place in the word) the map has this to say; "
A reliable geodetic positional determination of the weather station is 81°36'09''N, 16°40'12''W.". The map provides a handful of real fixed points to compare against which was invaluable when deciding which datasets to use as sources and which are wrong.
After a lot of double-checking, I realised that several datasets lined up with each other and the old coordinates listed on the DoD map. It appears that satellite mapping has solved the old cartographic inaccuracies and producing decent basemaps that aling with the handful of good coordinates listed in the DoD map, but that not every dataset got the memo, with quite a few still using traces from the flawed old physical maps.
This NASA map lies up, though some of the datasets I have toggled on in that link apparently don't. The source for that coastline layer is listed as
openstreetmap, and while I spent way too long trying to overlay a graticule on openstreetmap to confirm things visually without success, I eventually resorted to manually selecting various points and checking their coordinates which seemed to confirm that it lined up too. A lot of the layers helpfully provided for the R-QBAM project by
@Rac98 here including the administrative and topographic layers line up as well. I was particularly pleased to learn that
the Robinson webmap I've been linking to passes the test with flying colours and lines up with the others, as does the ESA 2020 land-cover data I've been using as an overlay (
description here,
downloadable here). Unfortunately the ESA dataset alas doesn't extend further north than 82.75°N, so I spent another day or so digging up sources that showed ice caps that far north and that lined up with the correct maps. I eventually found a collection of GIS layers covering Greenland downloadable
here, and was very happy to see that ice sheets and base geography lined up with the 2020 ESA data once I ported them all into QGIS.
Once I'd figured out which sources are more trustworthy than others, I got to mapping, which proved difficult, painful and time-consuming, however I somehow got it all done quickly. I think I got sucked in by the challenge Greenland presented, and once I'd figured out the dataset problems I had gotten so much done that it made sense to keep up the good pace and push on to completion.
Major note, my suspicions that the effects of global warming are visible in the R-QBAM
were bang on the money. The 2020 ESA data shows the fronts of several glaciers further inland than other older datasets, a fact confirmed by comparing the areas in question with the 60's DoD maps
linked to previously and some even earlier maps from the 50's
here. The
Q-Greenland GIS data I've already linked to also provides layers showing the retreat of several glaciers between 2001 and 2021, corroborating the ESA data and painting a depressing picture. Hell, the
Petermann Glacier in the far north near Ellesmere island has receded inland far enough to add about 30 pixels of new sea and coastline.
Long story short, when I expand the 1914 map to cover Greenland, I'll have to do a patch for Greenland's pre-industrial coastline. Typical.
On another note, while making this I ended up doing quite a lot of reading around Greenland and Arctic exploration at the turn of the 20th century. I find it darkly amusing just how blindly nationalist some of the 19th and early 20th century explorers got when naming new geographical features they discovered. Anyone want to guess the nationality of the man who named
Ile-de-France?; What about
Germania Land (complete with the
Cape Bismark)?; Or how about the
Roosevelt mountains and
Independence Fjord?. And that's what happened when they weren't naming places after themselves, for example
Nares Land,
Peary Land, or after minor Danish royals (
the Danish royal family being the Danish royal family, there are a lot of King Christan and King Frederick Lands) and many more I won't bother to list here.
Also, I can't not mention it now that I've learned it, but there's a place in the far north of Greenland called
Antarctic Bay. At first when I saw the name on a map I suspected some century-old trolling from a bored explorer, until I learned on further reading that the bay was named after
a 19th century polar exploration ship that saw use in both the Arctic and the Antarctic, so it at least makes some sense.
Final point,
one of Greenland's islands was only discovered as late as 1993.
With all that out of the way, on to ...
Patch 113 - Greenland 2
- Added The remaining two thirds of Greenland, finishing Denmark.
- Mildly tweaked several other arctic island groups, namely Iceland, Svalbard, Novaya Zemlya and the Franz Joseph archipelago.
- Moved Rockall.
View attachment 893928