Just a note that the old range map was a lot worse. @cmcheatle, maybe better to start with http://maps.iucnredlist.org/map.html?id=22694497 ?

The new map is better for the areas it does cover, but it appears to be missing the entire east Asian (and much of the high Arctic) range of the species. Black-legged Kittiwake also breeds along the western and eastern coasts of Greenland, most of the Arctic Ocean islands north of Russia (Novaya Zemlya, Severnaya Zemlya [and the adjacent northern Taymyr Peninsula], [apparently at least some of] the New Siberian Islands, Wrangel and Herald Islands, etc.), along the coast of the Russian mainland from at least the Kolyma River delta (but apparently also farther west along most of the coast of the Sakha Republic?) east and south around the Chukchi and Kamchatka peninsulas into the northern Sea of Okhotsk, and on the Commander Islands, the Kuril Islands south to Urup Island, and on Cape Patience (Terpeniya) on Sakhalin Island (as well as a couple small colonies in northwestern Spain). It winters south along the coasts of Asia at least to South Korea and southern Japan (along both coasts of Japan); and south in the east Atlantic to nw Africa (and in smaller numbers into the western Mediterranean and apparently the Black Sea); and well offshore throughout much of the North Atlantic (maybe to around 35°N, at least in the east Atlantic) and North Pacific (to around maybe 41°N).

I wouldn't only rely on GBIF if I were you (it's a great resource, but global coverage is highly uneven, so just because you don't see many records in an area doesn't mean a species is necessarily rare there) - I'd consult authoritative works such as Gulls of Europe, Asia and North America (Olsen and Larsson 2004); Birds of North America Online; Handbook of the Birds of the World Alive; relevant papers if you can find them; good field guides to the regions involved (if such field guides exist); etc.

A couple notes: Black-legged Kittiwake is not a vagrant to Japan, despite what it might look like in GBIF (and looking at the GBIF map myself, I'm not sure I see it as indicating that the species is a vagrant to Japan) - it's fairly common (or at least not that uncommon) off the coasts of northern Japan in winter, and less common but still regular farther south. On the east coast of North America, birds south of Delaware aren't considered vagrants (contra your statement above). And sure, they're rare and localized breeders in Spain, but they're still not rare in winter off the Atlantic coast (they're never super localized in winter where they occur). (Essentially, much more of the coast of Asia, Europe/nw Africa, and the east coast of North America should be included in the map - they're certainly more common in a lot of these areas than they are on e.g. Lake Michigan, Lake Ontario, or the coast of Florida, which are included in the map.)

Also, kittiwakes are truly pelagic during the non-breeding season - they're not "still effectively a land bird" or "simply passing over"; they forage, rest, etc. offshore (often well offshore, not in sight of land) for several months of the year - see the BNA habitat description a couple comments down. (So the map should extend much farther offshore than it currently does, too.)

A couple relevant range descriptions are below (too long to include in the same comment, sorry)...

From chapter 3 ("The breeding seabirds of the Russian Far East") of the publication Seabirds of the Russian Far East, for precise breeding distribution in eastern Russia:
"In the Arctic Far East seas, the western boundary of the kittiwakes’ nesting range is Chetrekhstolbovoy Island near the Kolyma River delta. Closer to the mainland coast, it does not breed west of Kittiwarken Rock in Chaun Bay (Kondratyev 1986)... The total number of kittiwakes nesting on Wrangel and Gerald [sic] islands ranges from 70 000 to 175 000 birds annually (Stishov et al. 1991). The total nesting population in the Far East Arctic basin as a whole could reach 250 000 or 300 000 (A.Ya. Kondratyev, unpubl. data). On Big Diomede, 3400 kittiwakes were counted on cliffs in 1991, but not more than 1500 pairs nested there. In that same year, the population of the northern part of the Bering Sea was estimated to be about 900 000–1 100 000 birds (Kondratyev 1993b). In the Kamchatka Region, about 202 000 pairs nest along the eastern coast of the peninsula, and perhaps another 80 000 pairs nest along the western coast (Vyatkin, in press b). Some 40 000 pairs nest on the Komandorskiye Islands (Vyatkin and Artyukhin 1994; Artyukhin, in press b). In the northern Sea of Okhotsk, the largest kittiwake colony is on Talan Island, which holds 30 000–35 000 pairs (A.Ya. Kondratyev, unpubl. data). About 90 000 nest on the Kuril Islands, with the largest colonies on Ptichy, Antziferov, Raikoke, Matua, and some other islands. The most southern colonies on the Kurils are on Urup Island (Velizhanin 1978). The only nesting area on Sakhalin is Cape Terpeniya, which had 5000 birds in 1981 (Nechaev 1991). A colony on Tyuleniy Island, decimated by uncontrolled exploitation in the 1950s, has recently recovered its status. In 1947–1948, it held 1500 birds (Gizenko 1955), but by 1963 there were not more than 200 birds (Benkovsky 1968). From 1974 to 1976, it held about 300 birds (Nechaev and Timofeeva 1980), but numbers increased to 1680 in 1991, 1800 in 1992, 1900 in 1993, and 2200 in 1994 (Trukhin and Kuzin 1996). It is also a common breeding bird on some islands of the Shantars (Yakhontov 1977).

(I don't have a copy of Gulls of Europe, Asia and North America, and the Amazon preview doesn't include the pages with the range map and description of the distribution, but I might be able to get access to a copy at some point in the next couple days.)

Would it be possible to combine parts of the range map you produced with the BirdLife/IUCN range map (which already covers all of the omitted areas)? (Also, I don't have any experience creating these sorts of maps so I have no idea how difficult/time-consuming this would be, but is there an easy way to trim a map so that it follows the coastline or at least doesn't extend so far inland, especially for the wintering range?)

The IUCN/BirdLife maps don't include the coast of South Carolina - the closest they get is a couple dozen km offshore, which seems roughly accurate (although it might indeed be a bit more accurate if they were moved a little farther offshore); Black-legged Kittiwake is evidently regular enough in offshore waters that it's an inland-only review species on the South Carolina checklist. (The most recent study of pelagic bird distributions off South Carolina that I could find seems to be this one, which describes kittiwakes as "...rare offshore, late December through mid-March... Feeding flocks of kittiwakes were observed on roughly two-thirds of winter trips to hard-ground reefs along the Continental Shelf when large commercial grouper/snapper boats were present. Counts ranged from three to 12 birds...")

If merging the maps, would it also be possible to maybe move the inland border on the west coast of North America so it isn't quite so far inland? (It probably makes sense for the map to include the coast since that's where most iNat observations of kittiwakes are [the IUCN/BirdLife map stops before it reaches the coast, which might be more accurate in terms of kittiwake density, but the kittiwakes that get seen the most by people are going to be those that make it to land, and while it wouldn't be the end of the world if those were marked as out of range, it wouldn't be the best thing either]. At the same time, though, the border of the new range map extends dozens of km inland from the coast [even up to a couple hundred km in some areas], which doesn't seem that realistic. If tracing along the coast is very time-consuming, I can try to work on it if you'd prefer - I'm busy through this Thursday, but I'd be free after that)

This is the workflow I have figured out:

Download a shapefile that represents the world coastlines
Run a geoprocessing task that creates a fixed distance buffer both inshore and offshore from the buffer - say 10km either way
Load in the hand-drawn inland ranges. Run another geoprocessing tasks that uses that range as an input, compares it to the buffer from step 2, and removes any areas in that range which are not also inside that 10km buffer, so as to remove any places it got too far inland.
Merge the result with the offshore range map from the ICUN.
Manually add back any remaining areas. The most obvious is the Great Lakes (unfortunately the shapefiles for step 1 I can find exclude them, so that disappears when I run step 3). I do think the Great Lakes should be in range for this species. I personally saw 10 this winter in 40 kilometers between my house and Toronto, and an 11th at Niagara Falls. They are annual, regular and in decent but low numbers.

That then gets a very close to ideal map. 2 issues remain.

in some places the ICUN only comes to say 20-30 km to shore, which means there are uncaptured "slivers" between the 2 maps. I can either extend the buffer used in step 2 to say 25km, which cuts those slivers down significantly, although a few would still remain. This of course means the inland "range" goes further inland. Or the slivers can be manually filled, which is doable, but tiresome. Or even just leave them which looks a little off
there is a bug in how GIS handles polygons very close to or which cross the international date line (when you try and create them, they reproject to wrap around the entire globe). Because of this I cant seem to get the last 100 or so or the most extreme northeast Russian mainland and the furthest out Aleutians incorporated. Maybe folks with higher end GIS skills or tools than I can do it, but all the web reading I did was well beyond me.

Thoughts ?

Cassi - I have your email, Max if you want to share yours, I can send pictures to show what it would look like.

My preference is that as long as the IUCN concept matches the IUCN concept (eg its not that the IUCN map refers to Rissa tridactyla sensu lato and iNat is using Rissa tridactyla sensu stricto) and there's no major error in the map (e.g. I've seen IUCN range maps where just a few small offshore islands are included and its clear that they accidentally omitted the continental polygon representing 99% of the range) and iNat has the most recent version (currently IUCN 2019-3) then even if the IUCN range isn't perfect they should be included because in my opinion (a) they are useful for getting a handle on what is meant by the taxon, (b) since they are external published content people can take them at face value / take up issues directly with IUCN, and (c) achieving a 'perfect' range map is impossible since they will always be abstractions of real distributions with polygons that by definition abstract and simplify the true distribution. Just my opinion curious what others think. But closing this since the IUCN range is current