Inversely, Ross Island just north of Manitoba’s Lake Winnipeg was missed by the Atlas of Canada (presumably for this same reason), but the data I was working with had it separated entirely from the mainland.

Now if finding the right islands can be tricky, deciding how big each one is, is truly fraught. The problems in determining a true and comparable area for each island are twofold. First, the cartographer has to wrangle a spherical Earth onto a two-dimensional plane (paper or computer screen). Second, the cartographer has to draw a precise line around the island, capturing all that is the island, and none of what isn’t the island, in the process.

The challenges of choosing the right map projection (or transformation from sphere to plane) are evidenced by the now vilified Mercator projection (below). You have likely seen this in countless online mapping applications — and it profoundly distorts the shape and area of landmasses as you get closer to the poles. Clearly not the map projection for comparing area (Greenland is not, as the Mercator projection would suggest, as big as Africa).

And while there are “equal area” projections that preserve a true relative area throughout a map (see example on right below), as I played with different equal area projections for this project, I realized that while they likely do preserve a relative equality, each one yields a slightly different absolute area on an island-by-island basis. Realizing this proves the impossibility of complete precision when calculating geographic area.

But surely, if an equal area projection yields comparable results, then at least we could successfully rank the islands by relative size. And we could, if only we didn’t have to accept that drawing the line around the island — the very heart of this exercise — is significantly more subjective and imprecise than one might think. This oldest of all cartographic exercises, putting pen to paper to trace a shoreline, has all of the pitfalls today that it did when explorers and colonists were busily drafting the shapes of the continents in the 16th and 17th centuries. Our instruments for collecting the raw data have improved immeasurably, but there remains a good deal of subjective decision making as the cartographer decides just where and how to draw the line between land and water.

Where to draw the line is simple enough for much of the shoreline, but becomes challenging when a large river on an island fans widely into a delta at its mouth. These are the ecotones of the cartographic world — the places that are two things at once, in this case river and ocean. And just where one ends and the other begins is in the hands of the cartographer and their “pen” as they draw this line. So, while the difference in area that will result from different choices here is very subtle, it is real, and will yield slightly different areal results. Greater differences still are latent in changing water levels, so as sea levels rise, or lake levels drop — an island grows or shrinks with every inch of water.

The core question about how to draw the line has to do with scale. As you zoom in closer to the Earth, more details are revealed. Little bays around an island that were not visible from 100 kilometres above the Earth are clear and present at 10 kilometres. Indeed, the complexity of the shoreline increases almost infinitely, as you get closer. Technically, you could get to the level of drawing around each pebble at the shoreline (for the truest true measure of the island’s size), adding waves and undulations to each metre of shoreline, and in so doing adding square centimetres of area with each jog. Clearly this level of detail is entirely unrealistic, but it proves the point that the scale you choose to draw the line at (see example below) has significant impacts on the final area calculated for the island.