Projecting the Curved Earth

For centuries, cartographers have grappled with an intractable mathematical dilemma: how to portray the surface of a three-dimensional sphere upon a two-dimensional plane. The geometry of a curved surface cannot be flattened onto a sheet of paper without introducing some degree of stretching, tearing, or compression. In mathematical terms, because the surface of a sphere possesses constant positive curvature while a flat plane has zero curvature, no projection can simultaneously preserve area, shape, distance, and direction across the entire globe. As a consequence, every flat map is inevitably an exercise in compromise, forcing its creator to decide which geographic properties must be maintained with precision and which may be sacrificed.

The most historically influential response to this challenge emerged in the sixteenth century with the Flemish geographer Gerardus Mercator. In 1569, Mercator introduced a cylindrical projection that revolutionised maritime navigation. On his map, lines of latitude and longitude intersect at right angles, forming a regular grid where any straight line drawn across the chart represents a line of constant compass bearing, known to mariners as a rhumb line or loxodrome. For sailors crossing vast, featureless oceans, this property was invaluable. A navigator could plot a course simply by drawing a single straight trajectory between origin and destination and holding that compass heading throughout the voyage, eliminating the need to recalculate headings continuously.

However, the mathematical formula that enabled Mercator to keep compass bearings straight came with a substantial cost in scale fidelity. As the projection approaches the poles, the spacing between parallels of latitude increases exponentially. Consequently, landmasses located far from the equator appear vastly enlarged relative to equatorial regions. Greenland, for example, appears on a standard Mercator map to be comparable in size to the continent of Africa, even though Africa is in reality roughly fourteen times larger. While this distortion posed little practical issue for ship captains working at sea, the widespread adoption of the projection in classrooms and general atlases over subsequent centuries led to persistent public misconceptions regarding global geography.

By the mid-twentieth century, growing dissatisfaction with Mercator's Eurocentric exaggeration prompted geographers to champion equal-area projections. These designs ensure that every landmass is depicted in exact proportion to its real-world surface area. The most widely debated of these was the Gall-Peters projection, which gained prominent advocacy in the 1970s. Promoted heavily as an equitable alternative that rectified historic biases against equatorial nations in Africa and South America, the projection accurately conveyed continental proportions. Nonetheless, cartographers criticised the projection for its extreme visual distortion of shapes, pointing out that tropical continents appeared excessively stretched vertically, while polar regions were severely flattened horizontally.

Recognising that both purely conformal maps (which preserve shapes locally) and purely equal-area maps introduce jarring visual artefacts, cartographers in the latter half of the twentieth century increasingly turned to compromise projections. Rather than attempting to eliminate one specific form of error entirely, compromise projections distribute distortions moderately across all properties. One notable example is the Robinson projection, designed in the 1960s to create a visually appealing world portrait for educational publishing. By allowing minor distortions in both area and shape, it produces an aesthetically balanced view of the Earth. A similar philosophy underpins the Winkel Tripel projection, which was later adopted by major geographical institutions as a standard reference map due to its exceptionally low overall distortion scores.

The rise of digital technology in the late twentieth and early twenty-first centuries brought unexpected developments to the field of cartography. When early digital platforms began developing interactive online maps for personal computing and mobile devices, engineers largely bypassed modern compromise projections. Instead, they resurrected a variation of the sixteenth-century cylindrical model, known as Web Mercator. This decision was primarily driven by computational efficiency: the projection simplifies rendering at different zoom levels and ensures that local street corners remain perpendicular, preventing rectangular buildings and road junctions from appearing sheared or warped when users zoom in closely.

Today, professional cartography relies on an extensive taxonomy of projections tailored to specific operational requirements. Aviation authorities, for instance, frequently utilise gnomonic or azimuthal projections centred on the poles to plot transcontinental flight paths, as these maps depict the shortest route between two points—a great circle—as a straight line. Interrupted projections, such as Goode’s homolosine, slice the oceans to display continental landmasses with minimal distortion of both shape and area, resembling an unpeeled orange laid flat. Ultimately, modern mapping software has not eliminated distortion, but rather empowered mapmakers to select precisely which mathematical trade-off best serves their intended audience.