The sinusoidal projection is an equal-area map projection in which parallels are straight, evenly spaced horizontal lines and meridians are generally sinusoidal curves. The central meridian is a straight line perpendicular to the parallels. It preserves the relative areas of regions but increasingly distorts their shapes away from the central meridian, especially at high latitudes. Its simple mathematics and area-preserving geometry make it useful for world maps, thematic mapping, and global environmental datasets.
Classification and appearance
The sinusoidal projection belongs to the pseudocylindrical family: its parallels are straight and parallel, while most meridians are curved. Unlike a cylindrical projection, it narrows toward the poles rather than maintaining a rectangular outline.
In its standard spherical form:
- The equator is twice as long as the central meridian from pole to pole.
- Parallels are spaced in direct proportion to latitude.
- Each parallel has its correct length at the nominal map scale.
- The poles are represented by points.
- Meridians are equally spaced along any given parallel.
The name describes the meridian curves. Because horizontal position varies with the cosine of latitude, each noncentral meridian traces part of a sinusoid when plotted against vertical position.
A complete world map has a broad middle and pointed polar ends. The choice of central meridian determines which regions lie near the least-distorted part of the map.
Spherical equations
For a spherical Earth, let:
- be longitude;
- be latitude;
- be the longitude of the central meridian;
- be the sphere’s radius;
- and be projected coordinates.
Angles are measured in radians. With the coordinate origin at the intersection of the equator and central meridian, the forward equations are
For a conventional uninterrupted world map, the longitude difference is normally taken within , placing the map’s cut along the meridian opposite the central meridian.
The inverse equations are
At either pole, , so the inverse longitude formula is undefined. This reflects the convergence of all longitudes at a pole rather than a defect unique to this projection.
False eastings and northings may be added to shift the coordinate origin without changing the projection’s geometry.
Why areas are preserved
The projection’s area property follows from its coordinate derivatives:
Their Jacobian determinant is
A sphere’s surface-area element in longitude and latitude coordinates is also
Thus corresponding infinitesimal areas on the sphere and map are equal before applying an overall display scale. Equal area does not imply preservation of angles, shapes, or arbitrary distances.
Scale and distortion
The projection has true scale along every parallel. At a fixed latitude, an east–west arc on the sphere and its projected horizontal segment have the same length at nominal scale.
Scale is also true along the central meridian. There, meridians and parallels intersect at right angles, and local shape distortion vanishes. Local distortion also vanishes on the equator.
Away from those lines, meridians cross parallels obliquely, producing shear. The scale factor along a meridian is
It increases with distance in longitude from the central meridian and, for a fixed longitude difference, with absolute latitude. Although the scale along parallels remains one, shear stretches and compresses features in other directions while preserving their areas.
Distances and directions
Several distinctions are important:
- Parallel distances are preserved: distances measured along a parallel have their correct scale.
- Vertical separation represents latitude difference: on a sphere, the difference in -coordinates equals the corresponding meridional arc length.
- Curved meridians are generally too long: the length of a noncentral meridian traced on the map exceeds its ground length.
- Arbitrary straight-line distances and bearings are not preserved.
The sinusoidal projection is therefore not a general-purpose distance or navigation map. Its equal-area property also does not make it conformal: small shapes are usually distorted except along the equator and central meridian. This contrasts with the Mercator projection, which preserves local angles but greatly enlarges high-latitude areas.
Historical development
An early documented use of the sinusoidal projection appears in a world map made by Jean Cossin of Dieppe in 1570. It was later associated with the French cartographer Nicolas Sanson and the English astronomer John Flamsteed, giving rise to the name Sanson–Flamsteed projection.
The alternative name Mercator equal-area projection occurs in cartographic literature, but it should not be confused with the familiar cylindrical Mercator projection. The two projections have fundamentally different geometries and distortion properties.
The sinusoidal projection became part of the standard repertoire of cartography because it combines exact area preservation with simple construction and calculation.
Interrupted forms and related projections
An uninterrupted sinusoidal world map strongly deforms regions far from its central meridian. An interrupted sinusoidal projection divides the world into lobes, each with its own central meridian. By placing cuts mainly through oceans, a map can reduce distortion within continental land areas.
Interruption sacrifices continuity. Routes and geographic features crossing a cut must be followed between separate lobes. Ocean-centered arrangements are also possible when marine areas, rather than continents, are the main subject.
The Goode homolosine projection combines sinusoidal portions at lower latitudes with Mollweide portions at higher latitudes. Both components preserve area. The combination retains the sinusoidal projection’s relatively favorable low-latitude geometry while using the Mollweide form to reduce high-latitude deformation. Goode homolosine maps commonly use interruptions as well.
Ellipsoidal form
The sinusoidal principle can also be applied to a reference ellipsoid. An ellipsoidal formulation uses the meridional arc distance for the vertical coordinate and the ellipsoid’s parallel radius for horizontal spacing:
where
is the prime-vertical radius of curvature, is the equatorial semiaxis, is eccentricity, and is meridional arc distance from the equator. This formulation is equal-area on the specified ellipsoid.
Spherical and ellipsoidal implementations are not interchangeable. A dataset described only as “sinusoidal” is incompletely specified: its Earth model, central meridian, units, and coordinate offsets are also needed to establish its coordinate reference system.
Applications and limitations
Thematic mapping
Because it preserves area, the sinusoidal projection is suitable for thematic maps comparing the extent of land-cover classes, ecosystems, or other geographic distributions. It avoids the systematic exaggeration of high-latitude areas found in some non-equal-area projections.
Its main visual limitation is shape distortion near the outer portions of a world map. Long, continuous geographic features may appear strongly bent, making the projection less suitable when familiar shapes or directional relationships are the priority.
Global environmental grids
The projection is used for global gridded Earth-observation products, including many land products from the Moderate Resolution Imaging Spectroradiometer (MODIS). Its simple equations and equal-area geometry support large-scale processing of land-surface observations.
In a regular projected raster grid, cells with fixed width and height have equal projected area and represent equal areas on the projection’s reference surface. This is useful for calculating area totals and area-weighted statistics.
Equal-area cells do not have identical ground shapes, nor does their area necessarily equal actual terrain surface area. Terrain slope, the Earth model, boundary-cell treatment, and data-processing methods remain relevant.
GIS processing
In a geographic information system, sinusoidal data can be transformed into other projections for display or analysis. Reprojection of raster data normally requires resampling. Consequently, exact area preservation by the original mathematical projection does not guarantee that every processing operation will preserve class totals, individual pixel values, or boundaries.