Find the question you need
Why degrees vary in length · Parallel radius · 60° worked example · Great circle · Full index
The same one-degree longitude difference can correspond to 111.2 or 55.6 km along a parallel under a spherical-Earth model. The latitude and path geometry determine the conversion from angle to ground length.
Chapter index
Advanced worked companion: great-circle and mapping scale
Chapters 1–4
1. The identical longitude coordinate gap with a different physical length · 2. Define latitude, longitude and the route being measured · 3. State the spherical model and its limits · 4. Derive a parallel’s effective radius
Chapters 5–8
5. Calculate one degree at the equator · 6. Calculate the matching one degree at sixty degrees north · 7. Compare thirty, forty-five and eighty degrees · 8. The pole exposes why longitude is not a fixed metre axis
Chapters 9–12
9. Do not confuse longitude differences with latitude differences · 10. A constant-latitude parallel is not usually the shortest surface path · 11. Check the great-circle central angle separately · 12. Map projections add another layer of scale
Chapters 13–15
13. Small coordinate differences can be converted locally · 14. A one-degree-by-one-degree grid cell changes physical area · 15. Reader checks and connected Geodesy learning
CHAPTER 1 OF 15
1. The identical longitude coordinate gap with a different physical length
Suppose a map shows two hypothetical point pairs, each separated by exactly one degree of longitude while remaining at one common latitude. One pair sits on the equator; the other at 60 degrees north. It is tempting to multiply both coordinate gaps by one fixed kilometres-per-degree constant.
That is incorrect for east–west distance. Under a spherical Earth with radius 6,371 km, a one-degree arc along the equator is approximately 111.195 km, while the corresponding one-degree arc along the 60° north parallel is approximately 55.597 km.
The two angular differences are identical, but the parallel circles have different radii. The examples are fictional locations on an ideal sphere, not measured transport routes or satellite survey results.
Latitude φ describes angular position north or south of the equatorial plane, and longitude λ describes angular position around Earth’s rotational axis relative to a reference meridian. Both are angles rather than distances.
For the first calculation we deliberately measure east–west arc length along a constant-latitude parallel. This is a different mathematical path from the globally shortest surface geodesic between two points at the same nonzero latitude.
The NOAA longitude explanation notes that meridians converge toward the poles, making one degree of longitude physically shorter at high latitudes.
Approximate Earth as a sphere of radius R=6,371 kilometres. This representative radius is intentionally rounded and not the equatorial or polar radius of a detailed reference ellipsoid.
Assume the point pair has the same latitude and a small difference Δλ=1°, expressed in radians as π/180≈0.0174533 before multiplication by radius. A degree cannot be substituted directly into a formula expecting radians.
Actual terrestrial geodesy uses a reference ellipsoid, datum and often elevation information. Our sphere makes the latitude effect transparent while keeping the numerical results clearly labelled as approximations.
On a sphere, the circle traced by a constant latitude φ has radius R cosφ about Earth’s rotational axis. At φ=0, cosφ=1 and the circle is the equator with full radius R.
At φ=60°, cosφ=0.5, so the radius of that east–west parallel is half the equatorial radius. Near the poles, cosφ approaches zero and the circle of latitude becomes very small.
The length along the parallel through longitude difference Δλ radians is s_parallel=R cosφ |Δλ|. This equation is about the length of a particular surface path, not any route through Earth’s interior.
Set φ=0°, R=6,371 km and Δλ=π/180 radians. The result is 6,371×1×π/180≈111.194927 km.
Round appropriately to approximately 111.2 km for the deliberately approximate Earth radius. This is east–west distance along the equator; the same circle is also a great circle, so the equatorial arc is a shortest surface route for a one-degree separation.
The result is consistent with USGS guidance on degree distances, which explains that longitude distance depends strongly on latitude.
At φ=60°, cos60°=0.5. Thus s_parallel=6,371×0.5×π/180≈55.597463 km.
The result is exactly half the equatorial parallel-arc length within this spherical model. The same longitude-coordinate difference produces about 55.6 km rather than 111.2 km.
These are model distances, not distances measured between specific actual addresses or coastlines. The chosen constant-latitude route must be stated before interpreting them.
At 30° latitude the cosine factor is about 0.8660, so one degree along the parallel spans approximately 96.298 km. At 45° it spans approximately 78.627 km.
At 80° latitude, cos80°≈0.17365, giving only 19.309 km per one-degree longitude difference under the same spherical model.
The resulting distances decrease as absolute latitude increases. North and south latitudes of equal magnitude have the same cosine factor on the sphere.
At a geographic pole, all longitude meridians meet at one point. In the spherical model cos90°=0, so the length of any one-degree east–west parallel arc tends to zero as the latitude approaches 90 degrees.
Longitude values at the exact pole are coordinate conventions around a location where meridians converge. Taking the difference between two longitude labels there does not represent movement across a positive physical distance.
This is a coordinate singularity, not an assertion that the Earth suddenly has a zero circumference. The relevant parallel circle contracts smoothly as latitude approaches the pole.
One degree of latitude represents movement north–south along a meridian. On a perfect sphere its arc length is R×π/180, independent of latitude; in our model that is the same roughly 111.2 km at every location.
One degree of longitude means movement east–west along a parallel whose radius is R cosφ. It is the longitude degree, not the latitude degree, that shrinks by a cosine factor in this simple comparison.
On a reference ellipsoid, even meridional metres per degree of latitude vary slightly with latitude. The spherical distinction is therefore a useful introductory approximation, not centimetre-accurate geodetic guidance.
CHAPTER 10 OF 15
10. A constant-latitude parallel is not usually the shortest surface path
Between two points at nonzero equal latitude, the shortest spherical great-circle arc generally bends away from their parallel rather than following it exactly. A parallel is a circle of radius R cosφ except at the equator, and it is not itself a great circle.
For close points, the difference can be extremely small, but it is a different requested geometry. We should not quote a parallel arc and call it an exact great-circle geodesic without checking.
A geographic distance API may use an ellipsoidal geodesic rather than a simple cosφ parallel approximation, so its output can differ legitimately from our hand calculation.
For points at equal latitude φ and longitude difference Δλ on a sphere, their central angle Δσ satisfies cosΔσ=sin²φ+cos²φ cosΔλ. The shortest great-circle distance is RΔσ for the minor arc.
An equivalent numerically stable expression for this special case is Δσ=2asin[cosφ sin(|Δλ|/2)] when Δλ is the smaller separation and latitude lies in its usual range.
The result agrees with the parallel length at the equator. At 60° for a one-degree gap, the great-circle distance is slightly less than the constant-latitude arc; both round to about 55.6 km at the precision of this article.
A planar map uses a projection to turn curved-surface coordinates into x–y positions. Different projections preserve different properties; no flat map can reproduce all distances, areas, angles and shapes across the whole Earth without distortion.
In a spherical Mercator representation, projected horizontal coordinate x=Rλ increases equally per radian of longitude, independent of latitude. A one-degree longitude difference therefore looks like the same x-coordinate separation everywhere in that projection.
But the true constant-latitude arc length scales with cosφ. At 60°, a Mercator map-distance measure must be interpreted with the appropriate local scale factor rather than mistaken for 111.2 km on the ground.
For a small angular displacement around latitude φ on a sphere, a local approximation is east–west distance dx≈R cosφ dλ and north–south distance dy≈R dφ, with angles expressed in radians.
The local ground distance for small combined changes can be approximated by sqrt(dx²+dy²), ignoring curvature terms of higher order and elevation. This is useful in simple reasoning and unit checking but may be inappropriate for long routes.
To calculate precise property or engineering distances, use suitable geodetic tools, reference systems and the actual path geometry rather than converting longitude degrees with a global fixed multiplier.
A geographic grid defined in angular degrees is not an equal-area grid. Near the equator, one-degree height and width represent approximately 111.2 km each on the sphere, giving around 12,364 km² for a small one-degree cell.
At 60° latitude the east–west dimension is about half as large, so a similarly centred angular cell occupies approximately half that area under the small-cell approximation.
This matters for population density, land-cover totals and environmental data. Equal numbers of latitude–longitude grid cells do not necessarily represent equal land or water areas; area-weighted aggregation may be necessary.
Does one degree longitude always equal 111 km? No. Its physical east–west parallel length decreases with |latitude|. Does one degree latitude behave identically? On the sphere it has a constant meridional arc length, though a real ellipsoid introduces modest latitude dependence.
Practice A: compute the one-degree parallel-arc length at 30° and 80° from s=R cosφ π/180. Practice B: determine the latitude at which that length is one quarter of the equatorial value: solve cosφ=0.25.
Continue with The Gold Standard of Geodesy, NOAA on longitude, USGS on degree distances and the Gold Standard hub. Our point pairs and spherical radius are illustrative.
Advanced worked companion: latitude scaling, geodesics and projection distortion
1. The radius of a parallel is the hidden length scale
Longitude measures angle around Earth’s rotational axis, but an arc length along a latitude circle depends on the circle’s physical radius. For a sphere of radius R, a parallel at latitude φ lies on a circle of radius Rcosφ. One degree around that circle is therefore 2πRcosφ/360.
This geometric expression is equivalent to Rcosφ×π/180. At the equator the cosine factor equals one, while at 60° it equals one-half. The coordinate gap has not changed; the physical circumference represented by a full revolution at that latitude has.
A fixed kilometres-per-degree-longitude shortcut can be useful locally, but it must refer to the appropriate latitude and Earth model. Applying an equatorial conversion globally is an information-loss error, because it silently removes the cosine factor.
2. Calculate one degree in four different latitude bands
For R=6,371 km, the equatorial one-degree parallel length is approximately 111.195 km. At latitude30° it is 96.298 km; at45° it is 78.627 km; at60° it is 55.597 km.
At80° the same calculation gives only 19.309 km. These values show a continuous decline toward the poles rather than an abrupt change at one special latitude. Symmetric north and south values agree on a sphere because cos(−φ)=cosφ.
The values are modelled constant-latitude surface arcs, rounded to a few decimals for arithmetic checking. They are not precise distances between surveyed points with terrain or altitude corrections.
3. Verify the radians step that often breaks distance calculations
One degree equals π/180 radians, approximately0.0174533. The arc-length law s=rθ assumes θ in radians. Multiplying Rcosφ by a bare numerical one, interpreted as a radian instead of a degree, would overstate the desired distance by roughly 180/π≈57.3 times.
At latitude60°, the parallel radius is 6,371×0.5=3,185.5 km. Multiplying by π/180 gives approximately55.597 km, not3,185.5 km. This unit conversion is a necessary part of the derivation, not a formatting detail.
A geospatial worksheet should label whether each angular column is stored in degrees or radians. Many software functions accept radians, and some interfaces convert automatically. Knowing the required convention prevents plausible-looking but incorrect kilometre outputs.
4. Changing the latitude changes the ground length even when Δlongitude stays one
Define the ratio of one-degree parallel lengths at latitude φ and the equator as s(φ)/s(0)=cosφ. At60° it is 0.5; at30° it is about0.8660; at80° it is about0.17365. The global angular longitude gap remains one degree in every row.
If a map report gives two one-degree longitude offsets, an east–west distance comparison is incomplete without the corresponding latitudes. A 1° difference does not by itself specify an absolute travel distance or physical separation.
This is a geometric effect independent of any country’s borders or road network. Administrative map features can make travel paths much longer than a short surface arc, but they are outside this mathematical exercise.
5. Near a pole a longitude label is not a local metre coordinate
As latitude approaches ninety degrees, cosφ approaches zero. Parallel length for any fixed small longitude gap likewise approaches zero in the spherical model. Meridians intersect at the pole, so multiple longitude labels can refer to the same physical endpoint there.
It would be inappropriate to represent one degree of longitude near the pole as the same east–west movement as one degree at the equator. A local coordinate system should account for the geometry and for the fact that the east direction is poorly defined exactly at the singular point.
This is not a claim that movement on Earth disappears near a pole. A north–south meridian arc still has nonzero length away from the exact pole; it is specifically the constant-latitude circle that contracts.
6. The shortest surface distance is not the same as a parallel arc
For two points at the same nonzero latitude, the shortest great-circle connection generally does not follow the constant-latitude parallel. Instead it curves slightly toward the nearer pole over a small longitude interval. The exception includes the equator, which is itself a great circle.
For the spherical model, an exact expression for the minor great-circle separation at equal latitude φ and longitude difference Δλ is δ=2asin(cosφ sin(|Δλ|/2)). The shortest surface distance is Rδ, with angles in radians.
At60° for a one-degree difference, the great-circle length is about 55.597 km, slightly less than the parallel arc 55.597 km. The two round similarly at ordinary teaching precision, but they are conceptually distinct paths.
7. Choose the right distance definition for a question
If a task asks for the length along the constant-latitude parallel, use the parallel-radius arc formula. If it asks for a shortest path over a spherical surface, use the great-circle distance. If it asks for a route through roads, tracks or legal corridors, neither ideal surface formula is a route network calculation.
Real geodesic software frequently uses an ellipsoidal Earth rather than a sphere, so an accurate geodesic output may differ slightly from either spherical result. This does not necessarily indicate that an instrument or calculation is faulty.
A defensible technical statement names the path type, coordinate reference system, distance units and Earth model before making a precision claim.
8. A map projection can make identical angular gaps appear misleading
A spherical Mercator map assigns horizontal coordinate x=Rλ, where λ is in radians. Thus the horizontal distance on the projected map for a one-degree longitude difference is Rπ/180 everywhere, even though the true parallel arc shrinks with cosφ.
At60°, the corresponding local horizontal scale factor is sec60°=2. The projected horizontal increment is twice the physical parallel length for this spherical model. Using map-coordinate differences directly as ground distances without accounting for projection scale therefore overstates the ground length.
Different map projections trade off shape, area, direction and distance fidelity. The USGS discussion of map projections explains why no flat projection preserves all surface properties globally.
9. A one-degree geographic grid is not an equal-area raster
An angular grid cell one degree tall and one degree wide near the equator has approximate dimensions111.2 km by111.2 km under our simple sphere. Its area is therefore about12,364 square kilometres. At60° its east–west width is halved, so an otherwise comparable cell has roughly half the physical area.
For a cell centred at latitude φ, with one-degree span in latitude and longitude, the exact spherical area is R²Δλ[sin(φ+Δφ/2)−sin(φ−Δφ/2)]. This expression factors into a common positive term multiplied by cosφ, showing that a centred cell at60° has exactly half the area of its equatorial counterpart in the spherical model.
This affects environmental maps and geographical statistics. Averaging values per grid cell without accounting for the cells’ different physical areas can misrepresent quantities that should be area-weighted.
10. Reverse the calculation: which latitude gives one quarter of equatorial width?
Set the parallel-length ratio to0.25, so cosφ=0.25. The positive latitude solving this is arccos0.25≈75.5225°. The corresponding southern latitude has equal magnitude and opposite sign.
At this latitude, a one-degree longitude arc is approximately27.8 km in the 6,371 km sphere, compared with111.2 km at the equator. The angular coordinate difference stays one degree while physical spacing has fallen to one quarter.
The inverse calculation demonstrates that latitude can be estimated from a geometric ratio under the assumed sphere. It is a mathematical example, not a practical way to determine an actual geographic coordinate from one rough ground distance.
11. An antimeridian crossing can conceal a tiny angular separation
Suppose one point has longitude179.5° east and another179.5° west at the same latitude. Subtracting signed coordinates naively can produce a difference of359°, even though the smaller angular separation across the antimeridian is just1°.
A correct local-distance calculation chooses the appropriately wrapped longitude difference, often mapped into the interval from−180° to180°, before applying a minor-arc formula. Otherwise the computed path can accidentally travel almost all the way around the parallel.
At the equator, the smaller1° arc is about111.2 km under our sphere, not a 359° circuit exceeding39,000 km. Coordinate conventions and geometry both matter to a meaningful distance report.
12. Latitude–longitude rectangles are not ordinary plane rectangles
A region bounded by two parallels and two meridians has curved edges on the Earth’s surface. Treating its angular side differences as the sides of an ordinary Euclidean metre rectangle can introduce error, particularly over large areas or at high latitudes.
For small areas, local easting and northing can be approximated by dx≈Rcosφ dλ and dy≈R dφ, with angles in radians. Using sqrt(dx²+dy²) then approximates short distances, but higher-order curvature, the exact path and ellipsoid effects become more important over longer routes.
A GIS workflow should choose a suitable coordinate reference system for metric operations, especially when calculating area, buffering geometries or allocating lengths across latitude bands.
13. The real Earth and its datum add uncertainty to a spherical approximation
Earth is better approximated by an oblate ellipsoid than a perfect sphere. Its radius of curvature varies by direction and latitude, so precise metres-per-degree calculations depend on the adopted ellipsoid and the geographic position.
Coordinate data also use reference frames or datums that determine how numerical coordinates relate to the Earth at a time. Combining incompatible coordinate systems can create position discrepancies separate from the cosine-latitude effect.
Our radius6,371 km is a deliberately representative rounded value. Reporting its calculations to a metre or centimetre would overstate model precision even though the arithmetic itself could be carried out to many digits.
14. Why elevation and terrain are not part of these surface arcs
Our examples measure ideal path lengths on a spherical reference surface. Travel over hilly terrain can involve longer physical paths than their projected horizontal separation, and a road or footpath can deviate greatly from a shortest surface connection.
A satellite image, navigational chart or GNSS coordinate pair does not automatically encode every terrain and route constraint. A practical travel-distance answer requires an explicit path model and appropriate data.
The purpose of this article is to distinguish angular coordinate differences from surface distance, not to recommend real routes or claim current travel times.
15. Audit a distance transformation before publishing it
Check that latitude and longitude are in the expected angular units, that the longitude difference has been wrapped appropriately, and that the intended route is a parallel arc, a geodesic or a transport path. Declare the Earth model and the output units.
For geographic grid analysis, verify whether cells represent equal physical area. If not, apply an appropriate area-weighting method where the question concerns physical totals or average values over land or ocean surface.
These checks keep a correct geometrical formula from being applied to the wrong physical question. A beautiful map can still conceal an unlabelled scale distortion or coordinate-convention error.
16. Learning checks and external foundations
Practice A: calculate one-degree parallel length at45° and80° using Rcosφ π/180. Practice B: calculate the ratio of one-degree parallel lengths at60° and30°. Practice C: identify why a 359° numerical difference across the antimeridian may represent a one-degree separation.
The NOAA longitude explanation gives the physical convergence of meridians, and USGS map-distance guidance describes the varying lengths of degrees of longitude. These are authoritative foundations for the spherical illustrations, not measurements of our fictional points.
Read The Gold Standard of Geodesy and the Gold Standard hub for connected questions in measurement, mapping and inference.
| Latitude | cos(latitude) | 1° longitude along parallel |
|---|---|---|
| 0° | 1.00000 | 111.195 km |
| 30° | 0.86603 | 96.298 km |
| 45° | 0.70711 | 78.627 km |
| 60° | 0.50000 | 55.597 km |
| 80° | 0.17365 | 19.309 km |
| 90° | 0.00000 | 0.000 km |