{
  "id": "coordinate-reference-system",
  "title": "Coordinate Reference System",
  "lang": "en",
  "source_lang": "en",
  "summary": "A coordinate reference system connects numerical coordinates to real-world locations through defined axes, units, and a reference datum or frame.",
  "category": "geography",
  "subcategory": "",
  "type": "concept",
  "aliases": [
    "CRS",
    "spatial reference system",
    "spatial referencing system"
  ],
  "generated_at": "2026-11-03",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/coordinate-reference-system/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/coordinate-reference-system/",
      "json": "https://aiwiki.page/zh/coordinate-reference-system.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "A coordinate reference system (CRS) is a framework that makes coordinates interpretable as positions in the real world. It specifies a coordinate system—its axes, their order, directions, and units—and connects that system to an object such as the Earth through a geodetic datum or reference frame. The same numerical pair can identify different places in different CRSs, while a single physical point can have many valid coordinate representations. Coordinates therefore describe a location unambiguously only when their reference system and, where necessary, their time of validity are known.\n\nCoordinate reference systems underpin cartography, surveying, navigation, and the integration of spatial information in a geographic information system (GIS). They allow satellite observations, property boundaries, aerial photographs, engineering designs, and environmental measurements to be compared within a common spatial framework. Their scope extends from local construction grids to global, time-dependent reference frames. Understanding a CRS requires distinguishing mathematical coordinates from their physical reference, map projection from datum, and horizontal position from height."
    },
    {
      "heading": "Components and defining properties",
      "text": "A coordinate system is a mathematical arrangement of axes and rules for expressing position. A CRS adds the information needed to relate those coordinates to a physical object. For a geodetic CRS, this relationship includes the origin, orientation, scale, and other defining properties of its datum or reference frame. For a projected CRS, it also includes a map projection and its parameters. A coordinate pair such as \\(500000,\\ 4500000\\) is consequently incomplete without context: the values might be metres in a national grid, feet in an engineering system, or coordinates in one of many projection zones.\n\nThe coordinate system specifies the number of dimensions and the meaning of each ordinate. A two-dimensional system might use latitude and longitude, or easting and northing. A three-dimensional system might add ellipsoidal height or use three Earth-centred Cartesian coordinates. Axis direction matters as much as axis name: an axis can increase eastward, northward, upward, downward, or along a locally defined direction. Axis order is also part of the definition. A latitude–longitude sequence is not interchangeable with longitude–latitude merely because both contain angular coordinates.\n\nUnits are properties of the individual axes rather than an assumption attached to all coordinates. Angular axes commonly use degrees, while linear axes commonly use metres, but historical datasets may use other units. A projected CRS can use metres for horizontal coordinates while an accompanying vertical CRS uses feet. The international foot is exactly \\(0.3048\\) metres; the former U.S. survey foot is exactly \\(1200/3937\\) metres. Their approximately two-parts-per-million difference becomes significant when coordinate magnitudes reach hundreds of thousands of units.\n\nA complete CRS definition is distinct from metadata describing where and how it should be used. An identifier, name, area of use, scope, and descriptive remarks help select and interpret a CRS, but do not replace its defining parameters. An identifier such as EPSG:4326 points to a registered definition; it is not itself a mathematical formula. Equally, a CRS does not determine the positional accuracy of every dataset expressed in it. Survey observations, image processing, transformations, and digitization each contribute their own errors.\n\nCoordinates also need an explicit relationship to the features they describe. A correctly defined CRS does not reveal whether a coordinate marks a building entrance, roof centre, parcel corner, or image-pixel centre. The broader process of georeferencing establishes the spatial relationship between data and a reference framework. CRS metadata supplies the language of position, while feature definitions and measurement procedures establish what has actually been positioned."
    },
    {
      "heading": "Geographic, geocentric, and engineering systems",
      "text": "A geographic CRS represents position using latitude and longitude, usually with respect to a reference ellipsoid. The ellipsoid is a smooth mathematical approximation to the Earth, defined by dimensions such as its semimajor axis \\(a\\) and flattening \\(f\\). Geodetic latitude is the angle between the equatorial plane and the normal to the ellipsoid at a point. It differs from geocentric latitude, which is measured using a line from the Earth's centre. Longitude is measured relative to a specified prime meridian, commonly Greenwich, although historical systems have used other origins.\n\nA two-dimensional geographic CRS contains latitude and longitude but no height. A three-dimensional geographic CRS adds ellipsoidal height \\(h\\), measured along the ellipsoidal normal. This height is geometrical rather than a direct measure of elevation above mean sea level. The WGS 84 ellipsoid, for example, has a semimajor axis of \\(6\\,378\\,137\\) metres and inverse flattening of \\(298.257223563\\). These dimensions define a surface; they do not, by themselves, specify the full physical realization of a global reference frame.\n\nA geocentric CRS instead uses three Cartesian coordinates \\(X,Y,Z\\), with an origin at or close to the Earth's centre of mass. Conventionally, the \\(Z\\)-axis follows the terrestrial reference pole, the \\(X\\)-axis lies in the equatorial plane toward the reference meridian, and the \\(Y\\)-axis completes a right-handed system. Geocentric coordinates are especially useful in satellite orbit determination and global geodesy because they avoid the angular singularities of geographic coordinates near the poles.\n\nFor a point expressed by geodetic latitude \\(\\varphi\\), longitude \\(\\lambda\\), and ellipsoidal height \\(h\\), the corresponding Cartesian coordinates are\n\n\\[\\begin{aligned} X&=(\\nu+h)\\cos\\varphi\\cos\\lambda,\\\\ Y&=(\\nu+h)\\cos\\varphi\\sin\\lambda,\\\\ Z&=\\bigl[\\nu(1-e^2)+h\\bigr]\\sin\\varphi, \\end{aligned}\\]\n\nwhere \\(e\\) is the ellipsoid's first eccentricity and\n\n\\[\\nu=\\frac{a}{\\sqrt{1-e^2\\sin^2\\varphi}}.\\]\n\nThese equations change coordinate representation on the same ellipsoid and datum. They do not constitute a datum transformation.\n\nAn engineering CRS serves a local purpose, such as positioning within a factory, tunnel, construction site, or vessel. Its origin and axes may be fixed to a convenient structure rather than directly to a global geodetic frame. A site grid may be designed so that coordinate distances closely match distances measured at ground level. Such a system can be rigorous and highly accurate within its intended area, even when its relationship to a national or global CRS requires a separately surveyed transformation."
    },
    {
      "heading": "Projected systems and map distortion",
      "text": "A projected CRS expresses positions on a curved reference surface as coordinates on a plane. It combines a base geographic CRS with a map projection, projection parameters, and a Cartesian coordinate system. Typical parameters include a central meridian, latitude of origin, scale factor, standard parallels, and false easting or northing. The false coordinates are conventional offsets added to avoid inconvenient negative values or to establish a useful grid origin. They do not represent an actual displacement of the Earth's surface.\n\nNo projection can preserve every geometrical property of a curved surface. Conformal projections preserve local angles, equal-area projections preserve area, and equidistant constructions preserve particular distances or scales along selected lines. These properties are not interchangeable. The Mercator projection is conformal, but its scale increases strongly toward the poles. A dataset expressed in metres in a Mercator-based CRS therefore does not automatically provide accurate ground distances or areas when ordinary planar calculations are applied.\n\nThe Universal Transverse Mercator system divides most of the world into 60 zones, nominally \\(6^\\circ\\) of longitude wide. Each zone uses a transverse Mercator projection with a central-meridian scale factor of \\(0.9996\\) and a false easting of \\(500\\,000\\) metres. Southern-hemisphere coordinates use a false northing of \\(10\\,000\\,000\\) metres at the equator. The conventional coverage extends from \\(80^\\circ\\) south to \\(84^\\circ\\) north, with special zone boundaries in parts of Norway and Svalbard. A UTM coordinate is incomplete without its zone, hemisphere, and underlying datum.\n\nNational systems often adopt projection designs suited to their territory. Lambert conformal conic projections are widely used for regions with a substantial east–west extent, while transverse Mercator projections suit narrower north–south zones. The British National Grid uses a transverse Mercator projection associated with OSGB36. In the United States, State Plane systems use several projection families and multiple zones. A familiar grid name alone may still conceal different datum realizations or unit conventions.\n\nWeb mapping commonly uses WGS 84 / Pseudo-Mercator, registered as EPSG:3857. Its projection equations use a spherical development of WGS 84 geographic coordinates, allowing convenient square map tiles and a consistent zoom hierarchy. Practical global tile schemes stop near \\(85.0511^\\circ\\) latitude because the Mercator projection becomes unbounded at the poles. This CRS is effective for interactive display but is not an equal-area or uniformly true-distance framework. Analytical measurements commonly require an appropriate local projection or calculations directly on an ellipsoid."
    },
    {
      "heading": "Reference frames, realizations, and coordinate epochs",
      "text": "A datum supplies the reference relationship between coordinates and the Earth, while a reference-frame realization gives that relationship operational form through observations, station coordinates, velocities, and conventions. In historical surveying, a datum might be anchored to an initial station and oriented through astronomical observations. Its ellipsoid could be fitted to the regional geoid rather than centred on the Earth's mass. Such a datum can serve its original region well while differing substantially from a modern geocentric frame.\n\nModern global frames draw on global navigation satellite systems (GNSS), very long baseline interferometry, satellite laser ranging, and Doppler orbitography. These techniques contribute different strengths: satellite laser ranging strongly constrains the geocentre, while very long baseline interferometry helps establish scale and Earth orientation. The International Terrestrial Reference Frame combines space-geodetic observations into successive realizations. Names such as ITRF2014 and ITRF2020 distinguish solutions, not simply dates at which all coordinates were measured.\n\nThe Earth's crust is not stationary. Plate tectonics, earthquakes, volcanic deformation, land subsidence, and glacial isostatic adjustment change station positions. A point fixed in a tectonic plate can move several centimetres per year in a global frame. At survey accuracy, coordinates without time information may therefore be incomplete. A coordinate epoch identifies the date at which coordinates describe a point's position. This differs from a frame reference epoch, which is a defining reference time for the frame or its associated model.\n\nA simple motion model represents position as\n\n\\[\\mathbf{x}(t)=\\mathbf{x}(t_0)+\\mathbf{v}(t-t_0),\\]\n\nwhere \\(\\mathbf{v}\\) is velocity. This linear approximation works for steady motion but does not describe every process. Earthquakes can cause abrupt coordinate discontinuities, and postseismic deformation can be nonlinear. Some reference-frame solutions consequently include discontinuity handling and additional deformation terms. A dynamic CRS requires time-dependent interpretation; the epoch normally accompanies the coordinates rather than being embedded in every individual CRS definition.\n\nPlate-fixed regional frames reduce coordinate changes for many mapping applications. The European Terrestrial Reference System 1989 is tied to the stable part of the Eurasian plate, while the Geocentric Datum of Australia 2020 provides a national framework referenced to epoch 2020.0. Their relationship to global frames evolves as plates move. WGS 84 likewise has successive realizations, so the name “WGS 84” alone is not a sufficient specification for all high-precision work. EPSG's WGS 84 datum ensemble accommodates several realizations for uses that do not require distinguishing among them; it does not make those realizations identical."
    },
    {
      "heading": "Vertical references and compound systems",
      "text": "Height requires a reference separate from the horizontal position, and different height definitions answer different physical questions. Ellipsoidal height measures geometrical separation from a reference ellipsoid. Gravity-related heights describe position using the Earth's gravity field and an adopted vertical reference surface. Their distinction matters because water flow, engineering levelling, flood modelling, and satellite positioning do not all use the same height quantity.\n\nThe geoid is a selected equipotential surface of the Earth's gravity field that approximately corresponds to global mean sea level. Orthometric height \\(H\\) is measured along the plumb line from the geoid to a point. Ellipsoidal height \\(h\\), orthometric height, and geoid undulation \\(N\\) are commonly related by\n\n\\[h \\approx H+N.\\]\n\nThe expression is widely useful, although rigorous treatment must account for the directions and definitions of the quantities involved. Normal heights use a different gravity-based construction and are related to ellipsoidal heights through a quasigeoid rather than a geoid. These height types should not be treated as synonyms.\n\nTraditional vertical datums often originated from tide-gauge observations combined with precise levelling. Mean sea level at one coastal station is not necessarily the same equipotential reference as mean sea level at another, because ocean circulation, atmospheric conditions, and other processes produce spatial differences. Levelling networks also accumulate observational error and can be affected by crustal movement. A national vertical datum is consequently a defined measurement framework, not simply an exact and universal “sea-level zero.”\n\nA vertical CRS specifies the vertical reference and a one-dimensional coordinate system, including the axis direction and unit. Height commonly increases upward, whereas depth commonly increases downward. Hydrographic chart datums, often based on specified low-water references, serve navigation and differ from many land-height datums. Tidal conventions also matter in precise geodesy because permanent tidal effects can be treated differently in coordinate and gravity-field products.\n\nA compound CRS combines independent component CRSs, such as a two-dimensional projected CRS and a gravity-related vertical CRS. A three-dimensional geographic CRS is not equivalent: its third coordinate is normally ellipsoidal height. Software that merely appends a height column to projected coordinates has not necessarily established a valid vertical reference. Converting GNSS-derived heights into a national height system generally requires an appropriate geoid or quasigeoid model and attention to the horizontal reference in which that model is defined. Digital elevation models therefore need vertical as well as horizontal metadata, especially when used for inundation analysis or comparison with surveyed infrastructure."
    },
    {
      "heading": "Coordinate conversions and transformations",
      "text": "A coordinate conversion changes coordinates without changing their underlying datum. Examples include converting geographic coordinates to geocentric coordinates, applying a map projection, changing axis order, or converting angular units. A coordinate transformation relates different datums or reference frames and commonly depends on empirically determined parameters. The distinction identifies whether the operation is principally a defined mathematical change of representation or also incorporates measured relationships between physical reference frameworks.\n\nA common three-dimensional transformation is the Helmert transformation. Its seven-parameter form uses three translations, three rotations, and one scale difference. Schematically,\n\n\\[\\mathbf{X}_{B} = \\mathbf{T}+(1+s)\\mathbf{R}\\mathbf{X}_{A},\\]\n\nwhere \\(\\mathbf{T}\\) is the translation vector, \\(s\\) is the scale difference, and \\(\\mathbf{R}\\) is a rotation matrix. Different rotation conventions exist, particularly the position-vector and coordinate-frame conventions. Applying parameters with the wrong convention can produce errors even when the numerical parameter values appear correct.\n\nA fourteen-parameter formulation adds rates of change for the seven parameters. It can describe the evolving relationship between reference frames, but it should not be confused with a complete model of individual point motion. Transforming coordinates between dynamic systems may require both a frame transformation and propagation between coordinate epochs using velocities or deformation models. The order of these operations and the epochs at which parameters apply must be explicit.\n\nRegional datum differences often contain spatially varying distortions inherited from historical survey networks. A single Helmert transformation cannot fully represent these patterns. Grid-based transformations provide position-dependent corrections, commonly as shifts in latitude and longitude or in projected coordinates. OSTN15, for example, supports transformation between ETRS89-derived positions and the OSGB36 National Grid in Great Britain. Its purpose is not merely to reproduce the map projection but also to accommodate the relationship between a modern geodetic reference and an established national mapping framework.\n\nA complete operation may therefore be a pipeline: inverse-project source coordinates, change units or axis order, transform the datum, apply a vertical correction, and project into the target CRS. Multiple operations can be valid for the same source and target systems, with different areas of use and expected accuracies. A high-quality local grid may outperform a global approximation within its coverage, but be unusable outside it. Missing grid files, incorrect interpolation, or silent substitution of a lower-accuracy operation can change results. Some inverse grid operations are solved iteratively, so simply reversing correction signs is not always an adequate inverse transformation."
    },
    {
      "heading": "Standards, identifiers, and machine-readable definitions",
      "text": "International standards provide a shared conceptual model for describing CRSs and coordinate operations. ISO 19111, Geographic information—Referencing by coordinates, distinguishes coordinate systems, datums, reference frames, coordinate reference systems, and operations between them. This separation allows software and data producers to describe the same spatial concepts consistently. The Open Geospatial Consortium develops related interoperability standards used in geographic services, files, and applications.\n\nThe EPSG Geodetic Parameter Dataset, maintained by the International Association of Oil & Gas Producers, registers CRSs, datums, ellipsoids, coordinate operations, and associated parameters. EPSG:4326 identifies a two-dimensional WGS 84 geographic CRS; EPSG:4979 identifies its three-dimensional geographic counterpart; EPSG:4978 identifies a WGS 84 geocentric CRS. These are different coordinate representations and should not be substituted without the appropriate conversion. EPSG:3857 identifies the widely used projected web-mapping CRS.\n\nAn authority code is a reference to a definition, not a complete embedded description. Systems may instead exchange Well-Known Text (WKT), a structured textual representation containing the CRS's components and parameters. The WKT representation standardized in ISO 19162 can express modern concepts such as dynamic reference frames and datum ensembles. Earlier WKT dialects and older software may not preserve all these distinctions. Consequently, two files that both contain a “WKT definition” can differ in how completely they record the intended reference framework.\n\nAxis order is a recurring interoperability issue. In the EPSG definition of EPSG:4326, latitude precedes longitude. Many graphics libraries and older GIS interfaces traditionally use \\(x,y\\), interpreted as longitude followed by latitude. OGC CRS84 uses longitude–latitude order, making it a distinct CRS specification despite sharing the familiar WGS 84 geographic setting. Protocol versions can also change expectations: Web Map Service 1.3.0 generally follows the axis order defined by the requested CRS. Reversing coordinates without examining the applicable convention can move data to a different continent or generate invalid latitudes.\n\nFormats preserve CRS information in different ways. GeoTIFF uses standardized keys for geospatial referencing; GeoPackage stores spatial reference definitions in database tables; shapefiles commonly rely on a separate projection file. Such sidecar files can be lost when datasets are copied. GeoJSON as standardized in RFC 7946 uses longitude–latitude coordinates based on WGS 84 and does not provide the former general-purpose mechanism for declaring arbitrary alternative CRSs. File format, identifier, axis convention, and software interpretation must therefore be considered together rather than assuming that every coordinate-bearing file communicates the same metadata."
    },
    {
      "heading": "Data integration, measurement, and uncertainty",
      "text": "CRS handling involves a fundamental distinction between assigning a reference system and transforming coordinates. Assignment states what existing numbers mean without changing them. Transformation calculates new numbers representing the same positions in another reference system. If a dataset already contains projected metre coordinates, assigning a geographic CRS does not turn those values into longitude and latitude. Conversely, transforming data from a wrongly assigned source CRS usually produces plausible-looking but incorrect results.\n\nDisplaying multiple layers together adds another level of complexity. GIS software can reproject layers dynamically into a display CRS without altering their stored coordinates. Visual agreement is useful but does not prove that underlying metadata or transformations are correct. At small map scales, displacements of several metres may be invisible. Base maps can also have their own georeferencing errors. High-accuracy comparisons require independent control, appropriate coordinate epochs, and a documented operation rather than alignment judged solely on a screen.\n\nMeasurement must match the geometry represented by the CRS. A Euclidean distance between projected points is a grid distance, not automatically the corresponding ground distance. Projection scale varies across the map, and elevation produces a further difference between distances on the reference surface and those at ground level. Likewise, an ordinary polygon-area calculation on longitude–latitude coordinates produces angular-coordinate area, not physical area in square metres. Spatial analysis can instead use suitable projected systems, ellipsoidal geodesic methods, or algorithms explicitly designed for spherical or ellipsoidal geometry.\n\nReprojection affects data models differently. Vector vertices can be transformed directly, but straight segments in one CRS may become curved in another, so additional vertices may be needed to preserve shape. Reprojecting raster data requires a new sampling grid and usually resampling. Nearest-neighbour methods preserve discrete class labels, while bilinear or cubic methods interpolate continuous values. These operations can change cell values, boundaries, and apparent resolution even when the coordinate transformation itself is accurate. Repeated reprojection and resampling can therefore introduce effects unrelated to datum accuracy.\n\nCoordinate uncertainty should be distinguished from the number of displayed decimal places. Six decimal places in latitude suggest a numerical increment of roughly a tenth of a metre, but they do not establish survey-quality accuracy. A rigorous uncertainty model can propagate an input covariance matrix through a transformation using\n\n\\[\\mathbf{C}_{\\mathrm{out}} \\approx \\mathbf{J}\\mathbf{C}_{\\mathrm{in}}\\mathbf{J}^{\\mathsf T} + \\mathbf{C}_{\\mathrm{op}},\\]\n\nwhere \\(\\mathbf{J}\\) is the Jacobian of the operation and \\(\\mathbf{C}_{\\mathrm{op}}\\) represents operation uncertainty under the adopted assumptions. Correlations and systematic errors require additional treatment. For practical data exchange, source accuracy, transformation accuracy, epoch, processing history, and vertical reference are often as consequential as the CRS identifier."
    },
    {
      "heading": "Historical development and specialized applications",
      "text": "Modern CRSs developed from efforts to connect astronomical observations, triangulation networks, and mathematical models of the Earth. Friedrich Wilhelm Bessel's ellipsoid of 1841, Alexander Ross Clarke's ellipsoids of the nineteenth century, and John Fillmore Hayford's 1909 ellipsoid reflect successive attempts to estimate the Earth's figure from available measurements. These surfaces supported many regional and national datums. Improved ellipsoid dimensions did not by themselves resolve differences in network origin, orientation, or survey distortion, so historical systems remained geographically diverse.\n\nThe growth of satellite geodesy changed the scale and physical basis of reference systems. Satellite observations made it possible to establish globally consistent, Earth-centred frames rather than connecting continents mainly through terrestrial triangulation and astronomical observations. WGS 84 became a major framework for navigation and global positioning, while the Geodetic Reference System 1980 supplied an influential ellipsoid used by numerous modern systems. These developments did not eliminate national grids: countries retained them for continuity in mapping, land records, engineering, and administration, while improving their connection to global reference frames.\n\nDifferent applications emphasize different aspects of a CRS. Cadastral mapping must maintain the relationship between coordinates, monuments, and legally defined boundaries; a new datum does not itself redefine the legal extent of a parcel. Tunnel construction and bridge engineering require control of local scale, orientation, and height. Aviation and marine navigation depend on consistent global positioning but also require specialized vertical references and operational conventions. Environmental analyses, including comparisons of sea-level rise with coastal terrain, can be highly sensitive to seemingly small mismatches between height systems.\n\nPrecision geodesy increasingly treats coordinates as time-dependent observations rather than permanently fixed labels. Continuous GNSS stations reveal seasonal ground motion, tectonic velocities, and deformation caused by changes in water storage or earthquakes. Interferometric synthetic-aperture radar measures surface displacement within its own observational geometry and must be related carefully to terrestrial reference frames. Changes once considered negligible in general mapping become central when monitoring millimetre-to-centimetre motion over years.\n\nCRSs also extend beyond terrestrial mapping. Planetary cartography uses reference surfaces, prime meridians, and body-fixed frames for the Moon, Mars, and other bodies. Planetographic and planetocentric latitude can differ, and longitude conventions must be stated explicitly. Within buildings, mines, and mobile platforms, engineering systems may move relative to an Earth-fixed frame. These applications retain the same essential separation between a coordinate representation, the reference to which it is attached, and the operations that connect it to other systems."
    }
  ],
  "links": [
    "earth",
    "cartography",
    "geographic-information-system",
    "georeferencing",
    "latitude-and-longitude",
    "prime-meridian",
    "map-projection",
    "mercator-projection",
    "plate-tectonics",
    "land-subsidence",
    "glacial-isostatic-adjustment",
    "digital-elevation-model",
    "united-kingdom",
    "open-geospatial-consortium",
    "epsg-geodetic-parameter-dataset",
    "well-known-text",
    "web-map-service",
    "geopackage",
    "map-scale",
    "spatial-analysis",
    "raster-data",
    "sea-level-rise"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}