Geodesy is the science of measuring and representing the geometry, gravitational field, and orientation of Earth, together with their changes through time. It establishes positions and heights, determines the size and shape of the planet, and connects observations made on the ground, at sea, and in space. Its central objects include the reference ellipsoid, a mathematical approximation to Earth’s flattened shape; the geoid, a selected surface of equal gravity potential; and reference systems that allow coordinates measured at different places and times to be compared consistently.
Modern geodesy treats Earth as a deformable, rotating body rather than a fixed surveying surface. The motion of tectonic plates, redistribution of water, loss of ice, atmospheric loading, and variations in rotation all affect the coordinates of observation stations. Conversely, precise measurements of these effects reveal processes within the solid Earth and its fluid envelopes. Geodesy therefore connects metrology, astronomy, geology, and Earth-system science.
Geodetic infrastructure supports navigation, mapping, engineering, satellite operations, and the monitoring of sea-level rise. Its measurements range from angles and distances between nearby monuments to intercontinental radio observations and satellite tracking. Their interpretation requires common spatial and temporal references, physical models of signal propagation and gravitation, and explicit treatment of measurement uncertainty.
Historical development
Attempts to determine Earth’s dimensions preceded modern instruments by many centuries. In the third century BCE, Eratosthenes estimated its circumference by comparing the Sun’s elevation at Alexandria with the reported near-zenith illumination at Syene. His reasoning joined an angular difference to an estimated surface distance. The surviving account gives a circumference of about 250,000 stadia, but uncertainty over the length of the stadion and the underlying distance measurement prevents a uniquely defensible conversion into modern units. The lasting achievement was the geometric method, not a securely known percentage accuracy.
Islamic scholars developed astronomical and geometric approaches to geographical position and Earth measurement. In the eleventh century, al-Biruni described determining Earth’s radius from the height of a mountain and the depression angle of the horizon. In early modern Europe, triangulation made it possible to extend a measured baseline across large regions through connected angular observations. Willebrord Snell’s early-seventeenth-century work in the Netherlands and Jean Picard’s meridian measurements in France exemplified the transition from isolated estimates to systematic geodetic networks.
The figure of Earth became a major test of mechanics. Isaac Newton argued that a rotating, self-gravitating Earth should be flattened at the poles. Eighteenth-century expeditions sponsored by the French Academy of Sciences measured meridian arcs near the equator in present-day Ecuador and at high latitude in Lapland. Their results supported an oblate rather than a prolate figure: a degree of latitude is longer near the poles because the meridian’s radius of curvature is greater there. Meridian-arc measurements also contributed to the original definition of the metre in the 1790s.
During the nineteenth century, national triangulations, precise leveling, astronomical positioning, and pendulum observations established extensive frameworks for maps and gravity studies. Carl Friedrich Gauss contributed methods of adjustment and differential geometry, while Friedrich Wilhelm Bessel and Friedrich Robert Helmert developed influential treatments of Earth’s figure and geodetic theory. International cooperation expanded through the Central European Arc Measurement initiative of 1862, an institutional predecessor of the International Association of Geodesy.
Artificial satellites transformed the field after 1957. Satellite orbits connected previously separate continental networks and exposed the large-scale structure of Earth’s gravity field. Laser ranging, radio interferometry, satellite navigation, and dedicated gravity missions subsequently made geodesy global and time-dependent. The historical progression was therefore not simply toward more accurate maps: it changed the object of measurement from a largely static surface to an interacting planetary system.
Earth’s figure and coordinate geometry
Earth has no single surface that serves every geodetic purpose. The physical surface includes mountains, seabeds, buildings, and vegetation. A reference ellipsoid is a smooth mathematical surface convenient for coordinate calculations. The geoid is defined physically through gravity potential and does not generally coincide with either topography or an ellipsoid. Confusing these surfaces produces errors in interpreting both positions and elevations.
A rotational reference ellipsoid is specified geometrically by its equatorial semimajor axis and polar semiminor axis , or by and the flattening
The WGS 84 ellipsoid has metres and inverse flattening . Its equatorial radius exceeds its polar radius by about 21.4 kilometres. These values describe an adopted reference surface, not a claim that every equatorial or polar point lies at those distances from Earth’s center.
Geodetic latitude is the angle between the ellipsoid’s outward normal and its equatorial plane. It differs from geocentric latitude, which uses the line from Earth’s center to the point. Longitude specifies angular position around the rotation axis, while ellipsoidal height is measured along the ellipsoidal normal. With eccentricity squared , the prime-vertical radius of curvature is
The corresponding Earth-centered Cartesian coordinates are
Cartesian coordinates are particularly useful for satellite orbit calculations and network adjustment. Latitude, longitude, and height are more convenient for many geographical applications, but their meaning depends on the adopted reference system.
Distances on an ellipsoid are commonly defined along geodesics, curves that locally minimize surface distance. Except in special cases such as meridians and the equator, these are not simple planar circles. Solving the direct geodesic problem determines an endpoint from an initial point, direction, and distance; solving the inverse problem determines distance and azimuth from two endpoints.
Local surveying often uses east, north, and up coordinates constructed at a chosen origin. Such coordinates approximate a tangent geometry over limited areas, but curvature matters as the survey expands. A map projection introduces another transformation, from the curved reference surface to a plane. Projection scale, elevation above the ellipsoid, and grid convergence must therefore be distinguished from actual ground distance and direction.
Reference systems, frames, and coordinate epochs
A terrestrial reference system specifies the conceptual rules for assigning coordinates to Earth. A reference frame is a practical realization of those rules through station positions, velocities, and associated models. This distinction resembles the difference between defining a measurement standard and realizing it through instruments. Coordinates are meaningful only when their frame, units, and relevant epoch are known.
The International Terrestrial Reference System is geocentric and rotates with Earth. Its origin is intended to coincide with the center of mass of the whole Earth system, including the oceans and atmosphere. Its scale is consistent with relativistic modeling conventions, and its orientation is maintained through agreed constraints. Successive realizations, called International Terrestrial Reference Frames, combine observations from several space-geodetic techniques. ITRF2020, for example, estimates station coordinates and motions from long observational histories rather than defining a new physical Earth.
A station position is often approximated by
where is its position at reference epoch and is velocity. Millimeter- and centimeter-level applications may additionally require seasonal terms, earthquake offsets, postseismic relaxation, and other nonlinear displacements. A coordinate without an epoch can therefore be ambiguous: a point moving 30 millimeters per year travels 0.6 metre in twenty years even if the monument itself remains intact.
Regional and national frames serve different operational needs. Some are approximately fixed to a tectonic plate, reducing the apparent motion of ordinary mapping coordinates within that region. Global frames instead represent motions relative to the Earth system as a whole. Transformations between frames commonly use three translations, three rotations, and a scale factor; time-dependent transformations also include their rates. A single rigid transformation cannot fully represent spatially variable deformation or distortions inherited from older surveys.
Older geodetic datums were often established by fixing the position and orientation of a local astronomical station and fitting an ellipsoid to regional observations. Their origins need not coincide with Earth’s center. Consequently, coordinates expressed in a historical datum can differ from modern geocentric coordinates by hundreds of metres without either set being internally meaningless.
Stable reference frames require both good observations and suitable station distribution. Sparse coverage in oceans and parts of the Southern Hemisphere weakens global geometry. Colocated instruments help connect techniques, but the surveyed vectors between their reference points must also be accurate. Frame maintenance is thus an ongoing observational activity, not merely the publication of a coordinate table.
Terrestrial surveying and astronomical observations
Classical terrestrial geodesy measures angles, distances, and height differences through networks of intervisible stations. Triangulation determines a network primarily from measured angles and one or more accurately measured baselines. Trilateration relies primarily on distances, while traversing connects a sequence of stations through measured directions and lengths. Modern total stations combine electronic angle measurement with electronic distance measurement, allowing the same observations to support engineering surveys and larger geodetic networks.
Electronic distance measurement uses the travel time or phase shift of electromagnetic radiation. If a pulse travels to a reflector and back, the geometric distance is approximately half the round-trip travel time multiplied by the propagation speed. The speed in air differs from the vacuum speed of light, so instrument reductions require atmospheric information. Temperature, pressure, humidity, and, for some methods, atmospheric composition affect the correction. Poorly represented atmospheric gradients can introduce systematic errors even when an instrument reports highly repeatable distances.
Angles are also affected by the propagation environment. Vertical refraction bends light paths through layers of differing refractive index. Near-ground observations can be strongly influenced by heating and turbulence, particularly over long lines. Instrument centering, target centering, axis misalignment, and monument motion contribute additional errors. Repeated observations in different instrument orientations help separate some instrumental biases, but repetition does not eliminate an unmodeled environmental effect.
Astronomical geodesy determines position and direction by observing celestial objects together with precise time. Astronomical latitude and longitude refer to the local direction of the plumb line rather than to the normal of a chosen ellipsoid. Their differences from geodetic latitude and longitude reveal the deflection of the vertical, which reflects variations in the surrounding mass distribution. Mountain ranges and subsurface density contrasts can therefore affect astronomical positioning without changing the geometry of the ellipsoid.
Precise differential leveling transfers height through approximately horizontal lines of sight to graduated staffs. Balanced sight lengths reduce several systematic effects, while repeated forward and backward runs provide checks on consistency. Long leveling routes require corrections for staff calibration, temperature, refraction, and gravity-related effects. They can also accumulate distortions when observations collected over decades are treated as if the ground had not moved.
Terrestrial observations remain essential despite satellite techniques. They connect instruments at fundamental observatories, transfer coordinates into tunnels and obstructed environments, and resolve local structural deformation. Their strongest role is often complementary: satellite methods supply global position and orientation, while local surveys establish detailed geometry between nearby objects that satellites cannot directly distinguish.
Satellite navigation and precise positioning
Global navigation satellite systems—including GPS, GLONASS, Galileo, and BeiDou—determine positions from signals transmitted by orbiting satellites. A receiver compares transmission and reception times to estimate apparent signal ranges. Because its clock is not perfectly synchronized with system time, an ordinary three-dimensional solution must estimate at least three position components and a receiver clock offset. Four satellites are therefore a basic minimum under simplified conditions, although practical processing generally uses many more observations.
A pseudorange observation can be represented schematically as
where is geometric range, is the speed of light, and are receiver and satellite clock offsets, is ionospheric delay, is neutral-atmosphere delay, collects instrumental biases, and represents remaining errors. Satellite orbit errors, reflected signals, antenna characteristics, and signal obstruction also affect the solution.
Geodetic receivers additionally measure carrier phase, whose precision is much finer than that of ordinary code measurements. The difficulty is that the receiver initially does not know the whole number of wavelengths between satellite and antenna. Estimating or fixing these integer ambiguities is central to high-precision positioning. Loss of tracking can introduce a cycle slip, requiring detection and renewed estimation.
Dual-frequency observations exploit the frequency dependence of ionospheric propagation to remove or estimate its dominant contribution. Neutral-atmosphere delay is largely nondispersive at navigation frequencies and is commonly divided into a hydrostatic component and a more variable wet component. Antenna phase-center offsets and variations must be calibrated because the electrical point measured by the receiver is not identical to the physical antenna reference point.
Relative positioning uses simultaneous observations at multiple receivers to suppress shared errors. Real-time kinematic methods can provide centimeter-level relative positions when ambiguities are resolved and environmental conditions are suitable. Precise point positioning instead uses accurate satellite orbit, clock, and bias products to estimate a receiver’s position without a nearby local reference receiver. Its convergence and performance depend on the observations and correction products available.
There is no single accuracy intrinsic to satellite navigation. A navigation receiver, a continuously operating geodetic station, and a moving survey antenna solve different measurement problems. Millimeter-scale long-term station velocities emerge from sustained observations, stable installations, consistent processing, and reference-frame control; they should not be confused with the accuracy of an isolated instantaneous position. Relativistic clock and propagation corrections are also indispensable parts of the system’s operation.
Space-geodetic techniques beyond satellite navigation
Several independent space-geodetic techniques provide information that satellite navigation alone cannot supply with equal strength. Their combination is important because each observes different physical quantities and carries different systematic errors. Very long baseline interferometry measures the difference in arrival time of radio signals from distant celestial sources at widely separated antennas. Most geodetic targets are quasars whose directions define an effectively external reference for terrestrial observations.
The fundamental observable in this method is a delay, not an ordinary angular reading. It depends on the baseline between antennas, source direction, station clocks, atmospheric propagation, and Earth’s orientation. Because the observations refer directly to distant radio sources, very long baseline interferometry is especially important for determining the relation between terrestrial and celestial frames. It also provides a direct observational basis for UT1, the time measure tied to Earth’s rotation.
Satellite laser ranging measures the round-trip flight time of short laser pulses reflected by satellites carrying retroreflectors. The LAGEOS satellites, launched in 1976 and 1992, are dense, passive targets designed for stable geodetic tracking. Laser ranging contributes strongly to the determination of Earth’s center of mass and the scale of terrestrial reference frames. Its accuracy depends on timing calibration, atmospheric corrections, station geometry, and models of the satellite’s optical response.
Lunar laser ranging uses reflectors placed on the Moon by Apollo missions and Soviet Lunokhod rovers. It measures the changing Earth–Moon distance and supports research on lunar dynamics, tidal evolution, and gravitation. Converting a photon travel time into a physical distance requires detailed modeling of observatory motion, lunar orientation, atmospheric delay, and relativistic light propagation. The observable is therefore embedded in a dynamical model rather than interpreted as a simple ruler reading.
DORIS, developed through French institutions, uses Doppler measurements of signals from a global network of ground beacons received by satellites. It supports precise orbit determination and contributes to terrestrial frame realization. Its observation geometry and instrument characteristics differ from those of navigation satellite systems, making it valuable in combined solutions.
At fundamental geodetic observatories, several techniques operate together. Local surveys connect the physical reference points of antennas, telescopes, and ranging instruments. These “local ties” are essential: two instruments at the same observatory do not automatically observe the same coordinate point. Disagreement among techniques can expose errors in antenna models, loading corrections, timing, or local surveys. The value of a multi-technique network lies partly in this capacity for independent verification, not simply in accumulating more measurements.
Earth rotation and celestial orientation
Earth’s rotation is irregular on timescales ranging from hours to millennia. Its instantaneous rotation axis moves relative to the crust, the rotation rate changes, and the axis changes direction in space. Geodesy measures these variations because they are required to transform terrestrial coordinates into celestial coordinates and because they contain information about exchanges of angular momentum within the Earth system.
Earth orientation parameters describe several distinct effects. Polar motion gives the position of the rotation pole relative to the terrestrial frame. UT1 describes rotation angle, while length-of-day variations express changes in rotational rate. Celestial pole offsets represent observed departures from adopted precession–nutation models. These quantities must not be treated as interchangeable descriptions of a single wobble: they concern different aspects of orientation and are constrained by different observations.
Polar motion contains an annual component and the Chandler wobble, a free rotational oscillation with a period of roughly 430 days in the real, deformable Earth. Its behavior depends on Earth’s elasticity, ocean response, dissipation, and excitation by atmospheric and oceanic processes. The pole also undergoes longer-term changes influenced by redistribution of mass, including changes in ice and continental water storage.
Variations in atmospheric winds and pressure alter the angular momentum of the atmosphere and its exchange with the solid Earth. Ocean currents, water mass redistribution, and interactions involving the core also affect rotation. The tides raised by the Moon and Sun produce periodic changes and contribute to long-term rotational evolution. Individual observed changes therefore require models that combine several physical reservoirs rather than attributing all rotation variability to one cause.
Precession and nutation describe changes in the orientation of Earth’s axis in space. Gravitational torques act on Earth’s nonspherical mass distribution, while the response depends on its internal structure. A deformable mantle and fluid core modify the motion predicted for an ideal rigid body. Very long baseline interferometry is especially valuable because its distant radio sources provide the celestial directions needed to observe these effects.
Earth rotation also connects geodesy with timekeeping. Atomic time provides a uniform physical timescale, whereas UT1 follows the variable rotation of Earth. Operational transformations require current orientation parameters and, where necessary, predictions. Inaccurate orientation propagates into satellite orbits, antenna pointing, and the interpretation of globally distributed observations, making this apparently astronomical subject part of practical terrestrial positioning.
Gravity-field determination and physical geodesy
Gravity in geodesy includes both gravitational attraction and the centrifugal effect associated with Earth’s rotation. In a conventional Earth-fixed description, the gravity potential is written
where is gravitational potential and is centrifugal potential. With the commonly used positive geodetic potential convention, the gravity vector is . Surfaces of constant are equipotential surfaces, and the local plumb line follows the gravity field rather than the normal to an arbitrary geometric surface.
Physical geodesy seeks to determine this field and its relation to Earth’s figure. A reference ellipsoid can be equipped with a mathematically defined normal potential . The difference is the disturbing potential. Under the assumptions of linearized theory, Bruns’s formula relates geoid height to disturbing potential:
where is normal gravity at the relevant reference location. This connects a physical property of the gravity field with a geometric separation between reference surfaces.
Gravity is measured using several instrument types. Absolute gravimeters determine acceleration from a physical trajectory, commonly the free fall of a test body measured interferometrically. Relative gravimeters measure changes between locations or times and require scale calibration and control of drift. Superconducting gravimeters provide highly sensitive continuous records at fixed observatories. Corrections may be needed for solid Earth tides, ocean loading, atmospheric effects, instrument height, and local environmental changes.
Satellite orbits reveal the gravity field because spatial variations in attraction perturb their motion. Global models commonly express the external gravitational potential in spherical harmonics. Low-degree terms describe broad features, such as the planet’s flattening-related quadrupole; higher degrees represent progressively finer spatial structure. Satellite-only observations become less sensitive to short wavelengths as orbital altitude increases, because these components diminish rapidly with distance.
The GOCE mission, operating from 2009 to 2013, used satellite gravity gradiometry to improve knowledge of the static gravity field. GRACE, launched in 2002, and GRACE Follow-On, launched in 2018, measure changes in the separation of paired satellites to recover time-variable gravity. Terrestrial, airborne, marine, and satellite observations are often combined because their spatial coverage and wavelength sensitivity differ.
Gravity interpretation is inherently nonunique. Different subsurface density distributions can generate similar external fields. A gravity anomaly may reflect crustal thickness, sedimentary structure, mantle variations, topography, or a combination of sources. Geological and seismic information therefore constrain interpretations that gravity observations alone cannot uniquely establish.
Height systems, the geoid, and vertical datums
A height is not fully defined by a number and a unit. It also requires a reference surface and a rule for measuring separation from that surface. Satellite navigation naturally provides ellipsoidal heights, whereas water movement and many engineering problems are governed by gravity potential. The distinction explains why a satellite-derived “altitude” may differ substantially from the elevation shown on a topographic map.
The geoid is an equipotential surface selected to approximate global mean sea level. It extends conceptually beneath the continents and is not identical to the instantaneous ocean surface. Winds, currents, temperature, salinity, and atmospheric pressure maintain ocean surface departures from a gravity equipotential. Mean sea level at a particular tide gauge consequently cannot be assumed to represent the same potential as mean sea level at every other coast.
The geopotential number of a point is
where is the reference potential and is potential at the point. It expresses a physically meaningful separation in potential. Orthometric height is obtained by dividing by an appropriate mean gravity along the plumb line between the geoid and the point. Because gravity inside the topography is not completely observed, practical orthometric heights depend on assumptions or models of internal mass distribution.
Normal heights avoid that particular dependence by using normal gravity from a reference field. Their associated reference surface is the quasigeoid rather than the geoid. The relationship between ellipsoidal height , orthometric height , and geoid height is commonly written
For normal height , the corresponding separation is the height anomaly , giving . These relations are extremely useful operationally, but precision applications must respect their geometric definitions and conventions.
Historical vertical datums commonly originated from tide-gauge records and leveling networks. Differences in observation periods, ocean conditions, gravity corrections, and subsequent land motion produced offsets among national systems. A leveling network may also contain accumulated error or represent measurements taken before significant subsidence or earthquakes.
International efforts toward a common height reference use an agreed reference potential and globally consistent observations. Realizing such a system requires gravity data, accurate geoid or quasigeoid models, and stable reference stations. Vertical-datum modernization is therefore more than replacing leveling with satellite positioning: it must connect geometric measurements with a physical potential field and account for the time dependence of both the land and the reference surface.
Estimation, uncertainty, and observational modeling
Geodetic quantities are rarely observed directly in their final published form. Instruments measure travel times, phases, angles, frequencies, or accelerations. Coordinates, velocities, gravity coefficients, and orientation parameters are inferred through mathematical models. The reliability of a result therefore depends jointly on instrument performance, observation geometry, environmental corrections, and statistical estimation.
A linearized observation model may be written
where contains observations or observation residuals, contains unknown parameters, is a design matrix, and represents errors. In weighted least squares, a common estimate is
provided the parameterization has sufficient rank and the inverse exists. The weight matrix represents the adopted observational covariance structure.
Networks often possess datum freedoms. Relative distances alone, for example, cannot determine a network’s absolute translation or orientation. Suitable constraints or generalized methods are needed to make the solution unique without imposing unintended distortion. Overly restrictive constraints can produce deceptively small formal errors while forcing genuine deformation into other parameters.
Error correlations are particularly important in geodetic time series. Atmospheric mismodeling, common satellite errors, loading effects, and regional environmental processes can correlate observations in time and space. Treating all daily coordinate errors as independent commonly understates velocity uncertainty. Seasonal signals, equipment changes, and offsets must also be distinguished from secular motion. A short record spanning part of an annual cycle can yield a misleading linear trend even when individual positions appear precise.
Instrument calibration establishes relationships between readings and physical quantities, while metrological traceability connects measurements through documented calibration chains to recognized standards. Geodetic processing additionally requires consistent conventions for tides, antenna models, relativistic effects, and atmospheric loading. Two technically competent analyses may disagree if they use incompatible conventions rather than different raw observations.
Independent validation is therefore essential. Examples include comparing satellite-derived heights with leveling and gravity-based models, checking coordinate changes against local surveys, and comparing independent space techniques at colocated stations. Formal covariance describes uncertainty under an adopted model; it does not prove the absence of systematic error. Reproducible processing, retained metadata, and reanalysis with improved models are integral parts of geodetic practice because the scientific interpretation can change without any change to the original measurements.
Crustal deformation and tectonic processes
Geodesy measures the kinematics of plate tectonics directly. A rigid plate’s surface velocity can be approximated by rotation about an Euler pole:
where is the plate’s angular velocity vector and is a position vector relative to Earth’s center. Observed departures from this model reveal deformation within plates, motion across boundaries, or nontectonic effects such as loading and subsidence.
Continuous satellite positioning networks measure interseismic strain accumulation across faults. Where a fault is locked at shallow depth, deeper relative motion produces a broad surface velocity gradient. Elastic dislocation models can relate that gradient to fault geometry, slip rate, and locking depth. These inferred parameters depend on model assumptions; a surface velocity field does not uniquely reveal the full distribution of stress or friction at depth.
Earthquakes produce abrupt displacements that can range from millimeters far from the source to metres nearby. The 2011 Tohoku earthquake in Japan caused large horizontal and vertical offsets recorded by the country’s dense geodetic network. High-rate positioning also captures parts of the dynamic displacement history. Postseismic observations reveal continuing deformation through afterslip, viscoelastic relaxation, and fluid-related processes, often operating simultaneously over different timescales.
Interferometric synthetic aperture radar measures changes in satellite-to-ground distance by comparing radar phase between acquisitions. It can map deformation across broad areas with dense spatial sampling. A single viewing geometry primarily measures displacement along the radar line of sight, not a complete three-dimensional vector. Atmospheric delay, orbital error, vegetation change, and loss of coherence can imitate or obscure motion, so interpretation benefits from multiple viewing directions and ground-based measurements.
At volcanoes, inflation or deflation can indicate changes in subsurface pressure or fluid storage. Source models range from idealized point cavities to distributed intrusions embedded in elastic or more complex media. Geodetic deformation is interpreted alongside seismicity, gas observations, and geological information because similar surface patterns can arise from different subsurface processes. Measurements constrain physical models but do not by themselves provide a deterministic timetable for eruption.
Longer-term vertical deformation includes glacial isostatic adjustment, the continuing response to past ice loading. Parts of Fennoscandia and Canada are still rising as the mantle and crust adjust to the loss of large ice sheets. Correcting for this motion is essential when interpreting modern ice-mass changes, sea-level records, and reference-station velocities. Tectonic and nontectonic deformation are therefore separated through combined spatial, temporal, and physical evidence.
Water, ice, and ocean monitoring
Changing water and ice loads affect Earth in two observable ways: they alter gravitational attraction and deform the solid surface. Geodesy uses both responses to investigate storage and transport within the Earth system. These measurements complement local observations because they integrate changes over large regions, including areas where gauges and field measurements are sparse.
GRACE and GRACE Follow-On recover broad-scale changes in gravity that can be expressed as equivalent water thickness. Over land, the signal includes groundwater, soil moisture, snow, surface water, and other mass changes. It is not a direct measurement of groundwater alone. Estimating groundwater storage requires separating the other components using observations and models. Spatial smoothing, leakage between neighboring regions, and the limited resolution of satellite gravity must be accounted for when interpreting individual drainage basins.
Changes in ice sheets and glaciers can be measured through gravity, satellite altimetry, and combinations of surface velocity and ice thickness. Each method observes a different property. Gravity responds to mass; altimetry responds to elevation, which must be converted using assumptions about snow, firn, and ice density. Bedrock movement and glacial isostatic adjustment also affect the interpretation. Agreement among independent methods is especially informative because their principal errors differ.
Satellite radar altimetry estimates sea-surface height from orbital position and the measured range to the ocean surface. The TOPEX/Poseidon mission, launched in 1992, established a major foundation for subsequent precision ocean-altimetry records. Corrections include propagation through the atmosphere, tides, sea-state effects, and instrument biases. Long-term continuity requires careful cross-calibration among successive missions and stable orbital reference frames.
Changes in global mean sea level arise mainly from changes in ocean mass and changes in seawater volume associated with temperature and salinity. Satellite gravity constrains the mass contribution, while ocean profiling observations constrain much of the steric contribution. Altimetry measures the combined surface-height change. Comparing these components forms a sea-level budget and provides a demanding consistency test across observing systems.
Tide gauges measure water level relative to the land supporting the instrument. Their records therefore combine ocean change with vertical land motion. A coast undergoing subsidence can experience faster relative sea-level rise than the global ocean average. Colocated or nearby geodetic stations help separate these effects, provided local motion between the station and gauge is understood. This distinction is crucial in deltas, sedimentary basins, and regions affected by groundwater extraction or tectonic deformation.
Mapping, engineering, and spatial infrastructure
Geodesy supplies the common spatial framework within which maps, property surveys, engineering designs, and navigation systems can be combined. A detailed dataset is not necessarily correctly located: internal precision and external reference accuracy are different properties. Two surveys may each describe local geometry well yet disagree because they use different datums, epochs, projections, or height systems.
In mapping, geodetic coordinates are transformed into projected coordinates suitable for display and computation. Every map projection distorts some combination of distance, area, shape, and direction. For engineering work, projection scale must also be combined with the effect of elevation above the reference ellipsoid. A distance computed on a map grid can consequently differ from the physical distance measured at a construction site. “Grid-to-ground” handling is a practical geodetic problem, not merely a cartographic detail.
Large infrastructure projects require consistent control across space and time. Long tunnels may be surveyed from opposite ends before underground networks meet. Bridges require monitoring of foundations, towers, and decks under thermal and mechanical loading. Dams and embankments require stable reference points outside the deforming structure. In each case, the reference network must be designed so that movement of its own monuments is not mistaken for structural displacement.
Hydrography introduces an additional vertical-reference problem. Seabed depths, vessel motion, tides, and satellite-derived positions must be related to an appropriate chart datum. Chart datums are often tied to low-water tidal definitions rather than the land-elevation datum used inland. Combining coastal topography and bathymetry therefore requires explicit transformations, particularly for inundation modeling and coastal engineering.
Satellite imagery, aerial photography, laser scanning, and mobile mapping depend on geodetic positioning and orientation. Ground control can constrain image geometry, while precise trajectory estimation locates sensors carried by aircraft or vehicles. Direct georeferencing reduces the need for extensive ground control but does not remove calibration errors in sensor alignment, timing, or camera geometry. Small angular errors can produce substantial positional errors when projected over long distances.
Geodetic information also underlies cadastral and legal boundary systems, although the legal definition of a boundary need not be identical to a coordinate estimate. Monuments, documentary descriptions, and jurisdiction-specific rules may carry legal significance independently of a modern positioning solution. Geodesy provides measurement consistency; it does not itself determine ownership or legal authority. Robust spatial infrastructure consequently includes technical standards, accessible reference services, transformation tools, and durable records of how coordinates were established.
International organization and emerging methods
Global geodesy depends on institutions that coordinate observations beyond the capacity or territory of any single state. The International Association of Geodesy organizes scientific work and associated services. The International Earth Rotation and Reference Systems Service maintains Earth-orientation and reference-system products, while services for satellite navigation, laser ranging, radio interferometry, and DORIS coordinate observations and analysis within their respective techniques. Their products are used together in global reference-frame determination.
The Global Geodetic Observing System promotes integration across geometric and gravity observations. A major challenge is maintaining observatories over decades: scientific value often increases with record length, while equipment replacement, land-use change, monument instability, and funding interruptions can break continuity. Better geographical distribution is also important. Adding precision at already well-observed sites cannot fully substitute for missing stations in poorly sampled regions.
In 2015, the United Nations General Assembly adopted a resolution on a global geodetic reference frame for sustainable development. This recognized that reference infrastructure supports disaster-risk management, environmental monitoring, and economic activities as well as basic science. International coordination concerns standards, data sharing, capacity development, and continuity of services; the frame itself remains a measurement-based scientific realization rather than a single physical object maintained in one location.
Emerging instruments include atom-interferometric gravimeters, improved spaceborne laser ranging, and optical clocks for relativistic geodesy. Under general relativity, clocks at different gravity potentials run at different rates. For nearby stationary clocks in the weak-field approximation, the fractional frequency difference is related approximately to the potential difference by
with the sign depending on the chosen ordering and potential convention. A fractional difference of order corresponds roughly to a centimeter-scale height difference near Earth’s surface. Turning this sensitivity into an operational height comparison also requires stable clock performance and accurate frequency transfer.
Future reference frames must represent nonlinear motion more completely and maintain consistency among gravity, geometry, and time. Improvements in observations can expose previously negligible effects, including thermal deformation of instruments, local hydrology, and differences among loading models. Automated processing and machine learning can assist with anomaly detection and data handling, but physical interpretability and independent validation remain necessary.
The enduring technical problem is continuity: observations made with changing instruments, satellite constellations, and processing models must remain comparable across decades. Preserving raw data, calibration histories, station logs, and analysis conventions enables later reprocessing and prevents apparent environmental trends from being artifacts of the observing system.