{
  "id": "moment-magnitude",
  "title": "Moment magnitude",
  "lang": "en",
  "source_lang": "en",
  "summary": "Moment magnitude measures earthquake size logarithmically from seismic moment, linking fault area, slip, and rock rigidity without the saturation characteristic of many amplitude-based scales.",
  "category": "science",
  "subcategory": "earth-science",
  "type": "concept",
  "aliases": [
    "Moment magnitude scale",
    "seismic moment magnitude",
    "Mw scale"
  ],
  "generated_at": "2026-11-05",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/moment-magnitude/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/moment-magnitude/",
      "json": "https://aiwiki.page/zh/moment-magnitude.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Moment magnitude, usually written \\(M_w\\), is a logarithmic measure of earthquake size derived from the earthquake’s seismic moment. For ordinary fault slip, seismic moment expresses the combined effects of the area that ruptures, the average displacement across the fault, and the rigidity of the surrounding rock. Moment magnitude therefore connects an earthquake’s reported size to a physical description of its source, rather than to the amplitude of a particular seismic wave alone.\n\nIntroduced by Thomas C. Hanks and Hiroo Kanamori in 1979, the scale became the principal measure used to compare large earthquakes worldwide. Its defining advantage is that it does not exhibit the strong magnitude saturation that affects several older scales when earthquakes become very large. It is nevertheless an estimate derived from observations and source models, not a direct reading from a single instrument.\n\nMoment magnitude describes overall source size, not the severity of shaking at a particular location. Earthquakes with the same \\(M_w\\) can differ substantially in duration, high-frequency radiation, local damage, and tsunami generation. Its interpretation consequently requires a distinction between seismic moment, radiated energy, rupture behavior, and the effects experienced at the surface."
    },
    {
      "heading": "Definition, numerical scale, and units",
      "text": "The widely used definition of moment magnitude is\n\n\\[M_w=\\frac{2}{3}\\left[\\log_{10}\\left(\\frac{M_0}{1\\ \\mathrm{N\\,m}}\\right)-9.1\\right],\\]\n\nwhere \\(M_0\\) is scalar seismic moment. Writing the ratio inside the logarithm makes explicit that its argument is dimensionless. An equivalent conventional expression, when moment is stated in dyne-centimetres, is\n\n\\[M_w=\\frac{2}{3}\\log_{10}M_0-10.7.\\]\n\nThe constants reflect the conversion between units and the historical calibration of the magnitude scale. Because \\(1\\ \\mathrm{N\\,m}=10^7\\) dyne-centimetres, the two expressions describe the same relationship to the precision implied by their rounded constants.\n\nThe inverse relation is particularly useful:\n\n\\[M_0=10^{1.5M_w+9.1}\\ \\mathrm{N\\,m}.\\]\n\nAn earthquake of magnitude 5 therefore has a moment of approximately \\(4.0\\times10^{16}\\ \\mathrm{N\\,m}\\); magnitude 7 corresponds to approximately \\(4.0\\times10^{19}\\ \\mathrm{N\\,m}\\); and magnitude 9 to approximately \\(4.0\\times10^{22}\\ \\mathrm{N\\,m}\\). Each increase of one magnitude unit represents a moment increase of \\(10^{1.5}\\), or about 31.6. An increase of two units represents a factor of 1,000.\n\nMoment magnitude itself has no physical unit. Seismic moment has the dimensions of force multiplied by length, expressed in the International System of Units as newton-metres. These are dimensionally equivalent to the joule, but seismic moment is not an amount of energy. Nor is it an ordinary mechanical torque, despite sharing the same dimensions. These distinctions arise from the different physical quantities represented, not from different unit algebra.\n\nThe scale has no mathematical lower or upper bound. Negative moment magnitudes are meaningful for sufficiently small seismic sources. At the upper end, possible earthquake sizes are constrained by the dimensions of available faults, their mechanical properties, and attainable slip, rather than by a ceiling built into the magnitude definition.\n\nReported values usually have one decimal place, but this does not imply exact knowledge to that precision. A difference of 0.1 magnitude unit corresponds to a moment ratio of approximately 1.41. Two estimates such as 7.8 and 7.9 may therefore represent substantially different inferred moments while still being compatible with observational and modeling uncertainties. Magnitude rounding also means that a published value represents a range of possible underlying moments rather than one exact value."
    },
    {
      "heading": "Historical development and the problem of saturation",
      "text": "The modern magnitude concept began with Charles F. Richter’s 1935 local magnitude scale, developed for earthquakes in southern California. Local magnitude, written \\(M_L\\), used the logarithm of the maximum amplitude recorded by a standardized Wood–Anderson seismograph, corrected for distance. Its practical achievement was to provide a reproducible numerical description of earthquake size from instrumental observations rather than from damage reports.\n\nAs seismic networks expanded, other magnitude scales were developed for waves recorded at greater distances. Body-wave magnitude, commonly \\(m_b\\), uses relatively short-period body waves, while surface-wave magnitude, commonly \\(M_s\\), traditionally uses surface waves with periods near 20 seconds. These scales made worldwide earthquake comparison possible before modern broadband digital recording became routine. Their calibration also helped establish the numerical range that later magnitude systems sought to preserve.\n\nA fundamental difficulty emerged for very large earthquakes. As rupture length and duration increase, more source information appears at longer periods. Amplitudes in a fixed short-period band cease to increase in proportion to the earthquake’s total size. This is magnitude saturation: the measuring procedure remains responsive to part of the radiation, but increasingly fails to capture the whole source. Short-period \\(m_b\\) commonly saturates around magnitude 6, while conventional \\(M_s\\) becomes strongly limited for the largest earthquakes, broadly around magnitude 8 and above. These are approximate ranges, not universal thresholds.\n\nSeismic moment supplied a physically grounded alternative. In 1966, Keiiti Aki’s analysis of the 1964 Niigata earthquake helped establish practical determination of seismic moment from long-period waves and its connection to fault area and slip. During the 1970s, analyses of great earthquakes showed that moment could distinguish events that older magnitude scales compressed into a relatively narrow numerical range.\n\nKanamori’s work in 1977 connected the size of great earthquakes to moment and energetics. Hanks and Kanamori then introduced the moment magnitude scale in 1979, choosing a logarithmic transformation that broadly preserved agreement with established magnitudes where those scales worked well. The result was not simply another instrument-specific amplitude formula, but a magnitude tied to a source parameter.\n\nThe 1960 Chile and 1964 Alaska earthquakes illustrate the historical importance of this change. Their commonly reported modern values are approximately \\(M_w\\,9.5\\) and \\(M_w\\,9.2\\), respectively. Long-period source analyses revealed their exceptional moments more effectively than conventional short-period magnitudes could. Adoption of \\(M_w\\) was gradual, however, and older scales remain useful for particular networks, historical catalogs, and rapid measurements."
    },
    {
      "heading": "Fault mechanics and the physical meaning of seismic moment",
      "text": "For a planar fault undergoing shear displacement in a medium of uniform rigidity, scalar seismic moment is\n\n\\[M_0=\\mu A\\overline{D},\\]\n\nwhere \\(\\mu\\) is the shear modulus, \\(A\\) is rupture area, and \\(\\overline{D}\\) is average slip. The shear modulus measures resistance to shear deformation and is expressed in pascals. A representative crustal value used in illustrative calculations is \\(30\\ \\mathrm{GPa}\\), although actual rigidity varies with rock type, depth, temperature, and tectonic setting.\n\nFor example, consider a rupture 100 kilometres long and 20 kilometres wide, with average slip of 1 metre and rigidity of \\(30\\ \\mathrm{GPa}\\). Its area is \\(2\\times10^9\\ \\mathrm{m^2}\\), giving\n\n\\[M_0=(3\\times10^{10})(2\\times10^9)(1) =6\\times10^{19}\\ \\mathrm{N\\,m}.\\]\n\nThis corresponds to \\(M_w\\approx7.1\\). The calculation shows why magnitude cannot identify fault length or slip separately: many combinations of area, displacement, and rigidity yield the same moment.\n\nReal ruptures are heterogeneous. Slip can vary sharply over a fault, and rigidity may differ among source regions. A more general expression is\n\n\\[M_0=\\int_A \\mu(\\mathbf{x})D(\\mathbf{x})\\,\\mathrm{d}A,\\]\n\nfor slip of consistent sense on a specified fault surface. Finite-fault models approximate this integral by dividing the rupture into patches. The summed moment is often more stable than the inferred slip on any individual patch, because observations can constrain broad source strength better than fine spatial detail.\n\nThe complete source description is commonly expressed through a seismic moment tensor. Ordinary shear faulting is represented by a double-couple component, whose geometry specifies fault orientation and slip direction. For a pure double couple, scalar moment can be recovered from the tensor through\n\n\\[M_0=\\left(\\frac{1}{2}\\sum_{i,j}M_{ij}^{\\,2}\\right)^{1/2}.\\]\n\nMagnitude retains this scalar measure while discarding directional information. Two earthquakes can therefore have equal \\(M_w\\) but very different fault orientations and radiation patterns.\n\nMoment also accumulates over time. If \\(M_0(t)\\) is cumulative moment, its derivative is the moment-rate function. A short rupture and a long rupture can reach the same final moment through very different histories. This distinction matters because the time dependence controls much of the emitted wave field. Slow-slip events can likewise have substantial geodetic moments while producing little ordinary seismic radiation. Assigning them a moment-equivalent magnitude describes deformation size, not an equivalence in shaking."
    },
    {
      "heading": "How seismic moment is measured",
      "text": "Seismic moment is inferred from recorded motion or deformation through models connecting the source to observations. A seismometer records ground motion filtered by its instrument response. Estimation therefore begins with response correction, timing control, and assessment of signal quality. Propagation effects must also be considered: geometrical spreading, attenuation, scattering, and the directional radiation pattern all alter amplitudes between source and receiver.\n\nOne important approach uses the displacement spectrum. In simplified source models, its low-frequency amplitude approaches a level proportional to seismic moment after corrections for distance, wave speed, density, and radiation pattern. At higher frequencies the spectrum decreases, with a transition commonly described by a corner frequency. Recovering the low-frequency level is central: if the recording band does not extend below the source’s characteristic frequency, inferred moment becomes more dependent on the assumed spectral shape.\n\nFor moderate and large events, waveform inversion is widely used. Synthetic waves are calculated for a trial source in an Earth model, and source parameters are adjusted to fit observed waveforms. Moment-tensor solutions can constrain scalar moment, mechanism, centroid location, and aspects of source duration. A centroid represents the effective center of moment release; it need not coincide with the hypocentre, which is the point where rupture begins. This distinction becomes especially important for ruptures extending hundreds of kilometres.\n\nDifferent inversion methods use different portions of the wave field. Regional procedures may fit body waves and surface waves recorded within a tectonic region. Global centroid-moment-tensor methods emphasize long-period records from widely distributed stations. Very large earthquakes can also excite normal modes of the Earth, whose long-period oscillations constrain source properties. Finite-fault inversions go beyond a single effective source and estimate how slip or moment release is distributed in space and time.\n\nPermanent deformation provides an independent route. Global navigation satellite system measurements record ground displacement, while interferometric synthetic-aperture radar maps changes in satellite-to-ground distance. These observations can constrain fault slip and hence moment without relying exclusively on seismic-wave amplitudes. Offshore deformation remains more difficult to observe directly, so seafloor geodesy and tsunami records can be important for major subduction earthquakes.\n\nThe methods are complementary rather than interchangeable. Seismic data resolve rapid rupture processes well; geodesy captures permanent displacement and is less directly affected by intense short-period wave amplitudes. Joint models can reduce ambiguities, but they must distinguish coseismic slip from later deformation and account for differences in spatial resolution. A published moment magnitude therefore reflects both the available observations and the assumptions used to translate them into a source model."
    },
    {
      "heading": "Comparison with other magnitudes and earthquake intensity",
      "text": "Moment magnitude belongs to a family of earthquake-size measures rather than replacing every other measure in every setting. Local magnitude remains useful for nearby small and moderate earthquakes because it can be obtained efficiently from regional networks. Body-wave and surface-wave magnitudes retain value in global monitoring and historical analysis. Duration magnitudes, based on the length of a seismic signal or coda, are also used by some local networks.\n\nThe notation matters. \\(M_L\\), \\(m_b\\), \\(M_s\\), and \\(M_w\\) do not denote identical measurements. Their numerical values often agree approximately over limited ranges because the scales were calibrated to maintain continuity, but systematic differences occur with depth, source mechanism, propagation path, and event size. A deep earthquake, for example, may generate weak conventional surface waves, making a surface-wave magnitude less suitable even when body-wave observations are strong.\n\nSaturation should not be confused with instrument clipping. Clipping occurs when recorded motion exceeds an instrument’s dynamic range. It can sometimes be avoided with a different sensor or recording setting. Magnitude saturation arises because a selected period band does not continue to track total source size. An unclipped short-period record can therefore yield a saturated magnitude. Moment magnitude avoids this particular limitation only when the observations and modeling adequately recover the full seismic moment.\n\nMagnitude is also distinct from seismic intensity. Intensity describes shaking and its effects at a specified location. The Modified Mercalli intensity scale, for example, uses observations of human perception, object movement, and damage. One earthquake has a source magnitude but a spatially varying field of intensities. Modern instrumental intensity estimates use measured or predicted ground motion to represent that variation.\n\nDistance, depth, rupture directivity, local geology, and building characteristics can all influence observed effects. Soft sediments may amplify particular frequencies; a sedimentary basin may prolong shaking; and rupture propagating toward a site can concentrate motion into a strong pulse. A nearby shallow earthquake of moderate magnitude can consequently cause more severe local effects than a larger but distant or deep earthquake.\n\nThe phrase “Richter scale” is often used informally for any reported earthquake magnitude, but it is technically inappropriate for a value determined as \\(M_w\\). Likewise, the statement that each magnitude unit means “ten times stronger” is ambiguous. A factor of ten belongs to the amplitude definition of certain traditional scales under specified conditions. For moment magnitude, one unit means approximately 31.6 times the seismic moment, while no single multiplier describes every aspect of shaking or damage."
    },
    {
      "heading": "Seismic moment, radiated energy, and stress drop",
      "text": "Seismic moment and radiated seismic energy describe different aspects of an earthquake. Moment measures the overall strength of the displacement source. Radiated energy, usually denoted \\(E_R\\), measures energy carried away by seismic waves. The earthquake’s mechanical energy budget also includes frictional heating, fracture-related processes, and other deformation. Neither moment nor radiated energy alone specifies that complete budget.\n\nA commonly used approximate scaling is\n\n\\[\\log_{10}\\left(\\frac{E_R}{1\\ \\mathrm{J}}\\right) \\approx1.5M_w+4.8.\\]\n\nCombined with the moment-magnitude relation, this implies a representative ratio \\(E_R/M_0\\) of about \\(5\\times10^{-5}\\). The relationship is useful for broad comparisons but is not an identity. It assumes an approximately typical relationship between moment and radiation; observed earthquakes can depart substantially from it.\n\nOne quantity expressing this difference is apparent stress,\n\n\\[\\sigma_a=\\mu\\frac{E_R}{M_0}.\\]\n\nIt has units of pressure and relates radiated energy to the source deformation represented by moment. Apparent stress is not the same as the actual stress acting on a fault, nor is it automatically equal to stress drop. It is a source parameter inferred from energy and moment estimates, both of which carry observational and modeling uncertainties.\n\nStress drop describes the decrease in shear stress associated with rupture. For an idealized circular crack of radius \\(r\\),\n\n\\[\\Delta\\sigma\\propto\\frac{M_0}{r^3},\\]\n\nwith the coefficient determined by the particular crack model. This relation explains why equal moments need not imply equal rupture dimensions. A smaller rupture producing the same moment generally requires greater average slip and, within the model, a higher stress drop. Such differences affect the source spectrum and can influence high-frequency shaking.\n\nUnder approximate constant-stress-drop scaling, rupture length grows as \\(M_0^{1/3}\\), rupture area as \\(M_0^{2/3}\\), and average slip as \\(M_0^{1/3}\\). These relationships help explain broad patterns in earthquake populations, but actual ruptures are constrained by fault width, depth, geometry, and heterogeneous material properties. Large strike-slip events, for example, cannot increase their downdip width indefinitely within a finite seismogenic layer.\n\nA tsunami earthquake provides an especially important example of the limits of energy comparisons. Such an event produces a tsunami unusually large relative to its conventional short-period seismic magnitude, often involving slow rupture near a subduction trench. Its long-period or geodetic moment can be much larger than short-period observations initially suggest. Moment magnitude improves the estimate of overall source size, but tsunami generation still depends on where and how displacement moves the seafloor and water column."
    },
    {
      "heading": "Rapid determination, magnitude variants, and uncertainty",
      "text": "Operational magnitude estimates change as information arrives. Immediately after an earthquake, a monitoring system may have only a few seconds of data from a small number of stations. Later estimates can incorporate additional wave types, longer periods, a wider station distribution, and geodetic observations. A revised magnitude therefore often reflects an improved representation of the source rather than an error in the arithmetic of the original estimate.\n\nThe problem is particularly acute during a great earthquake. Rupture can continue for several minutes, so the first observations necessarily describe an incomplete source. No method based solely on signals already emitted can directly measure moment that has not yet been released. Early algorithms may infer likely final size from empirical patterns, but such inference is distinct from observing the completed rupture.\n\nOperational catalogs sometimes attach method-specific suffixes to moment magnitudes. Labels such as \\(M_{ww}\\), \\(M_{wc}\\), and \\(M_{wb}\\) distinguish particular W-phase, centroid, and body-wave procedures. \\(M_{wp}\\) denotes a rapid moment-related estimate derived from P-wave processing. These are not fundamentally different logarithmic scales; they identify how source size was estimated and help users interpret differences among reported values.\n\nThe W phase is useful for rapid assessment of large earthquakes because it is a very-long-period seismic signal observed in the interval associated with early body-wave arrivals. W-phase inversion can estimate moment and mechanism without waiting for the complete surface-wave train used in some long-period methods. Its effectiveness still depends on station coverage, signal quality, filtering, and appropriate source modeling.\n\nSeveral sources contribute to measurement uncertainty. Instrument calibration and noise affect the data; imperfect Earth models affect predicted propagation; sparse coverage weakens constraints on source orientation and depth; and complex ruptures may not be represented adequately by a single point source. Some errors are correlated. A systematic velocity-model bias, for instance, is not necessarily removed simply by adding more stations.\n\nFor small relative moment errors, uncertainty propagation gives\n\n\\[\\delta M_w\\approx\\frac{2}{3\\ln10}\\frac{\\delta M_0}{M_0}.\\]\n\nThus a 20 percent relative uncertainty in moment corresponds locally to about 0.06 magnitude unit, before allowing for additional systematic effects. The logarithmic transformation compresses fractional moment uncertainty, but it does not remove it.\n\nAn uncertainty interval from one inversion may describe only the variability allowed within that model. Differences between reasonable models can be larger. Comparisons among agencies therefore require attention to data windows, source assumptions, and method labels, not merely to the last decimal digit."
    },
    {
      "heading": "Uses in tectonics, hazard analysis, and earthquake catalogs",
      "text": "Moment magnitude provides a common basis for comparing earthquakes across tectonic environments and historical periods. Large events in plate tectonics can differ greatly in geometry: subduction interfaces may rupture broad, gently dipping areas, while major strike-slip earthquakes often rupture long, comparatively narrow zones. Expressing both through seismic moment makes their total source sizes comparable without assuming identical fault shapes or shaking characteristics.\n\nSeismic moment is also used in deformation budgets. Geodetic observations measure ongoing strain accumulation, while earthquake catalogs describe moment released in seismic events. Under suitable assumptions about fault geometry and rigidity, the two can be compared to investigate how plate motion is accommodated. The interpretation is not a simple accounting exercise: aseismic creep, slow slip, distributed deformation, and incomplete historical records can all affect the apparent balance.\n\nIn Gutenberg–Richter statistics, earthquake occurrence is commonly described by\n\n\\[\\log_{10}N=a-bM,\\]\n\nwhere \\(N\\) is the number of earthquakes at or above magnitude \\(M\\) within a specified region and interval. For many datasets, \\(b\\) is near 1, although it varies and is sensitive to catalog completeness and analysis choices. Because moment increases by about 31.6 per magnitude unit while event frequency often decreases by roughly a factor of ten, rare large earthquakes can dominate total seismic moment release.\n\nCatalog consistency is therefore important. A historical catalog may mix local, body-wave, surface-wave, and moment magnitudes. Converting all entries to an equivalent \\(M_w\\) scale requires empirical relationships, whose applicability depends on magnitude range, tectonic setting, and source data. Such conversions introduce uncertainty and can be biased where the original scale saturates. Direct reanalysis of old seismograms can provide stronger constraints when suitable records survive.\n\nIn seismic hazard analysis, moment magnitude is widely used as a source-size variable in models of earthquake occurrence and ground motion. It is combined with source-to-site distance, faulting style, depth, and local site conditions to estimate measures such as peak ground acceleration or spectral acceleration. Magnitude alone cannot specify those motions, and a hazard calculation must represent both source variability and uncertainty in wave propagation and site response.\n\nFor tsunami assessment, \\(M_w\\) helps identify earthquakes capable of producing large ocean disturbances, but source depth, mechanism, and offshore location remain essential. The 2004 Sumatra–Andaman and 2011 Tohoku earthquakes, commonly assigned moment magnitudes around 9.1, demonstrated the importance of resolving extensive rupture and heterogeneous slip. Models of their tsunami sources require the distribution of seafloor displacement, not merely a single magnitude value; offshore observations and time-dependent source estimates supply information that the scalar magnitude necessarily omits."
    }
  ],
  "links": [
    "force",
    "international-system-of-units",
    "joule",
    "energy",
    "torque",
    "scattering",
    "density",
    "earth",
    "pressure",
    "measurement-uncertainty",
    "plate-tectonics"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}