When a M7.0 earthquake struck Haiti on January 12, 2010, at a depth of approximately 13 km, it killed an estimated 100,000 to 316,000 people and destroyed much of Port-au-Prince. Just three years later, a M8.3 earthquake struck deep beneath the Sea of Okhotsk off Russia's Kamchatka Peninsula at a depth of 609 km — releasing roughly 45 times more energy than the Haiti event — and caused no reported damage or casualties. The difference was depth.
An earthquake's location in three dimensions — its latitude and longitude on Earth's surface (the epicenter) and its depth below the surface (which defines the hypocenter) — is among the most critical factors determining its impact. Two earthquakes of identical magnitude can produce vastly different outcomes depending on whether the rupture occurs 10 km or 600 km below the surface, and whether the epicenter lies beneath a major city or under a remote stretch of ocean.
This article explains how seismologists determine where an earthquake occurs, why depth matters so profoundly, and how the spatial relationship between an earthquake's source and the surface shapes the shaking experienced by people and structures.
Definitions: Hypocenter vs. Epicenter
These two terms describe the same earthquake's location in different ways:
The hypocenter (also called the focus) is the point within Earth's interior where the earthquake rupture initiates. It is defined by three coordinates: latitude, longitude, and depth. In reality, large earthquakes don't occur at a single point — the fault rupture can extend for tens to hundreds of kilometers. The hypocenter represents the nucleation point where the rupture begins, not the full extent of the fault that slips. For the 1906 San Francisco earthquake (M7.9), the hypocenter was near San Francisco, but the fault ruptured approximately 477 km along the San Andreas Fault.
The epicenter is the point on Earth's surface directly above the hypocenter. It is the earthquake location reported on maps and in news reports — the latitude and longitude without the depth component. For shallow earthquakes, the epicenter is approximately where the strongest shaking occurs (though this is not always the case if the fault rupture propagates directionally). For deep earthquakes, the epicenter may be far from the area of strongest surface shaking.
The distinction matters because a shallow earthquake's epicenter closely approximates the point of maximum impact, while a deep earthquake's effects are distributed over a much wider area with lower peak intensity.
How Seismologists Locate Earthquakes
Triangulation Using P-S Arrival Times
The primary method for locating an earthquake is based on the difference in arrival times of P-waves (primary/compressional waves) and S-waves (secondary/shear waves) at seismograph stations. P-waves travel through Earth at approximately 5–8 km/s in the crust and faster in the mantle, while S-waves travel at roughly 60% of the P-wave speed. This velocity difference means that P-waves arrive at a seismograph station before S-waves, and the time gap between them increases with distance from the earthquake.
By measuring the P-S arrival time difference at a single station, a seismologist can calculate the distance from that station to the earthquake, but not the direction. This defines a sphere (in three dimensions) or circle (in two dimensions on a map) of possible locations centered on the station. Using data from a minimum of three stations produces three spheres that intersect at a single point — the hypocenter. In practice, data from dozens or hundreds of stations are used, and the earthquake location is determined by a least-squares inversion that finds the point in three-dimensional space that best fits all arrival time observations.
Computerized Location Algorithms
Modern earthquake location relies on automated algorithms that process seismic data in near-real-time. The USGS uses systems including Hydra, which combines automatic phase picks from seismograph stations with analyst review. Global earthquake location is also performed by the International Seismological Centre (ISC) and various national agencies.
The accuracy of earthquake location depends on the density and distribution of seismograph stations surrounding the event. Earthquakes in well-instrumented regions (California, Japan, Italy) can be located with precision of 1–2 km horizontally and 2–5 km in depth. Earthquakes in poorly instrumented regions (mid-ocean ridges, remote subduction zones) may have uncertainties of 10–20 km or more.
How Depth Is Determined
Determining earthquake depth is more challenging than determining the epicenter. Horizontal location is well-constrained by a distributed network of surface stations, but depth resolution requires either very close stations (within a distance comparable to the earthquake depth) or the identification of specific seismic phases that are sensitive to depth.
Depth phases are seismic waves that travel upward from the hypocenter, reflect off Earth's surface, and then travel downward to the recording station. The two most important depth phases are pP (a P-wave traveling upward from the hypocenter, reflecting as a P-wave off the surface) and sP (an S-wave traveling upward, converting to a P-wave at the surface reflection point). The time difference between the direct P-wave and the pP or sP phase is directly related to the earthquake depth: the deeper the earthquake, the greater the time difference.
Waveform modeling compares the recorded seismograms to synthetic seismograms computed for various assumed depths. By systematically varying the depth in the model and comparing the synthetic waveforms to the observed data, seismologists can determine the depth that produces the best fit. This technique is particularly valuable for moderate-to-large earthquakes where the waveforms are well-recorded at teleseismic distances (greater than ~1,000 km).
Depth Classifications
The USGS classifies earthquakes into three depth categories:
Shallow: 0–70 km — These earthquakes occur within the crust or the uppermost mantle. They include all earthquakes on continental transform faults (like the San Andreas), all intraplate earthquakes in continental interiors, and the shallowest portion of subduction zone seismicity. Shallow earthquakes account for approximately 70% of all earthquakes and produce the vast majority of earthquake damage worldwide.
Intermediate: 70–300 km — These earthquakes occur exclusively in subduction zones, within the downgoing oceanic plate as it descends into the mantle. They mark the transition zone where the subducting slab is heating and dehydrating. Intermediate-depth earthquakes are felt over large areas but produce less intense shaking at the surface than shallow earthquakes of equivalent magnitude.
Deep: 300–700 km — Deep earthquakes also occur only in subduction zones, in the deepest portions of subducting slabs. They are concentrated in several specific regions worldwide, including the Tonga-Kermadec subduction zone, the Java-Sumatra subduction zone, and beneath South America. The deepest earthquakes ever recorded have occurred at approximately 700 km — near the boundary between the upper and lower mantle (the 660-km discontinuity). No earthquakes have been reliably recorded below approximately 700 km.
| Depth Classification | Depth Range | % of Global Earthquakes | Typical Locations | Damage Potential |
|---|---|---|---|---|
| Shallow | 0–70 km | ~70% | All plate boundaries, intraplate regions | Highest — concentrated energy, intense local shaking |
| Intermediate | 70–300 km | ~22% | Subduction zones only | Moderate — felt widely, generally moderate damage |
| Deep | 300–700 km | ~8% | Deep subduction zones (Tonga, Java, South America) | Low — energy attenuated over long path to surface |
Why Depth Matters: The Physics of Attenuation
The reason depth is so consequential for earthquake damage is geometric spreading and attenuation. Seismic waves radiate outward from the hypocenter in all directions, and their energy spreads over an expanding spherical wavefront. The energy per unit area of the wavefront decreases with the square of the distance from the source (in a simplified homogeneous model). Additionally, the Earth's interior absorbs seismic energy through anelastic processes (internal friction), further reducing wave amplitude with distance.
For a shallow earthquake at 10 km depth, the surface directly above is only 10 km from the energy source. The seismic waves arrive with very little geometric spreading or attenuation, producing intense, concentrated shaking. For an earthquake at 600 km depth, the surface is 600 km from the source. The waves have spread over a vastly larger area and have been significantly attenuated by passage through the mantle. The result is weak-to-moderate shaking distributed over a very large area, rather than intense shaking over a small area.
This is why a moderate shallow earthquake can be far more destructive than a great deep earthquake. The energy that matters for damage is not the total energy released (which is related to magnitude) but the energy density at the Earth's surface — which is strongly controlled by depth.
Cross-section diagram — Hypocenter, Epicenter, and Seismic Wave Propagation. Show a vertical cross-section of Earth's crust and upper mantle with three earthquake scenarios at different depths (e.g., 10 km, 70 km, 300 km). For each, show the hypocenter location, the epicenter directly above on the surface, and expanding concentric wavefronts. Annotate the surface shaking intensity (MMI or qualitative) for each, showing that the shallow earthquake produces intense local shaking, the intermediate earthquake produces moderate shaking over a wider area, and the deep earthquake produces weak shaking over the widest area. Data source: USGS, general seismological principles.
How Depth Affects Earthquake Impact: Case Studies
Shallow and Devastating
2010 Haiti (M7.0, depth ~13 km): The earthquake's hypocenter was approximately 25 km west-southwest of Port-au-Prince at a depth of approximately 13 km. The shallow depth concentrated intense shaking directly beneath a densely populated city with almost no seismic building code enforcement. An estimated 100,000 to 316,000 people were killed, and approximately 1.5 million were displaced. The destruction was compounded by poverty, poor construction quality, and a lack of emergency response infrastructure — but the shallow depth ensured that the shaking was severe enough to destroy even some reasonably constructed buildings.
2015 Nepal (M7.8, depth ~8 km): The Gorkha earthquake ruptured at a very shallow depth of approximately 8 km, with the rupture propagating eastward beneath the Kathmandu Valley. The shallow depth and the directivity of the rupture toward Kathmandu amplified shaking intensity in the capital. Approximately 8,900 people were killed and over 600,000 structures were destroyed or damaged in the Kathmandu Valley and surrounding districts. Numerous UNESCO World Heritage monuments were severely damaged or collapsed.
1994 Northridge (M6.7, depth ~18 km): A "moderate" earthquake by global standards, the Northridge earthquake occurred at 18 km depth beneath a densely populated area of the San Fernando Valley in Los Angeles. The shallow depth and location directly beneath urban infrastructure produced peak ground accelerations exceeding 1.0g at some locations — among the strongest ground motions ever recorded in an urban area. The earthquake caused approximately $20 billion in damage (1994 dollars; over $40 billion in 2024 dollars), killed 57 people, and collapsed freeway overpasses and parking structures throughout the region.
Deep and Felt but Not Destructive
2013 Sea of Okhotsk (M8.3, depth ~609 km): This earthquake occurred deep within the subducting Pacific Plate beneath the Sea of Okhotsk, east of Russia's Kamchatka Peninsula. Despite being one of the largest deep earthquakes ever recorded, it caused no damage or casualties. The earthquake was felt across a huge area — including Moscow, more than 7,000 km away — but the shaking was nowhere intense enough to damage structures. The event is scientifically significant as one of the deepest M8+ earthquakes ever recorded.
2015 Bonin Islands (M7.9, depth ~680 km): One of the deepest large earthquakes recorded, this event occurred approximately 680 km beneath the Bonin (Ogasawara) Islands south of Japan. Despite the very large magnitude, the extreme depth meant that surface shaking was light. The earthquake was felt throughout Japan but caused no damage.
2001 Nisqually, Washington (M6.8, depth ~52 km): The Nisqually earthquake occurred within the subducting Juan de Fuca Plate at a depth of approximately 52 km beneath the southern Puget Sound region. This intermediate depth significantly reduced the surface shaking intensity compared to what would have occurred if the same earthquake had been shallow. Damage was moderate — estimated at about $2 billion — with some structural damage to unreinforced masonry buildings and liquefaction-related damage in fill areas. One person died. Had the same earthquake occurred at 10–15 km depth, damage and casualties would likely have been far greater.
| Earthquake | Magnitude | Depth (km) | Deaths | Damage | Key Lesson |
|---|---|---|---|---|---|
| 2010 Haiti | M7.0 | ~13 | Est. 100,000–316,000 | Catastrophic; capital largely destroyed | Shallow + poor construction + dense population = worst case |
| 2015 Nepal (Gorkha) | M7.8 | ~8 | ~8,900 | $7B+; 600,000+ structures damaged | Very shallow rupture beneath populated valley |
| 1994 Northridge, CA | M6.7 | ~18 | 57 | ~$20B (1994 USD) | Moderate earthquake at shallow depth under a major city |
| 2001 Nisqually, WA | M6.8 | ~52 | 1 | ~$2B | Intermediate depth greatly reduced surface intensity |
| 2013 Sea of Okhotsk | M8.3 | ~609 | 0 | None | Extreme depth = no damage despite massive energy release |
| 2015 Bonin Islands | M7.9 | ~680 | 0 | None | Among deepest large earthquakes ever — negligible surface effects |
| 2018 Anchorage, AK | M7.1 | ~46 | 0 | Moderate infrastructure damage | Intermediate depth limited severity in populated area |
Deep Earthquakes: How They're Possible
The existence of deep earthquakes posed a paradox for seismologists for decades. At depths below about 50–70 km, the pressure and temperature conditions in the mantle should prevent brittle fracture — the mechanism by which shallow earthquakes occur. Under these conditions, rock should deform by ductile flow (creeping slowly) rather than snapping in sudden brittle failure.
Deep earthquakes occur exclusively in subducting oceanic lithosphere — the cold, rigid slabs of ocean floor that descend into the mantle at subduction zones. The slab remains cooler than the surrounding mantle for millions of years as it sinks, preserving conditions that can permit brittle-like failure.
The leading explanation for deep earthquakes involves mineral phase transformations. As the subducting slab descends, the increasing pressure causes the mineral olivine (the dominant mineral in the upper mantle) to transform into a denser crystal structure called wadsleyite (at ~410 km depth) and then ringwoodite (at ~520 km depth). These phase transitions involve a sudden volume reduction. In the cold interior of the slab, where temperatures lag behind equilibrium, the olivine-to-wadsleyite transition may occur suddenly in localized zones, creating a mechanical instability that triggers seismic slip. This is called the transformational faulting mechanism, proposed by geophysicists Harry Green and Chris Marone, among others.
The deepest earthquakes cluster around 660–700 km, corresponding to the boundary where ringwoodite transforms to the lower mantle minerals bridgmanite and ferropericlase. Below this depth, no further phase transformations are available to trigger seismicity, and the slab has heated sufficiently to deform ductilely. This explains the sharp cutoff of seismicity at approximately 700 km.
Epicentral Distance: Why a Closer M6 Can Be Worse Than a Farther M7
The distance from the earthquake source to a given location — called the epicentral distance for surface measurement or hypocentral distance for the 3D distance — is a primary determinant of shaking intensity at that location. Seismic wave amplitude decreases with distance due to geometric spreading and material attenuation.
This means that a M6.0 earthquake with an epicenter directly beneath a city can produce more intense shaking at that city than a M7.0 earthquake with an epicenter 100 km away, even though the M7.0 releases approximately 31.6 times more total energy. What matters locally is the energy that arrives at the surface at that specific location.
Ground motion prediction equations (GMPEs), also called attenuation relationships, quantify how seismic shaking intensity varies with magnitude, distance, depth, and local site conditions. These equations are the foundation of seismic hazard analysis and building code design maps. The USGS National Seismic Hazard Maps, which inform the seismic design requirements in the International Building Code, are built on ground motion models that account for the complex relationships between source, path, and site effects.
For seismic hazard in a particular city, the most dangerous earthquake is not necessarily the largest one that could occur in the region. It is the earthquake that delivers the highest shaking intensity to that specific location — which depends on the combination of magnitude, depth, distance, and local soil conditions. According to the Uniform California Earthquake Rupture Forecast (UCERF3), the probability of a M6.7 or larger earthquake in the San Francisco Bay Area between 2014 and 2043 is approximately 72%, and the probability of a similar event in Southern California is approximately 93%.
Fixed-Depth Assignments: Interpreting USGS Data
When the USGS reports an earthquake, the listed depth is sometimes a measured value and sometimes a default assignment. Understanding the difference is important for interpreting earthquake reports accurately.
Default 10 km depth: When a shallow earthquake occurs in a region with sparse seismograph coverage, the depth may be poorly constrained. Rather than report a depth with very large uncertainty, the USGS assigns a default depth of 10 km. This indicates that the earthquake is known to be shallow but its exact depth is uncertain. Many earthquake reports for events in remote oceanic regions or developing countries carry this default depth.
Default 33 km depth: Historically, a default depth of 33 km was used for shallow earthquakes when the depth was unconstrained. This value corresponds to a conventional estimate of average crustal thickness. While this default is less commonly used in modern practice (replaced in most cases by 10 km), it still appears in some older catalogs and for some current events.
When evaluating an earthquake report, if the listed depth is exactly 10.0 km or 33.0 km, it is worth checking whether this is a measured depth or a default assignment. The USGS typically notes the uncertainty or quality of the depth determination in its detailed event pages.
Depth Distribution of Global Earthquakes
The depth distribution of earthquakes is not uniform. Most of Earth's seismic energy is released at shallow depths, even though earthquakes occur throughout the upper 700 km of the mantle.
| Depth Range | Classification | Approximate % of Earthquakes | Approximate % of Seismic Energy Released | Tectonic Setting |
|---|---|---|---|---|
| 0–33 km | Very shallow | ~55% | ~70% | All plate boundaries, intraplate |
| 33–70 km | Shallow | ~15% | ~15% | Subduction zones, deep continental |
| 70–300 km | Intermediate | ~22% | ~12% | Subduction zones only |
| 300–500 km | Deep | ~5% | ~2% | Deep subduction (Tonga, S. America, Java) |
| 500–700 km | Very deep | ~3% | ~1% | Deepest subduction (Tonga-Kermadec, Bonin) |
The USGS locates approximately 20,000 earthquakes per year worldwide — roughly 55 per day. Of these, the vast majority are too small or too remote to cause damage. Approximately 15 large earthquakes (M7.0–7.9) and one or two major earthquakes (M8.0+) occur globally each year.
What causes earthquakes Understanding earthquake waves Earthquake magnitude scales explained Check your earthquake risk
USGS Earthquake Glossary IRIS Earthquake Depth Educational Resources