California's Next Big One: San Andreas Fault Scenarios

Somewhere beneath the Coachella Valley — the golf-course desert east of Los Angeles where Palm Springs sits in the shadow of the San Jacinto Mountains — a section of the San Andreas Fault has been locked for approximately 300 years. The Pacific plate is moving northwest relative to North America at roughly 25 millimeters per year. In 300 years at that rate, the two sides of the fault should have slipped about 7.5 meters relative to each other. They have not. That 7.5 meters of displacement is stored as elastic strain in the rock surrounding the fault — compressed into the crust like a coiled spring, waiting for the frictional resistance holding the fault surfaces together to finally give way.

When it does, the energy release will be roughly equivalent to the simultaneous detonation of several hundred nuclear weapons. The ground will move violently for 50 to 100 seconds. A fault scarp will open across the desert floor, through the Coachella Valley, through the San Bernardino Mountains, past the outskirts of San Bernardino and Ontario, and northwest through the Cajon Pass — a linear gash in the landscape running for 300 kilometers or more, displacing roads, rupturing pipelines, shearing off bridges and freeway overpasses mid-span. In Los Angeles, 60 to 100 kilometers from the fault trace, shaking will persist for more than a minute — far longer than the 1994 Northridge earthquake that caused $40 billion in damage from 15 seconds of shaking on a fault nobody knew existed. And then, in a city crisscrossed by aging natural gas infrastructure, fires will start. Many of them. Simultaneously. In neighborhoods where water pressure has dropped to zero because the mains serving the hydrants have been severed by ground displacement.

This is not a worst-case scenario conjured from speculation. It is the consensus scientific projection for what happens when the southern section of the San Andreas Fault ruptures — the Uniform California Earthquake Rupture Forecast (UCERF3) places the probability of such a rupture at approximately 60% within the next 30 years. The ShakeOut scenario, developed by USGS and California Geological Survey scientists, provides the most detailed engineering and consequence analysis of any probable earthquake event in American history. The question is not really whether this earthquake will happen. The question is whether California will be ready when it does.

The Fault: Anatomy of a 1,300-Kilometer Right-Lateral System

The San Andreas Fault runs the full length of California — approximately 1,300 kilometers from the Salton Sea in the south to Point Arena in the north, where it dives offshore and connects to the Mendocino Triple Junction and the Cascadia subduction zone. It is the principal boundary between the Pacific plate (moving northwest) and the North American plate (moving southeast relative to the Pacific), accommodating approximately 75–80% of the total plate motion — the remainder being distributed across a family of parallel faults including the Hayward, Calaveras, San Jacinto, and Elsinore fault systems.

The fault does not behave uniformly along its length. Three fundamentally different behavioral regimes exist, each with distinct seismic hazard implications. The northern section, from San Juan Bautista to Point Arena, is locked and last ruptured catastrophically in the 1906 San Francisco earthquake — an M7.9 that ruptured approximately 477 kilometers of fault with offsets of up to 6 meters and killed an estimated 3,000 people. The central section, from Parkfield south to San Juan Bautista, creeps aseismically at approximately the full plate rate — it accommodates slip continuously without storing elastic strain and rarely produces large earthquakes. The southern section, from roughly Parkfield to the Salton Sea, is fully locked — it has not produced a great earthquake in the instrumental record and has been accumulating elastic strain since its last major rupture in approximately 1680.

Paleoseismology: Reading the Fault's History

The most direct evidence for the southern San Andreas fault's behavior comes from paleoseismology — the excavation and analysis of trenches dug across the fault trace to expose and date the sedimentary record of past surface ruptures. The definitive paleoseismic record for the southern San Andreas comes from Pallett Creek, a site in the Mojave Desert northeast of Los Angeles where the geologist Kerry Sieh began excavating in the 1970s and where subsequent work has extended the earthquake history back approximately 1,800 years.

The Pallett Creek record documents ten to twelve major surface-rupturing earthquakes over the past 1,800 years — a remarkable dataset that represents the longest and most precisely dated paleoseismic record available for any fault on Earth. The dates of the most recent events are approximately 1680 CE, 1470 CE, 1245 CE, 1100 CE, and 1000 CE — a rough average recurrence of one great earthquake every 130–150 years, though with substantial variability (some intervals as short as 45 years, others as long as 330 years). The penultimate event in 1470 and the most recent in 1680 define the current interseismic gap: as of today, the southern San Andreas has not ruptured for approximately 345 years — already at the longer end of the historical recurrence range.

Multiple Site Constraints and Along-Strike Variability

Pallett Creek is the most celebrated paleoseismic site on the southern San Andreas, but it is not the only one. Additional sites including Wrightwood, Indio, and Thousand Palms have extended the record and revealed important along-strike variability in rupture behavior. Not every earthquake in the Pallett Creek record is present at every site — some events apparently ruptured only part of the southern San Andreas, while others appear to have ruptured the full length from the Salton Sea to Cajon Pass or beyond. This variability is consistent with the segmented rupture behavior documented in the instrumental record of other large fault systems and complicates the characterization of the next event: will it rupture the full 350-kilometer length of the southern San Andreas, or a shorter segment?

The answer matters enormously for magnitude estimation. A full southern San Andreas rupture — from the Salton Sea through Cajon Pass and potentially continuing onto the central section toward Parkfield — could reach M7.9–8.0. A rupture limited to the Coachella Valley and San Bernardino Mountain segments might produce an M7.5–7.7. The ShakeOut scenario uses an M7.8 as its planning earthquake, based on the most likely partial-to-full southern section rupture, which is consistent with the typical rupture length for events of that magnitude on the fault geometry defined by geologic mapping and geodetic modeling.

The Slip Deficit: Measuring Stored Strain

Geodetic measurements — primarily continuous GPS networks deployed across Southern California — provide a quantitative picture of where elastic strain is accumulating on the southern San Andreas. The pattern is unambiguous: the fault trace is locked from the Salton Sea to at least the Cajon Pass, accumulating slip deficit at approximately 20–25 mm/year. With the last major rupture approximately 345 years ago, the total accumulated slip deficit is estimated at 6–9 meters — consistent with the surface displacements measured at paleoseismic sites for comparable events in the fault's history.

GPS measurements also reveal the three-dimensional deformation field around the locked fault: stations east of the fault (on the North American plate) are being dragged westward and northwestward by the coupling; stations west of the fault (on the Pacific plate) are moving northwestward freely. The gradient of velocity between the two sides — the velocity gradient that relaxes instantly when the fault ruptures — provides a direct measurement of the stored elastic energy. Strain energy density models based on these GPS measurements indicate that the stored elastic energy on the southern San Andreas is equivalent to approximately 10–30% of the total energy that would be released in an M7.8 earthquake.

UCERF3: California's Earthquake Probability Framework

The Uniform California Earthquake Rupture Forecast, Version 3 (UCERF3), published by the USGS and California Geological Survey in 2013, is the authoritative scientific assessment of earthquake probabilities in California. It synthesizes fault geometry, slip rates, paleoseismic recurrence data, geodetic strain measurements, and historical seismicity into a comprehensive probabilistic framework for estimating the likelihood of earthquakes of various magnitudes throughout the state over 30-year and longer time horizons.

UCERF3's key findings for the southern San Andreas include a 60% probability of a M6.7 or greater earthquake on the southern San Andreas system (including the San Jacinto and Elsinore faults) within the next 30 years. For the full southern San Andreas fault specifically, the 30-year probability of an M7.5+ rupture is estimated at approximately 19–22%. The probability of an M6.7+ somewhere on the combined southern California fault network — including the San Andreas, Hayward, San Jacinto, Elsinore, Garlock, and dozens of secondary faults — is approximately 93% over 30 years. Southern California is not asking whether a major earthquake will occur. It is asking which fault breaks first and how large it will be.

The Hazard Map: Ground Motion Predictions

UCERF3 feeds into probabilistic seismic hazard analysis (PSHA) that produces the National Seismic Hazard Maps — maps of peak ground acceleration (PGA) and spectral acceleration expected to be exceeded with specified probabilities over 50-year periods. For Southern California, the 2% probability of exceedance in 50 years (a roughly 2,500-year return period) ground motion map shows expected PGA values of 0.6–1.0g throughout the greater Los Angeles area — enough to cause severe damage to all but the most modern, well-engineered structures. In the near-field of the southern San Andreas (within 10–20 km of the fault trace), PGA values for the characteristic earthquake could reach 1.5–2.0g, with velocity pulses that can knock down structures that survive peak acceleration alone.

The ShakeOut Scenario: A Detailed Rupture Simulation

The ShakeOut scenario, first published in 2008 by a team led by Lucy Jones at the USGS Pasadena office, is the most comprehensive earthquake consequence analysis ever performed for a North American city. It describes a specific hypothetical M7.8 earthquake nucleating on the southern San Andreas near Bombay Beach at the northern end of the Salton Sea and propagating northwest through the Coachella Valley, through the San Bernardino Mountains via the San Gorgonio Pass area, through the Cajon Pass, and ending near Wrightwood — a rupture of approximately 300 kilometers with maximum surface displacement of 6–8 meters.

Ground Motion Modeling

The ShakeOut ground motion simulations used physics-based wave propagation models incorporating the three-dimensional velocity structure of the Los Angeles basin and surrounding region — including the deep sedimentary basin fill of the LA basin that amplifies long-period waves by factors of 3–5 relative to surrounding bedrock. The results show Modified Mercalli Intensity (MMI) values of IX (violent shaking) in the near-field along and adjacent to the fault trace, VIII (severe shaking) throughout the San Bernardino and Inland Empire region, VII (very strong shaking) across most of metropolitan Los Angeles, and VI (strong shaking) extending to San Diego, Phoenix, and Las Vegas. The duration of strong shaking in the Los Angeles basin exceeds 60 seconds in the simulation — three to four times longer than the 1994 Northridge event, which itself caused $40 billion in damage in 15 seconds.

Immediate Casualties and Injuries

The ShakeOut scenario projects approximately 1,800 deaths and 50,000 injuries under the baseline daytime scenario — 33 million people exposed to the shaking at a time when most are at work or school rather than sleeping in unreinforced masonry structures. A nighttime scenario, with populations in residential buildings, produces higher casualty estimates because residential building quality in Southern California is more variable than commercial construction and because people are less able to take protective action while asleep. The deaths are concentrated in areas with high densities of unreinforced masonry buildings — predominantly older neighborhoods in the eastern Inland Empire and portions of Los Angeles and San Bernardino counties along the fault corridor.

These casualty projections are often cited as evidence that the "Big One" will be survivable — 1,800 deaths in a metro area of 18 million is a much lower casualty rate than the 2010 Haiti earthquake or the 2023 Turkey-Syria events. But the projection depends critically on the time of day, on building occupancy, and — most importantly — on what happens in the hours and days after the initial shaking, when fire following earthquake, water system failure, and disrupted emergency medical services may produce as many or more deaths and injuries as the direct structural damage.

Fire Following Earthquake: The Hidden Multiplier

Of all the secondary hazards in the ShakeOut scenario, fire following earthquake is the most consequential and the least discussed in public preparedness messaging. The scenario projects that the earthquake will ignite approximately 1,600 fires across Los Angeles County simultaneously — from ruptured gas lines, overturned appliances, downed electrical wires, and chemical spills in industrial areas. Under typical Southern California conditions (low humidity, Santa Ana winds, dry vegetation), the city's fire department has the suppression capacity to handle approximately 10–15 simultaneous significant fires.

With 1,600 fires igniting in the first hour and no water pressure in large portions of the distribution system (because the water mains have been ruptured), the scenario projects that fires will merge into conflagrations burning unchecked through residential neighborhoods for days. The ShakeOut analysis estimated that fire could ultimately destroy more square footage of buildings than the direct earthquake shaking — in the worst-case fire scenario, accounting for more deaths than the structural collapses in the first minutes of the event. This is not hypothetical: the 1906 San Francisco earthquake killed roughly 300 people directly from structural collapse and approximately 2,700 from fire that burned for three days afterward, destroying 25,000 buildings across 500 city blocks.

Infrastructure: The Cascading Failures

Beyond the immediate structural damage and fire hazard, the ShakeOut scenario traces a web of cascading infrastructure failures that would make the post-earthquake environment in Southern California unlike anything the region has experienced.

Transportation

The scenario projects collapse or severe damage to approximately 300 freeway bridges and overpasses in the most heavily shaken areas, including multiple crossings of the fault trace itself on Interstate 10, Interstate 15, and State Route 62. The Cajon Pass — through which all rail lines and most highways connecting Southern California to the rest of the country pass — runs directly along the San Andreas Fault for several kilometers and would be closed indefinitely by fault surface rupture and slope failures. Southern California would be functionally severed from overland freight and passenger connection to the rest of the continental United States for weeks to months, until the fault crossing could be cleared and rebuilt.

Water Systems

The Metropolitan Water District of Southern California imports approximately 50% of the region's water from Northern California and the Colorado River through aqueducts that cross the San Andreas Fault multiple times. The State Water Project's Inland Feeder pipeline crosses the fault at multiple points; the Colorado River Aqueduct crosses at several locations in the Coachella Valley that would experience some of the highest shaking intensities in the scenario. The ShakeOut analysis estimated that Southern California could face a regional water shortage within two to three days of the earthquake, with full restoration of imported water supply taking six months to over a year.

Local water distribution would fail even faster. In the near-field shaking zone, liquefaction of alluvial soils and direct fault displacement would rupture water mains at hundreds of points, draining reservoirs and eliminating pressure from the distribution system. The scenario projects that approximately 50% of households in the most affected areas would be without running water for at least two weeks — a figure that interacts catastrophically with the fire hazard and creates a sanitation crisis that would strain the public health system independently of earthquake injuries.

Natural Gas and Utilities

Southern California's natural gas distribution system is an aging network of steel and cast-iron pipes running beneath streets that will experience severe ground deformation in the scenario. The scenario projects tens of thousands of gas leaks developing in the hours following the earthquake, with SoCalGas estimating it would take weeks to locate and shut off all leaking service lines. Each leak is a potential fire ignition source; in aggregate, they constitute the primary driver of the fire following earthquake hazard. SoCalGas and the California Public Utilities Commission have been implementing automatic shutoff valves and replacing older pipe materials for years, but the extent of remaining vulnerable infrastructure remains large.

The Hayward Fault: Northern California's Comparable Threat

While the southern San Andreas dominates discussion of California's "next big one," the Hayward Fault in the East Bay of the San Francisco Bay Area represents a comparably severe scenario for Northern California. The Hayward last ruptured in 1868, producing an estimated M6.8 that destroyed the small East Bay communities of the era. With 150 years of subsequent development on and adjacent to the fault, the same rupture today — the USGS HayWired scenario models an M7.0 on the Hayward — would affect approximately 7 million people in the Bay Area, projecting 800 deaths, 18,000 injuries, and approximately $82 billion in losses. The Hayward Fault runs directly beneath or adjacent to Oakland, Berkeley, Hayward, Fremont, and portions of San Jose — some of the most densely populated urban zones in the state.

The Hayward also has the unsettling distinction of being the only fault in California that creeps at the surface in some sections while remaining locked at seismogenic depth — a mechanical configuration that makes it particularly capable of hosting a large earthquake without much obvious precursory seismic activity. The creeping portion at the surface actually helps maintain a stress concentration at the locked-to-creeping transition at depth, which is thought to be one of the preferred nucleation points for a future large event.

The Southern San Andreas vs. Cascadia: Two Very Different "Big Ones"

California's "Big One" is often discussed in the same breath as the Cascadia subduction zone earthquake facing the Pacific Northwest, but the two events — while both catastrophic — have fundamentally different characters that lead to very different hazard profiles.

The Cascadia scenario involves an M9.0+ megathrust rupture spanning 1,100 kilometers from Northern California to Vancouver Island, generating 3–15 meter tsunamis that would reach coastal communities within 15–30 minutes and kill thousands of people who cannot evacuate in time. Cascadia's population exposure is dominated by coastal communities and the Portland and Seattle metro areas at moderate distance from the fault. The total energy release would dwarf the San Andreas event by a factor of 30–100. But Cascadia's lower population density in the near-field and the distributed nature of the shaking mean that the initial structural collapse deaths, while severe, may be lower than a direct San Andreas hit on the Los Angeles basin.

The San Andreas scenario involves a shorter, shallower fault — the locked zone is only 15–20 km deep — with less total energy but far greater population density in the near-field. The 18 million people in the greater Los Angeles metro area represent the largest urban concentration adjacent to a locked strike-slip fault anywhere in the United States. No tsunami threat. But fire, water failure, and the specific vulnerability of the LA basin's soft sediment amplification zones make the near-field consequences of a San Andreas rupture potentially more concentrated in casualties and economic disruption than a Cascadia event of ten times the energy.

What Has Changed Since 1906: Progress and Remaining Gaps

A century of seismic engineering and policy development has substantially improved Southern California's earthquake resilience relative to what existed in 1906. The Uniform Building Code adopted mandatory seismic design requirements in the 1930s, progressively strengthened after each major earthquake. California's Field Act (1933), passed after the Long Beach earthquake, set seismic standards for school construction that have held up remarkably well in subsequent earthquakes. The USGS ShakeAlert earthquake early warning system now provides seconds to tens of seconds of warning before shaking arrives at a given location — enough time to stop trains, open fire station doors automatically, and give people in low-shaking areas time to take protective action.

Unreinforced Masonry: The Remaining Vulnerability

The most significant remaining structural vulnerability in Southern California is unreinforced masonry (URM) — brick and concrete block buildings constructed without steel reinforcement, built before modern seismic codes and concentrated in older commercial and residential districts throughout the region. California has no statewide mandatory URM retrofit law — the 1986 URM law required only that cities inventory their URM buildings, not that they mandate retrofit. Individual cities have moved at very different speeds: Los Angeles completed its URM retrofit program in 2015, retrofitting or demolishing approximately 11,000 buildings. San Bernardino and other Inland Empire cities with high concentrations of URM in the near-fault zone have retrofitted far fewer.

The ShakeOut scenario projects that URM buildings will account for a disproportionate share of structural collapses and deaths — not because they are the most common building type but because their failure mode (sudden, catastrophic wall collapse with little warning and no ability for occupants to take cover in time) is uniquely lethal. A retrofitted or modern building may suffer significant damage in the scenario but will generally remain standing long enough for occupants to escape. A URM building may collapse within the first 10 seconds of shaking, trapping everyone inside.

Soft-Story Wood Frame Buildings

A second category of seismically vulnerable buildings in the Los Angeles region is the soft-story wood frame apartment building — typically a two- to five-story structure with parking or commercial space on the ground floor that creates a structurally weak "soft story" at the base. During the 1994 Northridge earthquake, hundreds of these buildings collapsed at the ground floor level, pancaking upper floors onto the parking level and killing residents in upper-floor apartments who had no warning of the structural failure beneath them.

Los Angeles passed a mandatory soft-story retrofit ordinance in 2015 — the most aggressive residential seismic retrofit program in American history — requiring approximately 13,500 buildings to be retrofitted within 25 years. Progress has been substantial, with tens of thousands of units retrofitted, but the program is not yet complete and many buildings in other Southern California jurisdictions without similar ordinances remain unretrofitted.

Preparedness: What It Actually Takes

The ShakeOut scenario's consequence analysis translates directly into practical preparedness requirements — not just for individuals but for the systems that will govern survival in the post-earthquake environment. The gap between California's current preparedness state and what the scenario demands is large but not unbridgeable.

Individual and Household Preparedness

The scenario projects that most of the Southern California region will be without reliable water supply for two weeks to several months. Every household should maintain a minimum two-week water supply — one gallon per person per day — and a food supply that does not depend on refrigeration or utilities. The California Governor's Office of Emergency Services recommends a 72-hour kit as a minimum; the ShakeOut scenario makes clear that two weeks is a more realistic minimum for water and food self-sufficiency in the most affected areas.

Structural survivability — actually being alive after the shaking — depends on building type, building location, and physical behavior during shaking. Drop, cover, and hold on under a sturdy table remains the correct response for most indoor locations: it minimizes injury from falling objects (the primary cause of indoor earthquake injuries) without the risks associated with running outside (falling exterior debris, glass) or standing in a doorway (a myth with no structural basis in modern construction).

Community and Regional Resilience

The most consequential preparedness investments at the regional scale are the infrastructure hardening programs that directly reduce the severity of cascading failures. Metropolitan Water District's Inland Feeder hardening program, SoCalGas's automatic shutoff valve installations, Caltrans's ongoing bridge seismic retrofit program, and the hospital retrofit requirements under SB 1953 each directly reduce the consequence chains that turn a survivable earthquake into a prolonged regional catastrophe. The rate at which these programs proceed — and the political will to fund them ahead of the event rather than after — is the primary determinant of how many people die in the 24 hours after the shaking stops, as opposed to the first 90 seconds of it.

The Probabilistic Reality: When, Not If

The framing of the San Andreas earthquake as "the Big One" — a singular, inevitable, apocalyptic event — is both useful and misleading. Useful because it communicates genuine urgency about a real and consequential hazard. Misleading because it implies a binary world in which nothing happens until this one specific event, after which everything changes. The reality is probabilistic and continuous: Southern California faces a graduated spectrum of earthquake hazards on dozens of faults, with the southern San Andreas being the highest-consequence single scenario but not the only one, and with moderate earthquakes on secondary faults occurring every several years in the region.

The UCERF3 probability of a M6.7+ earthquake somewhere in the Los Angeles region within 30 years is approximately 93%. That means a damaging but not catastrophic earthquake is more likely than not within the professional career of anyone reading this today. It will test building retrofits, stress infrastructure, strain emergency services, and either validate or expose the preparedness investments made in the interseismic quiet period. Whether the southern San Andreas goes in the next thirty years or the next hundred, the probability calculus is unambiguous: the region is not safe, the risk is quantifiable, and the gap between current preparedness and what the scenario requires is a policy choice that California makes every year it defers the next round of investment.

Conclusion

The southern San Andreas Fault is not a geological curiosity or a distant threat. It is a fully loaded mechanical system — a 300-kilometer-long fault interface carrying 300 years of accumulated plate motion in elastic storage — positioned adjacent to one of the largest metropolitan areas on Earth. The paleoseismic record establishes that it ruptures on a timescale of centuries, not millennia. The geodetic record establishes that it has stored sufficient elastic strain for an M7.8 event. The probabilistic forecast gives it a 19–22% chance of delivering that event in the next 30 years.

The ShakeOut scenario translates that probability into operational specificity: 1,800 deaths, 50,000 injuries, 1,600 simultaneous fires, months without reliable water, closed freeways, disrupted hospitals, and an economic shock projected at $200 billion or more. Each of those numbers is reducible by investments made before the event — in building retrofits, infrastructure hardening, water system redundancy, and community preparedness. The mathematics of earthquake risk are clear, the policy tools are available, and the only genuinely open question is whether the political and individual will to act on probabilistic hazard will keep pace with the fault's slow, relentless accumulation of the energy it will eventually release.