Baldwin Hills Reservoir — Built Over an Active Fault and Oilfield Subsidence, Breached in 1963

Summary

On the afternoon of 14 December 1963, the Baldwin Hills Reservoir — a lined earthfill impoundment perched on a low hilltop above the Baldwin Hills district of south Los Angeles — tore open and released the better part of its 250-million-gallon contents into the streets below, killing five people and destroying 277 homes. The embankment did not fail by overtopping, slope instability, or poor compaction. It failed because the ground beneath it moved. An active branch of the Newport–Inglewood fault system, running directly under the reservoir floor, offset roughly seven inches over the structure's twelve-year life, and on that day the displacement finally cracked the thin asphaltic membrane that was the reservoir's only barrier against its own erodible foundation. Water found the crack, piped through the soil beneath the lining, scoured a channel under the embankment, and blew through the dam in hours.

The reservoir was completed in 1951 by the Los Angeles Department of Water and Power, an agency whose memory still carried the 1928 St. Francis Dam catastrophe. Its designers knew the site sat within the Inglewood fault zone and did not treat that as disqualifying. They engineered around it: a compacted-earth bowl, a brittle asphalt-membrane liner over a gravel drainage blanket, and an underdrain system threaded with inspection pipes, all meant to catch and channel any seepage before it reached the loose, sandy, highly erodible foundation soils. It was a monitoring strategy substituted for a geological one, resting on two assumptions — that the fault would not move enough to matter, and that the underdrains would buy time if it did. Both failed on the same day. The fault ruptured the liner; the underdrains performed exactly as designed, with the caretaker seeing muddy water at the pipes at about 11:15 that morning, but the internal erosion was already beyond stopping. Operators dropped the reservoir and police evacuated some 1,600 residents within roughly four hours — the reason five died instead of a figure estimates placed in the hundreds. At about 15:38 the embankment gave way, and the reservoir emptied in seventy-seven minutes.

What made Baldwin Hills a landmark was the second half of its diagnosis. The fault had not been moving on its own geological schedule. Decades of oil extraction from the adjacent Inglewood Oil Field — and, critically, the high-pressure waterflooding injected to drive out the remaining crude — had withdrawn support from the strata, induced regional subsidence on the order of feet, and reactivated the very faults the reservoir straddled. A 1976 U.S. Geological Survey study concluded that 90 percent or more of the surface displacement around the dam was caused by exploitation of the oil field. Baldwin Hills became the canonical American case binding human-induced ground subsidence to the failure of a major water-retaining structure, and it ended the faith that a flexible lining could absorb a fault offset across an active trace.

---

Timeline

1924

Inglewood field discovered

Oil production begins on the Inglewood field flanking the Baldwin Hills, initiating decades of fluid withdrawal and the slow regional subsidence later implicated in the failure.

1947–1951

Reservoir constructed

The Los Angeles Department of Water and Power builds the Baldwin Hills Reservoir on a hilltop: a compacted-earth bowl roughly 232 ft high and 650 ft long impounding about 250 million gallons, lined with an asphaltic membrane over a gravel drainage blanket and underdrain system.

1951

Reservoir placed in service

The reservoir enters service supplying south and southwest Los Angeles. Designers acknowledge the Inglewood fault zone beneath the site but judge any fault movement too small to threaten the lining.

1950s

Waterflooding intensifies

Operators on the Inglewood field expand high-pressure water injection (waterflooding) to boost oil recovery, increasing pore pressures and ground movement along nearby faults.

1951–1963

Progressive fault offset

Over the reservoir's service life an active branch of the Newport–Inglewood fault beneath the floor offsets roughly 7 inches, while regional subsidence stretches the ground; cracks appear at the surface across the district.

Early 1963

Surface distress observed

Cracking and offset are noted in pavements and structures around the Baldwin Hills, evidence of accelerating ground deformation that is not connected to the reservoir's safety in time.

14 Dec 1963, ~11:15

Leak detected at underdrains

During a routine inspection the caretaker observes muddy water discharging from the underdrain pipes — the design's intended early warning. The increasing, sediment-laden flow signals that soil is being eroded beneath the lining.

14 Dec 1963, ~midday

Emergency drawdown ordered

DWP begins lowering the reservoir as fast as the outlet works permit and notifies authorities. Police and firefighters start evacuating the downstream neighborhoods.

14 Dec 1963, early afternoon

Evacuation completed

Roughly 1,600 residents are cleared from the path of a potential flood within about four hours — the action that holds the eventual death toll to five.

14 Dec 1963, ~15:38

Embankment breaches

A crack opens through the earth dam and rapidly enlarges as water pipes through the foundation; the embankment gives way and releases the impounded water down into the streets below.

14 Dec 1963, ~16:55

Reservoir emptied

The breach drains the reservoir in about 77 minutes. Five people are killed, 277 homes destroyed, and property damage runs to roughly $11–12 million.

1964–1976

Investigations and verdict

State and county boards of inquiry attribute the failure to fault offset and induced subsidence rupturing the lining; a 1976 USGS study assigns 90%+ of the ground displacement to oil-field exploitation. The reservoir is never rebuilt.

The Build: A Monitoring System Where a Geology Was Needed

Baldwin Hills was a hilltop reservoir, not a valley dam. The Department of Water and Power needed elevated storage to serve the southwest of the city by gravity, so it excavated a bowl into the crest of the Baldwin Hills and ringed it with a compacted-earth embankment standing about 232 feet at its highest and some 650 feet long, impounding roughly 250 million gallons. Built between 1947 and 1951, it was the work of the same agency that had owned the St. Francis Dam when that structure collapsed in 1928 and killed hundreds — and the design reflected an agency determined to make its earthwork leak-proof and watched.

The defining feature was the lining. The foundation soils were loose and sandy, the kind that erodes catastrophically once moving water reaches them. To keep water off that soil, the designers laid a thin, brittle asphaltic membrane across the reservoir floor and inner slopes, bedded on a gravel drainage blanket, and beneath it ran an underdrain system — a grid of perforated pipes meant to intercept any seepage and carry it, visibly, to inspection outlets a caretaker could read. The philosophy was explicit: the membrane was the barrier, the underdrains were the alarm. The one fact the design accommodated rather than confronted was the geology. The site lay squarely within the Inglewood fault zone, a branch of the Newport–Inglewood system, with active faults beneath the floor. The designers knew this and reasoned the movement would be too small and slow to damage the lining. In that judgment the reservoir's safety rested on a membrane a few inches thick spanning ground everyone agreed could move — the investigation had answered "is there a fault here?" with "yes," then proceeded as though the answer changed nothing.

The Failure Sequence: A Fault Reactivated, a Lining Torn, a Channel Piped Open

For twelve years the reservoir held, and for twelve years the ground under it deformed. The Inglewood Oil Field, immediately adjacent, had produced since the 1920s, and by the 1950s its operators were waterflooding — injecting water at high pressure to sweep additional oil toward the wells. Fluid withdrawal pulled the strata down into regional subsidence measured in feet; the injection raised pore pressures and reactivated faults in the same rock mass. The net effect over the reservoir's life was ground that subsided, stretched, and offset along the faults beneath the floor by roughly seven inches. The membrane meant to be the structure's only barrier was being slowly pulled apart by the earth it rested on.

On 14 December 1963 the accumulated displacement reached the membrane's limit. The brittle asphaltic lining cracked along the fault, and reservoir water drove into the gravel blanket and reached the erodible foundation soil. At about 11:15 that morning the caretaker saw the result exactly where the design said it would appear: muddy water at the underdrain pipes. The drains worked — but the sediment was the tell. It meant the foundation was already washing away, that an open seepage channel had formed beneath the lining, and that internal erosion — piping — was carving a conduit backward toward the reservoir. Piping does not wait: once a continuous channel exists, every cubic foot of escaping water enlarges it and accelerates the flow that enlarges it further. Operators dropped the reservoir as fast as the outlets allowed and the neighborhoods were cleared, but the dam could not be saved. In the early afternoon a crack opened through the embankment crest above the channel, widened in minutes, and at about 15:38 the dam breached, draining in roughly seventy-seven minutes and sending a flood through the streets that destroyed 277 homes. Five people who had not gotten clear were killed; without the underdrain warning and the four-hour evacuation, estimates of the toll ran into the hundreds and beyond.

The Reckoning: Naming the Man-Made Ground Beneath the Failure

The investigations — by the California Department of Water Resources, a county board, and engineers including the geotechnical specialist Thomas Leps — converged on a mechanism that was, in its proximate stages, unremarkable: foundation movement cracked a lining, water reached erodible soil, piping carved a channel, and the embankment breached. What made Baldwin Hills exceptional was the ultimate cause. The fault had not simply crept on a natural clock; it had been driven, the subsidence and ground extension that reactivated it being products of oil extraction and waterflooding next door. That conclusion was contested at the time, because it assigned a dam failure to industrial activity off the site and run by another party. The geological evidence hardened over the following years: surveys documented subsidence across the district measured in feet, fresh cracking radiating from the oil field, and offset that tracked production and injection history rather than regional seismicity. In 1976 a U.S. Geological Survey study settled the matter for the record — 90 percent or more of the surface displacement around the dam was caused by exploitation of the Inglewood oil field. The fault was real, but the increment of movement that killed the reservoir was, in large part, man-made.

That verdict reframed the original design error. The mistake was not that the engineers missed the fault; they mapped and acknowledged it. The mistake was the decision to retain water above an erodible foundation, across an active fault, behind a brittle membrane, in a basin where a neighboring operator was actively deforming the ground — and to treat an underdrain alarm as a substitute for not building there at all. The lining and drains were a monitoring system standing in for a geological judgment. They reported the failure faithfully; they could not prevent it. Baldwin Hills entered the literature as the case in which a reservoir's safety depended on ground its owners did not control and on movement another industry was inducing.

---

Contributing Factors

01

A reservoir sited across an active fault

An active branch of the Newport–Inglewood fault zone ran directly beneath the reservoir floor. The designers identified it and judged its likely movement too small to threaten the structure, then placed the impoundment's only barrier — a brittle lining — across the trace of a feature certain to move. The most hazardous geological element on the site was crossed rather than avoided.

02

Induced subsidence and fault reactivation from oil-field operations

Decades of extraction and high-pressure waterflooding on the adjacent Inglewood field withdrew support from the strata, caused regional subsidence on the order of feet, and reactivated the faults beneath the reservoir. The seven inches of offset that ruptured the lining were, by the later USGS finding, 90%-plus man-made — a deformation source entirely outside the reservoir's design envelope and outside its owner's control.

03

A brittle membrane as the sole barrier over erodible soil

The foundation was loose, sandy, highly erodible material, and the only thing keeping water off it was a thin asphaltic lining incapable of accommodating fault offset without cracking. Once the membrane fractured along the moving fault, nothing stood between the reservoir's head and a foundation that erodes the instant water reaches it. The barrier had no ductility to match the ground it spanned.

04

Detection substituted for prevention

The underdrain and inspection-pipe system was a sophisticated alarm, designed to make any seepage visible. It worked exactly as intended on the day of failure. But a system that reports erosion cannot stop it once piping has begun. The design relied on early warning where it needed an unconditional barrier, and the warning arrived too late in the erosion process to save the dam.

05

Unarrestable internal erosion (piping)

When fault offset cracked the lining, water reached the erodible foundation and began carving a seepage channel backward from the underdrain outlets — piping. The mechanism is self-accelerating: each increment of flow enlarges the channel, which increases the flow. By the time muddy discharge confirmed the process, the channel was established and the breach was hours away regardless of drawdown. ---

Aftermath

The Baldwin Hills failure killed five people, destroyed 277 homes, drove roughly 1,600 from their neighborhoods, caused some $11–12 million in damage, and disrupted water service to a large part of south Los Angeles. The reservoir was never rebuilt; its scarred basin was eventually reshaped into the Kenneth Hahn State Recreation Area. Its lasting weight, however, was technical and regulatory. Baldwin Hills became the American textbook case linking human-induced ground subsidence to the failure of a major water-retaining structure, forcing dam engineering and dam-safety regulation to treat anthropogenic ground movement — from oil, gas, groundwater, and mineral extraction — as a foundation hazard on par with natural geology. It reinforced, alongside the much larger contemporaneous Vajont and Malpasset cases, the principle that an embankment's competence is inseparable from the ground beneath it and the activities deforming that ground. In California, where the Division of Safety of Dams already carried the memory of St. Francis, the disaster hardened the prohibition against siting impoundments over active faults and against trusting a flexible lining to bridge an offset that the foundation could not. The reservoir's name became shorthand for a specific, uncomfortable lesson: a dam can be killed by ground its builders never touched.

---

Lessons

Do not impound water across an active fault on the assumption it will not move enough to matter: the offset you dismiss as small is measured against a brittle barrier that has no tolerance for any offset at all.
Treat induced ground deformation as a design load: map the extraction, injection, and subsidence happening around a site, and assume neighboring industry can reactivate faults and stretch the ground beneath your structure regardless of your property line.
Never let a monitoring system stand in for a barrier: an underdrain that reports seepage is not a defense against piping — once erosion of an erodible foundation begins, detection only tells you how the structure will fail, not whether.
Match the ductility of the barrier to the movement of the ground: a thin, brittle lining over erodible soil is a liability across any feature that can deform; if the foundation can move, the barrier must either accommodate that movement or the foundation must be made non-erodible.
Once piping starts in a loose foundation, assume the structure is lost and evacuate immediately: the early, decisive evacuation at Baldwin Hills is the reason five died instead of hundreds — treat the first muddy discharge as a breach in progress, not a problem to be managed. ---