The rise and fall of the St. Francis Dam

Material Information

The rise and fall of the St. Francis Dam
Paling, Dawn Michelle
Publication Date:
Physical Description:
ix, 168 leaves : illustrations ; 28 cm


Subjects / Keywords:
Dam failures -- History -- California -- Los Angeles County ( lcsh )
Dam failures ( fast )
History -- Saint Francis Dam (Calif.) ( lcsh )
California -- Los Angeles County ( fast )
California -- Saint Francis Dam ( fast )
History. ( fast )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )
History ( fast )


Includes bibliographical references (leaves 166-168).
General Note:
Department of Civil Engineering
Statement of Responsibility:
by Dawn Michelle Paling.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
60403901 ( OCLC )
LD1190.E53 2004m P34 ( lcc )

Full Text
Dawn Michelle Paling
B.S., Colorado School of Mines, 2001
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
for the degree of
Master of Science
Civil Engineering

This thesis for the Master of Science
degree by
Dawn Michelle Paling
has been approved
Brian Brady
7 Xcc-y

Paling, Dawn Michelle (M.S., Civil Engineering)
The Rise and Fall of the St. Francis Dam
Thesis directed by Assistant Professor Kevin L. Rens
On March 12, 1928, one of the worst civil engineering disasters of the 20th
century occurred suddenly. The St. Francis Dam, located northwest of Los Angeles,
California, released 15 billion gallons of water into San Francisquito Canyon, taking
the lives of over 400 people and leaving a wake of destruction. With the collapse
came a myriad of questions about how a failure so catastrophic could occur. This
paper summarizes the history of the St. Francis dam, the failure theories developed,
and the modifications in engineering practice.
The St. Francis Dam was designed and constructed by William Mulholland.
The Dam was designed as a stepped concrete gravity arch with a reservoir capacity of
30,000 acre-feet. Although the site chosen for the dam had questionable geological
characteristics, the topography and location greatly influenced the decision to build
the dam in San Francisquito Canyon. Construction of the dam began in the spring of
1924 and was completed two years later. However, the life of the dam was short
lived after failing suddenly on March 12, 1928.

A collaboration of theories of why the St. Francis Dam failed is brought
together in this report. Initially, geology was to blame due to the differing rock
formations, historical landslides, and fault line located underneath the dam. Design
modifications during construction were also considered questionable. In addition to
the theories above, this investigation will cover other conjectures that have been
developed since the failure.
The lessons learned from the St. Francis Dam failure brought about several
changes in the way future dams were to be designed. Geological surveys of potential
dam sites became an integral part of the design process. Uplift acting on the dam
base became a major design consideration resulting in deeper foundations and
seepage prevention. California added new laws that required any proposed dam to be
evaluated by an independent review panel. The changes brought about by the St.
Francis Dam failure has made the design of future dams safer, and prompted the
retrofit of dozens of existing dams in the effort to prevent similar disasters from
This abstract accurately represents the content of the candidates thesis. I recommend
its publication.

1. Introduction.........................................................1
2. Water Crisis.........................................................3
2.1 Alternative Water Sources............................................6
2.2 Location of the St. Francis Dam......................................8
2.3 Geology.............................................................10
3. William Mulholland..................................................15
4. Design..............................................................21
4.1 Design Concrete vs. Earth.........................................21
4.2 Design Specifications...............................................23
5. Initial Dam Distress................................................29
5.1 Events Prior to Failure.............................................32
5.2 Failure of the Dam..................................................35
6. Theories Behind the Failure.........................................43
6.1 Sabatoge Theory.....................................................43
6.2 Unstable Design Theory..............................................48
6.3 Inadequate Construction Techniques Theory...........................54

6.4 Inadequate Geology
6.4.1 West Abutment Failure Theory........................................61
6.4.2 East Abutment Failure Theory........................................64
7. Lessons Learned.....................................................71
7.1 California Law......................................................71
7.2 Dam Retrofits.......................................................75
7.3 Uplift Forces.......................................................76
8. Other Dam Failures..................................................78
8.1 Malpasset Dam Failure...............................................78
8.2 Vaiont (Vajont) Dam Disaster........................................85
8.3 Teton Dam Failure...................................................89
8.4 United States Dam Failures..........................................93
9. Conclusion.........................................................102
A. United States Dam Failures.........................................105

2.1 St. Francis Dam Area Map........................................10
2.2 St. Francis Dam Topography and Geology Map......................11
2.3 Geological Cross Section........................................14
3.1 William Mulholland..............................................18
4.1 St. Francis Dam Construction....................................24
4.2 St. Francis Dam Cross Sections..................................27
5.1 St. Francis Dam Crack Locations.................................30
5.2 Downstream Face of the St. Francis Dam. March 12, 1928..........33
5.3 William Mulholland's Final Inspection...........................34
5.4 Stevens Water Stage Gage Chart, March 12-13.....................37
5.5 Flood Path......................................................39
5.6 Wreckage in Piru Fillmore Area..................................40
5.7 St. Francis Dam Failure Aftermath...............................41
5.8 Remaining Section...............................................42
6.1 Final Dam Design Derived from Remaining Cross Section...........50
6.2 Final Dam Design as Provided by Los Angeles Bureau of Water
Works and Supply................................................52

6.3 Ladder Location......................................................66
6.4 Block Location Before Failure........................................69
6.5 Block Location After Failure.........................................69
8.1 Malpasset Dam During Construction....................................80
8.2 Remains of Malpasset Dam.............................................81
8.3 Teton Dam Collapse...................................................91
8.4 Remaining Teton Dam..................................................92
8.5 Number of Failures per Type of Dam...................................95
8.6 US Dam Type Distribution.............................................96
8.7 Number of Failures per Failure Mode..................................98
8.8 Number of Failures per Decade.......................................100

Dams Designed and Built Between 1920-1926 by Los Angeles
Bureau of Waterworks and Supply...............................8

On March 12, 1928 one of the worst civil engineering disasters of the 20th
century occurred suddenly and with little warning. The St. Francis Dam failed
unexpectedly unleashing millions of acre-feet of water into the canyon and
communities below. This failure brought about an extensive number of questions
with very few answers. However, with the collapse came a series of engineering
design changes and construction method modifications that would alter how dams
would be built and engineered in the future.
Located 40 miles northwest of Los Angeles in the San Francisquito Canyon,
the St. Francis Dam was intended to supply Los Angeles with over a years supply of
water. Although tension and opposition from the local communities threatened
construction of the dam, the need for water in the city prevailed and construction
began in 1924. After several changes and modifications of the dam during
construction as requested by the engineer, William Mulholland, the dam was
completed in 1926.
During the two-year life of the dam, cracks and concern of the stability of the
dam overshadowed the engineering achievement. Although periodical inspections by
Mulholland, including one completed on the day of the failure, deemed the dam safe,
the St. Francis Dam failed catastrophically only two years after completion and upon
its first filling. As the wall of water made its way to

the ocean, the aftermath proved to be devastating with over 400 lives lost, miles of
land destroyed, and communities tom apart.
With the failure came an inundation of questions as to how such a massive
structure could fail so suddenly without warning. Engineers and geologists flocked to
the disaster site to examine the remnants of the St. Francis Dam in hopes of piecing
together the answer. Based on evidence and rumors, several theories erupted to
explain the collapse. Ranging from sabotage by the local residents to insufficient
design and inadequate geology, a single cause for the failure could not be agreed
upon by the investigators.
Though catastrophic, the failure brought much needed attention to the
engineering methods used in concrete dam design. Knowledge learned through the
St. Francis Dam failure aided in changing the importance of the geology of the site,
the influence of uplift on dam foundations, as well as construction and design
methods. The result was a change in the design of future dams and retrofitting those
existing dams with questionable stability.

Water Crisis
During the late 1800s and early 1900s, Los Angeles was experiencing a
sudden boom in population. With the exponentially growing population, the citys
water supply was taking a substantial hit. Estimated at being able to provide enough
water to support a population of 200,000, the water supply was steadily shrinking at
the hands of the Los Angeles people.
At the end of the 19th century, the Los Angeles population exceeded 100,000.
Only four years later in 1904, the population nearly doubled to 175,000, nearing the
capacity of the current water supply. It was at this time, that concern arose as to
where the city would obtain water to support the ever-increasing population. William
Mulholland made the following statement regarding the need for water in the city, If
Los Angeles runs out of water for one week, the city within a year will not have a
population of 100,000 people. A city quickly finds its level, and that level is its water
supply. It was clear that a new source of water was required to keep the city alive.
(Rogers, 1995)
With a city prone to drought, the lack of water was not a new dilemma for Los
Angeles. However, the General Manager of the Los Angeles City Water Company,
William Mulholland, proposed a solution that would supply the city with a substantial
amount of water that would cease the concern for water for years to come.

Mulholland looked over 200 miles away to the Owens Valley to bring water to the
drought-ridden city by way of an aqueduct.
Although other sources for water were investigated and would have been
easier to acquire, Mulholland saw the extensive watershed that contributed to the
Owens River as an optimum source of water. Though the task was daunting,
Mulholland proposed to construct an aqueduct nearly 250 miles long that would cross
a mountain range, drop several thousand feet, and tunnel under the ground. However,
before plans to use the Owens River water were in place, obstacles had to be
overcome to acquire the rights to the water.
The ranchers and farmers dependent on the water provided by the Owens
River resisted the invasion of the aqueduct, concerned that their already limited
supply of water was going to significantly decrease and threaten their livelihood. At
the time, the Reclamation Service was investigating the Owens Valley for an
irrigation project that would aid in a substantially increasing the economy of the local
communities. With this knowledge known, Mulholland was pressured to acquire the
land and water rights immediately before the opportunity he was seeking was lost.
(PBS, 2001)
Before the Reclamation Service was able to determine any conclusions for the
Owens Valley irrigation project, Mulholland and his acquaintances were able to
access the land and water rights documents to determine how the city could claim
Owens River as their own. Through the purchase of enough land and water rights,

the Reclamation Service irrigation project was put to a halt allowing the Owens
Valley-Los Angeles aqueduct project to move forward by the end of 1905. (Rogers,
Although never before taking on such an enormous feat, Mulhollands vision
came to life once the acquisition of the land and water was complete. The initial task
of the Owens-Los Angeles Aqueduct was to construct the Haiwee Dam and Reservoir
to serve as a regulation reservoir. Once the dam was in place, construction of the
aqueduct followed immediately following. The selected inlet point was located north
of Independence, California where it began its long journey. Traversing along the
countryside, over a mountain range, and below the earth, the aqueduct eventually
made its way to San Fernando Reservoir.
Utilizing thousands of workers over the years of construction, Mulholland
managed to complete the project on time and within budget. The aqueducts opening
day occurred on November 5, 1913, nearly 11 years prior to the groundbreaking of
the St. Francis Dam, with the intent of supplying enough water to provide for a
population of 390,000 consuming a total of 58 million gallons per day. (Rogers,
With a maximum capacity to provide 485 cubic feet per second (cfs) of water,
the supply seemed unlimited (Rogers, 1995). The aqueduct quenched the thirst of the
people of Los Angeles allowing the city to continue its growth. But with the rainfall
each year diminishing and the drought worsening, the water supplied by the aqueduct

was providing less then expected while the population continued to exceed growth
predictions. During times of drought, flow of the aqueduct decreased significantly
from an average flow of 355 cfs to only 262 cfs (Rogers, 1995). Once again, the
question arose as to how, through times of drought, the city of Los Angeles was going
to survive with an ever-increasing population and constant lack of water.
2.1 Alternative Water Sources
With the Owens Valley-Los Angeles aqueduct supplying less water than
needed by the city in times of drought, other sources of water were investigated
during the early 1920s. At the time, the population of Los Angeles exceeded
576,000 and was increasing at a rate of more than 100,000 people per year, stretching
the limits of the aqueduct supply even when water was available. Mulholland had
significantly underestimated the population growth and the variation in the amount of
water the aqueduct would provide which was resulting in a water shortage for the
city. Estimating that the watershed upstream of the inlet of the aqueduct would
provide the additional required water for the growing population, the city bought
more land and water rights from local ranchers for an exorbitant price. For a cost of
$3 million to the city in attaining the necessary land, there was little increase in the
amount of water provided by the aqueduct. (Rogers, 1995)
When expanding the watershed area of the aqueduct did not solve the citys
water problem, another option for a source of water was sought. Mulholland was

given the task of determining what method of acquiring water would accommodate
the growing population during drought conditions. In determining the amount of
water required to support the people of Los Angeles, Mulholland concluded that the
most ideal alternative would be to build several water storage facilities in the area.
By constructing numerous dams and reservoirs that were capable of storing tens of
thousands of acre-feet of water, the city could survive those times when the aqueduct
provided an insufficient amount of water.
In 1921, Mulholland and the Los Angeles Water Bureau began design and
construction of several dams and reservoirs surrounding Los Angeles. The
construction of these dams resulted in providing storage of as much as 67,000 acre
feet of water, increasing the current local water storage by three times. Table 2.1
summarizes the construction of the dams in the Los Angeles area between 1920 and
1926. (Rogers, 1995)

Table 2.1 Dams Designed and Built Between 1920-1926 by Los Angeles Bureau
of Waterworks and Supply (Rogers, 1995)
Reservoir Name Height Dam Type Reservoir Capacity (acre-feet) Years of Construction
Lower Franklin 96 feet Hydraulic fill & rolled earth 1,050 1921-22
Stone Canyon 147 feet Rolled earth 8,000 1921-24
Upper San Fernando 82 feet Hydraulic fill 1,850 1921-22
Lower San Fernando Raised 7 feet Rolled fill Additional 3,800, to 14,670 1924-25
Encino 135 feet Hydraulic fill 3,229 1921-24
Sawtelle 34 feet Rolled earth 103 1923-24
Ascot 73 feet Rolled earth 219 1925-26
Hollywood 200 feet Concrete gravity 7,500 1923-25
Ultimately, Mulholland wanted to create enough water storage to supply Los
Angeles with a years supply of water if the need ever arose. In order to accomplish
this goal, five new dams were to be constructed along with expanding two existing
dams. Of the five new dams, the largest dam to be constructed would be located
between the Los Angeles Owens Valley aqueduct powerhouses located in the San
Francisquito Canyon. This dam would later be known as the St. Francis Dam.
2.2 Location of the St. Francis Dam
The original site considered for the location of the St. Francis Dam was in the
Big Tujunga Canyon. However, this plan was laid to rest quickly when an outpouring

of opposition was brought to court during condemnation proceedings. The cost of
obtaining the land at the proposed site was inflated by opposing landowners and
considerably more expensive then the city initially thought or was willing to pay.
With the original location of the Big Tujunga Canyon out of reach of the city,
Mulholland brought attention to an alternative dam site, the San Francisquito Canyon.
(Rogers, 1995)
During construction of the Los Angeles-Owens Valley aqueduct, Mulholland
had noted the unique topography of the San Francisquito Canyon which he
considered to be an ideal location for a dam. Once the opportunity arose to bring this
site to the attention of the Bureau of Waterworks Supply, Mulholland expressed his
extensive knowledge of the site to support his opinion of this ideal location and bring
his dream into reality.
The specific location in question was located along the San Francisquito
Creek, approximately halfway up the canyon and below where the San Francisquito
Canyon Construction Camp was located during the construction of the Los Angeles-
Owens Valley aqueduct. The site consisted of steep canyon walls that narrowed to
form a natural dam site and was located downstream from a broad wooded platform.
The expansive flat upstream would create an extensive reservoir to store an adequate
amount of water and the steep canyon walls would aid in creating an ideal dam

Figure 2.1 St. Francis Dam Area Map (Johnson, 1995)
With the location in the near vicinity of Los Angeles, the topography ideal for
a dam and reservoir, and the cost more reasonable then other options, the San
Francisquito Canyon became the most feasible and optimum option of the St. Francis
2.3 Geology
Before the decision of the St. Francis Dam site could be finalized, a survey
and geological investigation of the San Francisquito Canyon at the proposed location
was to be conducted. Although geological explorations at the time were uncommon,
Mulholland was considered to have extensive knowledge on the subject and chose to
conduct his own geological tests when he first discovered the site and the potential it

The initial geological exploration was performed in 1911, eleven years prior
to Mulholland proposing the site to the Los Angeles Bureau of Waterworks and
Supply. The location chosen in the San Francisquito Canyon that held the most
potential for a future dam determined to also be the site of a fault line that split the
two banks of the canyon. On the southeast side of the fault, the geology consisted of
mica schist; while on the opposite bank, Sespe conglomerate was the predominant
geology. (Rogers, 1995)
Figure 2.2 St. Francis Dam Topography and Geology Map (Rogers, 1995)

Exploratory adits, or horizontal tunnels, were excavated into the Sespe
conglomerate and water percolation tests were performed within these. The results of
these tests determined that the crumbly, reddish sandstone passed the percolation
tests. (Rogers, 1995)
Mulholland evaluated the mica schist on the opposite bank during his
investigation of the area at the time of the aqueduct construction. Previous
experience and knowledge determined that the mica schist was often found
treacherous in nature for a foundation of any structure. As stated in his Sixth Annual
Report of the Bureau of the Los Angeles Aqueduct, Mulholland made the following
report regarding the mica schist of the area (Outland, 2002):
As the face of the canyon opposite the lower power line [the aqueduct
line between Powerhouse No. 1 and Powerhouse No. 2] is exceedingly
rough, and the dip and strike of the slate such as to threaten slips, in
case side-hill excavation were made, this portion of the line was also
placed well back under the mountain and will be constructed from
adits [tunnels] run in from the canyons [parallel and beneath the
canyon walls]
Although the mica schist was questionable in its ability to provide adequate
support for the dam foundation, Mulholland felt there were construction methods that
could be utilized so that the stability of the dam would not be compromised.
In the 1920s, the technology required to accurately determine the properties
of the geology along with the exact make-up of the topography of the site was
limited. It was also unknown at the time that geology played a significant role in the
stability of a dam foundation. Years after the St. Francis Dam was built when there

was a greater knowledge of the significance of geology and more advanced
technology available, studies were conducted on the geology of the site to determine a
more accurate depiction of the location.
Unknown to Mulholland, the favorable topography of the site for a dam and
reservoir were due to prehistoric landslides that at one point created a natural
reservoir by damming the San Francisquito Creek. The prehistoric landslides were
developed within the Pelona Schist along the southeastern canyon wall. Outland
(2002) stated:
The seemingly-intact Pelona Schist had actually rotated downward
onto the opposing bank of Sespe conglomerate, thereby blocking the
canyon and creating a large landslide dam. The waters of San
Francisquito Creek had eventually overtopped the landslide dam and
re-excavated a channel. The broad flat area, seen by Mulholland as an
excellent reservoir site, had actually been created through
sedimentation behind the paleo landslide dam.
However, due to the ancient landslides, the site might not have been as
favorable as Mulholland had anticipated. The following figure depicts the geological
cross section through the slope showing the paleoslide.

Figure 2.3 Geological Cross Section (Rogers, 1995)
At the time, Mulholland made an educated and knowledgeable decision in
recommending the San Francisquito Canyon for a dam site. With the limited
resources of the time, it would have been difficult for Mulholland to know the extent
of the problematic geology of the site. Seeing the topography as ideal for a dam and
the geology to be sufficient as confirmed by the testing conducted, it was obvious to
Mulholland and the Bureau of Waterworks and Supply to place the future dam within
the walls of the San Francisquito Canyon.

William Mulholland
The designer and construction overseer of the St. Francis Dam, William
Mulholland, launched his engineering career in the late 1800s in the United States
after spending his childhood and adolescence in Europe. Born in Belfast, Ireland in
1855, Mulholland sought opportunity at the age of 15 and made the voyage across the
Atlantic Ocean as a journeyman sailor. Arriving in New York City in 1874, he
traveled across America working at various jobs including Michigan lumber camps,
Pittsburgh dry-goods business, drilling wells in California, and mining in Arizona.
Spring of 1878 brought Mulholland to Los Angeles where he began his career with
the Los Angeles Water Company, a private water provider for the city. (PBS, 2001)
After discovering opportunity in California, Mulholland worked his way from
the bottom of the water business working as a ditch cleaner in 1878 to achieving the
position of superintendent only eight years later in 1886 for the Los Angeles Water
Company. While advancing rapidly in his career, his status in the community also
progressed, as he became a prominent figure in the water supply industry. Due to his
highly regarded reputation, Mulholland was promoted to the head of the Department
of Water and Power once the city of Los Angeles took over the privately owned
company. (PBS, 2001).

During the start of his career at the Los Angeles Water Company, William
Mulholland became a self-educated engineer, as many of the engineers of that era
were, by reading technical literature and working with some of the most
knowledgeable water resource engineers of the time. Through his reading and field
experience, Mulholland began changing how natures water was harnessed and
brought to the people of Los Angeles and surrounding communities. (PBS, 2001)
One of Mulhollands defining endeavors was the Owens Valley Los
Angeles Aqueduct. The methods used by Mulholland and fellow cohorts in acquiring
the water rights brought about tension and turmoil to the local communities of Owens
Valley. Many of the residents felt that deception was the only reason why the water
offered to them by the Bureau of Reclamation to improve irrigation was taken away.
The decision to allow the water to be brought into Los Angeles from the Owens River
and taken away from the residents of Owens Valley was made by President Theodore
Roosevelt by federal approval in 1906. This decision was encouraged by Fred Eaton
after traveling to Washington to convince the government that the water would
benefit more people if it were moved to Los Angeles. (Mulholland, 1995)
Once the decision to allow the construction of the Los Angeles Owens
Valley aqueduct was final, Mulholland began the arduous trek of construction across
and through over 200 miles of land. With very few complications or setbacks,
Mulholland was able to complete the project on time and under budget in 1913
(Rogers, 1995). This accomplishment brought Mulholland into the spotlight as an

accomplished engineer and was highly regarded by many in his field. William had
the honor of being presented with an honorary doctorate from the University of
California at Berkley, gained membership in several engineering fraternities, and
broadened his engineering connections with people around the world. Not only was
he respected by the engineering society, the residents of Los Angeles revered
Mulholland and many civic leaders encouraged him to run for mayor. However,
Mulholland chose to stay out of the headlines and commit his life to engineering.
(Mulholland, 1995)

Figure 3.1 William Mulholland (Outland, 2002)
A few years after the completion of the Los Angeles Owens Valley
aqueduct, Mulholland began the construction of Mulholland Dam, later known as
Hollywood Dam. Knowing that a larger reservoir would be required to support the
needs of Los Angeles, construction of the St. Francis Dam began one year after the
start of Mulholland Dam. The St. Francis Dam was one of Mulhollands greatest
accomplishments, but became his greatest failure only two years after completion.

By the time Mulholland designed and constructed the St. Francis Dam, he was
well known for his knowledge in hydraulics and the design of several earth dams
surrounding the city of Los Angeles. However, Mulhollands knowledge of geology
was limited as was his experience with the design of concrete dams. Though his
confidence persuaded many that he was capable of constructing such a structure as
the St. Francis Dam, after the failure of the dam, it is uncertain if Mulholland was
qualified for the project. However, the design and construction of the structure might
have been sound given the fact that Mulholland Dam still stands today leading to the
belief that the geology of the site was the primary cause of the failure of the St.
Francis Dam.
After the collapse of the St. Francis Dam, Mulholland searched for answers to
explain why the dam failed so suddenly and with little, if any, warning. As stated in
Mulholland and the St. Francis Dam (Mulholland, 1995),
There was his [Mulhollands] progression of words as first he said
some terrific earth movement (not an earthquake) must have caused
the collapse; then in testimony at the coroners inquest, after hinting at
the possibility of sabotage, with tears coursing down his face and at
the point of breakdown, he said he envied the dead. His assertion that
if there were human error, it was his and his alone.
With few answers after the failure, Mulholland, as well as much of the
community, continued to blame himself for the death and destruction of so many
lives. A year after the disaster, Van Norman took over Mulhollands position of chief

engineer and general manager of the Department of Water and Power (DWP), while
Mulholland became a consultant for them and later resigned. (Mulholland, 1995)
Mulholland died in 1935 without a definitive answer as to why his dam failed.
From becoming one of the most well known engineers for his capability of capturing
the forces of water, to being known for the death of hundreds of people, Mulholland
died a broken man. Though Mulholland contributed greatly to the growth of Los
Angeles, for many he will only be known for the disaster of the St. Francis Dam.

4. Design
The design of any dam takes into account several factors including the reason
for the dam, the topography and geology of the chosen site, and available resources.
Based on these factors, an engineer can deduce the most appropriate dam for the
project. It is these factors that Mulholland investigated to determine that the St.
Francis Dam would be most ideal as a concrete gravity arch dam.
In the following sections, the reason for the dam to be designed as it was is
detailed. Modifications of the design are also specified along with the implications
these changes made in the final stability of the structure. However, with each
decision made about the design of the St. Francis Dam, Mulholland had reasons to
support his choices. From the lack of materials for an earth dam, to an increase in
height to enlarge the capacity of the reservoir, Mulholland made decisions that
seemed reasonable for the construction of the St. Francis Dam.
4.1 Design Concrete vs. Earth
With extensive experience in the construction of earth dams and limited
experience in concrete dams, Mulholland veered from his previously designed dams
and chose to construct the St. Francis Dam as a concrete gravity dam. At the time, it
was uncommon for dams to be constructed of concrete, and therefore, extensive
knowledge of the design of this type of dam was limited. However, there were

several reasons why Mulholland determined that constructing the St. Francis Dam of
concrete rather then earth fill would be a more optimal choice for this situation.
Before any concrete was placed for the St. Francis Dam, construction of the
Weid Canyon Dam, later known as Mulholland Dam, was underway. This dam, also
designed by Mulholland, was a curved concrete gravity arch dam, and eventually
served as the basis of design for the St. Francis Dam. The design of this dam was a
first of its kind for both Los Angeles and Mulholland. With this gained knowledge
and experience, Mulholland chose to duplicate the design for the future dam located
in San Francisquito Canyon. (Rogers, 1995)
Besides having the limited experience from the design and initial construction
of Mulholland Dam, the characteristics of the dam site also promoted the use of a
concrete dam. Similar to the site of Mulholland Dam, the San Francisquito Canyon
lacked the sufficient features required to construct an earth dam. The construction of
a hydraulic earth fill dam required a sufficient amount of clay and water to form a
semi-impervious core in which the remaining dam is formed around. In the case of
the St. Francis Dam site, the water and soil requirements could not easily be met to
construct an earth dam. Due to the lack of materials, an earth dam would not be a
viable option and was no longer an option for Mulholland. (Rogers, 1995)
Although concrete dams were uncommon for the time and minimal design
knowledge was available, a concrete arched gravity dam was chosen for the St.
Francis Dam. With the experience gained in the design of Mulholland Dam and the

site characteristics as they were, Mulholland was confident in the proposed structure
of a concrete gravity arch dam.
4.2 Design Specifications
With Mulholland Dam already designed and in the process of being built, it
was a matter of duplicating the design with minimal modifications to construct the St.
Francis Dam to meet the topography of the San Francisquito Canyon. The design of
the St. Francis Dam, originally prepared in May 1923, described the dam as a stepped
concrete gravity arch dam. The dam was arched upstream, on a 500-foot radius, with
an initial design height of 175 feet from the floor of the canyon. The resulting
reservoir capacity would be approximately 30,000 acre-feet of water. (Outland, 2002)
When it was determined that the city of Los Angeles was consuming more
water then anticipated during the initial dam design phase, changes in the design took
place before concrete was placed. A wing dike extending from the west abutment
was incorporated into the design in July 1924 that would increase the reservoir
capacity to 32,000 acre-feet. This modification was the first of many to come.
(Rogers, 1995)
In August of 1924, the first block of concrete to form the foundation of the St.
Francis Dam was placed on the floor of San Francisquito Canyon. Each block to
follow would eventually create the steps of the dam that would work collectively to
hold back the forces of water contained in the future reservoir. At this time, it was

also decided to increase the height of the dam by 10 feet to a total height of 185 feet
above the canyon floor. This would create an increase in the capacity of the reservoir
and the amount of water that may be supplied to the city of Los Angeles. The
following photograph shows the St. Francis Dam while under construction. (Outland,
Figure 4.1 St. Francis Dam Construction (Johnson, 1995)
After months of placing concrete to form the base of the St. Francis Dam,
more modifications were proposed and approved. Though the first alteration in the
dams design provided an increase in the reservoir capacity, it was determined that

the water being used by the people of Los Angeles was increasing even more rapidly
then anticipated. By altering the dam design for a second time in July 1925, the
reservoirs capacity would increase to 38,168 acre-feet. Enlarging the capacity of the
reservoir was a result of increasing the dam height by 10 feet, to a final dam height of
195 feet above the floor of the canyon. The increase in capacity was also a result of
raising and extending the dike 1,300 feet northwest of the dams right abutment.
(Rogers, 1995)
By this time, the dam had been raised 20 feet, or 11 percent, of the original
design. Though the dam had been increased substantially by engineering standards,
modifications to the base of the dam had not been made. Mulholland had stated that
the design of the St. Francis Dam had a safety factor of three or four; however,
revised calculations were never produced to verify the safety of the dam after the
changes occurred during construction. (Rogers, 1995)
Another noticeable change in the design of the St. Francis Dam occurred early
on during the construction phase. Probably noticed by few at the time, historian
Charles Outland (2002) noted nearly 40 years later through photographs taken during
construction that the base of the dam was built significantly different then depicted in
the original plans. It appeared that the bottom four steps of the design were never
constructed. A cross section of the remaining part of the dam after the failure verified
this discrepancy in that the width of the base was nearly 30 feet less than that of the
original design. The following figures depicts the cross sections of the dam including

what was likely the original design, the cross section provided to the governors
inquiry board, and the cross section of the block remaining after the collapse of the
St. Francis Dam. It can be noticed that each of the three cross sections varies
significantly from one another in terms of stability.

h---------------176.1 of----------------H
Figure 4.2 St. Francis Dam Cross Sections (Rogers, 1995)
In comparison with todays engineering standards, several basic dam design
procedures of today were neglected during the construction of the St. Francis Dam.

Expansion joints used to relieve stress during expansion and contraction of the
concrete were neglected in the main section of the St. Francis Dam. Also, minimal
uplift relief wells beneath the central core were engaged and limited seepage relief
was implemented in the design.
Though implemented in todays dam designs, features that relieve uplift
forces were rarely used and unknown during the time of the construction of the St.
Francis Dam. Such components as cutoff walls/trenches and grout curtains were not
part of the design. However, eight uplift relief wells were placed beneath the main
portion of the dam which was the only section that remained after the disaster.
Because little was known at the time about uplift forces and research was only
beginning on the subject, there was little, if any, concern about uplift relief measures
and therefore preventative measures were not implemented. (Rogers, 1995)

Initial Dam Distress
The initial filling of the St. Francis Dam occurred on March 1, 1926 when the
Los Angeles-Owens Valley aqueduct was diverted into the St. Francis reservoir. At
the time, the St. Francis Dam was not yet complete and would not be for another two
months. However, the dam was considered safe and worthy of resisting the forces
induced by the water without the upper section of the dam yet in place. (Outland,
With the dam filling at an initial rate of 1.8 feet per day over the first three
months, the water in the reservoir was rising more rapidly then expected. The results
were dramatic and hopes were rising for the potential the dam held for the future of
Los Angeles. (Rogers, 1995)
By May 10, 1927, the St. Francis Dam was complete and the water elevation
in the reservoir reached three feet below the spillway, or 177 feet above the canyon
floor. After this point in time, the spring run-off was coming to an end and the water
level began dropping. (Rogers, 1995)
However, during the initial filling between March of 1926 and May of 1927,
several cracks had formed along the face of the St. Francis Dam. These cracks
formed transverse to the axis of the dam and were located primarily on the
downstream face of the main structure. After inspection of these cracks, Mulholland
concluded that these cracks were transverse contraction cracks and

were formed due to the thermal stresses caused during the curing process of the
concrete. (Rogers, 1995)
Two sets of cracks, large enough to be noticed by onlookers and depicted in
photographs, were formed on the steeply rising flanks of the structure by the end of
1927. The following figure depicts the location of the most notable cracks.
Figure 5.1 St. Francis Dam Crack Locations (Rogers, 1995)
The cracks located on the flanks of the dam appeared to be wider at their
juncture with the abutments and narrowing upwards, while the cracks located near the
main section of the dam were wider at the top. The cracks located in the main
structure of the dam were filled with hemp and sealed with wedges of oakum and
then backfilled with cement grout. This method of sealing the cracks was meant to
deter further seepage that could further expand the cracks. (Rogers, 1995)

By February 1928, the water level in the reservoir once again reached the top
of the dam crest. By this time, various springs had formed near the foundation of the
St. Francis Dam. The primary location of these springs was on the west side of the
dam where the dam was built into the Sespe formation. (Outland, 2002)
On March 7, 1928, the water height in the reservoir reached its maximum
elevation of 1834.75 feet, or three inches below the top of the spillway. At this point,
the aqueduct was no longer diverted into the reservoir in order to help prevent water
from spilling over the top of the dam. With the water pressure acting on the dam
structure greater then ever before, the springs already present released more water and
new leaks were being formed on both abutments. One of the more notable leaks
formed along the wing dike, releasing water at a rate of approximately 0.60 cubic feet
per second (cfs). To remedy the problem, Mulholland had an 8 inch diameter
concrete pipe underdrain installed at the location of the leak, and eastward along the
base of the dike, where it would discharge along the west abutment contact of the
main dam section. The addition of the underdrain was meant to relieve pressure and
to prevent erosion from occurring at the foundation of the dam. (Rogers, 1995)
After five days of the reservoir being at capacity with a water level three
inches below the crest of the spillway, it was beginning to become difficult to observe
where leaks were forming due to the windswept water flowing over the top of the
dam which obscured any visible evidence of leaks. Little could be done at this time
to prevent or repair any cracks or leaks that were forming due to the amount of water

flowing over the spillway crest. It was now March 12, 1928, the day of the St.
Francis Dam Failure and little could be done to prevent the catastrophe from
occurring. (Rogers, 1995)
5.1 Events Prior to Failure
On March 12, 1928, the reservoir had been filled to capacity for five days with
water reaching three inches below the crest of the spillway. With the wind blowing
across the water, evidence of spillage over the crest could be observed on the face of
the dam. The following photograph shows the St. Francis Dam on March 12, 1928
depicting the darker concrete as the location of where water had spilled over the top
of the dam due to the high water elevation and strong winds across the reservoir.

Figure 5.2 Downstream Face of the St. Francis Dam, March 12,1928 (Nichols,
For the first time in almost two years, the Los Angeles-Owens aqueduct
released water into the San Francisquito Canyon after filling the reservoir and the
downstream storage areas to capacity. The water was released from the aqueduct at a
rate of 30 cfs, below Powerhouse No. 2. The remaining water that was spilling over
St. Francis Dam was allowed to flow down the canyon unimpeded due to the lack of
storage areas. (Outland, 2002)
While the diversion of the aqueduct was occurring, the damkeeper, Tony
Hamischfeger, contacted Mulholland to report a new leak that he had discovered in
the morning. This leak was reported as dirty, which is often an indication of

hydraulic piping indicating a cause for concern of the stability of the structures
foundation. (Rogers, 1995)
Mulholland and the Assistant Chief Engineer, Harvey Van Norman, made a
personal inspection of the dam around 10:30 a.m. on March 12, 1928 for
approximately two hours. The following figure shows the two men crossing the dam
during this inspection, only hours before the failure.
Figure 5.3 William Mulhollands Final Inspection (Nichols, 2003)
At this time, the water level had dropped to 1834.3 feet over the previous 48
hours, but the wind-swept water was still spilling over the face of the dam. The two
engineers made an inspection of the west abutment and noted that approximatly 2 cfs
of water was flowing from the Sespe beds, but was flowing inconsistently, with a

surging style of flow. However, unlike what was reported by the damkeeper,
Mulholland noted that the water was clear from the leak located beneath the concrete
wing dike which eliminated the concern of hydraulic piping at that location. This
discrepancy might have been due to the fact that the water would become muddy
where it washed across the uncompacted sidecast fill of the access road. It was
therefore concluded that a mistake was made by the damkeeper and there was no
potential danger at the site examined. (Rogers, 1995)
Besides the initial leak found by Hamischfeger, Mulholland discovered
another small stream of clear water at the juncture of the dams east abutment
against the schist. At the time, this leak was considered insignificant and Mulholland
felt that further inspection of the dam was not warranted. Both Mulholland and Van
Norman left around 12:30 after assuring the damkeeper that the observed leak of was
no concern and the dam was stable and safe. (Rogers, 1995)
The observation of the leak was the last one made by Hamischfeger and
reported to Mulholland. On the day of March 12, 1928, no further observations
concerning the stability of the dam were noted.
5.2 Failure of the Dam
In the hour prior to the failure of the St. Francis Dam, at least four separate
parties drove by the dam and reported what they had witnessed. In each case, very

little was noticed to indicate that there were problems with the dam or that a failure
was imminent.
One witness reported crossing over a scarp of approximately 12 inches that
cut across the roadway upstream of the dam. This scarp would have occurred within
the Pelona Schist bedrock as the road was constructed as a cut excavation along the
dams eastern abutment. The scarp that was witnessed could have been an indication
of a landslide occurring on the east abutment, resulting in forces acting on the left
side of the dam. (Rogers, 1995)
Another witness recalled seeing lights at the bottom of the canyon, below the
downstream face of the St. Francis Dam. This could have been an indication that the
damkeeper was investigating something unusual that he had heard or felt coming
from the dam. (Hogan and Laresen-Cleaves, 2001)
The final witness prior to the failure was a man who rode by the St. Francis
Dam on a motorcycle and indicated hearing or feeling something ominous after
traveling approximately a mile past the dam. This portion of roadway was known for
landslides and the witness verified that the sound heard was similar to rocks rolling
down the mountainside. After taking a few moments to realize the sound was
approximately a mile behind him, the witness continued on his way. This was the
most significant account of those witnesses who passed by the dam prior to the
failure. (Rogers, 1995)

Not only were the witnesses able to account events prior to the failure, the
Stevens Water Stage Gage located within the dam was also able to give an account of
what occurred previous to the collapse. Immediately before the St. Francis Dam
collapsed, the Stevens Water Stage Gage indicated that the water level of the dam
was dropping at a considerably high rate. Within the forty minutes prior to the
failure of the dam, the gage showed that the level of the lake had dropped 3.6 inches
in an accelerating manner. The following figure shows the trace of the Stevens Water
Stage Gage during the last day of the St. Francis Dam. (Rogers, 1995)

i I !
\ V ti
1 V-
1 .2 *
1 1 0 >
i i 1 ui

I2.M, 6RM. VI Mt March 12. ~ *-^-March 13
Figure 5.4 Stevens Water Stage Gage Chart, March 12-13 (Rogers, 1995)
The Stevens Water Stage Gage was located in the center of the main section
of the dam. After the collapse, this was the only section to survive, which enabled the
retrieval of the gage. As stated in A Man, A Dam and A Disaster (Rogers, 1995),

The Stevens Gage consisted of a 12-inch diameter pipe affixed to the
upstream face of the dam. Water fed into this stilling well through a
1-inch diameter hole at its lower extremity. The intent of a stilling
well was to filter out the oscillatory effects of wind-whipped waves on
the back of the dam.
The water elevation drop shown on the Stevens Gage that occurred before the
failure, equated to a release of 934,580 cubic feet of water. The employees of
Powerhouse No. 2, who were on watch the night of March 12, would more than likely
would have noticed this large amount of water being released in a short period of
time. However, surviving witnesses and the verbal communication between the two
powerhouses before the dam failure gave no indication that an excessively high
amount of water ever flowed through the canyon that night. Because there were no
witnesses to the water flowing down the canyon at the rate indicated by gage, the
accuracy of the Stevens Water Stage Gage is questionable. (Rogers, 1995)
At this point in time, the Stevens Water Stage Gage indicated the sudden drop
of water as the St. Francis Dam collapsed almost immediately. Millions of acre-feet
of water were released into the canyons and the communities below. With the failure
occurring in the middle of the night, the people downstream of the dam had little, if
any, notification that the dam failed and water was rushing towards them at an
accelerating rate. The water carried with it debris and rock which made chances of
survival even less.

The wave of water eventually traveled down through the canyon and
eventually made its way to the ocean. The following figure shows the time frame and
the communities that were hit in the flood of water.
The Path OF Destruction. This map shows the 54-mitc path of the flood from die dam site
in Los Angeles County and down the Santa Clara River Valley through Ventura County to the
Pacific Ocean. The docks show the time that the flood reached each community. The speed is
also indicated.
Figure 5.5 Flood Path (Nichols, 2003)
Though warnings were sent out, there was little to be done to prevent the
deaths of many of the residences of the communities the flood waters hit. When
daylight touched the towns, the true effect of the water was visible. Bridges were
destroyed, trees were uprooted, debris lay scattered across the land, and homes were
either moved from their foundation or splintered apart. There was little that went
untouched by the floodwaters. The following photographs show some of the
destruction that was incurred on the communities downstream of the St. Francis Dam.

Figure 5.6 Wreckage in Piru Fillmore Area (Santa Clarita Valley Historical

Figure 5.7 St. Francis Dam Failure Aftermath (Santa Clarita Valley Historical
It was a disaster that some predicted would occur, but caught all off guard.
After the failure occurred, Mulholland had very few explanations for the disaster and
only blamed himself for the failure if human error was to blame.
When inspections of the dam site began, only the center section of the dam
remained as evidence that a dam ever existed. The following photograph shows what
was left after the flood of water carried most of the concrete blocks downstream.

Figure 5.8 Remaining Section (Nichols, 2003)
With the water gone into the sea and the concrete blocks scattered
downstream, all that remained were questions as to why such a failure could occur so
suddenly without warning.

Theories Behind the Failure
The time immediately after the failure was full of confusion and speculation
as to the reasons behind the failure. Due to the limitations of knowledge and
technology of the time, a definite determination of the cause of the failure was never
found. Several theories proposed by people with various backgrounds investigated
the disaster at the time and during the following 75 years. With evidence provided to
support each theory, deciphering what actually happened to the St. Francis Dam on
March 12, 1928 has not been conclusively decided.
The theories surrounding the failure of the St. Francis Dam varied greatly.
Sabotage committed by the local residents, poor geological conditions, and drastic
changes in the design were just a few of the theories presented by investigators of the
time and researchers in the years to follow. The following sections present the
various theories that have evolved over the past 75 years concerning the disaster of
the St. Francis Dam.
6.1 Sabotage Theory
Of the various theories surrounding the failure of the St. Francis Dam, the
sabotage theory is considered the least likely reason for the disaster. However, with
the evidence found immediately after the collapse and the plausible motives of the

residents of the Santa Clara Valley, sabotage could not be completely ruled out.
Prior to the St. Francis Dam being built, the construction and completion of
the Los Angeles-0wens Valley aqueduct brought about several incidents of sabotage
against the aqueduct. The residents of Owens Valley were strongly opposed to the
building of the aqueduct for fear of losing the water they desperately needed to
continue their farming livelihood. Disputing over the water rights and necessity of
the water, the residents eventually lost their battle against the construction of the
aqueduct and plans by the city moved forward.
As the aqueduct was being built, numerous threats and occasional violent
actions against the aqueduct occurred over a period of three years. Several incidents
occurred in which local residents were involved in including capturing of the
headgates, kidnapping of Los Angeles representatives, illegal stopping and searching
of vehicles, and dynamiting of the aqueduct (Outland, 2002).
The war over the water rights of the Santa Clara Valley did not end when the
aqueduct and the St. Francis Dam were completed. Many farmers and residents were
still opposed to the structures and were in constant legal battles in hopes of regaining
the water they once had control of. The day the St. Francis Dam failed, a lawsuit
was currently under way over who owned the rights to the water held in the reservoir.
It was because of the long history of the battle over the much-needed water, that the
question of sabotage came into play while investigators sought a reason for the failure
(Rogers, 1995).

During initial investigations, hundreds of dead fish were discovered
downstream of the St. Francis Dam. There was not a single live fish left in the many
pools of water that were formed by the release of the reservoir. This fact did not go
by unnoticed and brought about attention from zoologist Edwin C. Starke. The fish
found in the pools were brought to Starke so that an examination of the fish could be
conducted to determine the cause of death. Starke concluded that the fish might have
died from concussions as a result of an under water explosion. However, he did not
rule out the possibility that the fish may have died as a result of the large quantities of
silt in the floodwater. The investigation into the death of the fish and the resulting
conclusion was the strongest argument in favor of sabotage as the reason for the
failure of the St. Francis Dam. (Outland, 2002)
However, the dead fish were not the only supporting evidence of sabotage. A
self proclaimed expert from Texas traveled to the site to make an inspection of the
remnants of the St. Francis Dam in hopes of determining scientifically the reason for
the failure. After Starkes investigation, he provided pieces of the dam that implied
dynamite was the cause of what brought down the dam. Further investigation
concerning this physical evidence was conducted by City Councilman Peirson M.
Hall, chairman of the Water and Power Committee of the City Council of Los
Angeles in 1928 (Outland, 2002).
After Halls investigation, he provided the following statement as found in A
Man, A Dam and A Disaster (Rogers, 1995):

I was at the dam site the day following the break. Shortly thereafter
the self-admitted expert from Texas examined the face of the blocks
lying immediately back of the center standing portion. He said that in
his opinion dynamite had been used because the effect of dynamite on
concrete was the same as weather on a log which lies out in the open -
it will sound and ring under a hammer at first, but with time it will
crumble so that it can be scraped off with the claws of a hammer, or if
left long enough, it can be clawed off with ones bare fingers.
I called the Hercules Powder Company and all other powder
companies in the United States to see if they had any reports on the
effect of dynamite on concrete, or if anyone in their employ would
have, or if they could give me the name of an expert who would be of
help in that respect. They finally came up with one man, whose name
I have forgotten, whose occupation was to destroy standing cement
smoke-stacks. I took him out to the dam and showed him the portion
which had been pointed out to me by the Texas expert, and he gave
me the same opinion. I was still not satisfied with that, so I had a
testing laboratory in Los Angeles conduct some experiments, and they
came up with the same answer.
With the expert testimony and extensive testing concerning the concrete
remnants, the sabotage theory was accepted by many people of the time.
However, with modem day technology and information, it was discovered that
the same markings on the concrete that was used as evidence supporting dynamite
could also be a result of reactive aggregate. Reactive aggregate is explained by
Charles Outland as types of sands and gravels which, when mixed with cement,
begin a chemical reaction in the finished concrete. In time, the appearance of the
surface of the concrete closely resembles the descriptions given by the experts who
examined the ruins at St. Francis Dam. Therefore, the concrete evidence used in the

testimony supporting sabotage could have been subjected to this reaction rather then
dynamite. (Outland, 1995)
The concrete was not the only cause to question the validity of the sabotage
theory. It was determined that in order to destroy a structure as massive as the St.
Francis Dam, over 6000 pounds of TNT would have to be utilized. To cause any sort
of destruction to the dam, the dynamite would have to be placed on the upstream face
of the dam at a depth greater then 30 feet. This seems nearly improbable given the
amount of dynamite and the location it would have to be placed. (Hogan and Larsen-
Cleaves, 2001)
Not only was the concrete evidence inconclusive and the method improbable,
there was also a lack of supporting evidence from the seismographs that were
recording earth movement in the area. In 1923, seismologists Charles Richter and
Harry Wood of the Carnegie Institute at California Institute of Technology (Cal Tech)
began recording blasts occurring in eleven rock quarries located in southern
California. These seismographs were recording ground movement from locations up
to 121 miles away. During their study, they concluded that the blasting they were
studying resembled that of earthquakes on the seismographs with the exception of
some phase interference due to their surficial origins. (Rogers, 1995)
During the investigation, the Governors Board of Inquiry requested that
Richter and Wood analyze the graphs produced by their seismographs on March 12
and 13, 1928. The St. Francis Dam was located only 35 Vi miles away from the

seismograph that was recording quarry blasts at Monolith 70 miles away. If there had
been any sort of significant blast at the St. Francis Dam site, it would have been
recorded by the seismograph. However, there was no sign of movement whatsoever
recorded by the seismograph during the night the failure occurred. (Rogers, 1995)
Though there was a lack of witnesses to support the sabotage theory and
nobody claimed responsibility, the evidence found after the failure leaves the door
open to the possibility that sabotage was the culprit for the collapse. The past history
of violence against the aqueduct, concrete remnants with evidence of dynamite, and
the cause of death of the fish downstream support the theory of sabotage. However
unlikely, sabotage could not be ruled out as a reason for the failure and continues to
remain a plausible theory.
6.2 Unstable Design Theory
Though the St. Francis Dam was considered to be well designed for its time,
the construction changes, methods and techniques used to build the dam came under
scrutiny after the failure. This section will examine in detail the various construction
methods that have theoretically been attributed to the reason the dam failed.
During the initial phase of construction, the formation of the St. Francis Dam
was already deviating from the original design. The initial base width of the dam was
designed for 176 feet. However, based on photographs taken during construction
along with measurements from the remaining section of the dam, the actual base

width was determined to be 148 feet. This is a significant change since the strength
and stability of a gravity dam is almost solely based on the self-weight of the
structure. By decreasing the width by 16 percent, the weight of the structure
decreased substantially. (Outland, 2002)
Not only did the width of the dam base change significantly from the initial
design, the height of the structure also changed drastically. During the construction
period, the dam was raised twice, resulting in a total vertical height change of 20 feet
or an 11 percent increase (Rogers, 1995). The initial dam design indicated a height of
175 feet above the canyon floor. Though the height increased substantially, there
were no modifications to the base of the structure. The following is a calculation of
the factor of safety of the dam based on the new height and constructed base width as
determined from the remaining dam cross section. A calculation of the factor of
safety of the design based on the drawing provided by the Los Angeles Bureau of
Water Works and Supply is also provided. A calculation of the factor of safety of
Mulhollands original design was not computed due to the lack of dimensions
regarding the initial design. Both calculations will not take into account the curvature
of the dam, similar to that of Mulhollands initial calculation. The following
calculations are also an estimate based on the drawings of Figure 4.2 or may be found
in A Man, A Dam, and A Disaster (Rogers, 1995). Due to the incomplete dimensions
of the drawing, the calculation of the factor of safety is an approximate value.

Figure 6.1 Final Dam Design Derived from Remaining Cross Section

Final Design Cross Section of Remaining Portion of Dam
Water Force
Density of Water, p = 62.4 pcf
Height of Water, Hw = 195.9 ft
Base of Water Force, bw = H*pw = 12224 psf
Total Water Force, Rw = bwH/2 1197356.5 pounds per foot width of dam
Location of Force = HJ3 = 65.3 ft
Overturning Force = HwR/3 = 78187378 lbs, overturning
Dam Force
Density of Concrete, pc = 120 pcf
b,= 10 ft
h-, = 199 ft
Area, A, = b,h, = 1990 ft2
Location of Force, L, = 143 ft
Total Force, R-i = A,pc = 238800 pounds per foot width of dam
Resisting Force = R, L, = 34148400 lbs, resisting
b2= 11.5 ft
h2 = 120 ft
Area, A2 = b2h2 = 1380 ft2
Location of Force, L2 = 132.3 ft
Total Force, R2 = A2pc = 165600 pounds per foot width of dam
Resisting Force = R2L2 = 21908880 lbs, resisting
b3= 126.5 ft
h3= 120 ft
Area, A3 = b3IV2= 7590 ft2
Location of Force, L3 = 84.3 ft
Total Force, R3 = A3pc = 910800 pounds per foot width of dam
Resisting Force = R3L3 = 76780440 lbs, resisting
b4 = 138 ft
h4 = 39 ft
Area, A* = b4h4 = 5382 ft2
Location of Force, U = 69 ft
Total Force, R4 = A4pc = 645840 pounds per foot width of dam
Resisting Force = FUL, = 44562960 lbs, resisting
Total Resisting Force = 177400680 lbs
Factor of Safety = = 2.27

Figure 6.2 Final Dam Design as Provided by Los Angeles Bureau of Water
Works and Supply

Final Design as Presented by Los Angeles Bureau of
Water Works and Supply
Water Force
Density of Water, pw = 62.4 pcf
Height of Water, Hw = 205.9 ft
Base of Water Force, bw = Hwpw = 12848 psf
Total Water Force, Rw = bJr\J2 1322718.07 pounds per foot width of dam
Location of Force = H*/3 = 68.63 ft
Overturning Force = HwRw/3 = 90782550 lbs, overturning
Dam Force
Density of Concrete, pc = 120 pcf
bi = 10 ft
hi = 209 ft
Area, A, = b^ = 2090 ft2
Location of Force, L! = 171 ft
Total Force, Ri = A^,, = 250800 pounds per foot width of dam
Resisting Force = RtL-i = 42886800 lbs, resisting
b2 = 11.5 ft
h2 = 166.5 ft
Area, A2 = b2h2 = 1914.75 ft2
Location of Force, L2 = 160.3 ft
Total Force, R2 = A2pc = 229770 pounds per foot width of dam
Resisting Force = R2L2 = 36832131 lbs, resisting
b3 = 154.5 ft
h3 = 166.5 ft
Area, A3 = b3h3/2= 12862.125 ft2
Location of Force, L3 = 103 ft
Total Force, R3 = A3pc = 1543455 pounds per foot width of dam
Resisting Force = R3L3 = 158975865 lbs, resisting
Total Resisting Force = 238694796 lbs
Factor of Safety = RResisting/Roverturning = 2.63
The two above hand calculations have shown that the factor of safety for
overturning for the dam is greater then 1.5, the allowable factor of safety of today

overturning. It is therefore determined that the dam was stable in the case of
overturning only. However, if uplift was present at the time of failure, the factor of
safety for overturning would drop significantly due to the buoyant forces acting on
the dam decreasing the resisting force. Had uplift forces been considered during the
design of the St. Francis Dam, either methods to prevent uplift would have been
incorporated into the design or the dam would have become more massive in weight
by increasing the width of the base. Although the calculations support the stability of
fMulhollands design, it is misleading since uplift forces incurred prior to the failure
are neglected in this calculation as well as Mulhollands calculation.
6.3 Inadequate Construction Techniques Theory
The changes to the dimensions of the design of the St. Francis Dam were not
the only construction issues that came into question after the failure. There were
many speculations as to the construction methods of the dam. The poor concrete
mixture and lack of reinforcing and contraction joints were only some of the many
missing components of the dam that some felt were necessary to have a safe and
stable dam structure.
Initially, it was felt by many that the concrete mixture used in the St. Francis
Dam was of poor quality due to the aggregate used. The quality of the cement was
deemed sufficient, while the aggregate was thought to have contained a large quantity
of dirt compared to standard concrete mixtures. The aggregate used was obtained

from the floor of the San Francisquito Canyon and mixed in as an unwashed
condition at approximately one barrel of cement per cubic yard of concrete. The only
specification concerning the aggregate was that rocks larger than six inches were to
be removed. (Outland, 2002)
Though what was considered a poor concrete mixture, core samples of the
concrete were taken and tested where it was determined by modem scientific methods
to be adequate and of good quality. Several samples were taken from the remnants of
the St. Francis Dam and each sample produced the same result. It was therefore
concluded that the concrete mixture was not to blame for the failure of the St. Francis
Dam. (Outland, 2002)
The lack of reinforcing in the entire structure also came up as a reason behind
the failure of the St. Francis Dam. Many thought with the lack of reinforcing, the
high stresses in the dam would become too great, resulting in the eventual dam
failure. However, it was common practice in gravity or arch-gravity dam design to
not include reinforcing in the design and construction. Because gravity dams use
their self-weight for their stability and to resist the stresses induced by the water,
reinforcing is unnecessary and therefore, not the reason behind the dam failure.
(Outland, 2002)
Yet another construction method came under scrutiny by investigators, the
lack of contraction joints. As with all concrete structures including sidewalks and
building foundations, concrete has a natural tendency to crack. The cracking occurs

due to the shrinking as it dries and both contracting and expanding as the temperature
changes. To mitigate the formation of objectionable cracks, shrinkage reinforcement,
temperature control systems, and contraction joints are often incorporated into the
construction and design of the structure. However, it is only in more recent times that
these crack prevention methods have been incorporated in almost all concrete
structures. At the time the St. Francis Dam was constructed, these methods were not
widely used and were not considered necessary. The following is a statement found
in the April 1, 1928 edition of Electrical World (Outland, 2002):
The dam itself contained no vertical joints as definite provision for
contraction because it is not the policy of the Los Angeles Bureau of
Water Works and Supply to make provision in dam design for
contraction joints, but it is the policy to allow contraction cracks to
occur where they will and subsequently to close them if necessary.
This procedure is followed because it is the opinion of the Bureaus
engineering department that cracks are very likely to occur elsewhere
anyway, and it is believed that it is the better policy to allow them to
occur where stresses dictate rather than to attempt to fix them by
expansion joints ... In a structure of the St. Francis type it is common,
according to the consensus of engineering opinion, that contraction
cracks should occur approximately every 100 feet. This is borne out
by the fact that the central portion of the dam that remains standing is
slightly more than 100 feet wide, indicating that perhaps contraction
cracks had something to do with the exact point at which rupture
occurred when the side foundations gave way ....
Some of the cracks noticed before the collapse of the St. Francis Dam was a
direct result of contraction occurring in the concrete. There were two cracks that
formed fifty-eight feet to the west of the outlet gates and another approximately the
same distance to the east. Though the exact time of when these cracks formed was

not known, it was evident in photographs taken in September of 1927 that a grout line
on the fracture to the west of the gates was present. After the failure, it was noticed
that this grout line matched the outline of the west side of the remaining section of the
dam (Rogers, 1995).
Though it was concluded by engineers that the lack of contraction joints did
not contribute to the failure of the dam, the question arose as to how a contraction
crack and a fracture resulting from stresses caused by lateral forces could be
differentiated. There were several cracks that had formed that were suspected of
being a result of stresses due to pressure rather than contraction cracks, yet it was
unclear as to the definite cause.
One of the other major construction flaws was the lack of uplift force relief
features. At the time the St. Francis Dam was constructed, little was known about the
affect uplift forces on a dam structure. Research was only becoming available
regarding uplift at the time of the failure. However, such structural features as uplift
relief wells, deep cutoff walls/trenches, or a grout curtain was neglected in the design
with the exception of the center section. At this location, eight uplift pressure relief
wells were present, which might have contributed to allowing this section to survive
the failure. Without any of these uplift relief methods, the dams resultant thrust was
shifted to the downstream third of the main dam block. This would have caused
instability of the dam if seepage were to occur directly below the dam foundation.
(Rogers, 1995)

Though uncommon at the time to include uplift pressure relief methods, the
dam design was criticized by several prominent dam engineers after the failure
occurred. These engineers realized the importance of uplift in dam design and
realized this major fault in the design of the St. Francis Dam. The reason these forces
are critical in dam design, especially gravity dams, is that these uplift forces offset the
dead weight of the structure. The uplift forces are due to water seepage occurring
beneath the dam. This force acts as a buoyancy force and therefore every pound of
uplift force result in neglecting one pound of deadweight. Because gravity dams
work by their self weight alone, a decrease in the dead weight due to uplift will
significantly decrease the stability of the dam. It was determined that approximately
11,300 psf of uplift force was applied to the dam which only had a deadweight force
of 24,975 psf. This would have decreased the self weight by 45 percent, causing the
resultant thrust to shift outside of the middle third of the dams base, causing
instability of the structure. (Rogers, 1995)
With several construction methods that were questioned, it has been
concluded by engineers that only the lack of uplift relief methods could have
contributed to the St. Francis Dam failure. Although todays construction practices
use construction techniques that have been researched and shown to contribute to the
stability and durability of a structure, it has also been determined by engineers that
those techniques that were questioned in the construction of the St. Francis Dam
played an insignificant role in the failure. However, it is still unknown as to how

much of a role the uplift forces played in the failure of the dam and if such features as
cutoff walls and grout curtains might have prevented the St. Francis Dam failure.
6.4 Inadequate Geology
When the Los Angeles Bureau of Waterworks and Supply were seeking a
location for the St. Francis Dam, the information regarding the importance of geology
was scarce. Although testing was conducted in the San Francisquito Canyon where
the St. Francis Dam was proposed to be constructed, the technology and knowledge
of the time did not allow Mulholland to realize the dangers the geology held.
Though cracks due to contraction and expansion of the concrete had occurred
throughout the dam structure, there were also cracks resulting from excessive
pressure. Two cracks were noticed and evaluated by Carl Grunsky who believed that
they were not a result of contraction or expansion, but due to the soil onto which the
St. Francis Dam was built upon. He based his opinion on the location of the cracks as
well as the characteristics of the cracks. (Rogers, 1995)
A crack due to lateral pressures was discovered before January 1928 on the
eastern portion of the dam. As stated by Charles Outland (Rogers, 1995), the crack
began at the crest near the last section of the spillway and extended obliquely to a
point some sixty-five feet below, where it reached the hillside, indicating the
possibility that the mountain of schist against which the dam abutted had begun to
move and exert pressure on the east wing.

The crack located on the east side of the dam was not the only one of its type
Carl Grunsky questioned. A similar crack formed on the opposite side of the dam
indicating that the soil on which the dam was built on the west side was also exerting
pressure on the dam. Geologist later indicated that the red conglomerate, in which the
west side of the dam was built on, had a tendency to swell when wet.
There were several tests conducted on the soils and rock that made up the
foundation of the St. Francis Dam. The conglomerate that was tested resulted in a
crushing strength less than a quarter of the strength of the concrete used in the dam.
Testing also included subjecting the conglomerate to water. The result was a soft,
granular mass that had little strength. While dry, the conglomerate resembles the
same characteristics of rock, which is why those who observed the site felt that the
foundation of the dam would be suitable. (Outland, 2002)
While the conglomerate made up the west side of the canyon, schist made up
the dams foundation of the east side. The schist found in the St. Francisquito
Canyon was found to be exceptionally hard, but was severely laminated which greatly
affects the stability of the schist when stress is applied parallel to the lines of
cleavage. Because of the lamination, slippage could easily occur when stress is
applied resulting in a significant decrease in the stability of the schist. Charles
Outland compared to the schist as a deck of cards in the following statement the east
abutment of the dam possessed the strength of a deck of cards that is pushed
obliquely on the table. (Outland, 2002)

Many experts deemed the foundation material as the root of the failure of the
St. Francis Dam. The following statement found in The Story of the St. Francis Dam
was presented in a report to Governor Young (Outland, 2002):
With such a formation, the ultimate failure of this dam was inevitable,
unless water could have been kept from reaching the foundation.
Inspection galleries, pressure grouting, drainage wells and cut-off
walls are commonly used to prevent or remove percolation, but it is
improbable that any or all of these devices would have been
adequately effective, though they would have ameliorated the
conditions and postponed the final failure.
The following two sections will detail two theories behind the St. Francis
Dam failure involving the geology of the area, the west abutment failure and east
abutment failure. These two theories both question the construction of the dam and
the choice of location of the St. Francis Dam due to the poor geology of the site.
6.4.1 West Abutment Failure Theory
The failure of the west abutment of the St. Francis Dam has been one of the
two most widely agreed upon theories among engineers, geologists, and researchers.
With the questionable construction and evidence found before and after the failure
concerning the west abutment, there is a substantial amount of proof to support this
The only evidence provided illustrating the construction of the west abutment
was from the construction photographs that were in the possession of the Bureau of

Water Works and Supply. These photographs showed little excavation of the west
canyon wall. As stated by Charles Outland (Outland, 2002),
The sides of the hill were stripped of all loose material and concrete
poured against the exposed rock. From the fault line to the top of the
hill considerable depth was attained before proceeding with the
concrete work. At the fault line itself a shallow cut-off trench was dug
into the hill, a feature that indicates a belated recognition by the
builders of the weakness of that formation.
A number of experts agreed that the initial break in the dam occurred at the
old fault line located on the west abutment. The evidence supporting this theory was
based upon the location of the dam pieces after the failure. Those blocks from the
west side of the dam were found furthest downstream from the original dam site.
Unlike the west side, the east side remnants remained near the still standing section of
dam. The location of the dam blocks suggested that the water initially rushed through
the west side of the dam, carrying those blocks the furthest downstream. (Outland,
It was theorized that the water that was percolating through the fault line of
the west abutment caused the foundation to be undermined or weakened which would
have either caused a blowout or the dam to collapse under its own weight. Once the
water was released, it began scouring the schist on the opposite wall of the canyon,
which caused the foundation of the east side of the dam to deteriorate swiftly. The
erosion of the east side abutment caused the failure of the east side of the St. Francis

Dam. By this time, there was less water in the reservoir to create enough energy to
displace the blocks as much as the west side blocks. (Outland, 2002)
Many agreed upon the west abutment failure, but there was one piece of
evidence that remained under scrutiny, the Stevens automatic water stage recorder.
Though this recorder supported the immediate failure theory due to the drastic drop it
recorded in a short period of time, it also recorded a drop in the reservoir of three-
tenths of a foot in the forty minutes prior to the failure. This drastic drop would
equate to a release of 2200 cubic feet per second of water 20 minutes before the
failure. Only five minutes before the failure, the Stevens gage indicated that a release
of 15,000 cubic feet per second was occurring. Had this occurred, there should have
been several witnesses to the large amount of water flowing down the canyon.
However, there were no reports of such an occurrence. (Outland, 2002)
Though the reliability of the Stevens gage is questionable, the remaining
physical evidence strongly supports the theory of the west abutment failure. With the
poor geology of the west side and the location of the remnants of the dam, there is
little evidence against this theory. However, it has also been theorized that the east
abutment failed initially, followed by the west side. The next section will provide
evidence in support of an initial east abutment failure.
6.4.1 East Abutment Failure Theory

The theory behind the east abutment failure was brought about by the
Grunskys and Dr. Bailey Willis, a geologist and professor emeritus of Stanford
University. Both the geology of the east abutment and the construction of the dam at
this location came into question after an investigation was conducted by the above
mentioned people.
Similar to the west abutment, the construction of the base of the dam was
determined from construction photographs taken throughout the time the St. Francis
Dam was being built. From these photographs, it was determined that the east
abutment was primarily shallowly keyed. If the rock at this location had been firm
and unyielding rather then fractured as the schist was, this minimal construction
might have been adequate. However, after the flood of water cascaded down the
canyon, this hillside of schist eroded away with the water. (Outland, 2002)
As stated in The Story of the St. Francis Dam (Outland, 2002),
According to Mr. Grunsky and Dr. Willis, the reservoir water had
permeated far into the schist formation, lubricating the rock until it
started to move. Slowly the weight of this moving mountain began to
exert itself against the dam. To further complicate matters, Dr. Willis
established that the first reaction of the conglomerate on the opposite
abutment would be to swell upon becoming wet, and raise any
structure built upon it.
With the force of the sliding schist acting on the east side and the uplift forces
from the expanding conglomerate on the west side, the dam could no longer resist the
forces being exerted on the structure. Before the dam collapsed, an indication of the

increase of stress was visible by the two diagonal cracks found on each side of the
dam. Though easily confused as cracks due to contraction or expansion of the
concrete, the cracks were more likely formed by an increase of stress on the dam
which was substantiated by the diagonal fractures having the greatest width where
they joined the hillside. The characteristics of these cracks indicated that the wings
were being raised by the movement of both the schist sliding and the conglomerate
canyon walls swelling. (Rogers, 1995)
In the 1928 Grunsky Report, Grusky and Dr. Willis made the following
statement (Outland, 2002):
As soon as the dam was loosened on its base the toe of the structure
spalled off. This was probably the beginning of its breaking up, and
probably occurred at some time after 11:30 p.m. during the 23 minutes
in which the water in the reservoir apparently fell 0.3 feet. Thereupon,
quite likely, a part of the east end of the dam, meanwhile undermined,
went out and the dam at this end lost its hillside support. Hydrostatic
uplift at the already loose west end and the weight of the remaining
portion of the undermined east end caused a temporary tilting of the
dam toward the east, accompanied by a rapid washing away of the
hillside under the dam at its west end which then also began to break
up. The reservoir water was now rushing with tremendous force
against both ends and against the upstream face of all that was
standing of the dam. This rush of water carried away huge blocks of
concrete from both ends of the dam.....
Grunskys theory was substantiated by the location of a ladder found after the
collapse of the St. Francis Dam. This ladder was wedged in a crack that must have at
one point been large enough during the rocking process of the collapse to allow the
ladder to be found in its final position, and then was clamped tightly once the

structure settled into its foundation. The following is a photograph of the ladder in
question (Outland, 2002):
Figure 6.3 Ladder Location (Outland 2002)
The way the ladder was wedged into position indicated that the substantial
amount of movement of the dam had occurred early in the failure and that the water
was destroying both the east and west side of the dam almost simultaneously.
The theory proposed by Grunsky was also supported by the final location of
the center section of the St. Francis Dam. Located at the top of the standing section
was a benchmark that gave the specific location of that point of the dam. After the
failure occurred, a survey of this point was conducted determined that the top of the
center section moved 5.5 inches downstream and 6 inches toward the eastern

abutment, resulting in a total displacement of 8.4 inches, on a bearing of south 3
degrees west, indicating a clockwise rotation. A survey of the remaining dike section
on the right abutment was also conducted and it was concluded that it had shifted 0.12
to 0.09 feet towards the canyon and had lifted upwards 0.24 to 0.31 feet. The
movement of the dam aids in supporting the theory presented by Grunsky. (Rogers,
Another portion of the theory presented by Grunsky helps in explaining the
drastic change in the water elevation that occurred during the last half hour before the
dam failed. Grunsky proposed that the dam was lifted upwards, resulting in the
Stevens gage reading a lower reservoir elevation. This would explain why witnesses
did not see water traveling down the canyon immediately prior to the failure.
(Rogers, 1995)
In summary, the following six key points developed by Lee, Grunsky, Willis
and Gillette are presented in A Man, A Dam and A Disaster to support the theory that
the failure initially occurred at the east abutment prior to the west abutment failure
(Rogers, 1995).
1. The Stevens Gage was pulled towards the east abutment when
the reservoir pool was at an elevation of between 1,800 and
1,815 feet.
2. Post-failure photos clearly show that the west abutment
construction road was not damaged above the 1,775 elevation,
despite being comprised of erodible sidecast fill dropped over
the hillside by steam shovel. Undisturbed brush is clearly seen
above this level in the post-failure photos. Lee had measured
the levels of erosive scour along the construction road fill,

noting that the highest level of erosion just below the west
abutment was 60 feet below reservoir pool level at the time of
3. Lee also asserted that the final resting position of Block 11
upstream of Block 12 was indicative that Block 12 had been
removed from east abutment before 11 from the west abutment
(See Figure 6.2 and 6.3).
4. Much greater volume of the east abutment, in excess of
100,000 cubic yards of material, was swept downstream than
within the relatively weak Sespe beds comprising the west
abutment, suggesting more water, and thereby erosive force
was placed upon the east side of the dam site.
5. Block 35 and 32 from the so-called missing section between
the main dam (Blocks 1/2/3/4) and the east abutment, were
found furthest downstream after the failure at a far greater
distance than any block identified from the west abutment (See
Figure 6.2 and 6.3).
6. A large pile of schist from the east abutment was noted beneath
Blocks 7 and 5 after the failure, as well as perched upon Blocks
5 and 7, downstream of Block 1, a good 30 feet above creek
level. This would suggest that the east abutment landslide
carried these pieces down and enveloped them with schist
detritus (See Figure 6.2 and 6.3).
The following drawing depicts the numbered blocks mentioned above as they
were when the dam was in place and where they were found after the failure.

Figure 6.4 Block Location Before Failure (Rogers, 1995)
Figure 6.5 Block Location After Failure (Rogers, 1995)

These six conclusions made by the investigators strongly support the theory of
an initial east abutment failure. With the facts presented and the evidence gathered,
this failure theory seems the most plausible and can explain many of the
inconsistencies found in other theories.

Lessons Learned
After the St. Francis Dam failure, the way dams were designed and
constructed came under extensive scrutiny. With several theories as to why the dam
failed, each possibility became a point to evaluate for future and existing dams. From
uplift theories to responsible parties for the design, several techniques were changed
and have continued to evolve to be able to create the most safe and stable dam
possible, while continuing to be cost effective.
The following sections will detail each of the changes that were a result of the
failure of the St. Francis Dam. Through research and experience following the
failure, dams have evolved to the point where disasters have become less common
and residents can feel safe living below dam structures. Though engineering failures
have brought about catastrophic results, those lessons learned have aided in making
future structures more stable and reliable for all those involved.
7.1 California Law
After the failure of the St. Francis Dam, the method as to which the design of
the dam was conducted was scrutinized. In the case of the St. Francis Dam, William
Mulholland was the only person involved in verifying that his own design was correct
and adequate for the conditions that the dam would experience. Though Mulholland

requested that an outside consultant review the design of his Los Angeles Owens
Valley Aqueduct, no such request was made for the design of the St. Francis Dam.
The only outside source that Mulholland had consulted with was Stanford Professor
John Branner who was brought to the site to view the area years before construction
began on the St. Francis Dam. However, this single visit was the only outside
consultant brought in on the design and construction of the dam. (Rogers, 1995)
After the St. Francis Dam failure occurred, the state of California began
questioning the lack of outside consultants in the design and construction of dams. It
was determined by the Governor's Board of Inquiry that the owners of all dams
should submit their plans to an independent consultant for review and approval. This
step in the design process would allow another source to judge the stability of any
proposed dam before construction could proceed. (Rogers, 1995)
Following the conclusion of the Governor's Board of Inquiry, the State
Engineer, Hyatt, and the President of the Railroad Commission, Whitsell, made a
proposal to legislation in 1928. As stated in A Man, A Dam and A Disaster (Rogers,
1995), Hyatt and Whitsell
Proposed legislation mandating state review of all but federally-
sponsored dams and reservoirs in California. A sweeping act, it was
the first of its kind in the United States, and many other states later
followed suit. The legislature enacted legislation to bring about a dam
control law the flowing year. This new law mandated that the owner
of proposed dams pay for a review of their projects by a board of
eminent civil engineers and geologists retained by the State Engineer.
This body subsequently became the Division of Safety of Dams
(DSOD), eventually absorbed within the State Department of Water

Resources (DWR), the first such agency created strictly for dam safety
review in the world.
With the new law in effect, there would be a greater chance that mistakes and
inadequate designs would be discovered before construction could commence. With
a minimum of two independent parties viewing the design plans, there would be a
smaller chance of future disasters occurring due to poor and insufficient dam designs.
Not long after this law passed, the San Gabriel Dam came under investigation
during the middle of construction. The dam, located in the San Gabriel Canyon, was
designed to be the largest dam in the world at the time with a height of 425 feet, a
length of over 2,000 feet, and a reservoir capacity of 240,000 acre-feet. The basic
dam design was similar to that of the St. Francis Dam, an arched concrete gravity
dam. However, it also encountered similar problems to that of St. Francis Dam, poor
geological characteristics. (Rogers, 1995)
Once the law passed and problems became apparent during the excavation of
the site of the San Gabriel Dam, the state review board was brought in to examine the
dam design and the geology of the area. The review board consisted of two
engineers and two geologists. Both the engineers and geologists independently came
to the same conclusion; the west abutment, which had been sliding during excavation,
was an ancient paleomegalandslide. The project came to an immediate halt and the
contractor and chief engineer were prosecuted on charges of conspiracy on the
grounds that they knew previously about the poor geology of the site. (Rogers, 1995)

With the law in effect for only months, the results were already apparent.
With the cancellation of the construction of the San Gabriel Dam, it could be seen
that the state board would be effective in future decisions as to the safety of the dams
that would be constructed in the state of California.
The following sections of the California Water Code were a direct result of
the St. Francis Dam failure and may still be found in the current code. (Legislative
Counsel of California, 2003)
6250. Unless application for approval of the dam has heretofore
been made, every owner of a dam completed prior to August
14, 1929 shall, immediately after the effective date of this part,
file an application for the approval of such dam.
6380. The department shall make inspections at State expense
of all dams in the State completed prior to August 14, 1929.
6450. Any dam which the department finds was not 90 percent
constructed on August 14, 1929 shall be subject to the same
provisions as a dam commenced after that date.
6452. Dams found to be 90 per cent or more constructed on
August 14, 1929 shall be subject to the same supervision as
dams which were completed prior to that date.

These sections of the California Water Code were included after the
failure of the St. Francis Dam to help ensure that those dams constructed
around or before the time of the collapse would be constructed and designed
in such a way to potentially prevent a similar failure.
California was not the only state to bring into affect laws regarding the
construction and design of dams. Several other states followed soon after with
similar laws requiring outside consultants to review plans of dam structures.
7.2 Dam Retrofits
With the sudden and unexpected failure of the St. Francis Dam, concern over
the safety of existing dams heightened drastically. Dams were now being scrutinized
for their design and construction techniques after the methods came under question
for the cause of the St. Francis Dam collapse.
One of the first dams to be inspected after the failure was Mulholland Dam.
The dam was designed by William Mulholland and was the basis for the St. Francis
Dam. Concerned that the nearly identical dam would fail as suddenly as the St.
Francis Dam, the reservoir level had been lowered shortly after the collapse by order
of William Mulholland. (Rogers, 1995)

In January and March of 1930, two separate boards were established to
inspect Mulholland Dam, located in the Cahuenga hills just above Hollywood.
During these initial inspections, it was determined that modification of the dam was
unnecessary. However, it was concluded that the dam lacked uplift relief, which
caused the resultant force to act outside the middle third of the structure. This was
deemed unacceptable by the both inspection boards should the dam be filled to
maximum capacity. In order to alleviate the potential failure of the structure,
spillways would have to be constructed at a lower elevation to bring the resultant
force back into the middle third of the structure. (Rogers, 1995)
By 1931, another board was brought together to inspect Mulholland Dam for a
second time. The resulting report concluded that the base width of the dam lacked the
ability to resist uplift forces or earthquake loading as well as basal sliding. A massive
retrofit was conducted in 1931 to alleviate the fears of the instability of the structure.
The retrofit included extending a conduit down Weid Canyon for the spillway and
placing over 300,000 cubic feet yards of earth fill against the downstream face of the
dam. These modifications to Mulholland Dam would allow the dam to survive the
forces exerted upon it by the reservoir it contained. (Rogers, 1995)
Following the collapse of the St. Francis Dam, dams continued to be updated
to meet new engineering guidelines that would allow dams to be safer and more
stable structures.

7.3 Uplift Forces
After the failure of the St. Francis Dam, the research on uplift forces on dams
became intensified. It was believed by many experts that uplift forces acting on the
St. Francis Dam was one of the major reasons for the failure. At the time the St.
Francis Dam was built, little was known about the affects of uplift on such a structure
and measures to alleviate the forces were not included in the design.
Since the St. Francis Dam failure, there has been extensive research on the
affects of uplift forces on dam structures. Because a gravity dam, such as the St.
Francis Dam, relies solely on the self weight of the structure to resist overturning
forces, it is essentially that all the weight of the dam can be accounted for to create
the greatest stability. It has been determined that the reason for these uplift forces is
due to water seepage occurring below the base of the dam. This seepage results in a
buoyancy force which acts as uplift on the dam. For each pound of uplift, one pound
of deadweight is negated, resulting in a decreased dam weight and less stability.
In order to decrease the affects of uplift on a dam, there are several structural
features that may be included in the dam design to prevent seepage from occurring.
These methods include cutoff walls/trenches, grout curtains, and uplift pressure relief
wells. A cutoff wall reduces seepage under a dam by forming a wall of impervious
material that is located in the foundation beneath and dam to form a water barrier. A
grout curtain decreases seepage beneath a dam by forming a thin vertical zone in the
foundation into which grout is injected. Uplift pressure relief wells are used to collect

and control seepage underneath the dam by constructing vertical wells or boreholes
downstream of the dam. With these methods, minimal seepage is allowed to occur,
and therefore uplift forces are decreased and less affect on the structure occurs. In
todays dam design, these preventative measures are incorporated to aid in the
stability of the dam so that the majority of the self weight of the dam may be used in
determining the factor of safety of the design.

Other Dam Failures
Though the failure of the St. Francis Dam taught engineers numerous lessons
about the design and construction of dams, there are still many unknowns in
attempting to control the forces of Mother Nature. The design of concrete dams, as
well as earth dams, have improved over the years following the St. Francis Dam
disaster, yet there have continued to be dam failures throughout the past century.
Concrete dams have only accounted for a portion of dam failures throughout
history. Earthen dams have also failed, causing death and destruction in the path of
its wake. The National Performance of Dams Program (NPDP) has recorded the
failures of both concrete and earthen dams throughout the United States during the
past 100 years and have documented the cause of the failure for each dam. A list of
these failures may be found in Appendix A.
The following sections will present some of the more notable concrete and
earth dam failures that have occurred throughout the world and over the past century.
Each section will include a brief history of the dam as well as the reasons and theories
as to why the dam had failed.
8.1 Malpasset Dam Failure
The Malpasset Dam, located in the south of France near the town of Frejus
and in the Reyran Valley, was constructed by Andre Coyne as a thin double curvature

concrete arch dam. Construction of the dam began on April 1, 1952 and was
completed 30 months later in 1954. With a height of nearly 197 feet, a foundation
base width of 22 feet, and a crest width of nearly 5 feet, the 635 foot span dam was a
significant presence in the area. Meant primarily for irrigation and to tame the river
due to the extreme weather conditions, the dam was a valuable resource for many of
the residents in the communities surrounding the dam. (SimScience)
The dam was constructed against a rock face on the right wall and terminated
into a wing wall on the opposite bank. The dam was curved differently on each side
of the dam face in order to save money and decrease the amount of concrete used for
the construction of the dam, however the result was an increase in complexity of the
construction. At the time, the Malpasset dam was the thinnest dam ever built with a
base width of only 22.2 feet. (SimScience) The following picture depicts the dam
nearing the end of its construction period.

Figure 8.1 Malpasset Dam During Construction (SimScience)
It was determined by the designer that the dam had a sufficient factor of safety
and was well within the limits of technological feasibility regardless of the thin nature
of the dam. The construction of the dam consisted of blocks approximately 44 feet
long and five feet high with steel or copper barriers placed between the joints which
were filled with a dense mortar. Once completed, Malpasset Dam would hold a
maximum capacity of over thirteen billion gallons of water. At the time, it seemed
that the dam would be indestructible given the experienced dam designer and the
methods used during construction. (Levy and Salvadori, 1992)

Though meant to provide for the communities below, the life of the dam was
short lived and caused destruction and death for those who depended on it for their
livelihood. Five years after the completion of the Malpasset Dam, the dam failed
suddenly on the night of December 2, 1959. After raining continuously for five days,
the dam cracked and burst, releasing the water it contained into the valley below. The
flood of water resulted in the death of at least 420 people, although some estimates
bring the death toll to over 500. Blocks of concrete from the dam were found nearly a
mile from the initial dam site. The following picture shows a portion of the
remaining dam located on the right bank of the canyon.
Figure 8.2 Remains of Malpasset Dam (Collins, 2004)

Similar to the St. Francis Dam, the exact reason for the failure of the
Malpasset Dam was not completely clear to engineers and experts. Preliminary
reports pointed to a failure resulting from a shift of one of the abutments. An initial
geological survey of the site in 1946 was conducted and deemed the land
questionable for the site of a dam. Due to lack of funding, further analysis of the
geology could not be conducted, therefore, the project continued to move forward;
unbeknownst of the impact the poor geology would have on the stability of the dam.
(Collins, 2004)
With time, came more knowledge as to the significance of a stable dam
foundation and the technology to determine the underlying geology. Research
conducted after the failure of the dam concluded that the geology of the area was
unsuitable due to the varying rock formations that lay beneath the location of the
dam. Current research showed that the underlying geology of the site included
anthracite, thin coal deposits mixed with shale and gray schists as well as areas of
fluorite. This combination of rock formations had incompatible qualities, especially
given the varying weather conditions of the site. During the summertime, the
differing rocks would react differently in the heat and dryness the weather brought.
While the winter brought significant rainfall in which only some of the rock
formations absorbed the water. This poor combination of rock formations and the
lack of knowledge of this geology resulted in a poor foundation for the Malpasset
Dam. A tectonic fault was also discovered downstream of the dam after the failure.

Although this fault was not located directly below the dam foundation, movement of
the fault might have contributed to the shifting of the surrounding geology and the
instability of the dam. (Collins, 2004)
The geology of the site was not the only questionable component in the failure
of the Malpasset Dam. The construction methods surrounding the dam were also
believed to have contributed to the failure of the dam. While the dam was being
constructed, there were several delays due to lack of funding and labor disputes.
These delays occurred throughout the placement of the concrete and caused the
construction of the dam to exceed what was initially anticipated. During each of the
postponements, the already placed concrete, in some cases, became completely
hardened, causing the newly poured concrete to not thoroughly bond with the existing
concrete, resulting in a potentially non-homogenous structure. This could result in
the structure not acting as a whole which would decrease the stability of the dam.
(Collins, 2004)
There is also speculation that the construction of the roadway, A8 autoroute,
only 200 meters away from the dam site, contributed to the instability of the dam. In
order to construct the roadway, explosives were used which may have resulted in
shifting of the rock located below and adjacent to the dam. The final result might
have been a loosening of the attachment of the dam to the shoulders of the rock.
Loosening of the dam concrete might have also occurred resulting in unseen cracks
that could have contributed to the final failure of the dam. (Collins, 2004)

An analysis of the dam after the failure was conducted which focused on
several key issues surrounding the design and stability of the Malpasset Dam. The
following is a list of those failure modes analyzed (SimScience):
Concrete stresses
Buckling of the arch
Sliding of the abutment block
Sliding of the dam at the contact with rock
Sliding on the downstream fault
Though thoroughly evaluated, there was no conclusive evidence indicating
whether any of these possible failure modes were actually the cause of the collapse.
Though there were no definitive answers as to why the Malpasset Dam failed,
there were several lessons learned for the design of future dams. As with the St.
Francis Dam, it was realized that the composition of the surrounding geology was an
essential step in choosing a stable dam location. This encouraged the study of rock
mechanics as well as the testing of the geological foundation for various
characteristics. Not only was the awareness of geology heightened, the failure also
brought about the need for safety monitoring for arch dams. With more precise and
regular inspections of dams, failures might be prevented and communities
downstream could be warned in advance before catastrophe could strike.
Though Malpasset Dam was constructed after the St. Francis Dam, not all
lessons were learned or appreciated the first time around. With time, knowledge,
experience, and technology, it is the hope that future dams will be constructed and

designed such that failures are rare and warnings occur before disaster can strike
communities downstream.
8.2 Vaiont (Vajont) Dam Disaster
The Vaiont Dam is located near Belluno, Italy, approximately 100km north of
Venice in a deep valley tributary to the Piave river valley in the Tyrolean Alps of
northeast Italy. Situated in a steep walled canyon, the dam was built to create a deep
reservoir with a water capacity of 5.3 billion cubic feet making it the third largest
reservoir in the world at the time. Meant to provide the surrounding communities
with hydroelectric power, the dam was a welcome addition to the area. (Vaiont
Reservoir Landslide Disaster, 2003)
The Vaiont Dam was designed as a thin arch concrete dam which was
completed in 1960. The dam itself was designed to be the second highest dam in the
world at a height of 860 feet and a crest span of 623 feet between the canyon walls.
During construction of the dam, several challenges arose in order to increase the
stability of the structure. Due to the fractured rock of the canyon, the stability of the
foundation came into question, but was attempted to be resolved by injecting large
amounts of concrete into the bedrock to contribute to the strength. Though actions
were taken to stabilize the earth surrounding the dam and instabilities of the geology
were known, this would not prevent the disaster that would occur three years after the
completion. (Vaiont Reservoir Landslide Disaster, 2003)

Before construction of the Vaiont Dam began, geological studies of the site
were conducted to determine the composition of the rock below the dam as well as
where the reservoir would be located. Early studies indicated several complications
with the location that was chosen that could be detrimental to the life of the dam. The
following list states the conclusions made after the geological study was conducted as
found in Vaiont Reservoir Landslide Disaster (Vaiont Reservoir Landslide Disaster,
Steep canyon walls composed of interbedded limestone and
shale; badly cracked and deformed structures; open fractures in
the shale are inclined toward the future reservoir body. The
steepness of the canyon walls enhance the strong driving forces
(gravitational) at work on rock structures.
High potential for bank storage (water absorption by canyon
walls) into the groundwater system that will increase water
pressure on all rocks in contact with the reservoir.
The nature of the shale beds is such that cohesion will be
reduced as its clay minerals become saturated.
Evidence of ancient rockfalls and landslides on the north side
of the canyon.
Evidence of creep activity along south side of the canyon.
With the knowledge determined from the geological survey, it should have
been apparent that the site would not be an ideal location for a dam. However, the
engineers and those supporting the construction of the dam chose to continue forward
with the plan to locate the dam in Vaiont Canyon.
Upon initial filling in 1960, 2.5 million cubic feet of earth from an ancient
landslide slid into the reservoir from the upper southern side, adjacent to the dam.

From this point on, the location was monitored and it was shown that the soil on the
south side was creeping, moving as much as 1 cm a week. As the reservoir level rose,
it was shown that the rate of creep increased. By September 1963, movement
exceeded 40 cm per day. (Vaiont Reservoir Landslide Disaster, 2003)
Beginning on September 28, 1963, heavy rains fell in the area, raising the
water level of the reservoir. Concerned that the excess amount water would result in
an increase in the rate of creep, engineers opened the outlets on October 8, 1963 to
release water to lower the water level. However, this attempt at alleviating the creep
of the soil came too late. By the next evening, an 8.5 billion cubic foot (500 ft by 1.2
mile by 1.0 mile slab) landslide crashed into the reservoir within 30 seconds creating
a 330 foot tsunami wave which swept over the top of the dam and into the valley
below. The rush of water that traveled downstream swept through the town of
Longarone, killing nearly 3000 people. (Vaiont Reservoir Landslide Disaster, 2003)
Though the rock and soil from the landslide piled up behind the dam resulting
in an increase of stress in the structure, the dam remained intact and remains standing
today (Natural Environment Research Council, 2000). Though the dam did not fail,
this disaster resulted in an important engineering lesson for dam design. After the
disaster, it was shown that the dam itself did not contain flaws, however, the geology
of the site did. Had the initial geological survey been considered more thoroughly,
this dam would not have been built at the location that was chosen. Though dams
such as the St. Francis Dam and Malpasset Dam proved that geology of a dam site

plays a significant role in the safety and stability of a dam, those lessons were not
headed in the case of the Vaiont Dam was built. At the time, politics took precedence
over the concerns the geology which allowed the dam to be constructed.
Not only was it realized that geology plays a significant role in the success of
a dam, the disaster also taught the importance of dam monitoring as well as warning
systems for towns that would be in the path of water should a failure occur. Though
engineers were monitoring the movement of the unstable land surrounding the dam
and reservoir, nothing was done to warn the townspeople downstream. With the
significant amount of movement that was occurring only days before the disaster, it
should have been realized that the safety of those towns downstream could be in
danger of a disastrous event. With the lack of warning, the people of Longarone had
no chance to escape the rush of water that traveled down the canyon in the middle of
the night.
The Vaiont Dam still stands today as a reminder of those lives lost on the
night of October 9, 1963. Since this disaster, more dams have become monitored for
movement and in several locations, warning systems have been put in place to warn
people located downstream. With these methods, it is the hope that a reduction in the
amount of disasters occurs and there are fewer fatalities in the event that a failure
does occur.

8.3 Teton Dam Failure
The Teton Dam was located 44 miles northeast of Idaho Falls and retained the
water of the Teton River for irrigation, flood protection, and electrical power.
Construction of this dam began in February 1972 and was completed over four years
later in June 1976. Designed as an earthfill dam, over 10 million cubic yards of select
earth was used to construct the Teton Dam. However, the life of the Teton Dam was
short lived. Upon initial filling, the dam collapsed suddenly releasing 300,000 acre
feet of water into the valley below. Flooding towns and farmland downstream, 14
lives were lost and resulted in a cost of nearly $1 billion. (H.G. Arthur, 1977)
The plan for the Teton Dam initially began in 1932 when investigation of a
suitable site was conducted. The initial site investigated was located 15 miles
upstream and east of the final dam location. Following this first investigation, two
other sites were evaluated but were determined to be uneconomical and that seepage
loses could be expected. By July 1957, the Corp of Engineers selected the final site
of the Teton Dam after drilling occurred to determine that the quality of the
foundation at the location was suitable. (H.G. Arthur, 1977)
The Teton Dam consisted of over 10 million cubic yards of earth that had
been extracted from the floor of the future reservoir. The core of the dam consisted
of clay, silt, sand, gravel, and cobbles tightly compacted by tamping rollers that was
considered by engineers to be impervious to water. This watertight core was overlain
by four additional layers of rock materials of various kinds. The final size of the dam

was 305 feet in height with a crest length of 3,200 feet, a crest width of 35 feet and a
maximum reservoir capacity of 288,250 acre-feet. (H.G. Arthur, 1977)
While the reservoir was filling in the early part of 1976, there were no
observations indicating that there were structural deficiencies in the Teton Dam. By
May 1976, the dam was being filled at an average rate of rise of 3 feet per day with
no adverse affects on the structural integrity of the dam. But by the beginning of
June, two spring areas had developed in the right abutment, releasing 40 to 60 gallons
of clear water per minute. The following day, on June 4, 1976, another spring was
discovered in the right abutment. At this time an inspection of the abutments and
downstream face of the dam occurred during the entire day, but evidence of seepage
could not be found. By the morning of June 5, a wet spot on the downstream face of
the dam was observed eroding its way into the embankment. To alleviate the
problem, bulldozers were sent to push rock material into the hole that was forming.
However, by 10:30 in the morning, one of the bulldozers fell into the hole that had
formed. (H.G. Arthur, 1977)
Only a half hour later, a whirlpool was observed 10 to 15 feet from the
intersection of the reservoir surface with the embankment, and 100 to 150 feet from
the right shoreline. As time passed, the whirlpool increased in diameter and depth,
but remained in the same location. At this time, both bulldozers were lost on the face
of the embankment as the erosion of the abutment was increasing at a rapid pace. By
11:55 the crest of the embankment dropped into the reservoir and only two minutes