Formation of California's Salton Sea in 1905-07 was not "accidental"

Abstract

It is widely thought that the Salton Sea was created accidentally in 1905-07 because of engineering negligence in the diversion of Colorado River water for agricultural use in California’s Imperial Valley. This is a misconception. Scientific data and historical records establish that formation of the Salton Sea was not accidental. The lake formed during 1905-07 in the same manner that numerous other large Salton Basin lakes did for at least tens of thousands of years from the Late Pleistocene through the late 19th century: as a result of the lower Colorado River’s natural hydrodynamic regime, floodplain morphodynamics, and established avulsion style in combination with changes in streamflow attributable to regional hydroclimate. A large body of scientific and historical evidence indicates the 1905-07 Colorado River flooding into the Salton Basin and the creation of a large lake there would have occurred regardless of man-made modifications to the river’s natural levee and distributary channels. In fact, the flooding would likely have been even worse in the absence of human intervention.

Full text

103
2020 desert symposium
Formation of California’s Salton Sea in 1905–07
was not “accidental”
Jenny E. Ross
Stout Research Center, Colorado Desert District, California State Parks, Borrego Springs, CA 92004; jenny@jennyross.com
Introduction—the creation ood
For over 100 years it has been widely accepted that
the Salton Sea, California’s largest lake, was created
accidentally in 1905–07 within an otherwise
desiccated desert basin (e.g., Nijhuis, 2000;
Barringer, 2014) as the result of engineering
negligence by the California Development
Company (CDC) as it struggled to keep
irrigation water owing from the lower
Colorado River into the edgling agricultural
community of the Imperial Valley within the
Salton Basin (Figure 1). e rst diversion
point for moving Colorado River water into
the Imperial Valley (“Heading No. 1” on
Figure 2) was constructed beginning in 1900
about 500 m north of the US–Mexico border.
e opening from the lower Colorado River
into the company’s diversion canal was cut
out of the river’s natural western levee at an
oblique angle, and ow into the main course
of the canal was controlled by a wooden
headgate placed a few hundred feet from the
river bank down the canal (Grunsky, 1907).
Below the headgate the canal connected to
one of the Colorado River’s natural delta
distributary channels, the Alamo River, which
was then dry. Diversion of water began in
June 1901, and soon the river’s heavy load of
silt began to repeatedly obstruct the headgate
and canal. As a consequence, water shortages
in the Imperial Valley agricultural area began
occurring in 1902 and continued through
1903 and into 1904, putting tremendous
pressure on the CDC to x the problems.
Eventually, in the fall of 1904, the company resorted to
making two unprotected cuts into the river’s natural levee
farther south in an attempt to achieve reliable diversion
of river water into the canal and onward into the Imperial
—It is widely thought that the Salton Sea was created accidentally in 1905–07 because of
engineering negligence in the diversion of Colorado River water for agricultural use in California’s
Imperial Valley. is is a misconception. Scientic data and historical records establish that
formation of the Salton Sea was not accidental. e lake formed during 1905–07 in the same manner
that numerous other large Salton Basin lakes did for at least tens of thousands of years from the
Late Pleistocene through the late 19th century: as a result of the lower Colorado River’s natural
hydrodynamic regime, oodplain morphodynamics, and established avulsion style in combination
with changes in streamow attributable to regional hydroclimate. A large body of scientic and
historical evidence indicates the 1905–07 Colorado River ooding into the Salton Basin and the
creation of a large lake there would have occurred regardless of man-made modications to the river’s
natural levee and distributary channels. In fact, the ooding would likely have been even worse in the
absence of human intervention.
Figure 1. Overview map. (1) Colorado River; (2) Gila River; (3) Yuma, AZ; (4)
approximate path of the modern, controlled Colorado River’s channel through
the lower delta (where the channel is mostly dry today); (5) lower delta region
below the crest; (6) Imperial Valley in the Salton Basin; (7) Salton Sea.

j. e. ross | formation of california’s salton sea in 1905–07 was not “accidental”
104 2020 desert symposium
Valley via the Alamo River (Grunsky, 1907; Cory, 1915;
Kennan, 1917; Brown, 1923).
e rst additional cut, known as the Upper Mexican
Heading, was made just below the US–Mexico border.
It quickly showed a pronounced tendency to silt up. e
CDC eventually decided to close the rst cut and try
again slightly farther south. e second cut, known as the
Lower Mexican Heading, was made about 4 miles (~6.4
km) below the international border. It was a simple cut
with a dredger, made about 40 to 50 feet (~12-15 m) wide
and 6 to 8 feet (~2-2.5 m) deep. It was connected to the
CDC canal, which in turn connected to the Alamo River
channel. ere was no headgate. rough this cut there
was sucient fall from the river to the canal for the water
to achieve scouring velocity, so silt did not accumulate.
Instead, natural erosion of the unprotected cut began
immediately. In the CDC’s haste to reinitiate the ow
of irrigation water into the valley, the company failed
to add a control structure at the new diversion point,
although they intended to do so eventually. e company’s
engineers perceived no urgency in
adding that structure because the
primary problem they had experienced
up until that time was too little ow
from the river into the valley rather
than too much (Grunsky, 1907; Cory,
1915; Kennan, 1917; Brown, 1923).
Unanticipated high streamow
on the lower Colorado River arrived
in early 1905, and oodwaters soon
rapidly eroded the unreinforced cut and
rushed through it. e river’s high ows
avulsed across the delta and streamed
north primarily through the Alamo
River channel and another previously
dry but well-established natural
distributary channel of the Colorado,
the New River, that headed into the
Salton Basin. Continuing to widen
the breach in the Colorado River’s
natural levee and erode and overtop
the river’s distributary channels, the
oodwaters coursed across the delta in
sheetow, rampaged through recently-
developed farm elds in the Imperial
Valley, poured into the central Salton
Basin, ooded the Southern Pacic
Company’s railroad tracks, and began
creating an enormous lake dubbed
the “Salton Sea.” e CDC, Southern
Pacic, and hundreds of workers made
many desperate attempts to block
the breach in the river’s levee. Each
time the eorts ultimately failed as
numerous large oods raced down
the lower Colorado River below Yuma
during 1905. By August 1905, the
entire ow of the Colorado River was
rushing through the breach, into the river’s distributary
channels, across the Imperial Valley, and into the growing
Salton Sea. Extremely high streamow continued on the
lower Colorado River in 1906 and washed away every
structure the CDC attempted to use to block the breach.
In November 1906, the river nally appeared to be
thwarted and human control achieved. But on December
5, 1906 another huge ood roared down the Colorado
past Yuma. New breaks occurred in the repaired levee,
and soon the river was once again owing uncontrollably
through the Imperial Valley and onward into the central
Salton Basin. Aer many additional eorts, in January
1907 the Southern Pacic Company was nally able to
block the oodwaters and turn the river toward the Gulf
of California by using millions of tons of quarried rock
dumped into the breach (Grunsky, 1907; Cory, 1915;
Kennan, 1917; Brown, 1923).
e story of the epic two-year battle to stanch the
raging ow of the wild Colorado River and redirect the
Figure 2. Sketch map of the lower Colorado River below Yuma, and the California
Development Company’s diversion headings below Pilot Knob. (Modied from
Grunsky (1907), Fig. 2., to specify the location of Yuma and designate the Alamo
River channel.)

j. e. ross | formation of california’s salton sea in 1905–07 was not “accidental”
105
2020 desert symposium
river toward the Gulf of California is a well-known tale
of man against nature described fully in many historical
accounts (e.g., Cory, 1915). e saga of this “Creation
Flood” that formed the Salton Sea, crippled the nascent
Imperial Valley, and led ultimately to the damming
and complete control of the Colorado River for human
purposes (e.g., LaRue 1916, 1925) is engaging and highly
memorable. At the time the events unfolded, they “were
so spectacular as to result in world-wide notoriety.” (Cory,
1915.) For the pioneers of the region who toiled to create
what they hoped would be an agricultural Eden in the
desert, as well as for others who followed in developing
the Imperial Valley into an extraordinarily productive
agricultural region, the story served as an inspirational
saga demonstrating the power of persistence and human
ingenuity to succeed despite seemingly insurmountable
odds. e tale vividly demonstrated the capacity of
mankind to triumph over and control wild nature (e.g.,
Larkin, 1907; Howe & Hall, 1910; Farr, 1918; Sperry, 1975).
But the memorable story and its appealing allegorical
aspects led to the widespread adoption of a fundamental
misconception that has colored opinions of the Salton Sea
ever since: namely, the misimpression that the 1905–07
ooding into the Salton Basin would not have occurred,
and the Salton Sea would not have been created, were it
not for the infamous series of incautious decisions made
by the California Development Company. e Creation
Flood story has resulted in the ingrained but mistaken
view that the Salton Sea is accidental and unnatural, a
man-made lake in a parched desert where such an expanse
of water should not be.
But historical records and scientic data of various
types indicate that formation of the Salton Sea in 1905–07
was not an accident, and engineering negligence was
not the cause. e lake formed in the same manner that
lakes had been forming in the Salton Basin, sustained
by Colorado River water, for at least tens of thousands
of years from the Late Pleistocene through the late 19th
century: as a result of the Colorado River’s natural
hydrodynamic regime, oodplain morphodynamics, and
established avulsion style in combination with changes
in streamow attributable to regional hydroclimate. A
large body of scientic and historical evidence indicates
the 1905–07 Colorado River ooding into the Salton
Basin and the creation of a large lake there would have
occurred regardless of man-made modications to the
river’s natural levee and distributary channels. In fact, the
ooding would likely have been even worse in the absence
of human intervention.
Geologic and geographic context
e Colorado River arrived at the proto-Gulf of California
approximately 4.8 Ma (Crow et al., 2019; Dorsey, 2012),
and began building a vast delta at the boundary of the
Pacic and North American tectonic plates. e Salton
Trough, the northwest landward extension of the Gulf
of California Shear Zone, was originally part of the
proto-Gulf and began accumulating Colorado River
sediments during the early Pliocene (Dibblee, 1954;
Muer and Doe, 1968; Winker and Kidwell, 1996). e
northern Salton Trough likely became cut o from marine
waters of the Gulf by latest Pliocene time, as the result of
aggradation of delta sediments and net plate movement to
the northwest along the San Andreas fault (Winker and
Kidwell, 1986; Winker, 1987; Winker and Kidwell, 1996;
Dorsey, et al., 2011); but marine incursions northward
may have occurred during periods of very elevated sea
level (Ross et al., 2020, this volume). e Salton Basin is a
below-sea-level, fault-bounded ri valley lying within the
northern Salton Trough north of the U.S.-Mexico border
and straddling the plate boundary (Figure 1). e lowest
elevation in the central Salton Basin was determined by
the Southern Pacic Company in 1891 to be 280.2 feet
(85.4 m) below sea level (McGlashan and Dean, 1913); in
1903 it was found to be -286 feet (-87 m) (MacDougal,
1907); and in 1907 it was measured at -278 feet (-84.7 m)
(Grunsky, 1907).
e Salton Basin was part of the Colorado River’s
delta and shiing oodplain, and received part or all of
the river’s ow at various times as a result of avulsion
and channel switching that delivered water to the north.
Aected by tectonic, sedimentary, hydrologic, and
climatic factors, the Colorado River adjusted its ow
sometimes into the Salton Basin, sometimes into the Gulf,
and sometimes to both regions (Cecil-Stephens, 1891;
Blake, 1914; MacDougal, 1915; Brown, 1923; Knien,
1932). When the Colorado River owed into the Salton
Basin, large lakes were oen created and sometimes
sustained for long periods (Blake 1854, 1858; LeConte
1855; Sykes 1914, 1937; Knien, 1932; Li et al., 2008a,b;
Rockwell et al., 2018). ick lacustrine and uvial-deltaic
sedimentary deposits with a Colorado River provenance
that accumulated in the Salton Basin from the Pleistocene
through the Holocene have a total thickness of several
thousand meters and include the Borrego Formation
(Tarbet, 1951), the Brawley Formation (Dibblee, 1954), and
the Lake Cahuilla beds (Blake, 1907).
roughout the Late Pleistocene and Holocene, the
path of the lower Colorado River through its oodplain
and delta was extremely variable. Until the river was
dammed and controlled in the mid-20th century, the
entire delta region was a maze of constantly shiing
distributary channels transporting heavily silt-laden
water. According to Ives (1861):
“e channel is circuitous . . . Slues branch in
every direction . . .e water is perfectly fresh, of
a dark red color, and opaque from the quantity
of mud held in suspension. e shiing of
the channel, the banks, the islands, the bars
is so continual and so rapid that a detailed
description, derived from the experiences of one
trip, would be found incorrect, not only during
the subsequent year, but perhaps in the course of
a week, or even a day. . .”

j. e. ross | formation of california’s salton sea in 1905–07 was not “accidental”
106 2020 desert symposium
e Salton Basin is topographically separated from
the Gulf of California by a region of aggraded deltaic
sediments where today the city of Mexicali, Mexico and
the Mexicali Valley agricultural area are located (Brown,
1923). e highest elevation portion of that region, the
crest of the Colorado River delta, is a broad linear zone
beginning east of the Sierra de Los Cucapah and Cerro
Prieto at Volcano Lake, and trending northeast across the
delta toward the mainstem of the Colorado River (Knien,
1932; Muer and Doe, 1968). ere is a pronounced
asymmetry in the slopes directing drainage in opposite
directions on either side of the delta crest—toward the
Gulf of California in the south and toward the Salton
Basin in the north (Grunsky,
1907; Brown, 1923; Knien,
1932). Knien (1932) explained,
“South of the crest the slope is
quite uniform to the gulf, with a
gradient of less than two feet to
the mile. North of the crest the
slope to the Salton Sea is much
greater.” At its east end, the
delta crest meets the course of
the Colorado River’s mainstem
where, prior to damming and
course modication, the river
turned sharply south-southwest
near Pilot Knob aer owing
due west from its junction
with the Gila River (Emory,
1848; Knien, 1932; Muer
and Doe, 1968). e low point
of the delta crest in the west,
where it reaches Volcano Lake,
was approximately 13 m when
measured in the mid-twentieth
century, but is now lower
as the result of tectonic and
anthropogenic subsidence
(Arnal, 1961; Sarychikhina et
al., 2 011).
e Imperial Valley
agricultural region is located
within the Salton Basin and
extends north from the US–
Mexico border to the southern
and southeastern shores of
the Salton Sea. e current
elevation of the Imperial
Valley along its southern
border with Mexico ranges
from approximately sea level
in the west to approximately
13 m above sea level in the
east, excluding the area farther
east covered by the Algodones
Dunes. e Salton Sea, which
is currently receding from water-deprivation, is located
in the central Salton Basin and presently has a surface
elevation of approximately 72.4 m below sea level (USGS,
2020). e lake has been sustained primarily by irrigation
drainwater from the Imperial Valley since the Colorado
River was dammed and controlled in the mid-twentieth
century and prevented from owing into the Salton Basin.
Delta distributaries and oodplain
morphodynamics
Just as it had for millions of years, in the 19th and early
20th centuries the lower Colorado River owed at will
through an extensive and complex network of distributary
Figure 3. Sketch map of the Colorado River delta and Salton Basin showing distributary
channels. (Modied from Fig. 1 in MacDougal (1907), drawn by Godfrey Sykes, to clarify the
label for Pilot Knob.)

j. e. ross | formation of california’s salton sea in 1905–07 was not “accidental”
107
2020 desert symposium
channels, sloughs, and lagoons in a delta and oodplain
covering thousands of square kilometers, including the
Salton Basin. Sykes (1937) estimated that the areal extent
of the Colorado River delta was approximately 8600 km2
in the early 20th century. e dynamic and capricious
course of the river in its lower reaches was described by
C.K. Clark in 1913: “e lower Colorado has no xed
channel, because of the character of the soil, which is a
deposit of silt, easily eroded. e current swings back
and forth, cutting the banks and changing the meander
line….” (Cory, 1913).
e Alamo River and the New River are Holocene delta
distributary channels of the Colorado River that were
established sometime prior to the mid-1800s (Emory,
1848; LeConte, 1855; Blake 1854, 1858). ey conveyed
ow from the Colorado River into the Salton Basin
during high-water periods (LeConte, 1855; Blake 1854,
1858; Knien, 1932) until damming and control of the
Colorado River in the mid-20th century. Descriptions of
the congurations of those streams during the 19th and
early 20th centuries make clear their ongoing relationship
with the mainstem of the lower Colorado River and with
other delta distributary channels (Figure 3). e Alamo
River branched o of the Colorado’s western levee near
where the mainstem turned sharply south-southwest
aer owing “perfectly straight” west for approximately
3–4 miles (~4.8–6.4 km) from its junction with the Gila
River (Emory, 1848). During periods of high water, the
Alamo owed southwest from the Colorado and then
meandered west and north as it followed the topographic
contours in the area north of the delta crest (Emory, 1848;
Sykes, 1914). When lled, the Alamo owed through a
series of sloughs and lagoons in the northern delta region
which sustained the wells of the Alamo Mocho Station,
well-known to travelers through the desert (Grunsky,
1907; Jonas, 2009); then it curved northward and owed
into the Salton Basin if it contained sucient streamow
(Grunsky, 1907; Sykes, 1914). e Rio Paradones, another
delta distributary channel of the Colorado River, branched
o the west levee of the mainstem south of the Alamo
River. e beginning of the Rio Paradones bifurcation
channel was situated at a place where the south-southwest
owing mainstem turned briey toward the east before
owing generally south (Grunsky, 1907). e Paradones
owed west-southwest atop the delta crest and ended at
the low point where Volcano Lake was perched adjacent
to the beginning of the New River (Grunsky, 1907; Sykes,
1914; Jonas, 2009). During high-water periods, the New
River collected Colorado River oodwater that was
delivered to it via the Paradones and Volcano Lake. In
addition, it received overbank discharge from the Alamo
River, and also accumulated water from sheetow across
other portions of the delta (Grunsky, 1907). e New
River owed north from the delta crest and delivered its
water into the Salton Basin (LeConte 1855; Sykes, 1914;
Knien 1932). In 1848, U.S. Army Lieutenant Colonel
W. H. Emory described additional arroyos north of the
Alamo heading west and then north from the bend of the
Colorado River where the mainstem turned sharply to the
south-southwest aer owing due west from its junction
with the Gila River (Emory, 1848).
Another distributary channel on the south slope of the
delta crest, the Rio Hardy, owed south from the divide at
Volcano Lake and ended at the Gulf of California (Hardy
1829; Howe and Hall, 1910; Blake, 1914). During periods
of normal ow, Volcano Lake emptied preferentially
into the Rio Hardy, but during high ows its waters were
distributed both into the Rio Hardy and the New River
(MacDougal, 1915; Cory, 1915). When the Rio Hardy
overowed during high-water periods, sheetow spread
west through a gap below the Sierra de Los Cucapah
and Sierra El Mayor to ll the shallow below-sea-level
basin lying between those mountains and the Peninsular
Ranges. e lake formed there, known as Laguna Salada
or Laguna Maquata, sometimes achieved a maximum size
of approximately 40 miles (~64 km) long by 20 miles (~32
km) wide, depending on available streamow (Cory, 1915;
MacDougal, 1915; Knien, 1932).
It was observed by Grunsky (1907) that the fall of
the lower Colorado River’s mainstem course southward
to the Gulf of California along the east side of the delta
region was signicantly less than the fall of other courses
through distributary channels that led into the Salton
Basin:
“e Colorado River ows southerly in a
direction in which the general fall of the ground
surface is only about 1.5 . per mile, which the
river in its meanderings cuts down to an eective
fall of about 1 . per mile. Toward Volcano Lake,
southwest from the river, the general surface
gradient is 2 . or more per mile; and westward,
in the direction paralleling Alamo River, it is
nearly 3 . per mile to a point near Calexico [at
the southern edge of the Imperial Valley]. ence
northward into Salton Basin, on lines of greatest
slope, the country falls away at the rate of from 4
to 5 . per mile.”
In the 19th century it was recognized that ow through
the Colorado River’s distributary channels would
bring water into the Salton Basin whenever high-water
conditions existed in the mainstem (Blake 1854, 1858;
LeConte, 1855; Cecil-Stephens, 1891; Grunsky, 1907;
LaRue, 1916; Brown, 1923). LaRue (1916) noted that the
geomorphology of the delta region was a crucial factor in
this process: “As the slope of the delta is greatest toward
the north and west, the river during ood periods is
continually seeking a new channel to Salton Sea.” It was
observed that “overows” of the river into the Salton Basin
would occur at two times of year: in the spring and early
summer, as a result of snowmelt in the headwaters of the
Colorado River and its tributaries, and in the winter as
a result of storms that brought heavy precipitation and
caused ash ooding in the lower Colorado River basin,

j. e. ross | formation of california’s salton sea in 1905–07 was not “accidental”
108 2020 desert symposium
particularly along the Gila River and its tributaries (Cecil-
Stephens, 1891; Brown, 1923). During high-water periods
the lower Colorado would overow its banks at many
points, particularly below Yuma beginning where the river
curved abruptly toward the south-southwest aer owing
due west. A relatively low natural levee on the west side of
the curve below Pilot Knob was especially vulnerable to
over-topping (Emory, 1848; LaRue, 1916). When the river’s
streamow was high and overows were of particularly
signicant volume, which happened numerous times
during the 19th century as described below, large lakes
were created in the Salton Basin via overow and avulsion
which shunted the mainstem’s water into the Alamo and
New Rivers (MacDougal, 1915; Brown, 1923; Knien
1932). In 1891, referring to the gage on the lower Colorado
River at Yuma, Cecil-Stephens (1891) noted, “Hitherto,
New River has always owed when the Colorado marked
19 feet at Yuma.” It was only during periods of drought
that distributary channels did not bring Colorado River
water into the Salton Basin (LeConte, 1855; Blake 1858;
Cecil-Stephens, 1891). Even then, there was a residual
vegetated salt marsh in the central basin stretching as
much as 25 miles (~40 km) long and 5 miles (~8 km) wide
holding Colorado River water that had become saline
from evaporation (Farr, 1918).
As is true of all rivers, and particularly those that
carry large loads of sediment and form fan deltas, the
opening, shiing, blocking, and reopening of the lower
Colorado River’s distributary channels was aected by
varying streamow and uctuating quantities of sediment
the river carried, deposited, and eroded (Andrews,
1991). A characteristic of dryland rivers generally and
the pre-dam Colorado River specically is the transport
of very large quantities of sediment, both as suspended
load and as bedload (Andrews, 1991; Tooth, 2000).
Knien (1932) noted, “As a carrier of silt the Colorado
is probably without a peer among the greater streams
of the world.” He explained that in 1904, during a dry
year preceding the 1905–07 ood, a researcher from
“the Arizona Experiment Station made a careful study
of the river silt. He found that an acre-foot of Colorado
River water contained on an average 9.62 tons of silt,
and that for the year the river’s burden amounted to over
120,000,000 tons—this for a year when the total discharge
was considerably under normal. e average annual load
passing Yuma is probably around 160,000,000 tons, which
translated into terms of volume of dry soil would be
approximately 80,000 acre-feet.”
During the portions of the 19th century for which
historical records exist, the lake-creating oods from the
Colorado River into the Salton Basin were all self-limiting
due to silt deposition when streamow slowed (LaRue,
1916). Extensive sedimentation occurring in the river’s
distributary channels as oodwaters slackened eventually
caused cessation of ow northward into the Salton Basin
(LaRue, 1916; Brown, 1923). During periods of drought
and chronically slack ow on the lower Colorado River,
openings from the mainstem to distributary channels
became blocked by deposition of silt, and the courses of
distributaries became clogged with sediments and blown
sand, and were sometimes overgrown with vegetation
(LaRue, 1916; Schyler, 1907). In addition, at those times
the sedimentation in the mainstem raised the streambed
considerably in relation to the river’s oodplain below
Yuma. en, if ash ooding occurred, the elevated
streambed of the mainstem could not hold the ow
(Schyler, 1907; LaRue, 1916). e river would rapidly
overtop its natural levees along the western edge below
Yuma. When such ooding ensued, sediment dams in
clogged distributary channels would quickly erode—
especially if drought had riddled them with mud cracks
through which oodwaters could penetrate—and avulsion
and bifurcation of the mainstem ow would occur
(MacDougal, 1915; LaRue 1916). When streamow slowed
following a period of high water, and sedimentation
once again blocked distributaries where they branched
o the mainstem, the closures le zones of weakness in
the Colorado’s natural levees at the former bifurcation
points, which encouraged reopening of the distributary
channels in the same spots during the next period of
high water (LaRue 1916; Andrews, 1991). e portion of
the lower Colorado River’s mainstem from the curve at
Pilot Knob to the Rio Paradones below the US–Mexico
border was known to be a stretch particularly vulnerable
to avulsion and bifurcation (Grunsky, 1907; Schyler, 1907;
MacDougal, 1915; LaRue, 1916; Brown 1923). ese early
reports of the lower Colorado River’s behavior are entirely
consistent with the modern understanding of oodplain
morphodynamics and processes initiating avulsion and
bifurcation (Slingerland and Smith, 1998; Kleinhans et al.,
2012; Hajek and Edmunds, 2014; Dean et al., 2016).
Pleistocene-to-Holocene Salton Basin lakes
Scientic data and historical records establish that many
large lakes occurred in the Salton Basin from the Late
Pleistocene through the Holocene, sustained by Colorado
River water. e most generally well-known among
them is Lake Cahuilla, an enormous Late Pleistocene-
to-Holocene lake with a highstand overow path south
across the delta crest that lled the Salton Basin to an
elevation of approximately 13 m above sea level beginning
at least 20.5 kya and continuing intermittently through
the 18th century (Blake 1854, 1858, 1907; Brown, 1923; Li et
al., 2008a,b; Rockwell et al., 2018). At various times Lake
Cahuilla was closed, through-owing, or overowing,
depending on climate conditions and the amount of
Colorado River streamow available (Li et al., 2008a,b).
At its southern end the giant lake was supported by the
elevated zone trending northeast across the delta. When
the lake’s level reached about 13 m above sea level it
overowed at the lowest point of that delta crest, sending a
stream to the Gulf of California.
Lake Cahuilla was rst described by geologist William
Phipps Blake in the mid-19th century, following his

j. e. ross | formation of california’s salton sea in 1905–07 was not “accidental”
109
2020 desert symposium
participation in the initial survey of the West for possible
railroad routes during which he visited the Salton Basin
in 1853 (Blake 1854, 1858). He noticed thick layers of
calcium carbonate encrusting the east face of the Santa
Rosa Mountains and terminating in a line at a consistent
elevation on the mountainside, and he realized that this
tufa deposit had been formed by an immense lake. He
later decided that ‘Lake Cahuilla’ was an appropriate
name for the body of water (Blake 1907), to honor one of
the Native American tribes that lived along its shores and
exploited its rich natural resources for thousands of years.
Blake was also the rst scientist to note that the elevation
of the Salton Basin extended far below sea level (Blake
1854, 1858). Research conducted a century later concluded
there is evidence at various locations in the Salton Trough
indicating that even larger lakes, including at least one
that reached an elevation of about 46 m above sea level,
existed during the late Pleistocene prior to Lake Cahuilla
(omas, 1963).
Focused scientic studies of the timing of Lake
Cahuilla were rst conducted in
the late 1970s and early 1980s
(Wilke 1978; Waters 1983). at
research concluded there were at
least six lengthy episodes during
the past 2000 years when the
13-m lake existed, and its last
occurrence was either in the 15th
(Wilke, 1978) or 16th (Waters,
1983) century. More recent
studies (Li et al., 2008a,b) based
on stable isotope analysis and
serial radiocarbon dating of Lake
Cahuilla’s thick tufa deposits,
determined that the giant lake
began depositing tufa on the east
face of the Santa Rosa Mountains
about 20.5 kya, and there was no
hiatus in that deposition through
at least 1300 years BP. Li et al.
(2008a,b) determined that the
lake was sustained primarily by
ow from the Colorado River,
and was either a full closed lake
or was overowing at its delta sill
intermittently or continuously as
the result of the Colorado River’s
high streamow during periods of
very wet regional hydroclimate.
An additional study by Rockwell
et al. (2018) used radiocarbon
dating, historical records, and
modeling of lake lling and
evaporation rates to place the
timing of the penultimate
occurrence of Lake Cahuilla
during approximately the rst
half of the 17th century and the nal incarnation of the
13-m lake during approximately the mid-18th cent ury.
Tree-ring based reconstructions of Colorado River
streamow indicate that the last two incarnations of Lake
Cahuilla as identied by Rockwell et al. (2018) correspond
with periods when extremely wet hydroclimate existed
in all or a portion of the Colorado River basin—in the
upper basin as reected in reconstructed streamow for
the river at Lees Ferry (Figure 4a; Woodhouse et al., 2006;
Meko et al., 2007), and/or in the lower basin as reected in
reconstructed streamow for tributaries of the Gila River
(Figure 4b; Meko et al., 2008). It’s important to note that
these tree-ring based data may understate very high ow
and ash-ooding (Woodhouse et al., 2006; Meko et al.,
2008). Nonetheless, the data indicate: (a) the penultimate
occurrence of Lake Cahuilla was during a long period
of extremely wet hydroclimate in the 17th century that
caused exceptionally high streamow in both the upper
and lower Colorado River basins; and (b) Lake Cahuilla’s
nal occurrence was during a long period of very wet
Figure 4. Tree-ring based streamow reconstructions: (a) Reconstructed streamow 1550-
1950 for three major upper Colorado River tributaries and for the mainstem at Lees Ferry,
smoothed with a 50-year spline to highlight low-frequency variability (modied from
Woo dho use et al. (2006), Fig. 9). (b) Reconstructed streamow for the Salt + Verde + Tonto
Rivers (tributaries of the Gila that are important sources of ow for the lower Colorado River
below Yuma), based on a 6-year running mean for 1330-2005. Solid black line represents
averaged ows plotted as % of normal, where normal is the median of all 6-year running
means. Dashed aqua line is reconstructed 6-year mean for 1999-2004, to serve as a baseline
comparison for the entire record. Gray areas dene the 80% condence interval. (Modied
from Meko et al. (2008), Fig. 13.)

j. e. ross | formation of california’s salton sea in 1905–07 was not “accidental”
110 2020 desert symposium
hydroclimate in the 18th century that most aected the
lower Colorado River basin and caused particularly high
streamow on the Gila River’s tributaries. Notably, the
reconstructions also conrm the occurrence of extremely
wet hydroclimate in the entire Colorado River basin
during the early 20th century, as discussed below.
Additional support for the conclusions of Rockwell et
al. (2018) regarding the timing of Lake Cahuilla’s nal
occurrence exists in the form of a large detailed map of
North America by John Rocque, topographer to England’s
King George III, published in 1762 (Figure 5). e map
shows the Colorado and Gila Rivers pouring their entire
ow into a giant lake that is separated from the northern
Gulf of California to its south by an expanse of land.
According to text on the map within the cartouche, the
details shown were “taken from Actual Surveys and
Observations Made in the Army employ’d there, From the
year 1754, to 1761.”
Although for most of the 19th century the Salton
Basin was only occasionally visited by people who kept
records of what they saw, historical reports indicate that
particularly heavy “overows” from the Colorado River
into the Salton Basin via the New and Alamo Rivers were
observed to occur and form lakes in the central basin
numerous times during the 1800s—in at least 1828, 1840,
1849, 1852, 1859, 1862, 1867, 1884, and 1891 (Grunsky,
1907; Cory 1913; MacDougal, 1914, 1915; Knien 1932).
At other times there was ow from the Colorado River
into the Salton Basin that lled large sloughs and lagoons
along the distributary courses, but was insucient to
form a large lake in the central basin (Blake 1854, 1858;
LeConte, 1855; Grunsky, 1907). For example, Grunsky
(1907) expla ine d :
“In the mesquite and arrow-weed thicket at
the original head of the Alamo, there was an
occasional accumulation of so much water, and
submersion of so much land, that the locality was
called ‘e Lagoons’ (Las Lagunas). Although
these lagoons received water at practically every
high-water stage of the river, they did not always
yield enough to the Alamo River to produce
a ow throughout the river’s entire length. In
other words, there were many years in which
the Alamo did not discharge any water into the
lowest portion of the Salton Basin. e lagoons,
in addition to feeding the Alamo, appear also to
have been one of the sources of supply for the Rio
Paradones.”
For a long period
in the late 19th century
there was at least
some ow of Colorado
River water into the
Salton Basin on an
annual basis as the
result of a recurring
breach in the river’s
natural levee along the
stretch below where
the mainstem curved
sharply toward the
south-southwest near
Pilot Knob (Brown,
1923; Knien, 1932).
Knien (1932) stated:
“During the
last decades of
the nineteenth
century there was
a minor break
in the Colorado
near Algodones,
occurring
annually at
the time of the
summer ood.
A portion of the
diverted water
went down to
the Salton Basin
Figure 5. A General Map of North America by John Rocque, published in 1762, with superimposed
enlargements showing (a) the Colorado and Gila Rivers emptying into a large lake that is separated from
the northern end of the Gulf of California and appears to be Lake Cahuilla; and (b) the text in the map’s
cartouche.

j. e. ross | formation of california’s salton sea in 1905–07 was not “accidental”
111
2020 desert symposium
in the channel of the Alamo. A greater portion
passed through the Paredones to Volcano Lake
and was there divided, the larger part passing
south through the Hardy, the smaller northward
through the New.”
In 1850, Dr. J.L. LeConte and U.S. Army Major General
S.P. Heintzelman traveled to the Salton Basin seeking
mysterious “boiling springs” and volcanic features
reported to be at the shore of a salt lake (LeConte 1852,
1855). Accompanied by an Indian guide, they went to
the southeast portion of the central Salton Basin where
they found several “volcanic mounds” about 100–150 feet
(~30.5–45.7 m) high that were near the shore of a large
salt lake and “arranged in the arc of a circle.” e features
LeConte (1855) described are the Salton Buttes, dormant
rhyolite domes (Wright et al., 2015). When the surface
elevation of the Salton Sea was approximately 227–231 feet
(~69.2–70.4 m) below sea level from 2005 to 2014 before its
more recent decline (Imperial Irrigation District, 2020),
the southeast shore of the lake was close to the Salton
Buttes (personal observations, 2005–2014). us, it is
apparent that in 1850 there was a large lake in the Salton
Basin that was the size of the modern Salton Sea, roughly
48 km (30 miles) long and 24 km (15 miles) wide.
From December 1861 to January 1862 an extraordinary
period of extremely heavy precipitation lasting for
approximately 43 days, likely caused by a series of major
atmospheric river events (Dettinger and Ingram, 2013),
caused a megaood aecting vast expanses of the western
and southwestern U.S., including Oregon, Washington,
California, Nevada, Idaho, Utah, Arizona, and New
Mexico. e Colorado River delta region was completely
inundated, the Army’s Fort Yuma at the junction of
the Colorado and Gila Rivers was transformed into
an island, entire settlements on the lower Gila River
and lower Colorado River were washed away, and a
large lake estimated to be 60 miles (96 km) long and 30
miles (48 km) wide formed in the Salton Basin (Rigg,
E.A., 1862; Wheeler, G.M., 1876). In the vicinity of Lees
Ferry, Arizona, between the upper and lower Colorado
River basins, the 1862 ood had an extraordinary peak
discharge that was estimated to be in excess of 400,000
cubic feet per second (second-feet) (Dickinson, 1944).
In February 1891, the lower Colorado River “rose to
an unusually high stage, the water at that time being
contributed mainly by the Gila and its tributaries. It
overtopped its banks below Yuma, and submerged large
areas along the Alamo and New Rivers.” (Grunsky, 1907.)
During a lengthy dry period preceding the February 1891
ood, those distributary channels became blocked with
sediment and thick deposits of blown sand; so in February
1891 the oodwaters pooled rather than owing onward
(Grunsky, 1907; Schuyler 1907; MacDougall, 1915).
But later when the usual spring high water caused the
Colorado to breach its western levee in the stretch south
of Pilot Knob below the U.S.-Mexico border, the already-
swollen distributary channels received enough additional
ow to fully erode the blockages along their courses, and
oodwaters poured into the Salton Basin (Schuyler 1907).
e ooding created a large lake in the central basin that
was estimated to cover approximately 150-160 square
miles (~388-414 km2) (Schuyler, 1907).
Discussion
Analysis of lower Colorado River hydrodynamics and
oodplain morphodynamics
By latest Pleistocene time, the lower Colorado River had
developed characteristic hydrodynamics, oodplain
morphodynamics, and avulsion style across its delta that
were contingent on regional hydroclimate. e unique
topography of the region played an important role.
ewell-established patterns of the river’s uvial-deltaic
behavior continued through the Holocene until the river
was dammed and controlled in the mid-twentieth century.
ere were three main patterns in the river’s behavior
based on dierent hydroclimate conditions:
(a) Average hydroclimate: During periods oftypical
spring high water, and sometimes as a result of large
winter storms briey yielding heavy precipitation
and ash ooding in the lower Colorado River basin,
the river overowed or occasionally broke through
its levee along the stretch below its junction with the
Gila River, avulsed moderately, and sent a portion of
its ow toward the Salton Basin. When streamow
naturally decreased, channel sedimentation increased,
distributaries became blocked, and the river’s ow was
once again conned within the mainstem.
(b) Temporarily very wet hydroclimate: During periods
of extremely high streamow mediated by short-
lived changes in the region’s hydroclimate, the lower
Colorado River overowed, avulsed, bifurcated, and
moved by sheetow across its oodplain, reopening
established distributary channels and creating new
ones. Because of delta topography and exceedingly
low base level on the north side of the delta crest,
once oodwaters were streaming into the Salton
Basin they became temporarily entrenched while
wet climate conditions continued. When streamow
subsided signicantly with a shi to a drier climate,
sedimentation blocked distributaries, and the river’s
ow was once again limited to its mainstem channel.
(c) Prolonged periods of extremely wet hydroclimate:
Lengthy periods of extraordinarily high streamow
lasting for decades, centuries, or millennia caused
major, long-lived modications to the river’s oodplain
geomorphology and delta. Cutbacks of distributary
channels, extreme erosion at points of bifurcation,
and long-term entrenchment of the river incourses
delivering ow into the Salton Basin resulted in the
creation and perpetuation of Lake Cahuilla. Once
established, the huge lake was lled to a through-
owing condition (i.e., it was constantly overowing

j. e. ross | formation of california’s salton sea in 1905–07 was not “accidental”
112 2020 desert symposium
at its delta sill), or was regularly lled to the point of
overowing, or was simply sustained near its highstand
level as a closed lake—depending on the amount of
Colorado River streamow available at any given time,
which was in turn dictated by variations in the region’s
overall extremely wet hydroclimate. When the climate
shied to drier conditions, Lake Cahuilla shrank,
became saline from evaporation, and may sometimes
have disappeared entirely when very lengthy droughts
occurred.
Although the position of the lower Colorado River’s
mainstem in the region below its current junction with the
Gila is not known with precision for much earlier periods,
historical documentation indicates that at least for the past
several hundred years (until the damming and control of
the river) there were four geomorphic factors crucial to
the Colorado’s oodplain morphodynamics in its lower
reaches: (1)the sharp curve to the south-southwest near
Pilot Knob made by the mainstem aer briey heading
due west from its junction with the Gila; (2) the existence
of an unusual elevated zone trending northeast across the
central delta, and forming a pronounced drainage divide
within that region; (3) the morphology and position of
the delta crest in relation to the big curve near Pilot Knob;
and (4) extraordinarily low regional base level north of the
delta crest. In combination, these factors ledwith virtual
inevitability to overow, avulsion, and bifurcation of
streamow along the river’s western levee at and below the
curve, and to ooding into the central Salton Basin during
periods of very high ow—such as occurred during
1905-07.
Other workers have suggested that base level and
gradient fully determined the issue of whether the river
owed to the Gulf or to the Salton Basin throughout the
Holocene. Howard and Lundstrom (2005) state, “In line
with evidence that in the late Holocene the huge Salton
basin lled several times to overowing (ancestral Lake
Cahuilla), an automatic delta-switching mechanism
governed by changing base levels is here proposed. In this
model, incised N-directed channels graded to below sea
level would tend to capture the river’s ow from other
delta distributaries until Lake Cahuilla lled to above sea
level. When the ow then switched back toward the Gulf,
the lake would evaporate and the cycle would renew.”
Similarly, Howard et al. (2007) assert, “We infer that
when Lake Cahuilla rose to its spillover level, the feeding
distributaries silted in and lowered their grade enough to
provide an impetus for the river to switch back to paths
down the south side [sic] of the delta to the Sea of Cortez.
Shut o from inow, evaporation of 1.8 m/yr would dry
Lake Cahuilla in a few decades, again lowering the base
level below sea level and setting the stage for another cycle
of northward diversion, downcutting, lake lling, and
spillover.”
However, this hypothesis is not consistent with the
dataonLake Cahuilla developed by Li et al. (2008a,b),
or with the historical record. Lake Cahuilla existed at its
highstand level for millennia, and could not have done so
if this cyclical switching model were correct. In addition,
there were numerous occasions when lakes much smaller
than Lake Cahuilla, with surface elevations very far
below sea level, were formed by Colorado River ow into
the Salton Basin; and then their lling was truncated
when—notwithstanding signicantly lower base level
in the Salton Basin—the river’s course switched back
toward the Gulf of California. us, while base level and
gradient have played important parts in the formation
of Salton Basin lakes, those factors have not been fully
determinative of the direction of the Colorado River’s
ow. Climate-related changes in streamow, erosional
capacity, sediment load and sedimentation, along with the
unique geomorphology of the oodplain, have all played
crucial roles.
Comparison of 1891 versus 1905
In order to understand whether the 1905 formation of the
Salton Sea was truly “accidental” and caused by human
negligence, or was the result of the lower Colorado River’s
well-established oodplain morphodynamics and avulsion
style combined with regional hydroclimate, it is useful to
compare what occurred in 1905 with what happened in
1891 prior to any man-made modication of the river’s
natural levee and distributary channels below Pilot Knob.
e 1891 ood is the event selected for this comparative
purpose because it is the only signicant lake-creating
ood that occurred prior to 1905 for which there is a
gaged discharge record of the river at Yuma.
A comparison of the river discharge that resulted in the
1891 ood into the Salton Basin with the discharge that
occurred during the 1905 ood shows that the conditions
on the lower Colorado River during 1905 were far more
hydrologically extreme than during the lake-forming
ood of 1891. e extraordinarily wet conditions during
1905, in combination with the nature of the river’s well-
established oodplain morphodynamics and avulsion
style, indicate that ooding into the Salton Basin and
initiation of Salton Sea formation would have occurred
that year even in the absence of man-made modications
to the river’s natural levee and distributary channel.
• e location where the Colorado River bifurcated in
1905 as a result of the man-made cut in the river’s
levee was along the same stretch of the lower course
below Pilot Knob, just below the US–Mexico border,
where the river broke through its natural levee in 1891.
In addition, prior to 1891 the river had previously
overtopped, avulsed, and bifurcated along that
same vulnerable stretch of its levee many times in
the absence of any human intervention. us, it’s
reasonable to conclude that the same underlying
hydrodynamic forces and characteristic oodplain
dynamics were operational in 1905 as in 1891 (and in
earlier oods).

j. e. ross | formation of california’s salton sea in 1905–07 was not “accidental”
113
2020 desert symposium
• ere was a delay between the river overtopping its
natural levee during ash ooding in the winter of 1891
and ooding into the central Salton Basin the following
spring because the river’s distributary channels had
previously become clogged with silt and huge deposits
of blown sand during a lengthy period of drought
preceding the February 1891 high water (Schuyler,
1907). For several months the winter oodwaters
soaked into those accumulated sediments and ponded
extensively. en, during the spring high water and
levee break, the obstructions in the distributary
channels were eroded and the accumulated oodwaters
were released into the central Salton Basin. In contrast,
in 1905 the river had a clear course through its
distributary channels into the central basin because the
major blockages in those channels had already been
scoured away by previous ooding (Schuyler, 1907).
us, it’s reasonable to conclude that the unblocked
conditions of the distributary channels in 1905 made it
even more likely that year than in 1891 that oodwaters
would make their way into the central Salton Basin.
• In the winter of 1891, there was a seven-day period of
extreme high water on the lower Colorado River below
Yuma, as the result of major ash ooding on the
Gila and its tributaries that occurred from February
23 through March 1 (Murphy, 1906; Schuyler, 1907;
McGlashan and Dean, 1913). Later, the spring high
water that occurred during two periods in May and
June was not in itself remarkable, but it was able to
break through the already-weakened western levee
below Pilot Knob (Schuyler, 1907; McGlashan and
Dean, 1913). In contrast, in 1905, an extraordinary
series of seven major back-to-back oods occurred
on the Gila River and its tributaries from January 15
to April 30 that pushed the lower Colorado River’s
discharge below Yuma to extreme levels for prolonged
periods (Murphy, 1906; USGS, 1906; McGlashan and
Dean, 1913). Following the January–April oods, very
high streamow (over 19,000 second-feet; details in the
next paragraph) continued during May and July, and
extreme discharge (over 50,000 second-feet; details in
the next paragraph) occurred again throughout June
and from the end of November to early December
(Murphy, 1906; USGS, 1906; McGlashan and Dean,
1913).
• According to Schuyler (1907), as a general matter
the stage of the river during which overow began
below Yuma was 22.0 feet above sea level, and at
that high-water stage the lower Colorado’s discharge
was typically about 19,000 second-feet. Using those
numbers as a low-threshold indicator for what
constitutes ‘very high’ streamow, and 50,000
second-feet as a low-threshold indicator for what
constitutes ‘extreme’ streamow, a comparison of the
circumstances in 1891 and 1905 yields the conclusion
that stage and discharge of the lower Colorado River
below Yuma in 1905 were exceedingly high for much
longer than they were in 1891. During the ood
from February 23 to March 1, 1891 when gage height
ranged from 23.9 to 33.2 feet, discharge signicantly
exceeded 19,000 second-feet for a total of seven days
(McGlashan and Dean, 1913). Discharge above an
extreme level of 50,000 second-feet occurred on ve
of the seven days. During that seven-day period the
highest daily discharge achieved was 101,000 second-
feet on one day (Schuyler, 1907). In May and June 1891,
the gage height slightly exceeded 22.0 feet for a total
of 45 days, ranging mostly from 22 to 23 feet on those
days but reaching 24 to 25 feet on a total of ve days
(McGlashan and Dean, 1913). No measurements of the
associated discharge are available for that spring high-
water period. In contrast, during 1905 the gage height
was above 22.0 feet on 151 days, and on 79 of those
days it exceeded 25 feet. During the entire month of
June it was above 27 feet, and in November it reached a
maximum of 31.3 feet (USGS, 1906). Of the 148 days in
1905 for which discharge measurements are available,
the river’s discharge signicantly exceeded 19,000
second-feet during a total of 76 days (USGS, 1906;
McGlashan and Dean, 1913). Discharge higher than an
extreme level of 50,000 second-feet occurred during
a total of 26 days, and extraordinarily high discharge
greater than 70,000 second-feet occurred on 15 of those
extreme-ow days. During January-April the highest
daily discharge achieved was 111,000 second-feet,
and two other days had discharges well above 90,000
second-feet (Dickinson, 1944). In addition, another
period of extreme streamow occurred throughout
June 1905, with discharge above 50,000 second-feet
every day for which measurements are available that
month (USGS 1906; McGlashan and Dean, 1913).
During June 1905 the highest daily discharges achieved
were 94,300 and 92,400 second-feet (Dickinson, 1944).
From November 30 to December 5, 1905 another
period of ash ooding with very high streamow
occurred with discharge above 19,000 second-feet on
each of those days (USGS, 1906; McGlashan and Dean,
1913), and two days had extreme discharges of 103,000
and 77,360 second-feet (Dickinson, 1944).
Inevitability of Salton Basin ooding during high ows
It is important to note that even aer the lower
Colorado River had been forced to ow toward the
Gulf of California in January 1907 by construction of
an enormous rock dam across the site of the breach in
the western levee, there was ongoing concern about the
possibility, and even likelihood, of additional ooding
into the Imperial Valley and central Salton Basin. is
concern demonstrates an understanding that ooding
into the Salton Basin during very high discharge
conditions was a characteristic and inevitable feature of
the lower Colorado River’s hydrodynamic regime and
oodplain morphodynamics in the event of suciently

j. e. ross | formation of california’s salton sea in 1905–07 was not “accidental”
114 2020 desert symposium
wet hydroclimate. In fact, this recognition and the desire
to prevent the river from ooding into the Imperial Valley
and central Salton Basin formed the primary motivation
for the construction of Hoover Dam and other control
structures built on the upper and lower Colorado River in
succeeding decades. LaRue (1925) stated:
“To protect these lands from oods extensive
levee systems have been built and must be
maintained… Although millions of dollars have
been spent in constructing the levees, these
works alone, however well maintained, cannot
assure protection from the ood menace. ere is
grave danger that during periods of high run-o
the levees will be breached and the entire ow
of the Colorado will nd its way into Imperial
Valley and the Salton Sea. If these valuable
properties on the lower river are to be protected,
dangerous stages must be prevented by holding
back a part of the ood-making waters. e need
for ood control is therefore urgent.”
Conclusion
Formation of the Salton Sea in 1905-07 was the result of
wet regional hydroclimate and the river’s characteristic
hydrodynamic regime, oodplain morphodynamics, and
avulsion style across its delta. As a result of extremely high
ows on the lower Colorado River during that period,
the river behaved in exactly the same manner it had since
at least the latest Pleistocene when streamow was high:
by overowing, avulsing, bifurcating, and ooding into
the Salton Basin. Formation of the Salton Sea was only
“accidental” from the standpoint of the people who were
trying very hard to prevent the river from owing into
the basin in the same manner it had for millennia. Were it
not for their strenuous and persistent eorts to block the
river’s ow, the ooding would likely have been far worse.
Acknowledgements
I thank George T. Jeerson for many constructive
conversations over the course of years about the Salton
Basin and Lake Cahuilla, David M. Meko for assistance
with Figure 4b, Christian Schoneman for logistical
support, and David M. Miller and George T. Jeerson for
helpful reviews of this manuscript.
References cited
Andrews, E.D., 1991. Sediment Transport in the Colorado River
Basin. In Colorado River Ecology and Dam Management:
Proceedings of a Symposium May 24-25, 1990, Santa Fe, New
Mexico. Washington: National Academies Press. pp.54-74.
Arnal, R.E., 1961. Limnology, sedimentation, and
microorganisms of the Salton Sea, California. Geological
Society of America Bulletin 72:427-478.
Barringer, F., 2014. Preserving an Accident, the Salton Sea
in California, for the Good of Nature. New York Times,
November 10, 2014. Retrieved from: https://www.nytimes.
com/2014/11/11/us/-salton-sea-migrating-birds-preserving-
a-mistake-made-by-our-meddling-with-nature-.html.
Berg, N., 2013. e Salton Sea: An accidental oasis turned
environmental tragedy. Retrieved from: https://medium.
com/changing-city/the-salton-sea-an-accidental-oasis-
turned-environmental-tragedy-4a92a650c94.
Blake, W. P., 1854. Ancient Lake in the Colorado Desert. e
American Journal of Science and Arts, Second Series,
17:435-438.
Blake, W.P., 1858. Report of a Geological Reconnaissance in
California. New York: H. Ballèrre, 367 pp.
Blake, W.P., 1907. Lake Cahuilla, the ancient lake of the
Colorado Desert. National Geographic Magazine 18:830.
Blake, W.P., 1914. e Cahuilla Basin and Desert of the
Colorado. In MacDougal, D. T. and collaborators, e Salton
Sea: A Study of the Geography, the Geology, the Floristics,
and the Ecology of a Desert Basin. Washington, D.C.: e
Carnegie Institution of Washington. pp. 1-12.
Brown, J.S., 1923. e Salton Sea Region, California - A
Geographic, Geologic, and Hydrologic Reconnaissance with
a Guide to Desert Watering Places. U. S. Geol. Survey Water-
Supply Paper 497, 292 pp.
Cecil-Stephens, B.A., 1891. e Colorado Desert and Its Recent
Flooding. Journal of the American Geographical Society of
New York, 23:367-377.
Cory, H. T., 1915. e Imperial Valley and the Salton Sink. San
Francisco: John J. Newbegin. 457 pp.
Crow, R., Schwing, J.E., Karlstrom, K.E., Heizler, M.T.,
Pearthree, P.A., House, P.K., Dulin, S.A., Stelten, M.E.,
Crossey, L.J., 2019. Redening the Age of the Colorado River:
Geological Society of America Abstracts with Programs,
51:(5), Paper No. 134-9.
Dean, D.J., Topping, D.J., Schmidt, J.C., Griths, R.E., Sabol,
T.A., 2016. Sediment supply versus local hydraulic controls
on sediment transport and storage in a river with large
sediment loads. J. Geophys. Res. − Earth Surf. 121 (1):82–110.
Dettinger, M. D.; Ingram, B. L., 2013. e Coming Megaoods.
Scientic American, 169: 64–71.
Dibblee, T.W., Jr. 1954. Geology of the Imperial Valley region,
California. In Geology of Southern California, edited by
R.H. Jahns, California Division of Mines and Geology
Bullet in, 170(2 , 2):21-81.
Dickinson, W. E. (1944), Summary of records of surface waters
at base stations in Colorado River Basin 1891–1938, U.S.
Geol. Surv. Water Supply Pap. 918. Washington: Government
Printing Oce. 272 pp.
Dorsey, R.J. 2012. Earliest delivery of sediment from the
Colorado River to the Salton Trough at 5.3 Ma: evidence
from Split Mountain Gorge. In Search for the Pliocene: the
southern exposures, edited by R.E. Reynolds, California
State University Desert Studies Consortium, e 2012 Desert
Research Symposium, pp. 88-93.
Dorsey, R.J., Housen, B.A., Janecke, S.U., Fanning, C.M.,
and Spears, A.L.F., 2011. Stratigraphic record of basin
development within the San Andreas fault system: late
Cenozoic Fish Creek-Vallecito basin, southern California.
Geological Society of America Bulletin, 123:771–793.

j. e. ross | formation of california’s salton sea in 1905–07 was not “accidental”
115
2020 desert symposium
Emory, W.H., 1848. Notes of a Military Reconnaissance from
Fort Leavenworth, in Missouri, to San Diego, in California.
Washington: Wendell and Van Benthuysen. 614 pp.
Farr, F.C., 1918. History of Imperial County, California.
Berkeley: Elms and Franks. 516 pp.
Grunsky, C.E., 1907. e Lower Colorado River and the Salton
Basin. Transactions of the American Society of Civil
Engineers, LIX:1-62.
Hajek, E.A. and Edmonds, D.A., 2014. Is river avulsion style
controlled by oodplain morphodynamics? Geology,
42(3):199-202.
Hardy, R.W.H., 1829. Travels in the Interior of Mexico. London:
Henry Colburn and Richard Bentley. 540 pp.
Howard, K.A., and Lundstrom, S.C., 2005. e Changing
Paths of the Lower Colorado River. In Geologic and Biotic
Perspectives on Late Cenozoic Drainage History of the
Southwestern Great Basin and Lower Colorado River Region:
Conference Abstracts, edited by M.C. Reheis, U.S. Geol.
Survey Open-File Report 2005–1404, pp.12-13.
Howard, K.A., Stock, G.M., Rockwell, T.K., Schafer, J., and
Webb, R.H., 2007. Holocene Cyclical Switching of Colorado
River Water Alternatively to the Sea of Cortez or to the
Salton Sink. American Geophysical Union, Spring Meeting
2007, abstract id. H44A-05.
Howe, E.F., and Hall, W.J., 1910. e Story of the First Decade in
Imperial Valley, California. Imperial: Edgar F. Howe & Sons.
291 pp.
Imperial Irrigation District Water Department, 2020. Historical
Monthly Elevations of Salton Sea at Fig Tree John, 1904-2020.
Obtained via personal correspondence with M. Kidwell, IID
Water Master.
Ives, J.C., 1861. Report upon the Colorado River of the West:
Explored in 1857 and1858. Washington: US Government
Printing Oce. 368 pp.
Kennan, G. 1917. e Salton Sea: An Account of Harriman’s
Fight With e Colorado River. New York: e MacMillan
Company. Retrieved from: https://books.google.com/
books?id=uw64AAAAIAAJ.
Kleinhans, M.G., Ferguson, R.I., Lane, S.N., and Hardy, R.J.,
2012. Splitting rivers at their seams: bifurcation and avulsion.
Earth Surface Processes and Landforms, 38(12):47-61.
Knien, F.B., 1932. e Natural Landscape of the Colorado
Delta. University of California Publications in Geography,
5:149-24 4 .
Larkin, E.L., 1907. A ousand Men Against A River: e
Engineering Victory Over e Colorado River And e
Salton Sea”. e World’s Work: A History of Our Time. XIII:
8606–10. Retrieved from: https://books.google.com/books?id
=3IfNAAAAMAAJ&pg=PA8606#v=onepage&q&f=false
LaRue, E. C., 1916. Colorado River and Its Utilization. U.S. Geol.
Survey Water-Supply Paper 395, 257 pp.
LaRue, E.C., 1925. Water Power and Flood Control of Colorado
River Below Green River, Utah. Washington: Government
Printing Oce. 171 pp.
LeConte, J.L., 1852. On some fossils from California.
Proceedings of the Academy of Natural Sciences of
Philadelphia, Vol. V, 1850 & 1851, p. 264.
LeConte, J.L., 1855. Account of Some Volcanic Springs in
the Desert of the Colorado, in Southern California. e
American Journal of Science and Arts, Second Series,
XI X(55):1-6.
Li, H.C., Xu, X.M., Ku, T.L., You, C.F., Buchheim, H.P.,
Peters, R., 2008a. Isotopic and geochemical evidence of
palaeoclimate changes in Salton Basin, California, during
the past 20 kyr: 1. δ18O and δ13C records in lake tufa deposits.
Palaeogeography, Palaeoclimatology, Palaeoecology,
25 9 :182–197.
Li, H.C., You, C.F., Ku, T.L., Xu, X.M., Buchheim, H.P.,
Wan, N.J., Wang, R.M., Shen, M.L. 2008b. Isotopic and
geochemical evidence of palaeoclimate changes in Salton
Basin, California, during the past 20 kyr: 2. 87Sr/86Sr ratio in
lake tufa as an indicator of connection between Colorado
River and Salton Basin. Palaeogeography, Palaeoclimatology,
Palaeoecology, 259:198–212.
MacDougal, D.T., 1907. e Desert Basins of the Colorado
River. Bulletin of the American Geographical Society,
39(12):705-729.
MacDougal, D. T., 1914. e Salton Sea: A Study of the
Geography, the Geology, the Floristics, and the Ecology of a
Desert Basin. Washington, D.C.: e Carnegie Institution of
Washington. 182 pp.
MacDougal, D.T., 1915. e Salton Sea. e American Journal of
Science, Fourth Series, XXXIX(231):231-250.
McGlashan, H.D., and Dean, H.J., 1913. Water Resources of
California, Part III, Stream Measurements in the Great Basin
and Pacic Coast River Basins. Washington: Government
Printing Oce. 922 pp.
Meko, D.M., Woodhouse, C.A., Baisan, C.A., Knight, T., Lukas,
J.J., Hughes, M.K., and Salzer, M.W., 2007. Medieval Drought
in the Upper Colorado River Basin. Geophysical Research
Letters, Vol. 34, L10705, doi:10.1029/2007GL029988.
Meko, D. M., and Hirschboeck, K.K., 2008. e current drought
in context: A tree-ring based evaluation of water supply
variability for the Salt–Verde River basin, Final Report.
Available at http://www.treeow.info/sites/default/les/SRP-
II-Final-Final-Report-08-08-08.pdf (last accessed January
2020).
Muer, L.P.J., and Doe, B.R., 1968. Composition and mean age
of detritus of the Colorado River delta in the Salton Trough,
southeastern California. Journal of Sedimentary Petrology
38:384-399.
Murphy, E.C. and others, 1906. Destructive Floods in the United
States in 1905. Washington: Government Printing Oce. 87
pp.
Nijhuis, M., 2000. Accidental Refuge: Should We Save the Salton
Sea? High Country News, June 19, 2000. Retrieved from:
https://www.hcn.org/issues/181/5865.
Rigg, E.A., 1862. Correspondence between Maj. E.A. Rigg and
Col. J.H. Carleton, January 23, 1862. U.S. Department of
War, Ocial Records of the Union and Confederate Armies,
Ser. I, 50(P. I, Ch. LXII):815–818.
Rockwell, T.K., Meltzner, A.J., and Haaker, E.C., 2018. Dates of
the Two Most Recent Surface Ruptures on the Southernmost
San Andreas Fault Recalculated by Precise Dating of Lake

j. e. ross | formation of california’s salton sea in 1905–07 was not “accidental”
116 2020 desert symposium
Cahuilla Dry Periods. Bulletin of the Seismological Society
of Americ a, 108(5A):2634–2649.
Rocque, J., 1762. A General Map of North America. London:
M. A. Rocque in the Strand & A. Dury in Dukes Court St.
Martins Lane. British Museum, Catalogue of Maps, Prints,
and Drawings, forming the geographical and topographical
collection attached to the Library of his late Majesty King
George III, London, 1829.
Ross, J.E., Kidwell, S.M., Dettman, D.L., Bright, J., Dorsey, R.J.,
and Jeerson, G.T., 2020. Evidence of Pleistocene Marine
Incursions into the Salton Basin. In Changing Facies, edited
by D.M. Miller, e 2020 Desert Research Symposium, this
volume.
Sarychikhina, O., Glowacka, E., Mellors, R., Suárez-Vidal, F.,
2011. Land subsidence in the Cerro Prieto Geothermal Field,
Baja California, Mexico, from 1994 to 2005—An integrated
analysis of DInSAR, leveling and geological data. Journal of
Volcanology and Geothermal Research 204:76–90.
Schuyler, J.D., 1907. e Overow of the Colorado River into
Salton Basin. Report of J.D. Schuyler, Consulting Hydraulic
Engineer, dated March 20, 1907. James D. Schuyler Papers,
University of California at Riverside, Water Resources
Collections and Archives, Collection Number WRCA 063.
Slingerland, R., and Smith, N.D., 1998. Necessary conditions for
a meandering river avulsion. Geology 26(5):435–438.
Sperry, R.L., 1975. When the Imperial Valley Fought for its
Life. e Journal of San Diego History, San Diego Historical
Society Quarterly, Vol. 21, No. 1. Retrieved from: https://
www.sandiegohistory.org/journal/1975/january/imperial-2/.
Sykes, G., 1914. Geographical Features of the Cahuilla Basin.
In MacDougal, D. T. and collaborators, e Salton Sea: A
Study of the Geography, the Geology, the Floristics, and the
Ecology of a Desert Basin. Washington, D.C.: e Carnegie
Institution of Washington. pp.13-20.
Sykes, G., 1937. e Colorado Delta. American Geographical
Society Special Publication No. 19. Washington: Carnegie
Institution. 193 pp.
Tarbet, L.A. 1951. Imperial Valley. American Association of
Petroleum Geologists Bulletin 35:260-263.
omas, R.G., 1963. e Late Pleistocene 150 Foot Fresh Water
Beach Line of the Salton Sea Area. Bulletin of the Southern
California Academy of Sciences, 62(1):9-17.
Tooth, S. 2000. Process, form and change in dryland rivers: a
review of recent research. Earth-Science Reviews 51:67-107.
USGS, 2020. National Water Information System, USGS
10254005 Salton Sea NR Westmorland CA, water
surface elevation above NGVD 1929 on 24 January 2020.
Retrieved from: https://waterdata.usgs.gov/ca/nwis/
uv?site_no=10254005.
Waters, M.R., 1983. Late Holocene lacustrine chronology and
archaeology of ancient Lake Cahuilla, California. Quat. Res.
19(3):373–387.
Wheeler, G.M., 1876. Annual report on the geographical surveys
West of the one-hundredth meridian, in California, Nevada,
Utah, Colorado, Wyoming, New Mexico, Arizona, and
Montana, Appendix JJ. In Annual Report of the Chief of
Engineers for 1876. Washington, D.C.: Government Printing
Oce. 355 pp.
Wilke, P.J., 1978. Late Prehistoric Human Ecology at Lake
Cahuilla, Coachella Valley, California. Berkeley: University
of California Archaeological Research Facility Contributions
No. 38. 168 pp.
Winker, C.D., and Kidwell, S.M., 1986. Paleocurrent evidence
for lateral displacement of the Pliocene Colorado River delta
by the San Andreas fault system, southeastern California.
Ge ol ogy, 14:78 8-791.
Winker, C.D., 1987. Neogene stratigraphy of the Fish Creek –
Vallecito section, southern California: implications for early
history of the northern Gulf of California and Colorado
delta. Ph.D. dissertation, University of Arizona, Tucson, 494
p.
Winker, C.D. and Kidwell, S.M., 1996. Stratigraphy of a
marine ri basin: Neogene of the western Salton Trough,
California. In Field Conference Guide, edited by P.L. Abbott
and J.D. Cooper, Pacic Section of American Association
of Petroleum Geologists, GB 73, Pacic Section Society of
Economic Paleontologists and Mineralogists, Book 80, pp.
295-336.
Woodhouse, C. A., Gray, S.T., and Meko, D.M., 2006. Updated
streamow reconstructions for the Upper Colorado
River Basin. Water Resour. Res. 42, W05415, 16 pp.
doi:10.1029/2005WR004455.
Wright, H. M., Vazquez, J.A., Champion, D.E., Calvert, A.T.,
Mangan, M.T., Stelten, M.E., Cooper, K.M., Herzig, C., and
Schriener, A.Jr., 2015. Episodic Holocene Eruption of the
Salton Buttes Rhyolites, California, from Paleomagnetic,
U-, and Ar/Ar Dating. Geochemistry, Geophysics,
Geosystems, 16:1198–1210.