Commentary: The flood that filled the Salton Sea began in earnest on November 29, 1905, but it was not a singular event. As a result of decisions to supply water to an important agricultural area, the disaster seemed to occur in slow motion. Development of the irrigation system for the Imperial Valley occurred over many years and resulted in the construction of a canal that existed in both Mexico and the U.S. In 1905, one of the intakes (“cuts” or “headings”) to take the water from the Colorado River into the canal system began to erode disastrously. The quoted material below is only part of the account. I refer you to the complete book which is available gratis on Google Books.
Reference: Kennan, G. 1917. The Salton Sea: An Account of Harriman’s Fight with the Colorado River. New York:MacMillan.
“Throughout the month of August 1905, the intake continued to widen, with the caving away of its banks, and in September Mr. Harriman and President Randolph decided that an other effort must be made either to close the break, or to regulate and control the flow of water through it. About the first of October, at the suggestion and under the supervision of Mr. E. S. Edinger, a Southern Pacific engineer, an attempt was made to close the channel west of the island by means of a six-hundred-foot barrier-dam of piling, brush-mattresses and sandbags. This dam, which was built in October and November at a cost of about $60,000, might perhaps have checked or lessened the flow through the crevasse if nothing unforeseen had happened; but on the 29th-30th of November a tremendous flood, carrying great masses of driftwood, came down the Gila and increased the discharge of the Colorado from 12,000 to 115,000 cubic feet per second. The dam could not withstand such pressure, and even before the peak of the flood was reached it went out altogether, leaving hardly a vestige behind. As a large part of the island was eroded and carried away at the same time, further operations in this locality were regarded as impracticable. The crevasse had then widened to six hundred feet, and nearly the whole of the river poured through it into the deepest part of the Sink, where there was already a lake with a surface area of one hundred and fifty square miles. The main line of the Southern Pacific, in many places, was almost awash, and the whole population of the Valley was alarmed by the prospect of being drowned out. If the break could not be closed and the river brought under control before the period of high water in the spring and summer of 1906, it seemed more than probable that sixty miles of the Southern Pacific track would be sub merged; that the irrigation system of the California Development Company would be destroyed; and that the whole basin of the Imperial Valley would ultimately become a fresh-water lake.
The difficulty of dealing with this menacing situation was greatly increased by the necessity of furnishing an uninterrupted supply of water to the farmers of the valley while engineering operations were in progress. It would not do to shut the river out altogether, because that would leave without irrigation nearly two hundred square miles of cultivated land. The Colorado must be controlled, but not wholly excluded. Several methods of solving this problem were suggested, but the only two that seemed likely to succeed were advocated by Consulting Engineer Schuyler and Chief Engineer Rockwood. Mr. Schuyler proposed that a new steel-and-concrete head-gate be put in near Pilot Knob, where a solid rock foundation could be secured; that the four miles of silted channel be re-excavated and enlarged by a powerful steam dredge specially built for the purpose; and that the whole low-water flow of the river be then turned through this head-gate into the enlarged canal and thence into the Alamo barranca [deep gully] west of the break. By this means the settlers would be continuously supplied with water, while the crevasse-opening would be left dry enough to close with a permanent levee or dam. The whole work, it was thought, could be finished in three months, or at least before the coming of the next summer flood….
The task [to close the eroding intake and put the Colorado River back on its previous course] set before Messrs. Randolph, Cory, Hind and Clarke was one that might well have daunted even engineers of their great ability and experience. As the  summer flood approached its maximum, in the latter part of June, the crevasse widened to more than half a mile, and the whole river, rushing through the break, spread out over an area eight or ten miles in width, and then, collecting in separate streams as it ran down the slope of the basin, discharged at last into the Salton Sea through the flooded channel of the New River barranca. Thousands of acres of land, covered with growing crops, were inundated, and thousands of acres more were so eroded and furrowed by the torrential streams that they never could be cultivated again. The works of the New Liver pool Salt Company were buried under sixty feet of water; the towns of Calexico and Mexican” were partially destroyed, and in many places the tracks of the Inter-California Rail road (a branch of the Southern Pacific) and the Holtville Interurban were deeply sub merged or wholly carried away. The wooden flumes which carried the irrigating water over the New River barranca were swept down into the Salton Sea, and 30,000 acres of cultivated land in the western part of the Valley became dry, barren and uninhabitable. At the height of the flood, the Colorado discharged through the crevasse more than 75,000 cubic feet of water per second, or six billion cubic feet every twenty four hours, while the Salton Sea, into which this immense volume of water was poured, rose at the rate of seven inches per day over an area of four hundred square miles. The main line of the Southern Pacific was soon inundated, and five times in the course of the summer the company had to move its track to higher ground.”
Reference: Kennan, G. (1917). The Salton Sea—An Account of Harriman’s Fight with the Colorado River. New York:MacMillian. beginning at page 50.