delivered the opinion of the court:
The plaintiff railroad sues the United States for just compensation for the alleged impairment of its right-of-way resulting from the construction of a dam in the Rio Grande River and the creation of a large reservoir by the impounding of the water of the river.
The line of railroad here involved runs from Albuquerque, New Mexico to El Paso, Texas, a distance of 254.1 miles. It was built in 1880 and acquired by the plaintiff in 1899. It roughly follows the course of the Rio Grande River which flows south through Albuquerque and El Paso. About 100 miles south of Albuquerque the railroad descends into the river valley at a point near the former town of San Marcial, New Mexico, and runs diagonally across the valley and the river, crossing the river on Bridge 1006-A. Prior to 1920 it crossed at Bridge 1005-A which was then located about a half-mile north of the present 1006-A. After crossing the bridge, the railroad climbs out of the valley.
By statutory authority granted in 1905, the United States constructed a dam at Elephant Butte, New Mexico, which *341is approximately 42 miles south of Bridge 1006-A. The Government acquired land and flowage rights extending from the site of the dam to a point just north of Bridge 1006-A, as to other owners of land in the valley, but did not acquire any flowage rights from the plaintiff.
The spillway elevation of the dam was only two feet lower than the level of the stream bed at Bridge 1006-A in 1915 when the dam first began to impound water. In 1924 the head of the reservoir came to within 4.6 miles of the plaintiff’s bridge. In May and June of 1942, though the level of water in the reservoir was much higher than in 1924, the head of the reservoir was 5.2 miles downstream from the bridge. This was because the stream bed downstream from the bridge was aggraded, i.e. built up with sediment, by 1942. The plaintiff says that the Government’s reservoir was and is the cause of the rise in the level of the stream bed which will make the plaintiff’s location in the valley untenable unless it spends some $5,000,000 to raise its bridge and approaches. Eather than making this expenditure it will, for some $3,000,000, move its track out of the valley of the river. Its suit is for the lesser sum, and for some $700,000 which it spent because of a special situation in 1949.
The plaintiff’s past and prospective difficulties are caused by the aggradation, by deposits of silt, of the bed of the river in the area of its bridge and approaches. When flowing water, laden with sediment, collides with still water, or water flowing with less velocity, the sediment is dropped and the channel is aggraded. The plaintiff says that that process, caused by the Government’s reservoir, is responsible for the aggradation of the stream bed of the Eio Grande in the plaintiff’s area.
The Government says that if any of the aggradation has been caused by the reservoir, it is only a small and unmeasurable part of the total aggradation. We have found that from the beginning of its operation the plaintiff has been engaged in a constant struggle to maintain its tracks, embankments, dikes and bridges against the effect of floods and sedimentation in the valley of the Eio Grande. Our finding 25 recites the history of this struggle. These events extended from 1881 to 1949, and the plaintiff does not claim *342that any of its expenditures, prior to 1949, were attributable to the reservoir.
The Eio Grande rises in the mountains of Colorado, where the melting snows and other precipitation amount to some 40 inches per annum. It is a clear stream until several tributaries in New Mexico, which tributaries flow through soil which is easily eroded, deposit great quantities of sediment at their mouths. These deposits, or deltas, are usually made during the summer floods in New Mexico, when the Eio Grande itself is carrying very little water. During the next spring run-off the Eio Grande scours these deltas away and carries the sediment down stream. In some of the tributary streams, the amount and rate of erosion have greatly increased during the time for which there are records. Erosion has also been increased because of the removal of vegetation from the uplands of the tributaries by overgrazing.
About the year 1930 a new plant appeared along the Eio Grande, the salt cedar or tamarisk. It thrives along the banks of the stream and on the deltas. Its branches begin close to the ground and it has a thick root structure. It entangles rubbish and slows down the flow of the water causing it to drop its sediment. Its growth has substantially affected the course and habits of the river.
In a state of nature the Eio Grande had a valley, in the area of the plaintiff’s installations, of some 3,500 feet, and it shifted its channel from time to time, as the existing channel filled up with sediment so that a flood would break through the banks to the lower land outside the banks. The plaintiff’s dikes and embankments directed the flow of the water through its bridge. As the river spread out below the bridge its velocity decreased and it dropped its sediment and aggraded the channel.
The parties are agreed that, if the Government’s dam and reservoir had not been built, there would have been aggradation of the bed of the river at an average rale of 0.27 of a foot per year. If this average rate had been maintained from 1895 to 1954, the elevation of the riverbed in 1954 at the plaintiff’s bridge would have been 4,464.6 feet. The actual elevation was 2% feet more than that. The average *343stated above bas little relation to what happened in any particular year. For example, between April 1937 and October 1941, there was a degradation of seven feet, because of a break-through of a bank; thereafter there was unusually heavy aggradation to 1949, then degradation to 1954.
The level of water in the reservoir fluctuates greatly from time to time. In consequence the location of the head of the reservoir, the place where the flowing water meets the still water, and tends to drop its sediment, fluctuates by many miles. If the delta at one time is formed up the stream, closer to the plaintiff’s bridge, and later the water in the reservoir is low, the flow of the water will tend to erode the former delta and carry the sediment down to the current head of the reservoir.
The Government has, since 1940, undertaken what are known as Middle Eio Grande Valley flood control and rehabilitation projects. A dam on the Eio Jemez, one of the lesser tributaries of the Eio Grande, is in operation. It will detain approximately 12 percent of the total sediment previously entering the Eio Grande from its tributaries. Other dams in the project will have a similar effect on a small scale. They will also, to some extent, even out the flow of water in the Eio Grande, and thus tend to prevent floods which cause aggradation.
From the above recital it is evident that there are numerous factors which have a part in producing the aggradation and degradation of the bed of the Eio Grande. It is a highly capricious stream, and it would be guesswork for us to attempt to determine how it will act in the future, or even the reasons why it has acted as it has acted in the past. When the plaintiff located its railroad in the valley of this river, it subjected itself to hazards and expenses which it has had to bear, almost from its first day in the valley. The plaintiff has not proved that the Government’s dam and reservoir have increased or will increase those hazards and expenses by any amount even approximately measurable.
The plaintiff’s action is premature. It may happen that in the future a condition may develop in which it is possible to determine, with reasonable certainty, that the defendant’s reservoir has produced an ascertainable harmful effect upon the plaintiff’s property. Our instant judgment is not in*344tended to operate as a bar to a future suit based upon such a condition.
The plaintiff’s petition will be dismissed without prejudice.
It is so ordered.
Littleton, Judge (Bet.) ; Laeamoee, Judge; Wbutakee, Judge; and JoNes, Chief Judge, concur.FINDINGS OF FACT
The court, having considered the evidence, the report of Trial Commissioner C. Murray Bernhardt, and the briefs and argument of counsel, makes findings of fact as follows:
1. Plaintiff railroad, a Kansas corporation, owns a right-of-way between Albuquerque, New Mexico, and El Paso, Texas, running through the valley of the Rio Grande River and through or near the former town of San Marcial, New Mexico, which lies 102.7 miles south of Albuquerque. The railroad between Albuquerque and San Marcial was built in 1880, was acquired by plaintiff in 1899, and has been used ever since by plaintiff for railway purposes.
2. The plaintiff is suing to recover damages in the amount of $3,705,910.81, of which $3,005,236.50 represents estimated costs for moving tracks and acquiring new right-of-way, and $700,674.31 represents costs incurred in 1949 and 1950 in raising grades, building a bridge, and running a silt survey. It bases its right to recover on the claim that, as a direct result of the construction of Elephant Butte Dam and the attainment in 1942 of the maximum level of the reservoir behind it, extraordinary deposits of silt and sediment accumulated in the area of plaintiff’s Bridge 1006-A and the former town of San Marcial, thereby interfering with and impairing the continued use of its present right-of-way in that area. It says that this condition is permanent.
3. At a point 42.4 miles south of San Marcial, New Mexico, and Bridge 1006-A, the Reclamation Service of the United States Department of the Interior built, and has since maintained, the Elephant Butte Dam across the Rio Grande and, to the north of it, the Elephant Butte Reservoir created by the dam. They were located and constructed from 1911 to completion in 1916, pursuant to the Act of February 25,1905, 33 Stat. 814. Water storage commenced January 6, 1915.
*3454. For the construction of the said dam and reservoir the defendant acquired, by purchase or condemnation, various tracts of land at the upper or northerly end of the reservoir and also certain flowage rights for the flooding of other lands adjacent to the reservoir. The lands and the rights originally acquired by the defendant did not include any land or right-of-way owned by the plaintiff at that time. The flowage rights so acquired by the defendant ended just south of plaintiff’s former Bridge 1005-A, which was immediately south of San Marcial, did not include the site of the present Bridge 1006-A located at the base of Black Mesa, and included lands on both sides of the Eio Grande below about elevation 4452.3 mean sea level.
5. The spillway level or spillway elevation of the dam, representing the lip of the structure of the dam, is 4,450.3 feet. With the reservoir at that level the dam discharges 10,400 cubic feet per second through its sluice gates. The maximum discharge capacity of the dam with a four-foot crest (i.e., depth of water passing over the spillway four feet in excess of the spillway level) is 33,000 cubic feet per second. Throughout its history the reservoir has been full only on one occasion, namely, during a flood in 1942, at which time the maximum crest was 2.19 feet over spillway level and the maximum discharge was 8,220 cubic feet per second. The elevation of the Eio Grande riverbed at Bridge 1005-A in 1915 was 4,452.3, or .19 feet lower than the peak crest elevation of the reservoir reached in the 1942 flood. However, at the height of the 1942 flood the head of the reservoir was 5.2 miles downstream from Bridge 1006-A due to the accumulation of sediment in intervening years through that five-mile stretch of the riverbed.
PREDICTABILITY OE RESERVOIR INFLUENCE ON AGGRADATION
6. On March 24,1914, an inspector for the Department of Justice stationed at San Marcial wrote his superior as follows:
The town of San Marcial is situated 40 miles above the Elephant Butte Dam project on the Eio Grande Eiver. When said dam is completed and filled with water, the upper end of the reservoir will reach to within a short distance of San Marcial. The Eio Grande *346River is silting up very rapidly. I believe the riverbed has raised at least 2 feet since I came to San Marcial in December 1910. The land on which the town of San Marcial is built is now lower than the bed of the river and consequently very wet. We are protected from constant inundation by dikes between us and the river. Before many seasons have passed it will be practically impossible for San Marcial to continue as a town owing to the constant silting up of the river. The property owners here realize this condition; they know it so they figure the dam below will so retard the flow of the river past San Marcial as to cause the river to silt up much more rapidly still, and the result will be that all the property owners will unite in a suit against the United States Government to secure damages because the dam has made (rather will have made) San Marcial untenable as aplace to longer live in, when as a fact the ordinary silting up of the river will soon bring about the same condition without the influence of the dam. * * * The Santa Fe Railroad may also be party to the suit. * * *
7. On April 15, 1914, the Supervising Engineer of Elephant Butte Dam advised the Chief Counsel of the United State Reclamation Service by letter that the Engle Reservoir (i.e., Elephant Butte Reservoir) might produce seepage and overflows into San Marcial due to effects of the reservoir on the gradient of the Rio Grande. He observed that “If it is found that the reservoir actually affects detrimentally the overflow and seepage conditions at San Marcial, it is doubtful, in my mind, whether the United States can escape obligations to pay for the damage done.”
8. By letter dated May 12, 1914, P. W. Dent, an attorney in the United States Reclamation Service (now Bureau of Reclamation), submitted an opinion to the Chief Counsel of that Service relating to the question whether the United States would be liable for damages if the channel of the Rio Grande near San Marcial should silt up as a result of the construction of Elephant Butte Dam. In that opinion Mr. Dent stated that a concrete conclusion could not be formulated because sufficient data was not available upon which to predicate such a conclusion. He, therefore, suggested that an investigation be made to determine “pertinent facts so as to be fortified at the proper time with evidence to show the true conditions should suit be instituted.”
*3479. By memorandum dated May 14, 1914, Will B. King, Chief Counsel of the United States Beclamation Service, suggested that the Project Manager carefully study the conditions relating to the silting of the river and obtain all necessary data so that the Government might have a continuous record of the conditions. The Project Manager informed the Chief Counsel that the collection of such data would be undertaken.
10. On November 4, 1914, an engineer for the United States Beclamation Service submitted a report to his superiors concerning the possible future effect of Elephant Butte Beservoir, stating in part as follows:
A Past River Ohcmges.
It is said that when the railway company located its shops at San Marcial the site was dry and well above the river, and it is obvious that they would not have chosen the site if it had then been as low and subject to overflow as it is now. In general the river appears to be gradually raising its bed or channel and overflowing its banks, and if not restricted, occasionally changes its course, making a new channel for some distance. This is the general tendency of the Bio Grande in southern New Mexico, and it may be due to the filling up of its channel by sand and gravel brought in by tributary arroyos, exceeding what the river itself is able to transport.
5. Dikes and, Ditches.
The railway company has built extensive dikes or levees to protect its tracks and yards from overflow by the river. They have raised the levees higher several times and riprapped them with rock. These levees also protect the town from overflow from the river. At an ordinary stage the river is two feet higher than the main streets of San Marcial. Seepage water stands in ponds and in the gutters throughout the town. The arroyos in the hills west of town sometimes bring down sudden, heavy floods which formerly caused water to stand a foot or two in depth in the town. A few years ago the town built a system of drainage ditches to carry off the water from the arroyos and to lower the water plane from seepage. The main ditch is over two miles in length with little fall, but the drainage system has considerably improved conditions.
6. The town has also built a dike from the railroad bridge to the foothills to the west, to prevent back *348water from the river flooding the town. The drain ditch passes through this dike and at this point gates are provided which can be closed when the river is in flood. The effect of closing these gates is that the town then has no possible drainage without pumping, and the river outside of the levees would be higher than the ground in the town.
J]í }& # ‡ ❖
10. The railroad runs through this swamp south of the bridge [Bridge 1005-A] and for 3000 feet or more the track is between elevations 4411 and 4412 [4454.3 and 4455.3 MSL1], while the right of way on each side is four or five feet lower. No lands or rights have been acquired from the railway company on account of the reservoir at this point, and it remains an open question whether any such action is necessary or advisable.
11. Plaintiff’s Vice President, W. B. Storey, on May 25, 1915, in a letter to Arthur P. Davis, Director of the United States Beclamation Service, requested information as to the relation between the elevation of the railroad bridge near San Marcial and the crest of the dam at Elephant Butte. Storey also expressed the belief that if the waters of the Bio Grande were backed up to San Marcial the plaintiff would have to abandon both its terminal facilities at San Marcial and many miles of road north of that point. Be-plying to Mr. Storey on May 28, 1915, Davis asserted that there was over 40 feet leeway between the elevation of the plaintiff’s track and the flood level of the reservoir and, therefore, seemingly no possibility of interference. However, on June 9,1915, Davis wrote to H. E. Baldwin, Through Chief of Construction, confiding that it was “startling” that a letter from Baldwin, giving the capacity of Elephant Butte Eeservoir up to 4410 feet (4453.3 m.s.l.), contained a notation directing that 42.512 feet be added for sea level datum, this amount being just the amount of leeway be*349tween the spillway level and the tracks of the Santa Fe at San Marcial.
12. In his letter to Mr. Storey of May 28,1915, referred to in the preceding finding, Mr. Davis stated that it was the Government’s plan to open the spillway gates if the water in the reservoir rose to an elevation of 4400 feet (4443.3 m.s.l.), the spillway elevation being 4401 feet (4450.3 m.s.l.), so as to prevent destructive floods below the dam and avoid the necessity of constructing levees along the river below the dam. This policy was not strictly followed during part of the 1942 flood (finding 40, infra) with the consequence that for one month water flowed over the spillway while the spillway gates were partially closed.
13. On March 21, 1916, the District Counsel at El Paso for the United States Eeclamation Service wrote to the Chief Counsel at Washington, with copies to the Construction Engineer and Project Manager, recommending that the reservoir be not entirely filled for two or three years longer as it might be that the reservoir would contribute to the damage at San Marcial when entirely filled. He further intimated that no information should be furnished to the Santa Fe Bailway.
14. In October 1916 Arthur P. Davis, Chief Engineer of the project, in an address given at the dedication of Elephant Butte Dam, predicted that the dam would entrap one million acre-feet of mud in its first 50 years of existence, the mud extending 40 miles upstream. The original (1915) capacity of the reservoir was 2,634,800 acre-feet. Siltation had reduced this capacity by 437,200 acre-feet by 1947.
15. The natural tendencies of the Eio Grande to aggrade in the San Marcial area and elsewhere were known to the designers, engineers, and officials charged with the planning and construction of Elephant Butte Dam, as well as to plaintiff’s engineers of that day. As early as 1917 the plaintiff was seriously considering the removal of its line from the river valley in the San Marcial area to another location paralleling the river. It was suspected that the filling of the Elephant Butte Eeservoir would increase sedimentation conditions in the San Marcial area, with attendant impairment to plaintiff’s right-of-way. There had been only limited stream measurement data in the San Marcial area prior *350to 1911 so that it was difficult to predict when or how often the reservoir would overflow. Compared to today, the knowledge of river behavior and sedimentation possessed by engineers, geologists and hydrologists prior to 1915 was imperfect. From the vantage point of today it is evident that, at least out of abundance of caution, more land or flowage rights should have been acquired by the Government for the reservoir.
STATISTICS OE AGGRADATION
16. The Rio Grande is, and has been at all times material to this action, a heavy silt- and sediment-bearing stream, and was known to be such prior to the construction of Elephant Butte Dam. It has been an aggrading stream for thousands of years.
17. Through the years relevant to the issues records have been maintained, somewhat incompletely, of stream flows and elevations of the river valley at various points across the valley above and below San Marcial. To systematize the taking of such data measurements were made by the defendant in 1914, 1916, 1917, 1920, 1925-26, 1929, 1936 and 1954 across the valley along range lines bisecting the river at roughly right angles. The plaintiff also made similar measurements in certain years, including 1950 and 1954. The range lines between which data was provided in detail in the present record cover a distance of about 13.2 miles measured along the axis of the valley, starting at range line N 9.5 miles north of Bridge 1006-A and ending at range line 15 about 3.7 miles below the southern end of the bridge. These range lines are coextensive with mileposts 997 southward to 1010 on plaintiff’s tracks. Reading from north to south the range lines are lettered or numbered as N, M, L, K, J, I, H, G, F, E, D, C, B, A, AA, 12, 13,14 and 15. Range lines F, G, H, I and J were first established in 1926; K, L, M and N in 1936; and the rest in 1914. Bridge 1006-A is about 150 feet north of range line B. Bridge 1005-A was, until its abandonment in 1920, about 500 feet south of range line C.
18. The following tables A through H summarize the ag-gradation and degradation which has occurred in the valley and the channel of the river from 1914 through late 1954 between range lines N and 15.
*351
*352
*353
*354
*35519. The elevations of the low points in the riverbed of the Eio Grande near San Marcial at approximately the location of Bridge 1006-A at various times were as follows:
Feet
January 1895_ 4448.4
January 1906- 4448.1
November 1908_ 4451.3
July 1915- 4452. 0
July 1920- 4452.3
December 1924_ 4456.6
September 1931_ 4457.4
April 1937- 4464.3
October 1941_ 4453.1
April 1943_ 4463.7
March 1949_ 4469.8
October 1954_ 4467.2
The reported elevations are taken from Defendant’s Exhibit 32 and differ in some respects from elevations at the same time and place reflected in table A in finding 18, which is based upon measurements made in particular years during surveys. Defendant’s Exhibit 32 is a continuous record of stream bed elevations derived from stream flow measurements, properly adjusted to reflect stream bed elevations at the location of Bridge 1006-A from 1895 to late 1954.
20. Based on the elevations reported in the preceding finding, it is apparent that the bed of the Rio Grande at or near Bridge 1006-A aggraded a net total of 21.4 feet from 1895 to March 1949. The increases did not occur at a uniform rate throughout the period, but in six identifiable phases of from three to twelve years each, with periods of slight or no increase being followed by periods of moderate or marked increase. Within each of the six principal phases of aggradation there were minor peaks and valleys indicating subordinate or seasonal fluctuations not significant enough to vary the trend of the phase. The following table, based on the data in finding 19, identifies each of the six aggradation phases and reports, as to each, the gross aggradation and the average annual rate of aggradation at or near the location of Bridge 1006-A:
*356
During tbe time gap in the preceding table between April 1937 and October 1941, the riverbed at Bridge 1006-A degraded sharply to elevation 4457.4 feet due to an avulsion of the river at Val Verde, just above San Marcial in the eastern part of the river valley, as described more fully in finding 23 a, infra. Except for this 1937-1941 episode of marked degradation, the period from 1931 to 1949 would appear to be one of continuous aggradation at an average amiual rate of about .7 feet. The rate of incline reported for the period October 1941-March 1949 is unduly accentuated for the reason that it starts at the depth of the Val Verde degradation period instead of a more normal level. (Cf. findings 23a, 40 and 44.)
21. It is reasonably concluded, and the parties are in agreement, that the natural rate of aggradation of the riverbed at or near the location of Bridge 1006-A, assuming the absence of the Elephant Butte Reservoir, would be about .27 feet per year. If this natural rate of aggradation had been uniformly experienced each year starting with 1895 the elevation of the riverbed in October 1954 at the bridge would theoretically have been 4464.6 feet, or 2.6 feet lower than its actual elevation at that time. But it must be remembered that the October 1954 elevation was itself 2.6 feet lower than the peak elevation in March 1949, due to the continuation of the latest degradation phase.
CONSEQUENCES OP AGGRADATION
22. The aggradation summarized in the tables in findings 18 through 20 affected both the regimen of the river and the existence of the plaintiff’s tracks, embankments, dikes *357and bridges in tlie valley of the Rio Grande in the San Marcial area.
23. a. The principal effect on the regimen of the river was to cause three local avulsions in the past 60 years. The first occurred in the period 1920-1928 when the river partially left its channel between the locations of Bridges 1005-A and 1006-A and formed an alternate channel on the eastern side of the valley bordering the Black Mesa, which channel it still occupies. During a major flood in 1929 the river avulted again, flooded San Marcial and the low ground to the south, broke the cross dike at the south end of Bridge 1005-A, and flowed into the new channel to Bridge 1006-A. The low ground in and around San Marcial was partially buried in sediment and aggraded until the swamping extended upstream some distance. In the major flood of 1937 the river avulted once more, this time above San Marcial at Yal Yerde, and adopted a new chamiel across the Val Yerde area to the south over an area of depression. In spreading out over this extensive area the river lost velocity and consequently a large amount of sediment was deposited in the Yal Verde area. By the time it rejoined the stream below Yal Yerde at the northern foot of the Black Mesa, somewhat above Bridge 1006-A, the water had lost much of its sediment load and its clarified condition restored its ability to transport sediment. Because of this the volume of water rejoining the river scoured the channel down through Bridge 1006-A and for several miles downstream. The scouring effect continued until late 1941 by which time the Val Verde area had been silted up and the river gradually returned to its previous channel. In late 1941 and 1942 the previously scoured channel filled up rapidly with large quantities of sediment carried by the heavy flows in those years. It is characteristic that sediment will deposit more rapidly in areas of the riverbed which have experienced rapid scouring action, for the river constantly seeks to maintain its equilibrium. The scouring or degradation accompanying the last avulsion commenced suddenly in April 1937, reached a maximum depth of 11.2 feet in late 1941, and thereafter filled up or aggraded even more rapidly than it had scoured so that by mid-1943 the *358riverbed had almost regained its peak elevation of April 1937 as it existed immediately prior to the avulsion, thereby nullifying the effects of the scour.
b. The mechanics of avulsion of a sedimentary, aggrad-ing stream such as the Eio Grande are as follows: In a time of high flows, as the water laps over the banks of the channel and spreads out over the flood plain, the first concentrated deposition of sediment occurs immediately at the hanks of the channel, because that is where the water first reduces velocity and drops its sediment. In time this builds a natural levee or dike on each bank of the river paralleling its course. Within these natural levees the bed of the river continues to aggrade until the channel bed itself is higher than portions of the flood plain on either side of the natural levees. Exceptionally high flows will then cause the river to break through the natural levees confining it and inundate the lower flood plains of the surrounding valley, forming a new channel at the lower elevations. The river cannot then return to its original channel until the process of aggradation repeats itself, natural levees again develop, the new stream bed aggrades to higher levels than the flood plain, an exceptional flood again breaches the levees, and the river avults once more. This process continues, now in one portion of the valley and again in another portion, back and forth across the valley, and eventually results in general aggradation of the entire valley. The process depends on periodic floods. There have been 54 flood flows in excess of 10,000 cubic feet per second from 1896 to 1954 at San Marcial, including 19 in excess of 15,000 cubic feet per second, which are considered major floods. At the time of trial an avulsion appeared imminent immediately below Bridge 1006-A because the riverbed was higher than areas of the flood plain lying west of it and awaited only a flood of sufficient proportions to breach the natural levees formed on the west bank of the channel.
24. The aggradation reported statistically in the tables in findings 18 through 20 has also affected the plaintiff’s structures in the valley above and below San Marcial, including its tracks, embankments, dikes and bridges. Plaintiff’s tracks and embankments must be constructed and maintained *359at sufficient levels above the surrounding terrain to avoid disruption by floods. Its dikes must be constructed and maintained to fend off flood waters and to control their direction. Its bridges crossing the channel must be maintained at levels adequate to permit the passage of maximum floods. When bridges are raised to higher elevations the tracks and embankments constituting their approaches must be raised accordingly for an appreciable distance to satisfy grade requirements. From the beginning the plaintiff has been engaged in a constant struggle to maintain its tracks, embankments, dikes and bridges from the effects of floods and sedimentation in the valley of the Eio Grande from milepost 993 to milepost 1009. The following finding details the dikes constructed by the plaintiff and others, the changes in locations and elevations of plaintiff’s tracks and bridges, and the events provoking these activities.
25. a. In 1881 the plaintiff or its predecessor constructed a dike extending 1,050 feet from the south end of Bridge 1005-A to the eastern foothills. The dike was designed to direct overflow water to Bridge 1005-A.
b. In 1882 the plaintiff or its predecessor converted Bridge 1005-A from a pile trestle to a steel bridge 501 feet long and the grade line was raised four feet so as to accommodate the bridge clearance over the river.
c. Prior to 1884 the plaintiff or its predecessor constructed another dike to the height of Bridge 1005-A extending along the east side of plaintiff’s tracks from Bridge 1005-A to about milepost 1004. The dike was designed to protect low areas of the plaintiff’s yard and tracks from floods.
d. A flood in 1904 caused washouts of plaintiff’s tracks at mileposts 1003.3 and 1003.8, and washed out part of the crossdike, but no waters entered the low area below Bridge 1005-A. The flood water topped the rails in the San Mar-cial yard by six inches, and flowed over plaintiff’s embankment at milepost 1006.4. In 1904 and 1905 the plaintiff filled the washouts at mileposts 1003.3 and 1003.8, strengthened the dike with riprap on the south end of Bridge 1005-A, and raised its grade and track a maximum of five feet from milepost 1006.2 to milepost 1006.8 to prevent a recurrence.
*360e. In 1912 San Marcial constructed a community dike extending the railroad dike beginning at the north end of Bridge 1005-A and extending along the west side of the river a short distance, then terminating at high ground on the west edge of the valley. In 1913 the Val Verde-La Mesa dike was constructed (not by the parties hereto) running for about three miles from the foot of Black Mesa northward and eastward along the east bank of the river to a point near Val Verde about two and one-half miles upstream from San Marcial.
f. Prior to 1914 San Marcial constructed a surface drain to relieve water accumulation. It drained into the river downstream and crossed the San Marcial community dike.
g. In 1920 a flood breached the railroad dike at the south side of Bridge 1005-A, flooded the low land below the bridge and washed out 2,000 feet of the plaintiff’s embankment from milepost 1006 to milepost 1006.4. After the flood the plaintiff raised the embankment a maximum of seven feet and constructed overflow Bridge 1006-A.
h. In 1923 a flood breached plaintiff’s crossdike and washed out about 80 feet of the embankment. After the flood Bridge 1006-A was extended 97 feet and its west bank was reinforced with riprap.
i. The large flood of August and September 1929 breached the railroad dike in 16 places between Bridge 1005-A and milepost 1002.8, flooded San Marcial and its rail facilities, and partially buried the town in sediment. Before the flood the plaintiff had commenced the replacement of the pile trestle Bridge 1006-A with a steel bridge. After the flood the bridge was raised and the plaintiff relocated its track between mileposts 1007 and 1007.9. Pursuant to a contract between the parties, the relocated track was placed in the reservoir bed and the plaintiff recognized the defendant’s flowage rights. The bridge was raised an additional eight feet and its approaches were raised.
]. San Marcial remained under water for a time after the 1929 flood, gradually became a swampy area, and by 1940 or 1941 it became a lake due to the drainage having become completely blocked by the deposition of sediment from the river channel.
*361k. A flood in 1937 breached the plaintiff’s dike just north of the former Bridge 1005-A, washed out the plaintiff’s embankment at mileposts 999.1 and 1003, and caused the river ta avult at Val Verde from 1937 to late 1941, at which latter date it commenced to revert to its former channel (see finding 23 a). Following the 1937 flood the plaintiff raised its grade from milepost 998.8 to 996.6 by four feet, and in 1938 raised the grade from milepost 998.6 to 1000.2 by five feet in order to get above the siltation.
l. In 1939 the plaintiff moved its track from milepost 994 to 1002.5 out of the valley to high ground because of the marked aggradation in the channel of the river in this stretch and resultant difficulty in maintaining the roadbed for the tracks. The plaintiff’s plans in 1937 to raise Bridge 1006-A still further because of the marked aggradation of the riverbed were postponed because of the rapid degradation occurring during the period 1939-1941 due to the Val Verde avulsion. However, this was a mere postponement, for by mid-1943 the riverbed had aggraded back to its original level and plaintiff raised Bridge 1006-A an additional 12 feet in that year. The approach grades were also raised from the bridge down to milepost 1007.6, and the embankment was raised for a distance of 4.5 miles.
m. The raises made in 1930 and 1943 to Bridge 1006-A added 20 feet to its original 1920 elevation. At an average annual rate of aggradation of .27 feet, which finding 21 establishes as the natural rate in the absence of the Elephant Butte Eeservoir, the 20-foot cumulative raise in the elevation of Bridge 1006-A, reached in 1943, theoretically should have permitted the bridge to accommodate until 1994 (74 years from 1920) flows of the river comparable to the capacity of the bridge as it was in 1920, except for transitory interim phases of aggradation which might temporarily reduce the flood clearance capacity of the bridge and even require it to be raised to meet a temporary situation. This estimate is further conditioned on the assumption that the river does not avult and form a new channel by-passing Bridge 1006-A. Actually, the 1943 raise increased the bridge’s capacity to pass flows of 50,000 cubic feet per second, much larger than any previous capacity. The raises in the elevation of Bridge *3621006-A through 1943 were thus made necessary by the continued aggradation of the riverbed and, in some part, to plaintiff’s new policy to maintain the bridge with a 50,000 cubic feet per second flood capacity.
n. The plaintiff does not consider the work described in paragraphs a through m of this finding to have been occasioned by or attributable to the influence of the Elephant Butte Beservoir. It considers its first expense attributable to the reservoir was in 1949, when it raised its grades from mileposts 1002.6 to 1004.5 and from mileposts 1006.6 to 1007.9.
o. For purposes of convenience, the various raises and relocations of the plaintiff’s tracks from 1905 through 1949 referred to in paragraphs a through n of this finding, together with data as to elevations and approximate locations, are summarized in the following table:
*363
*36426. While making the raises of grade in 1949 a slide occurred on March. 17, 1949, in the railroad embankment at milepost 1004.3, and Engine No. 905 overturned into San Marcial Lake. The fireman was killed and the engineer was injured. Engine No. 905 was a worktrain engine that handled dumpcars loaded with gravel from a pit track and was engaged in dumping the gravel on the fill to raise the grade. On March 22, 1949, another slide occurred at milepost 1004.4, approximately 500 feet south of the first slide, and Engine No. 1646 overturned into San Marcial Lake. No one was injured on that occasion. At the time of both slides the flood waters on the east side of the embankment were 7% feet higher than the waters on the west side of the embankment where lay San Marcial Lake.
27. The slides referred to in the preceding finding were caused by the saturated foundation of the soft clay and silt which gave way under the weight of the locomotives and under the conditions existing in March 1949. In mending the embankment the plaintiff constructed a 20-foot berm at the base of the fill so as to flatten the angle of the side slopes, thus providing adequate support for the increased height of the embankment. The slides would not have occurred had this type of construction been employed in the grade raise work performed in 1949 prior to the slides as described in finding 26. But it is not established that the plaintiff was negligent in employing the engineering methods it followed in the grade raise work preceding the slides.
28. In 1950 the plaintiff constructed overflow Bridge 1002-A, which consists of two 60-foot deck girders and one 48-foot deck girder. The purpose of that bridge was to provide an opening in the railroad embankment on the north end of San Marcial Lake. Bridge 1002-A is located just north of milepost 1003. By providing the opening, part of the flow of the Bio Grande is diverted through the railroad embankment into San Marcial Lake on the west. The effect of dividing the flow permits San Marcial Lake to become filled with sediment and thus reduces the differential in the height of the water on the two sides of the embankment. Had this been done sometime prior to the grade change in 1949, the embankment would have been stable, but when *365plaintiff built overflow Bridge 1002-A and acquired flowage rights in San Marcial Lake, it did not know that the defendant would subsequently commence the channelization project described in finding 48, infra. The cost of the 1949 raise of grade, plus the 1950 construction of Bridge 1002-A, and acquisition of flowage rights in San Marcial Lake, was $700,674.31.
CAUSES OP AGGRADATION
29. The plaintiff contends that the existence of Elephant Butte Eeservoir has been the chief cause of sedimentation increase in the valley of the Eio Grande at and in the vicinity of Bridge 1006-A, that this condition was first initiated as a result of the 1942 flood when the reservoir approached closer to the bridge than it has before or since, that the condition first became stabilized in 1949, and that the sediment deltas are permanent. The defendant attributes the sedimentation to the following causes:
a. Artificial restrictions on flow of river
b. Salt cedar infestation of the valley
c. Tributary erosion
d. Irrigation and overgrazing.
e. Influence of Elephant Butte Eeservoir
In the analysis which follows an effort has been made to preserve a distinction where possible between the effect of aggradation on plaintiff’s Bridge 1002-A, embankments, tracks and dikes in the valley on the one hand, and its effect on the continued existence of Bridge 1006-A on the other.
30. Artificial restrictions on flow of river.
a. North and south of San Marcial the Eio Grande occupies a comparatively narrow channel, only sufficiently wide and deep to confine the river in low or moderate flows. In high flows the river overflows its banks, spreads out on the flood plain of the valley floor, reduces velocity, and deposits its sediment which the reduced velocity can no longer support, all to the extent permitted by artificial barriers which have been erected over the years to direct or channel the direction and flow of the river. The flood plain in the vicinity of San Marcial varies from less than one to over two miles. The varying widths of the valley across the range lines from N down to 15 are as follows:
*366
Thus, Bridge 1006-A, which is located immediately above range line B, lies at the narrow waist of the valley, which widens considerably both above and below the bridge. In times of extreme flood the water will extend entirely across the valley floor from bank to bank if it breaches or tops the embankments and dikes erected at various times since 1881.
b. The plaintiff’s dikes referred to in paragraphs a and c of finding 25, which were constructed between 1881 and 1884 and bordered the old channel of the river in the vicinity of old Bridge 1005-A, were designed to direct the flood waters of the main channel to pass through that bridge, and had the effect of confining the flood waters at that point to about 800 feet instead of the whole 3,500-foot width of the valley. Other dikes were constructed, embankments were raised and reinforced, and bridges were built, lengthened, and raised as described in paragraphs e through 1 of finding 25. These were designed for the protection of the railroad, the former town of San Mar-cial, and the Val Verde area lying on the east side of the river above San Marcial.
c. These structures imposed an unnatural constraint upon the river, prevented it from achieving its normal development and from adopting the land forms and channel configurations which it would have sought naturally. But this general result is immaterial unless it affected or affects the security and permanence of plaintiff’s tracks, bridges, etc. From 1881 to 1913 the plaintiff’s railroad embankment between range lines F to C sealed off the western fringe of the valley containing the San Marcial area, from the river and confined the normal sedimentation effects of the river *367to tbe much broader eastern part of the valley. Adequate statistics are lacking as to the total sedimentation effect in this period, but as of 1914 the average elevation of the valley east of the plaintiff’s embankment varied from two to four feet higher than elevations at the same latitudes west of the embankment. From 1913 to 1937 the Yal Yerde dike and the plaintiff’s embankment confined the valley on the east and west, respectively, between range lines G and C from a natural maximum width of 13,750 feet to an artificial maximum width of 2,200 feet. The aggradation in this narrowed valley did not exceed a normal rate except at range line C. This may have been due to the fact that the narrowed valley increased the river’s capacity to transport suspended sediment further downstream to a point where conditions caused a decreased velocity. Nevertheless, the normal aggradation which did occur raised the level of the narrow valley as well as the channel bed above those parts of the valley shielded by the plaintiff’s embankment and the Yal Verde dike, thereby producing conditions favorable to an avulsion. The narrowed valley also, at flood times, increased the depth of the water and its pressure against the restraining embankment and dike. In the 1929 flood the plaintiff’s embankment ruptured in numerous places and San Marcial was obliterated permanently, eventually becoming a lake bordering the west side of the plaintiff’s embankment between range lines F and C. The 1937 flood breached the Yal Yerde dike and the river avulted, creating a new channel through the Val Yerde area. The 1937 Yal Yerde avulsion caused rapid aggradation in the valley east of the plaintiff’s embankment between range lines F and C, while the repaired embankment kept the valley west of it at unchanged elevations, thus creating an imbalance in elevations east and west of the embankment. Because of this the 1949 flood waters east of the embankment were 714 feet higher than the San'Marcial Lake level west of the embankment, thereby saturating and weakening the embankment and causing the slides described in finding 26, sufra, when the locomotives toppled into the lake. Thus, the plaintiff’s embankment did not appreciably increase the aggradation in the valley opposite its location, but it did *368contribute to an unbalanced distribution of the sediment so as to affect its own security. The plaintiff’s Tiffany channel project described in finding 28, supra, has been serving effectively since 1950, to restore a balance which should prolong the usefulness of the embankment and permit it to be raised indefinitely.
d. The defendant contends that Bridge 1005-A and, later, Bridge 1006-A, were responsible for the formation of fan-shaped sediment deltas immediately below their locations because they constricted the channel. The record does not support this contention as to Bridge 1005-A. A delta does exist, however, on the west bank of the river between range lines B and A immediately below Bridge 1006-A. The riverbed at this point is higher than low areas on the west side of the valley. The stage is set for an avulsion should a heavy flood break through or over the delta, which acts as a natural levee, and cause the river to change its course to the lower ground on the west. The bridge itself was raised in 1943 to accommodate flows in excess of the largest floods on record, and so by itself does not contribute to the delta. But the plaintiff’s embankment approaching the bridge narrows the valley down from a natural width of 3,500 feet to a constricted width of less than 1,000 feet. Flood waters passing under the bridge at considerable velocity escape from their confinement in the suddenly widened valley immediately below the bridge. The sudden dispersion causes a drop in velocity of the flood waters and they deposit quantities of their sediment load on the delta, thereby contributing to its increase in an unmeasurable degree. Should an avulsion occur as intimated, whether or not the permanency of Bridge 1006-A and the embankment approaching it from the north will be affected depends on the course of the avulsion. Should the river revert to its old channel which, until 1920, led through the old Bridge 1005-A just below range line C, obviously Bridge 1006-A would in time span a dry gully and the plaintiff would have to build a new bridge where its tracks would have to cross the new channel. If the avulsion were to occur below Bridge 1006-A, which appears to be a more *369logical development, the evidence does not reveal the effect, if any, on plaintiff’s facilities.
31. Salt cedar infestation of the valley.
a. Prior to 1930 the principal vegetation in the middle Eio Grande valley consisted of cottonwoods and willows. These trees have a dense canopy of leaves which, by intercepting sunlight, prevents the growth of smaller, sediment-catching brush underneath. Where present they impede the normal passage of sediment-laden water. Since approximately 1930 another type of plant known as salt cedar, or tamarisk, has appeared from unknown sources in great and growing quantities in the valleys and flood plains of the Eio Grande and other southwestern streams. By 1937 there were eight widely separated areas infested with salt cedar bordering the Eio Grande for several miles above Bridge 1006-A, and by 1947 the infestation had become virtually continuous in this stretch. Photographs and other inconclusive references in the record indicate strongly that salt cedar infestation is also prevalent below Bridge 1006-A.
b. Salt cedar is a fast-growing evergreen shrub which grows as much as two to three feet in its first year and attains a height of 12 to 15 feet. Its branches begin close to the ground and it has a thick root structure which, together with its dense growth of branches, present a greater obstacle to the flow of sediment-laden water than cottonwoods and willows.
c. Salt cedar will not take root in the active channel of the Eio Grande subjected to the flow of water variously during each year, such as in the channel under Bridge 1006-A, but only in those parts of the valley and flood plain which are not frequently covered by water. Delta formations at the head of reservoirs create conditions ideal for the growth of vegetation. Therefore, the effectiveness of salt cedar as a sediment trap occurs only during high flows when the river exceeds its normal channel and flows over surrounding areas infested with salt cedar growth. In these instances the salt cedar slows down the velocity of the water sharply and causes it to drop its sediment, thus causing the valley to aggrade perceptibly in those areas on such occasions. As was evidenced in the Yal Yerde avulsion of 1937-1941, *370the passage of the flood waters through dense stands of salt cedar so clarifies it that, in rejoining the main channel of the river, the water in its relatively clarified condition recoups its power to transport sediment and scours out the stream bed at the point of reunion and for some distance below, provided the episode is of sufficient duration. It is therefore concluded that since at least 1930 the presence of salt cedar has contributed in an unmeasurable degree to the aggradation which has occurred in the flood plain of the river above and below Bridge 1006-A, but not to the aggradation in the channel of the river. Undoubtedly the sediment delta on the west bank of the river immediately below Bridge 1006-A is attributable in part to the sediment-detention propensities of the salt cedar which covers the mounds. Much less sediment has entered the Elephant Butte Eeservoir to date than had been originally anticipated at the inception of the dam and reservoir. At least part of this loss is represented by sediment which has been detained in the valley above and below Bridge 1006-A since 1930 by the abundant presence of salt cedar.
32. Tributary erosion.
a. The Eio Grande is essentially a clear stream from its source in the Colorado mountains down to its junction with the Eio Chama in northern New Mexico. From that point downstream it carries an increasing load of sediment, large portions of it being contributed by the Eio Chama, the Eio Jemez, the Eio Puerco and the Eio Salado, the last two being the heaviest contributors. The sediment brought in to the Eio Grande by the tributaries arrives during the summer floods in the drainage basin. Because the Eio Grande habitually carries so little water during the summer months the sediment brought down by the tributaries during flash floods deposits in the parent stream at the mouth of each tributary in fanshaped or wedgeshaped formations and remains there until the next flows occur on the main river large enough to move it along downstream, usually occurring during the spring runoff. The proportion of sediment carried by the Eio Grande contributed by each of the respective tributaries has been roughly estimated as follows:
*371Percent
From Bio Grande above Cocbiti- 19
From Santa Fe Greet- 1
From Galisteo Greek_ 3
From Bio Jemez_ 12
From Bio Puerco_ 35
From Bio Salado_ 13
From minor tributaries — assumed_ 17
b. Prior to about 1880, each of the tributaries was characterized by a channel and a flood plain. The flood plains were covered with grass and other vegetation used for grazing or cut for hay. Beginning about 1885 a period of accelerated channel erosion or gullying commenced in the tributaries. Their channels were greatly deepened and broadened. As an example, in 1882 the channel of the Rio Salado at Riley was 50 feet wide and from three to ten feet deep. In 1883, during a severe flood, channel trenching and severe erosion began and continued until the channel at present is 400 to 600 feet wide and 10 to 20 feet deep at that location. Again, the Rio Puerco west of Albuquerque was, prior to 1890, 100 feet wide and 10 feet deep, while now at the same point it is 300 to 600 feet wide and 30 to' 40 feet deep. The original main channel of the Rio Puerco was only 5 percent of its eventual 1927 size. While there is some indication that the rate of tributary erosion has passed its zenith and is now in a descending trend, the end of the cycle may be as much as 100 years in the future. In the meantime tributary erosion is and will continue to be the major source of sediment entering the valley of the Rio Grande, including the San Marcial area. A dam has been built at the mouth of the Rio Jemez which will prevent sediment from that source entering the Rio Grande. Proposed dams on certain other tributaries may have a similar effect. No dams are presently proposed for the Rio Puerco or the Rio Salado, the two principal sources of sediment. The effect of the dams in existence or planned on the aggradation conditions in the San Marcial area is conjectured in finding 47, infra.
33. Irrigation and overgrazing.
The erosion of the tributaries described in the preceding finding synchronized with a period of increased climatic aridity, increased grazing, and, at least from 1880 to 1915, *372increased use of irrigation for agriculture. The first two factors tended to remove vegetation on the flood plains of the tributaries both by consumption and prevention of new or normal growth. Between 1880 and 1915 it has been estimated that as much as one-third to one-half of the river flow was removed for irrigation purposes, thus reducing the amount of water that would otherwise have been available to transport sediment in the Eio Grande, particularly in the summer months when large sediment loads were brought down into the river by its eroding tributaries. But these facts were available to and presumably known by those who planned the construction of the Elephant Butte Dam and its reservoir, prior to 1915.
34. Influence of Elephant Butte Reservoir.
The tables in finding 18, if plotted on graphs,2 would present more clearly the description of certain phases of ag-gradation in the text of succeeding findings 35 through 44, which explore the influence of the reservoir as a cause of aggradation.
35. It is a characteristic of streams carrying a large load of silt and sediment to deposit sediment at and near the point where they flow into lakes, reservoirs, or other still waters, because still water causes the velocity of the stream to decrease, thus lowering its capacity to carry silt and sediment. Due to a slight backwater effect the velocity of the Eio Grande decreases gradually for some distance above the point of actual entry into the reservoir, thus causing a deposit of sediment proportionate to the decreased velocity for a distance above the reservoir. Through the years the river has flowed into the reservoir at various pool levels and left large deposits of sediment at and above the various points where the reservoir level intersected the stream bed. Multiple deltas have been laid down in the reservoir as a result of several floods entering the reservoir at varying pool levels. The greatest aggradation occurs at times of heavy or flood flows, which also cause scouring and degradation under particular circumstances.
*37336. In order for a delta to form at the head of a reservoir the level of the reservoir must remain fairly constant for a sufficient period of time. The delta that forms is not permanent in the sense that subsequent flows cannot erode it, but it acquires in the course of time an increasing resistance to erosion as it compacts. When the level of the reservoir retreats below the delta due to low water conditions, there is a tendency for the sediment constituting the delta to be washed downstream by subsequent flows to the adjusted location of the reservoir head, for the natural tendency of the river is to maintain its equilibrium, i.e., to regain and maintain a slope parallel to its original bed. If the reservoir remains at approximately the same level for a considerable length of time so that the delta which has formed at its head has had no opportunity to dissipate, then the delta interrupts the downward passage of further sediment increments from upstream sources. In this case a wedge of sediment, originating with the delta at the head of the reservoir, builds back upstream at a greater rate than normal aggradation of the river, without existence of the reservoir, would produce. When the reservoir finally recedes, as it does and must in the case of the Elephant Butte Reservoir, a process of degradation sets in starting at the delta and working upstream, again with the object of restoring the equilibrium of the stream.3
37. From the completion in 1915-1916 of the Elephant Butte Dam, down to 1920, a larger amount of sediment was deposited throughout the length of the reservoir area, from the dam up to Bridge 1006-A, than during any other comparable period of time since then. The greatest accumulations of this sediment occurred in the area from 17 to 27 miles below the bridge. During three of these years the head of the reservoir hovered between 15 and 25 miles below the bridge, thus coinciding with the most concentrated areas of sedimentation. Aggradation at the location of Bridge 1006-A during this period was less than the normal rate, indicating that the delta which formed in *374the Narrows 22 miles below bad not affected aggradation at tbe bridge location.
38. From 1920 through 1925 heavy deposits oí sediment occurred in the river from 10 to 18 miles below the bridge, while the head of the reservoir fluctuated in that same area most of the time. A lesser, but still considerable amount of sediment was deposited during the same period from four to eight miles below the bridge location, while on three separate occasions in that time interval the reservoir head was less than five miles below the bridge location. From 1920 through 1924 the riverbed at the location of Bridge 1006-A aggraded 4.2 feet (ignoring a sudden one-foot rise and fall in January 1924), an average annual rate of .84 feet, considerably in excess of the .27 foot norm. This sudden increase at the bridge may have been the result of reservoir influence, but the period involved was relatively short and none of the highly qualified expert witnesses who testified could feel reasonably sure that the increase over normal was the product of reservoir influence.
39. From 1925 to 1935 the bed of the reservoir and valley received additional depositions of sediment from the bridge down to the dam. The heaviest part of this sedimentation occurred in the area 14 to 20 miles below the bridge, an area which also marked the location of the fluctuating reservoir head during flood flows occurring in 1926, 1927, 1929, and 1932. The sediment delta which formed during this period at the area indicated caused a wedge of sediment to extend upstream which eventually, but not immediately, reached Bridge 1006-A. For example, from 1924 to 1931 the river at the bridge location aggraded 1.3 feet, an average annual rate of .16 feet, considerably less than normal. But from 1931 to April 1937 the riverbed at the same location aggraded 4.4 feet at an average annual rate of .70 feet, a sharp increase and about 2.7 feet more than a normal aggradation rate of .27 feet per year would have produced. It is probable that the reservoir influence contributed to the increased rate of aggradation from 1931 to 1937, together with other causes heretofore discussed.
40. Between 1935 and 1947, despite the sharp but temporary period of degradation in the immediate vicinity of *375Bridge 1006-A caused by the Yal Verde avulsion of 1937-1941, further net aggradation occurred from 30 miles below the bridge to 10 miles and more above the bridge. The delta at the Narrows, 22 to 28 miles below the bridge, was considerably enlarged, but this is of no direct significance to the present issues. A considerable deposition of sediment also occurred from a point about 18 miles below the bridge which thickened substantially five miles below the bridge and continued up past the bridge for 10 miles or more in substantial but varying depths. In May and June 1941 the area experienced 39 consecutive days of extremely high flows, a condition unprecedented since prior to the construction of the dam. The reservoir filled abruptly to within seven or eight feet of capacity and the head of the reservoir reached a point about seven miles below Bridge 1006-A. In April, May and June of the following year, 1942, the area experienced an additional 32 days of flood flows. The 1942 floods poured into the reservoir, already at high levels from the floods of the preceding year, overflowed the reservoir 2.19 feet in excess of its crest elevation (the first time in its history that the reservoir had been full), and brought the head of the reservoir to within 5.2 miles below Bridge 1006-A. The level of the reservoir at its peak was .19 feet above the 1915 elevation of the stream bed at the'bridge location, but accumulations of sediment in the intervening years prevented the head of the reservoir from reaching the bridge. For V/2 years the reservoir level remained at the highest elevation in its history, and the head of the reservoir remained in closer sustained proximity to the bridge than at any previous time. The 1942 flood situation could have been partially relieved if the defendant had permitted water to discharge from the dam at a faster rate by opening the spillway gates, but for a period of one month during the 1942 flood the water flowed over the spillway while the spillway gates were partially closed, contrary to defendant’s original flood policy (finding 12, supra). The 1941 and 1942 floods deposited a large amount of sediment within a short period of time at and above the reservoir head, which fluctuated above and below seven miles from the bridge. The sediment rapidly filled in the deep *376trenches which had been eroded in the stream bed by the Val Verde avulsion of 1937-1941, another illustration of the river’s natural tendency to regain its equilibrium. The reservoir remained at relatively high levels through mid-1945, and the accumulations of sediment which had occurred five miles below the bridge when the reservoir was at its peak height hindered the downward passage of sediment continuing to come from upstream sources. From 1943, by which time the degradation caused by the Val Verde avulsion had been compensated, to March 1949, the bed of the river at the bridge aggraded 7.2 feet, or an average annual rate of 1.15 feet. The period 1936 to 1946 marked a turning point in the aggradation trend of the river. Prior to that time the riverbed had aggraded predominantly southward of the bridge. Subsequently the trend of aggradation shifted and become more pronounced upstream from the bridge. The belt of sediment deposited from 1936 to 1946 was surprisingly uniform from four miles below to ten miles above the bridge.
41. From its peak elevation of 4469.8 feet in March 1949 the riverbed at the bridge commenced to degrade until, by the time of the last recorded data in late 1954, it stood at elevation 4467.2, which was 2.6 feet higher than what its theoretical elevation would have been if the riverbed had aggraded at a constant normal rate of .27 foot per year since 1895. By 1954 a general degradation had occurred in the area from four miles below to 22 miles below the bridge. This phase of degradation eliminated broad areas of the 1935-1947 sediment strata as well as portions of the 1925-1935 sediment strata in that distance, eroding an estimated 7 percent of the total sediment deposited since 1916 and cutting deep trenches in the stream bed. The evidence does not established when or where this phase of degradation commenced.
42. a. The reservoir may have played some role in the ag-gradation pattern affecting plaintiff’s structures in the valley of the Rio Grande in the San Marcial area by interfering with the normal ability of the river to dispose of its sediment load derived from upstream sources and aggravated by the artificial and natural conditions in the valley *377heretofore discussed. Such an influence of the reservoir would be transitory but repetitive. Deltas created at its head, and upstream sediment wedges caused by the deltas, are not permanent but are subject to erosion whenever the reservoir retreats to lower levels and esposes the deltas for a suflicient time interval to degradation influences. However, whenever floods enter the reservoir standing at high levels for protracted periods4 aggradation will occur commencing at the reservoir head and working back upstream to and beyond Bridge 1006-A so as to affect conditions there. If conditions are severe and aggradation sufficiently pronounced, Bridge 1006-A may at some time in the future he adversely affected and require raising in order to maintain its flood capacity, even though the condition may be transitory in the sense of enduring only through several years.
b. As noted in finding 14, supra, at the dedication ceremonies of the dam in 1916 the project’s Chief Engineer predicted that by 1966 37 percent of the reservoir’s original capacity would be lost through siltation. Actually, by 1947 about 17 percent of the reservoir’s original capacity had been lost through siltation, representing a somewhat lower rate than predicted. The record does not explore the future effect of increasing sediment saturation of the reservoir on the ability of the river between Bridge 1006-A and the reservoir head to maintain its equilibrium. It may happen that at some point the reservoir will become so congested with sediment that the slope of the riverbed between the bridge and the reservoir head will differ radically and permanently from the original bed of the stream and the river will no longer be able to maintain its equilibrium. If this happens, the reservoir will become an increasing threat to the plaintiff’s facilities in the San Marcial area, for the never-ending silt coming down the river, having less room to lodge in the reservoir, must go somewhere, and the river will progressively lose its ability to carry it well past Bridge 1006-A. Thus, while phases of excessive aggradation at Bridge 1006-A are transitory now and may be so for some years *378Fence, the time may come when aggradation at and near the bridge will assume more of a permanent nature. The record leaves this problem in a condition of hypothesis and conjecture.
43. The plaintiff contends that the reservoir influence was responsible for the embankment collapse in 1949 described in findings 26 and 27, supra, and also necessitated the construction of Bridge 1002-A as described in finding 28. San Marcial Lake contributed to the saturation of the embankment. The lake was the product of the 1929 flood. The plaintiff does not contend that the influence of the Elephant Butte Reservoir evidenced itself prior to 1942. Hence, to the extent the plaintiff’s theory is correct, the reservoir could not have precipitated the 1929 flood which created San Mar-cial Lake. The lake contributed to the saturation of plaintiff’s embankment which collapsed in 1949. The river flood waters on the east of the embankment completed the saturation. If the reservoir had not been in place the 1949 flood might have been of lesser extent because at least part of the siltation east of the embankment was attributable to the reservoir influence, but the flood would still have been sufficiently high to have weakened the embankment. Thus it is not established that the influence of the reservoir was responsible for the 1949 damage to the embankment. The circumstances of the building of Bridge 1002-A are closely allied. The bridge was built in order to silt up San Mar-cial Lake and even up the valley levels on either side of the embankment. Prior to 1950 the embankment was solely responsible for excluding sediment from depositing in the area west of the embankment, and concentrating what sediment there was in the valley east of the embankment. Even though the reservoir contributed to some of the aggradation east of the embankment, normal aggradation without the contribution of the reservoir influence created by 1950 an imbalance in valley elevations which the building of Bridge 1002-A in 1950 was designed and needed to correct, and did. Thus the cost to plaintiff of building Bridge 1002-A was merely part of the price it had to pay to maintain facilities in an unstable valley, irrespective of the reservoir. The costs of acquiring flowage rights into San Marcial Lake and *379of running a siltation survey in 1950 cannot be attributed to the reservoir for the same reasons.
44. The plaintiff contends that its cause of action arose in 1949 when the trend of reservoir-induced aggradation in the stream bed at Bridge 1006-A first became stabilized, marking the culmination of the aggradation influence of the reservoir which started during the 1941-1942 floods by building a delta at the reservoir headwaters 5.2 miles below the bridge, which delta progressed upstream and, by 1949, permanently encroached upon plaintiff’s railroad properties in the San Marcial area. Defendant’s Exhibit 32 is a line graph portraying the fluctuating elevations of the riverbed at the location of Bridge 1006-A from 1895 through most of 1954. By superimposing a straight line on this graph to portray what the elevation of the stream bed would have been through those years had aggradation proceeded at a constant rate of .27 foot per year, which has been found to be the normal rate of aggradation, the deviations from theoretically normal aggradation become apparent. Thus, the stream bed elevations at the bridge location were consistently below normal until late 1922.5 From then to the middle of 1932 the actual and normal elevations approximately coincided, with minor or temporary variations above and below. From the middle of 1932 to early 1937 the actual elevations aggraded in the latter year to a peak of 4.7 feet above the normal. From mid-1942 to early 1949 the actual elevations again aggraded to a peak of 7 feet above the normal. Both the 1932-1937 and the 1942-1949 episodes of aggradation proceeded at approximately the. same rate in excess of the normal, but the earlier phase endured for a shorter period than the later and so produced a gross excess over normal of somewhat less than the 1949 peak. Insofar as the causes of aggradation are concerned, both of these episodes of rapid aggradation had approximately the same physical circumstances, with the exceptions that the later episode was preceded first, by the Val Verde avulsion of 1937-1941, and second, by the 1941-*3801942 floods which, brought the headwaters of the reservoir closer to the bridge than at any time since 1924. The Val Verde avulsion and its concurrent degradation of the stream bed at the bridge did not contribute to the sharp aggradation episode which followed it, at least not after the effect of the degradation had been nullified. But the plaintiff contends that it could not attribute the earlier aggradation phase to reservoir influence, while it could attribute the later phase to reservoir influence for the sole reason of the proximity in 1941 and 1942 of the reservoir headwaters to the bridge which, augmented by the 1941 and 1942 floods, initiated the upstream formation of deltas. The plaintiff anticipated when Elephant Butte Dam was built that its reservoir when full would cause sedimentation in the San Marcial area. It had seen that the stream bed at the bridge location had aggraded sharply from 1920 to 1924, and knew that in 1920,1921, and again in 1924 the reservoir headwaters had reached to within less than five miles of the bridge location, which may have given it some corroboration of the prevailing theory that high water levels in the reservoir aggravated sedimentation in the San Marcial area. For the 13 years preceding early 1937, when the stream bed at the bridge rose to a record level, the reservoir headwaters had been no closer than 11 miles from the bridge location, and usually 15 to 20 miles downstream from the bridge. Thus, while the plaintiff knew in 1937 that some influences had caused the stream bed elevation at the bridge location to aggrade sharply for the preceding five years to a point where the bridge required a substantial raise to obtain the desired clearance, the plaintiff could not at that time attribute the increased aggradation with reasonable assurance to the reservoir influence. When the stream bed elevation at the bridge again rose to an unprecedented height in 1949, the plaintiff for the first time could logically attribute at least part of the aggradation of the preceding seven years in excess of the normal to the 1942 proximity of the reservoir to the bridge. The plaintiff could not have reasonably attributed the 1949 aggradation peak to the reservoir influence at any interim time between 1942 and 1949 because from past performances it had reason to believe that the trend might be merely temporary and might be reversed *381through natural causes as had happened on previous occasions. In short, the causal relation, assuming that there was a causal relation, was not manifest until 1949. The silt survey made by the Eeclamation Service in 1947, which evidenced the sedimentation trend above and below the bridge, was not available to plaintiff until 1949. For this reason plaintiff was not fully aware until 1949 of the sedimentation trends from 1942 to-1947. A downward trend occurred from the 1949 peak to late 1954, the last data in the record. Whether or when the downward trend will erase the preceding excesses so as to return the stream bed to elevations approximating the normal line arising at a constant rate of .27 foot per year is unpredictable.
CONTROL AND CORRECTION OP AGGRADATION
45. Finding 28, supra, describes the construction by plaintiff in 1950 of Bridge 1002-A, which provides an opening in the railroad embankment at the north end of San Marcial Lake and diverts part of the flow of the Eio Grande through the lake lying west of the embankment. The finding also describes the sedimentation effect of this diversion. It provides a means for the plaintiff to maintain an equality in elevations of the valley on either side of the embankment between approximately ranges F and C, and the safety of plaintiff’s embankment in that area is now greater than it was in 1949. How permanent this effect will be is speculative, since eventually San Marcial Lake will become entirely silted, but it is reasonable to conclude that so long as the valley levels on either side of the embankment remain in equilibrium there will be no limit to the height plaintiff may safely raise its embankment. However, the embankment elevation must remain fairly consistent with the elevation of Bridge 1006-A, or at least to the extent made necessary by the approach grade requirements. To raise Bridge 1006-A another 20 feet would involve building new foundations and would cost about $500,000. It would also require raising the embankment on the approaches to the bridge and also raising Bridge 1002-A, for an estimated total cost of about $5,000,000, including the $500,000 for raising Bridge 1006-A,
*38246. Beginning about 1940 the Bureau of Eeclamation of the Department of the Interior, and the Corps of Engineers of the Department of the Army, began an investigation which resulted in the creation of the Middle Eio Grande Valley flood control and rehabilitation projects. A comprehensive plan was adopted, authorized by Congress, and has been partially put into operation.
47. Within the comprehensive plan two projects have been authorized in New Mexico under the jurisdiction of the Corps of Engineers. A dam on the Eio Jemez, a tributary of the Eio Grande, has been completed and is in operation. The Jemez dam and reservoir will detain approximately 12 percent of the total sediment previously entering the Eio Grande from its tributaries. (Cf. finding 32, supra.) The authorized Abique dam on the Eio Chama, for which appropriations have been made for construction plans and specifications, will detain a smaller but unknown percentage of sediment from entering the Eio Grande, when and if the dam is in operation. The Platoro Dam on the Conejos Eiver, which is now in operation, will probably prevent only a small amount of sediment from reaching San Marcial, since the Conejos tributary of the Eio Grande is north of the Chama Eiver and pronounced sedimentation of the Eio Grande commences at and below the junction of the Chama. No dams are contemplated for the Eio Puerco and Eio Salado, which are the two heaviest silt-contributing tributaries of the Eio Grande. The dams in operation and those in prospect, mentioned above, will partially relieve the Eio Grande of a minor portion of its sediment load, and will partially control the release of waters so as to even out and diminish the flows down the Eio Grande. Since severe aggradation occurs primarily during floods, these improvements will eventually have a salutary effect on aggradation at San Marcial, but the degree of relief is probably minor and speculative.
48. In 1954 the Bureau of Eeclamation completed a flood-way project in the vicinity of San Marcial. The project consisted principally of a 50,000 cfs floodway channel, a 2,000 cfs diversionary conveyance channel, and a dike. The dike ran from the west side of the valley at milepost 999 down to a point east of Bridge 1002-A, where it connected with a *383cross dike leading to the bridge. At the junction of the dike and cross dike, headgates restricted the flow of water from the river through the conveyance channel to 2,000 cfs, representing on the average the volume of one-half of the total annual flows into the Elephant Butte Eeservoir. The conveyance channel leads the water into San Marcial Lake via plaintiff’s Bridge 1002-A. From the lake the water remains on the west side of the valley as it passes Bridge 1006-A, and finally rejoins the main stream or the reservoir about 20 miles below the bridge. In going through San Marcial Lake the diverted water deposits much of its sediment, which has accelerated the sedimentation of the lake. Any excess in the flow of the river over 2,000 cfs flows through the floodway channel, which is confined to the east side of the valley down to a point due east of Bridge 1002-A by the dike mentioned above, and below that down to Bridge 1006-A by plaintiff’s railroad embankment bordering San Marcial Lake. The floodway channel has been cleared of all salt cedar and other vegetation impeding the flow of water. It has a capacity of 50,000 cfs as far as Bridge 1006-A. Below that it has a capacity of 25,000 cfs and is confined to the east side of the valley by another dike. Except at one point the dikes which are part of the floodway project are of unprotected earth without riprapping. In heavy prolonged floods they are vulnerable to being breached.
49. The channelization of the river, which is a part of the comprehensive plan for the improvement of the valley, is primarily designed to control floods and theoretically will tend to expedite the transportation of sediment from upstream down into Elephant Butte Eeservoir. The effects of this program are not immediately apparent but competent engineers advise that beneficial results are inevitable.
50. It is reasonable to conclude that the projects described in findings 45 through 49 will help collectively to curtail some of the aggradation in the bed of the Eio Grande at and near the vicinity of Bridge 1006-A, and that the plaintiff’s embankments and bridges from milepost 999 down to Bridge 1006-A enjoy at present a greater immunity to damage from floods and aggradation than prior to 1950. It cannot be established from the record whether the improvements are *384a temporary palliative or, if they are maintained, will adequately protect the plaintiff’s present right-of-way in the San Marcial area of the valley indefinitely. This will depend in large measure on the size and frequency of future floods and on the security of the flood way dikes.
DAMAGES
51. The raise of grade in 1949 referred to in finding 25 n, supra, the construction and protection in 1949 of Bridge 1002-A referred to in finding 28, supra, and a siltation survey performed in 1950, cost the plaintiff $700,674.31.6 The cost included $25,000 for acquisition by plaintiff of flowage rights in San Marcial Lake.
52. The plaintiff contends that with normal aggradation of the stream (i.e., in the absence of reservoir influence) it could maintain its present facilities in the valley for 125 years with ordinary raises of grade as necessary and contemplated to compensate for normal aggradation, that the reservoir influence has shortened the life of its present facilities by one-half, that as a result it will either have to spend $5,000,000 to raise Bridges 1006-A and 1002-A and the intervening embankment by 20 feet, or to construct a new line out of the valley. This new line, which it was stipulated between the parties would cost $3,005,236.50, would cross the valley at milepost 999, and would run along the plain east of the valley down to Pope, New Mexico, where it would rejoin the present line at milepost 1012. Plaintiff advocates this latter solution as being both cheaper and more immune from the vagaries of the river, since milepost 999 is conceded to be above reservoir influence. A new line would last about 100 years. In addition to the above, the plaintiff contends it has been damaged in the stipulated amount of $700,674.31, described in finding 51, for the 1949 raise in grade, the construction in 1950 of Bridge 1002-A, and the 1950 siltation survey.
53. The defendant contends that the reservoir contributes little if any to the total sedimentation conditions in the San *385Marcial area, that most of the sedimentation is attributable to the other factors detailed in findings 30 through 33, supra, that the influence of the reservoir in contributing to phases of excessive aggradation is merely transitory as illustrated by the fact that in October 1954 the riverbed had degraded to an elevation only 2.6 feet higher than the line of a normal .27 foot per year aggradation and was still going down, and that defendant’s comprehensive project for flood control will substantially relieve sedimentation and flood conditions affecting plaintiff’s facilities in the area.
54. The plaintiff’s contention that, in the absence of the reservoir, by making minor adjustments from time to time, its present facilities would last 125 years is in error. Assuming that the riverbed were to aggrade at a constant normal rate of .27 foot per year from 1943, when Bridge 1006-A was raised 12 feet to obtain a flow capacity of 50,000 cfs, by 1987 the 12-foot raise would be fully nullified even without reservoir influence. If the plaintiff were in 1987 to raise the bridge again in order to restore the 50,000 cfs flow capacity, it would have to install new foundations. In doing so it would reportedly raise the bridge 20 feet instead of 12 feet at a cost of $500,000. But theoretically in the intervening years the normal .27 foot per year aggradation would have occurred simultaneously at the location of Bridge 1002-A and the embankment between the two bridges. For this reason and for the additional reason that a 20-foot raise in Bridge 1006-A would require corresponding raises in the approaches, in 1987 the plaintiff, if it wanted to remain in its present location, would have to spend $5,000,000 (plaintiff’s estimate) to effect a 20-foot raise in Bridges 1006-A and 1002-A and the intervening embankment and approaches. By so doing Bridge 1006-A would theoretically retain a 50,000 cfs flow capacity until about the year 2017, and by 2061 normal aggradation at .27 feet per year would have reduced its flow capacity down to the 1943 capacity prior to the 12-foot raise in that year. It does not seem plausible that the plaintiff would expend $5,000,000 in 1987 for an improvement lasting 74 years or less until 2061, when it could, at a cost of $3,005,236.50, reroute its line out of the valley between mileposts 999 and 1012 and secure a rout© free of *386river and reservoir hazards which would last for a claimed 100 years.
CONCLUSION OE LAW
Upon the foregoing findings of fact, which are made a part of the judgment herein, the court concludes that as a matter of law the plaintiff is not entitled to recover and the petition is therefore dismissed without prejudice.
There is some confusion in elevations in the various exhibits in the case. This arises from, the fact that, prior to 1903, elevations used by both plaintiff and the United States Reclamation Service were based on so-called Engle datum, sometimes referred to as U.S.R.S. datum, which was lower than mean sea level datum. In 1911 the united States Geological Survey changed the datum to mean sea level. By virtue of subsequent adjustments, conversion of Engle datum, to mean sea level datum requires the addition to the former of 43.3 feet. In these findings wherever the Engle datum is referred to it will be followed parenthetically by the converted mean sea level datum. Unless otherwise specified, all elevations given will be mean sea level.
The data contained in the voluminous graphs submitted in evidence by the parties at trial has been reduced by the commissioner to the tables reported in finding 18, a process requiring considerable reconciliation and interpretation. The graphs would be too unwieldy to reproduce in these findings.
But see finding 42 b, infra, which predicts a gradual diminution of the river’s capacity to regain its equilibrium as the reservoir progressively congests with sediment.
Even -with the flood control and rehabilitation project discussed in findings 45-48, infraf, the reservoir may fill to capacity in the future on possibly rare occasions.
Plaintiff's Exhibit 4a, which purports to portray roughly the same type of data as Defendant’s Exhibit 32, but on a somewhat different basis, shows that from 1895 to 1922 the actual elevations of the stream bed varied from above to below the normal on numerous occasions, thus differing sharply from Defendants’ Exhibit 32. But Plaintiff’s Exhibit 4a suffers from certain defects and thus has been disregarded for the purpose of this finding.
This was stipulated by the parties. They also stipulated that the plaintiff spent $897.25 in addition to the above in operating a work train, but plaintiff waived “all right and claim” to that sum.