F 1563 .5 B38 BECKER MECHANICS OF THE PANAMA CANAL SLIDES BANCROFT wm^- University of California Berkeley DEPARTMENT OF THE INTERIOR FRANKLIN K. LANE, Secretary UNITED STATES GEOLOGICAL SURVEY GEORGE OTIS SMITH, Director PROFESSIONAL PAPER 98 N MECHANICS OF THE PANAMA CANAL SLIDES BY GEORGE F. BECKER Published July 25, 1916 Shorter contributions to general geology, 1916 (Pages 253-261) WASHINGTON GOVERNMENT PRINTING OFFICE 1916 DEPARTMENT OF THE INTERIOR . FRANKLIN K. LANE, Secretary UNITED STATES GEOLOGICAL SURVEY GEORGE OTIS SMITH, Director Professional Paper 98 N MECHANICS OP THE PANAMA CANAL SLIDES BY GEORGE F. BECKER Published July 25, 1916 Shorter contributions to general geology, 1916 (Pages 253-261) WASHINGTON GOVKKNMENT PRINTING OFFICE 1916 CONTENTS. Page. Prefatory note, by George Otis Smith 253 Observations on the slides 253 Limiting depth of disturbance 254 Conditions in a wide cut 254 Limiting values of k 256 Examples of slide curves 256 Hydrostatic analogy 257 Formation of ruptures 258 Bulging of canal bottom 258 Effect of the form of the banks 258 Note on finite strains 259 Summary 261 ILLUSTRATIONS. Page. FIGURE 22. Curve of uniform tangential strain 255 23. Elastic curves for a=75, 80, 85, 89 257 24. Diagram illustrating simple shear and shear , each of ratio 5/4 260 MECHANICS OF THE PANAMA CANAL SLIDES. By GEORGE F. BECKER. PREFATORY NOTE. By GEORGE OTIS SMITH. This geophysical study of the Panama Canal slides is now presented for the reasons set forth in the following letter of transmittal: The DIRECTOR. SIR: I have the honor to transmit herewith a paper on the mechanics of the Panama Canal elides. It was prepared as a contribution to the report of the committee of the National Academy of Sciences on the Panama Canal slides, appointed at the instance of the President of the United States. As some delay 1 is anticipated in completing the full re- port, this chapter is now submitted for publication with the sanction of President C. R. Van Hise, chairman of the committee. Very truly, yours, GEORGE F. BECKER. Dr. Becker visited the Canal Zone in 1913 as a geologist of- the United States Geological Survey and since that time has given the prob- lem the benefit of his study. His appointment as a member of the committee of the National Academy of Sciences has made it appropriate for his conclusions, based upon his personal observations and already reported in part to the Canal Commission, to be stated for the benefit of his associates and other American scientists and engineers. OBSERVATIONS ON THE SLIDES. Early in 1913, before water was admitted, I spent some weeks hi examining the geology of the Culebra Cut, now officially known as the Gaillard Cut, with special reference to the origin of the landslides. 1 These appear to me to be of two kinds mere superficial slips on joint planes or other slippery surfaces and deeper- seated "breaks," as they are known by the I 1 had the great advantage of Mr. Donald MacDonald's companion- ship throughout these field studies. 45498 16 engineers. It is only with the latter that this paper is concerned. The breaks in their inception are marked on comparatively level banks by groups of cracks or narrow fissures nearly parallel to the cut, and these almost immediately develop into series of step faults with small throws, many of them only a fraction of an inch hi height, the hade where not vertical being invariably to- ward the canal so far as I could observe. 2 Many of the steps of these faults are only a yard or two in width. There seems little order in the time of formation of the cracks; in some breaks groups of small faults first appear rather close to the cut, those at a greater distance from it developing later. In others the earliest cracks are hundreds of feet from the canal and the intermediate ground splits up afterward. In all the breaks which I could examine the first small movements involved no perceptible gap- ing, or none of the same order of magnitude as the throws of the faults. At or about the same time as the cracks on the bank were formed nearly horizontal cracks also appeared in the cut near the bottom of the bank, but which of these were the earlier it seemed impossible to decide. After a break has made a fair start the cracks more remote from the cut gape and show under- lying curved surfaces which reach the general level of the top of the bank nearly at right angles or crop out almost vertically, and at the outcrop the vertical cross section of such a surface shows a very moderate radius of curva- ture. The surfaces of rupture are fairly smooth, many of them slickenslided a little 2 Mr. MacDonald records that "some of the blocks sank a little in front and tilted up in the rear, so that they were a yard above the front part of the block behind." This behavior was unusual, and I saw no in- stances of it. Local inhomogeneities in the bank might perhaps bring about irregularities in surfaces of rupture which would account for excep- tional throws of 2 or 3 feet. No other suggestion on this subject occurs to me. 253 254 SHORTER CONTRIBUTIONS TO GENERAL GEOLOGY, 1916. below the outcrop, but not smooth enough to make accurate measurements of their radii of curvature practicable. As nearly as I could discover these radii measured between 100 and 200 feet. Where these underlying ^urf aces are exposed to a considerable extent it is apparent that the radii of curvature increase rapidly with increasing depth, and some exposures from which disintegrated material had been removed appeared to prove that as the cut is approached the radius of curvature becomes very large indeed. Movement of the slides perhaps never en- tirely ceases, but it varies greatly in velocity, from a fraction of an inch a day to many yards. After considerable motion has taken place the sheets of rock are broken up and the external surface of the slide becomes as rough as a choppy sea. A certain amount of consolidation and of what might be called secondary cohesion some- times occurs in a slowly moving slide of large dimensions after the material has been reduced to a chaotic condition. In such cases well- developed curved surfaces of rupture and step faults form, indistinguishable in general char- acter from the initial disturbances in the solid bank. This surprising fact indicates that definite mechanical laws of wide applicability underlie the formation of slides. I was witness to these phenomena in the Cucaracha slide, and they have made their appearance in other and more recent breaks. During the progress of a large slide upheaval of the bottom of the canal may take place from time to time, showing that deformation of the rocks extends to a certain depth below the deepest excavation ; but this upheaval does not attend every spasm of activity in the slide, nor does the amount of material thrust up in- dicate that deformation extends more than a few yards beneath the bottom of the canal. A layer of rock say a hundred feet in width, buckled by nearly horizontal pressure, would show, even if it were only a couple of yards in thickness, mounds of rubble as much as 20 or 30 feet in height, or of the order of magnitude of the observed upthrusts. LIMITING DEPTH OF DISTURBANCE. To simplify the mechanical problem as much as possible, suppose the case of a level plain underlain to a great depth by an ideally ho- mogeneous rock. At any depth in this rock the pressure will be hydrostatic and equal to the depth multiplied by the density. Suppose a narrow trench to be sunk vertically in this rock, the width being so small that caving of the sides can be prevented by mine timbering. Then, because of the one-sided relief of pressure there will be at the bottom of the cut a hori- zontal stress, directed from the wall into the cut, which is equal to the product of the depth and the density. This stress will tend to pro- duce a horizontal shear and to drive the bot- tom of the wall into the cut. If the cut is sunk deep enough, so deep that the stress is equal to the resistance of the rock to simple shearing stress at the elastic limit, this defor- mation will occur and the wall will bulge. This seems a rather hasty statement, but in the last section of this paper the strains are considered in detail; it is there shown that the elastic limit for simple shear would be reached long before the limit for mere linear compres- sion, and that of all elementary resistances that resistance which opposes stress such as is ex- erted by a pair of scissors is the weakest. Let the limiting depth at which this one species of flow makes its appearance be denoted by VD so that if > 'is the density the hydro- static pressure is py l} which is also the value of the shearing stress. CONDITIONS IN A WIDE CUT. The hypothesis of a narrow timbered cut was employed in finding the limiting depth, y l} in order to avoid the complication of a caving bank. Let a wide cut be substituted, one a mile wide if the reader chooses, but let the bank be vertical. Then even before the depth y t is attained any real rock wall would break down or cave. But imagine for a moment the rock replaced by a substance so tough that, though it would undergo permanent deformation at the same limit as the rock, it would hang on long enough to be studied. A ductile substance, such as wrought iron, would act in this way. Consider a surface of uniform deformation nearly as deep as y^ and extending into the wall. This surface will surely not be horizon- tal, for such a strain would imply the expendi- ture of an infinite amount of energy. Before caving can take place in a homogene- ous bank the material of the bank must be strained to its elastic limit. The vertical cross MECHANICS OF THE PANAMA CANAL SLIDES. 255 section of the bank must therefore include a line along which the strain is uniform. This line must reach the top of the bank somewhere, and it may be assumed that the line is curved, because that is a far more general hypothesis than that it is straight, besides being in har- mony with observation. In fig. 22 OBC represents the bank and ABOD a part of the cut. The x axis, or OX, is taken at a depth y t from the original surface, and EC is a curved line along which the shearing stress is uniform. The problem is to find its equation. At any point the original hydrostatic pressure was (]/i y)p, but excavation of the cut, having disturbed the original equilibrium and brought about strain reaching the elastic limit, has developed a shearing stress which is equal to FIGURE 22. Curve of uniform tangential strain. (y\~y}p P er un it length and which is of itself inadequate to cause flow. But there is another manifestation of stress to be considered. The shearing stress is equivalent to a tension in the direction of the curve, say T per unit length. Let ^ be the angle which the tangent to the curve makes at xy; let 8$ be an elementary angle and 8s a corresponding arc. Then ele- mentary mechanics shows that the tension, T, acting along the arc 5s is equivalent to a normal pressure 1 T8\f/. It has already been explained that pi/ t per unit length is the shearing stress needful to strain the mass to its elastic limit for simple shear. Hence if stress of this intensity is to be set up along the curve EC the following equa- tion must hold good: or, more briefly, T 8s i See Tait, P. Q., Properties of matter, p. 253, 1894; or Lamb, H., Statics, p. 276, 1912. Here 8s/8\f/ is the radius of curvature, say R, while Tfp is a constant characteristic of the material and essentially positive. It may therefore be replaced by 6 2 , and then which is the most general equation of the elastic curve. Replacing R by its value in terms of dy/dx and d 2 y/dx 2 and integrating once gives = C2l 2 cos (1) where C is a constant of integration. The form of the curve depends on the value of C. For the present problem it is evident that the curve can not cross the x axis and that y can not become negative, so that C must equal or ex- ceed 2& 2 . It is easily proved that if C= 2b 2 the equation represents a curve coinciding with the x axis for an infinite distance. This is not a case to be considered, and therefore C>2b~. The equation then represents the elastic curve of Euler's eighth class, a diagram of which is given in Thomson and Tait's "Natural philos- ophy," 611, figure 7. For some purposes equation (1) is conven- ient enough. Thus if the ordinate of the point at which the tangent of the curve is vertical is called T/V, then O= y v 2 ; while if the ordinate at the point where the tangent is horizontal is y , then i/ 2 = y v 2 2i 2 . But values of the abscissae are not so simple. It is needless to say that the geometry of the elastic curve has been thoroughly known for a century and that this is no place to expound the subject. A few results, however, must be set down. By substituting where