• D.96/czs Aft kiik, RNATI Jblic geautng file Energy 149 S. Devirtnient of Office Deformation Accompanying the echo Operations Hebgen Lake Earthquake of August 17, 1959

By W. BRADLEY Iv1YERS and WARREN HAMILTON

THE HEBGEN LAKE, MONTANA, EARTHQUAKE OF AUGUST 17, 1959

GEOLOGICAL SURVEY PROFESSIONAL PAPER 435—I

UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1964 •

CONTENTS

Pun Abstract 55 Introduction 55 Acknowledgments 56 Deformation in West Yellowstone baaia 56 1939 structures 56 Madison Arm fault and tsonocline 56 Faults at Home Butte so Graben involving faulted bedrock 59 Warping of the ground at Madieon Fork Ranch_ 56 En echelon fisaures and 'runlets 60 Preexisting structures 62 Mechanics of structures 88 Slightly reactivated structures 68 Old conical depressions 68 Ring-fault depressions 69 Effects of earthquake on Ilehgen Lake 70 Surges 70 Overtopping of the dam 70 Surges along the lakeshore 70 Period of the surges 71 Profile of the highest surge 72 Bank slump 72 Absolute subsidence 74 Measurement of subsidence 74 Releveling 74 Highway profile determination 75 Shoreline measurement 75 Scarp heights 76 Fathometet survey 76 Character of subsidence 76 Deformation west of Madison Range Madison Valley 77 Cliff Lake and Wade Lake 78 Madison Range fault 78 Deformation in Madison Range 78 Deformation east of Madison Range so Geometry of deformation 80 Subsidence pattern 80 Relation between faulting and wszping 81 Relation to older structures 85 Structures northeast of Hebeen Lake 85 Relation of modern faulting to Cenozoic faulting 85 Control of modern fainting by learamida structure 88 Structures south of Hebgen Lake 67 Earlier warping of West Yellowstone basin 88 Surficial deposits 88 Changes in drainage 88 Relation to regional structure 89 Madison Valley 89 Henrys Lake basin 91 Centennial Mountains and Centennial Valley 91 Superposition of structures in the Madison-Centennial region 95 Seismic evidence of movement pattern 95 Geodetic evidence of active zones 96 Conekuaioni 93 Reform= cited 97 to THE HEBGEH LAKE, MONTANA, EARTHQUAKE OF AUGUST 17, 1959

DEFORMATION ACCOMPANYING THE HEDGES LAKE EARTHQUAKE OF AUGUST 17, 1959'

By W. Bs ABUT MrEns and Winntx HsaarriroN

ABSTRACT easy faulting. The bedding that controls these two faults at the surface can hardly extend to a depth of more than a mile The gently sloping obsidian sand plain that forms the sur- or two, so these scarps cannot directly portray the pattern at face of most of the West Yellowstone baain south of Hebgen deformation in deep basement racks. Lake is the won flank of a broad basin of new subsidence Releveling of bench marks and comparison of the postquake produced during the 1959 earthquake. An area 43 miles long road profile with the prequake profile show that the lower and 14 miles wide subsided measurably; there was a mean- part of the Madison River canyon sod the adjacent part of tared maximum elevation above previous levels of 1.7 feet. Madison Valley subsided evenly about 7 to 8 feet and that The maximum subsidence, 22 feet. was In part of Hebgen there teas no new warping at the Madison Range front, East Lake basin where a tract of about CO square miles dropped of Madison Slide and Earthquake Lake, which buried prequake more than 10 feet Subsidence was determined by relereling reference points, bench mark, and road profiles prove that of bench marks and road profiles and by measurements of the upper part of the Madison Rorer canyon subsided 6 to lakeshore changes and of heights of new fault scarps. Con- 34 feet The subsidence of Madison Valley decreased to a tours of subsidence can be drawn with high reliability and small fraction of a foot in a few miles northwest of the mouth precision over a considerable part of the affected area_ Ex. of the Madison River canyon. South of the canyon a segment tensions of the contours across areas of few or less reliable of the Madison Range front fault was reactivated, bot the data seem to indicate a broad basin that plunges gently amount of absolute subsidence is not known. Observed shore- eastward. directly across Madison Valley and the Madison line changes at Cliff Lake, west of the upper Madison valley. Rance to the West Yellowstone basin and Heinen Lake. I Indicate that the lake was tilted. A belt 10 miles wide north Lessened subsidence continued eastward to the Gibbon River. of the subsided tract was elevated by from 0.1 to at least L7 The obsidian sand plain at the West Yellowstone basin ia feet. broken by scarps 0.1 foot to 3 feet high along faults that were NortAtrending Madison Valley wee much deformed during reactivated in the 1959 earthquake. Older displacements along late Quaternary time by eandreading faults and warps that three same structures. produced by normal faulting In the extend from Centennial Valley and the Centennial Mountains; basement rocks, had formed monocilnes and faults io upper other east-trending structures have been superimposed across Pleistocene surficial deposits, with a maximum structural the older structures of the Madison Range, West Yeliowatons relief of 50 feet. Some new surficial structures that appear basin. and the northwestern part of the Yellowstone Plateau. on casual inspection to indicate compressional deformation The 1959 earthquake seems to be part of a relatively new actually are due to refraction in the unconsolidated fill of structural system in which older north-trending blocks are reactivated normal faults in the basement rocks, being complexly aegignented and deformed by the eastward The begin of new subsidence ends obliquely and abruptly extension of structures from the Centennial region. no the northeast against reactivated faults with scarps 5 to southeastward near the northeast aide 20 feet high that trend INTRODUCTION of Hebgen Lake. Subsidence is nearly constant for many miles along the zone of major scarps. Two of the three molar scarps The violent earthquake of August 17, 1959, which occur in areas where bedrock Is exposed and are limited to centered. near Hebgen Lake in southwestern Montana, Paleozoic rocks dips steeply to- places where bedding in the formation of several large ward the basin and Is thus oriented favorably to permit slid- was accompanied by the ing down the dip; where the bedding attitude flattens, the fault scarps and many smaller ones, and by the abrupt scarps dwindle and die out. The subsidence is accounted for subsidence and warping of a very large area. The almost entirely by new scarps where the bedded rocks hare high fault scarps. many miles long. that formed north- optimum attitudes. It was accomplished by o combination of east of Hebgen Lake are described in detail by Wit- warping and faulting where attitudes are less favorable, and (chapter G). Other aspects of the deformation, by warping alone to anfaulted areas where bedded rocks would kind otherwise hare been broken at high apples to directions of as revealed by detailed studies of the area south of the lake and of the lake basin are described in this paper. Also considered is the whole deformation pat- ..atriiat IT by local date: Atigniit IS by Greenwich date. 55 DEFORMATION ACCONITANYLNG THE E:EBGEN LAKE EARTHQUAKE 56 THE BERGEN LAKE, MONTANA, EARTHQUAKE OF AUTCST 1T, 1955 57 tern in relation to older structures and to the com- Major streams heading in the mountains cross the plex structural evolution of the region. The geology plain; but due to its high permeability, no streams of part of the area discussed is shown on plate 5 rise on its surface. Near the widely spaced streams and described in chapters R. S, and T. Scarps and the plain has been cut into steplike erosional terraces, other features formed during the earthquake and con- but over most of its extent it is a monotonous con- tours illustrating the amount of accompanying subsi- structional surface, essentially unmodified since iLs dence are shown on plate 5. A brief note that sum- deposition after the Bull Lake glaciation. The sedi- marizes some of the deformation data and our inter- ments and the dating of the sand plain are described pretation of them has been published (Myers std by G. M. Richmond (chapter T). Hamilton, 1961). The sand plain had been deformed before 1959 by a number of small faults and monoclines with ACKNOWLEDGMENTS struc- tural relief of as much as 20 feet each. Monoclines Incorporated in this report are the results of studies predominate at the surface, but they are probably by a number of men both within and without the related to normal faults at depth. Almost all these Geological Survey. Releveling of bench marks north- structures were reactivated, with minor surficial fault- westward from West Yellowstone, which provided ing and tilting, during the 1959 earthquake, and the the critical information regarding absolute changes of entire plain was tilted gently northward. The new elevation, was made possible by the cooperation of tilt steepens gradually northward from West Yellow- Capt. J. H. Brittain, Chief of the Geodesy Division, stone. Most of the reactivated structures are in the U.S. Coast and Geodetic Survey, and was accom- least tilted part of the flank of the major structure. plished by a field party under the supervision of W. N. Grabler. W. K. Cloud, of the Seismological 1959 STRUCTURES Field Surrey of the Coast and Geodetic Survey, per- MADISON ARM FAULT .11.2:13 MONOCL/N15 formed many services of liaison between that organi- Tile zation and the Geological Surrey. structure with the greatest local relief formed south of Hebgen Lake Personnel of the Montana Power Co. provided much during the 1959 earthquake lies midway along the information regarding Hebgen Luke and Hebgen Dam south shore of Madison Arm (pl. 21. before and after the earthquake. Chief Engineer TIE.F Madison Arm structure—partly a mono- dine. partly a Ray M. Ball and Engineers James Kreitzberg and fault, and partly a combination of hot three-quarters Glenn Jones are among those to whom particular of a mile long and trends east- norfeastward. Tilts and thanks are due. displacements during the aliwavida earthquake were greater here The U.S. Forest Service took low-altitude vertical than elsewhere on the sand plain, and much Flamm 3a.New scarp la the nod Wale. Dip at fault ft eorfnee I. 85' bats Hader scarp: apposite Melts are right together ea that dere- aerial photographs along the new fault scarps and of the interpretation of the structures of the sand plain at depth is dropped blank la stl.ghtly a...Ohne by the relatively oplitrawa blade- This la a refracted normal fault aol a revere! lama. 1St. Si. supplied prints of these photographs for our use based upon 35.1 South of Itaillsoa Arm, 5.3 tale. ovrthareat at west 1".ellaraaorte. knowledge of this well less than a week after the earthquake. Helicopter -developed structure. The 1959 displacement flights provided by the Forest Service permitted in- was greatest along the west- warning at others, and by a combination of both The base of the monocline is commonly broken by ern half of the structure. spection of many earthquake features from the air. The most obvious new elsewhere. a nearly continuous miniature underthrust that at the features are north-facing scarps Releveling of road profiles in the Madison River as high as 27 inches Along the western half of the structure, where it surface dips gently back under the incline of the that dip very steeply southward canyon by the Bureau of Public Roads during the beneath the relatively affects the main surface of the sand plain, flexure and larger structure (fig. 321. The small structure is year upthrown block (fig. 30). Although there are gen- after the earthquake demonstrated that local sub- I scarps are superimposed on an older structural rise called a mole-track thrust. a name adapted from the erally cracks an inch or two sidence continued there: we are indebted to Lynn D. wide along the faults, that is about 5 feet high and as steep as 8°. The total term "mole track scarp- which T. W. Dibblee (in the Tingey of the Bureau for these data. - downdropped block in places is driven tightly in ground-surface tilt is several degrees steeper than the Oakeshott, 1955, p. 24, fig. 1) applied to a similar- under the relatively upthrown block so that the latter new tilting shown by the trees. and the difference appearing surface structure. These features do indeed DEFORMATION er IN WEST YELLOWSTONE BASIN slightly overhangs; this is clear evidence that this represents tilting before the trees grew. resemble the little ridges made by moles (figs. 31. 32). near The surface of most of the West Yellowstone basin, -shore feature is not a lakeward slump. The eastern part of the Madison Arm structure The feature at Madison Arm is a small welt several including the southern pert of the Hebgen Lake basin Monoclinal warping was also produced along the crosses river terraces cut to several levels into the inches high and marked by an abrupt upward roll of as well as the area to the south. is a broad expanse of structure during the 1959 earthquake. This flexure sand plain. Tire lowest of these was under shallow the ground surface that curves irregularly along the obsidian sand and fine grave? that slopes gently north- is made obvious. where its dips are more than several water in Hebgen Lake before the emergence of Madi- base of the monocline in the upper plate of a surficial westward (pl. 5). These surficial deposits are 200 decrees, by a zone of correspondingly tilted lodge- son Arm that resulted from warping of the lake bed thrust fault, Thrust displacement is typically 4 to feet thick where penetrated by an oil-test hole 2 miles pole pines (fig. 311, and the monocline can in places during the earthquake (see p. 76). Only the earth- 6 inches. The mole-track thrust was dug out in a northwest of West Yellowstone and more than 240 be traced most easily by following the tilted trees. quake structures affect these terrace surfaces; any number of places: the gentle surface dip continues for feet thick on the south abruptly shore of the Madison Arm The new structural relief is generally 2 to 21,..; feet, pre-1959 warping and faulting in this eastern segment a foot or a foot and a half and then steepens where a water well of this depth occurred before the cutting of the the chief com- bottomed in sand. represented chiefly by faulting at some places, by terraces. to 45° or more. Pine-needle duff is as THE. HEBGEN LAKE. MONTANA, EARTHQUAKE OF AUGUST t7, 155Y DEFORMATION Acco..%EpANT -Ec.G THE FIEBGEI: LAKE EARTHQUAKE 59

tr.: als1:4- ,11 f2. Aki . -- L •

V"4,,, • • . , --- ..641.1 . ,1 1!Clir:74...

"'•" .•-- • ,„ - • • - --40' - 173113= "--1".3t

t•••••, -•••• •er • ,•• ** 1 •W;" V>2.- 11•4X,4"

• , "1-kr- •

ROM! ES—Mole-crack thrust at base of menoeithe formed daring 1559 earthquake. White tape, indicate horizontal Sit feet! and rercleal Flouts 51.—yew tears 5.3 miler northwest Of West Yellowstone. Photograph by lad Mono-cline lo the sand Wain. Beall Heft! with figbely slued tasure ted way to an atypically awn end 5 feet, componeula of nen [Donnell./ slope. South of Mara. Arm. abrupt monnelint reenter; Which Firth way In turn to n geurie monoclint. A mole.track throat mom along the base Of the mono/Una. loan Stacy. pass the head of the tawnier. South of Matson Arm, 5.3 all.. northwest of Weal. Yellowstone Fbothyroplz by John fit Stacy. v y FAULTS AT HORSE BUTTE 1950) found this spring to have a much higher radon ponent of the material offset at the surface. The cuts across a bar of Hebgen Lake that was emergent Numerous new scarplets along the south and south- content than did the others of the group, which is similar mole-track thrusts, described by Dibblee from after the earthquake. and offset of the prequake shore- west sides of Horse Butte formed during the earth- possibly further evidence of a deeper source. the area deformed during the Arvin-Tehachapi earth- line demonstrates that large blocks moved along this quake. The scarps form two sets; most strike west- The shape of Horse Butte is suggestive of a fault- quake in California, have been generally accepted as small structure during the earthquake. Levels were northwest. and fewer strike east-northeast. The west- block origin. indicators of compression. Although this may be ran with a telescopic alidade along the prequake high- northwest scarps lie in a zone a mile wide that es- GRABEN INVOL,ING FAULTED BEDROCK a valid interpretation in California. at Hebgen Lake water line for tends from Horse Butte to Edwards Island. Along they are products of the length of the bay in both directions At only one place south of Hebgen Lake was an normal tensional faulting, as is nearly these faults, the south or southwest side is from the new scarp. a total length of about 300 yards all area of exposed bedrock seen to be displaced by 1959 brought out later. The original analogy with the relatively dowodropped. track of a mole was obliquely across the strike of the fault. The prequake scarplets. No actual outcrop was cut by scarplets, made by Hotel (1893, p. :32S-333) The zone of northwest-trending scarplets is the site for a similar but shoreline of the southern- block was uniformly 13.0 ± but bedrock must be involved at a depth of a few much larger rounded ridge of soil of more titan a dozen springs formed at the rime of that 0.2 feet above the level of Hebe•en Lake of the moment, feet.. This area is in the western part of a small developed along a major strike.siip fault and the 1950 earthquake. Most of these springs are small and the Basin Ranger Station (west of the "* •• resembles very much the pathway of a gigantic shoreline of the northern (downdropped) and cold. and their temperature (about 4 °Fl indi- graben near the block south of mole." was 11.5 0.3 feet above the lake. Most of cates a shallow ground-water source. One spring has South Fork of the 3ladison River. and The eastern half of the the variations in the readings on the Madison Arm, along the south edge of sec. 9, T. 13 S., Madison Arm structure. de- north block a larger flow nod a warmer temperature .55°F) than veloped on stream-cut terraces. is a nearly continuous probably represent the broad sag typically developed the others. and this water must have come from a R. 4 E.). new fault scarp several inches to 1,1 feet high and in front of each fault and monocline of the sand deeper source: presumably it has migrated upward A. parallel-sided block of rhyolite. 300 to 401 feet without apparent associated warping. This scarp plain (see page 62 I. along a fault zone. A. Tanner (oral communication. wide and 1.500 feet long. extends eastward from the TEE B1BOEN LAKE, MONTANA, FARTERYCARE OF AUGUST 17, 1959 IJOrr EARTEIQIIAIS-E 60 DEFORMATION ACCOMPANTDIC THE RUMEN 61 bedrock hillside and projects through the Bull Lake three tiers of logs end-to-end across the projection of moraine. It is apparently a horst uplifted before the fault. The north and south tiers are still nearly deposition of the moraine. Along the south boundary level, but the center tier was skewed down to the of this horst, a scarplet with 4 to 6 inches of dis- north so that the north tier is now lower than the placement farmed during the earthquake. West of the south. This effect had partly accumulated over many .... ---- horst. an area underlain by rhyolite was cut by sev- years before the earthquake—showing a surprising ee ----- eral shorter new scarplets that LIE on the same trend speed of deformation unaccompanied by earthquakes— ... but en echelon, stepping to the left; the vertical die- and was partly an effect of the ground shift accompany. placement on these was also about 4 inches. These ing the quake. The inside of the lodge was much -"--- occurred along earlier small topographic escarpments damaged. that are visible on the ground and on prequake aerial East of the lodge, and still on strike from the fault. photographs. East of the horst, and samin stepped to the bed of the Sonth Fork of the Madison River was the left, another new scarplet as much as 10 inches steepened fora distance of about 100 feet by a gentle high formed along an obvious old scarp where the monocline. and the water now runs more rapidly than moraine had previously been displaced 5 feet. it did before the earthquake. The water now piles up About 500 feet south of the horst, several other against the bank at the base of the monocline and new scarplets formed in 19e9. They are downthrown partly spills over, which it did not do before the on their north sides by as much as a foot and thus earthquake. define a graben flanking the horst. The easternmost West of the lodge and just east of the end of the of these new scarplers follows a prominent old scarp northermfault, a ditch previously carried water north- on the surface of the moraine that is as much as 15 ward through a horse corral 100 yards beyond the feet high. (This scarplet, like another longer one 4 faults. After the earthquake, the water, instead of miles to the east, was discovered by field investiga- flowing through the corral, reached only to the south tion of the site of an old scarp, that was recognized end of it. and the ditch spilled out into a newly on prequake aerial photographs.) The western scarps created broad depression in the sagebrush fiat midway follow obscure topographic alinements which may or between fault and corral. Another new pond formed may not be of structural origin. by tilting of the ground just above the upper fault. B' Several established springs issue from the moraine Springs in the vicinity of the ranch were drastically A A' B Fissures 1959 scarp at points on projections of the lines of new scarplets changed by the earthquake: springs previously large Mole-track One spring is at the southeast end of the exposed- became small, ones previously small became large, theuSt bedrock part of the rhyolite horst, less than 50 feet and several new ones were formed. beyond the last quake fissure. Another lies at about Because of the presence of buildings, river, and the south margin of the graben, although at a point ditch at the Madison Fork Ranch. warping of the where the graben cannot be precisely defined. ground. with an amplitude of about a foot, was quite between Lod Soars. termed Owing 1550 earthQ¢hkt. slid earlier Flames ad.—Schematic amp ad merlons .hawing typical re aciono •carp• The close RIR(11,14Cd. relationship of earthquake scarplets to obvious. Such relatively slight deformation would • soononloes of rho hand pls.... finch-L. of structures earlier topographic features, whose forms alone are be unrecognizable in most settings, but similar warp- suggestive of fault origin, indicates that the old fea- ing and changes of level presumably accompanied to the monoclinal slope. and as the rises dwindle, so do the scarps. and both lap and are confined tures are fault scarps and that the new most of the new displacements on the faults of the strike west-northwestward scarplets commonly disappear together; but there are conspicu- Almost all the fissures also probably represent movement on bedrock faults. West Yellowstone basin. Relatively rapid warping W.. but a few swing a few degrees ous exceptions. Mole-track thrusts like those along from N. 60° to 80° quadrant. All en echelon fissures without earthquakes has presumably occurred else- the Madison Arm structure formed at a number of into the southwest WARPING OF THE GROUND AT MADISON PORN RANCE in strike in a clockwise where in the basin also. places at the base of monoclines where new scarps were along monoclines diverge controlling structures: the fissures Changes of level ...striated with movement on sev- relatively high. direction from the mx ECHELON FisSONES AND OCARPLETS places where fissures break eral of the small faults in surficial material are well Most of the structures face north. Fissures and step to the left. At two monoclines, however, the demonstrated by effects at the Madison Fork Ranch, During the 1955 earthquake, discontinuous fissures faults dip steeply south at the surface. so that the level ground undeformed by from the controlling which is just west of the South Fork of the Madison and scarplets formed along the upper slopes of pre- downdropped blocks are in many places overhung fissures diverge counterclockwise exceptions is west of the River and south of U.S. Highway 191. 4.3 miles north existing monoclines throughout their length. Vertical slightly by the upraised blocks. Dips tend to be structures. One of these 35), and the other of west from West Yellowstone. Several low scarps offsets on the new scarps are almost torariably down intermediate between vertical and perpendicular to the South Fork of Madison River [fig. bank of Grayling Creek upstream here received new surface-breaking offsets of a few toward the topographic basin. The offsets amount to monoclinal slopes. is along the west Blarneystone Ranch. inches to a 1 foot The newly developed structures range from mark- from the foot, with the north sides downthrown; as much as lt e feet but are commonly less than clear. Their en oblique fissures a few yards long that show no The origin of the fissures is not these faults are indicated on plate e. Both old and and do not exceed half a. foot along considerable edly evidence for displacement, through longer and less oblique echelon pattern seems at first glance to be new scarps are short. About 75 yards east of the end lengths of the monoclines. New displacement is gen- vertical fissures. continuous or minor strike-slip fault displacements. The of the northernmost of the faults mapped, a erally related to the height of the monocline: the scarplets of small displacement, to it large log however, show no preferential strike-slip displacement building, used nearly continuous low scarp parallel to the monocline as a dude-ranch lodge, was built in highest scarps tend to be on the highest monoclines; Most simply opened directly. (fig. 33). The en echelon structures commonly over- of opposite walls. 62 TEE HEBGEN LAKE. MONTANA, EARTHQUAKE OF ACCOST 17, 1957 DEFORMATION ACCOMPANTDM TEM REBGEN LAME EARTHQUAKE 63

Where one fissure had an inch or two of strike-slip gentler slopes, from the irregularly curving fronts of N Porte-rataIv developer! Md. S component (as shown, for example. by the offset of stream terraces. Broad. very shallow sags 100 feet termed kern gravel near Bedded rhoselegre or base and Weis Iowan:1MP hin soil in sucks embedded in the duff), the next might have the more wide are developed in front of some of the mono- Tander aces same amount of opposite displacement. More signifi- clinal rises. As the margins of the sags are poorly Sol thickened by slumping, cantly, even the highest of the new scarps showed no defined, their depths cannot be measured precisely, - k more than this same maximum of an inch or two of but they are commonly about a foot. Most of the rises Crewed k strike-slip displacement, and that in inconsistent direc- lie in a west-trending arcuate zone, which is gently GreemsTinned crtrasbettellull Gray etti.ddlil WOG at the tions. Were there a strike-slip component in the convex in plan to Laminated silt el Structureless gravel. the north and 6 miles long, that obsidian sand of the sand plain. not stained no sane plain Slerrong basement faulting, the highest surface scarps should extends from the Madison River to the west edge of lOrr"dd be sand plain show the most lateral daring POST tersories displacement. the sand plain. The individual structures are sep- DI faulting A. more likely explanation is that the en echelon arable into two groups that trend about N. 70° W. fissures were produced by the passage of longitudinal- and N. 75° E., respectively. S ID ai D b seST compressional earthquake waves through the uncon- A section of one of these old structures in the sand solidated sand. All fissures are subparallel to the long plain is exposed in the east bank of the South Fork The scan non. 84 —Fault In the nand plain. Tile Morelachst obsidian mad one faulted without washing; affect was mare than 2 feet. axis of the basin of new subsidence, De...loped and therefore to of the Madison River, 2.5 miles south of Madison wY beetled by a styes-eta serene, and a [efface pcseol deposited: this woe in own mantled by loess. after .5115 1 soil we. structural relief of an the presumed controlling regional structures at Arm The fault was then reactivated with souther foot of offset. accompanied by the' ,espies of a mocoollac with a depth, of Hebgen Lake (fig. 34). The ground profile esposure, during Me 14,60 sairsbdualer. edditional 6 Sect. This Sara mu orals reactivated slang meat of Its tenets, though not at this and to the earthquake-generating fault phatoneants. as determined is that of a stream terrace deformed by a combined Elroiure in bank of South Pork of Madison Saver. SEy ses 14, T. to S.. IL 4 E.. drawn from a seines of by Ryan's (1962) first-motion study. Over most of fault and monocline and slightly modified by erosion. the basin the passage of waves of alternate compres- The terrace surface is underlain by about 5 feet of Two different structural effects are displayed in the been composites of multiple separate offsets. Further sion and rarefaction (perhaps the long-period Ray- sandy pebble gravel composed of rhvolite reworked river cut. The deformation of the terrace surface offset occurred during the 1959 earthquake. as the leigh waves) left no surface trace: but where the ob- from the upper part of the obsidian sand plain, whose produced a monocline, and this was slightly aug- structure is followed throughout its length by discon- sidian sand was actually being folded during the earth- original surface lay about 30 feet above the present mented by faulting. The older preterrace deforma- tinuous new scarplets, small fissures, and mole-track quake, the topmost layers were in tension, and the terrace. Loess covers the terrace, and a weak soil is tion of the sand plain sediments. which occurred when thrusts. rarefaction impulses of the passing waves may have developed on the loess and the upper part of the they were buried beneath 30 feet of similar material, Another old structure of the sand plain deforms an opened the surface fissures. A retracting effect of gravel. The terrace surface, its soil zones, and the produced a steep normal fault without monoclinal abandoned channel of Cougar Creek 6 miles north of divergent monoclinal trends could have produced the loess and gravel veneers have been warped and warping. This difference in behavior between sur- West Yellowstone_ The channel is shown on the exceptional variations of individual fissures from the faulted into a rise having 7 feet of structural relief. Of face and buried materials is a critical factor in the topographic base map of plates 2 and 5 as the unin- general strike. tlus amount, about 15 inches represents offset along interpretation of the mechanics of the structures of terrupted course of an intermittent stream, and its a norciel fault that dips 75° north, and the re- surface monocline pre- essential continuity is evident on the aerial photo- PREEXISTLNG STRUCTSTP.ES the sand plain (p. 68). The mainder is due to monoclinal flexing that produced sumably formed above a fault that did not break graphs from which the topographic map was made.. Structures south of Hebgen Lake that were acti- structural dips in the surficial material as steep as 9° through to the surface. At the park boundary, however. a structural rise with vated during the earthquake all follow preexisting near the fault. No structure. old or new, is visible in the modern a slope of 15° to 20°, apparently a monocline, devel- structures, along at least some of which there has Earlier movement on the same fault had offset the swampy flood plain of the river aloug, strike west of oped obliquely across this channel and formed a been- repeated deformation during late Quaternary underlying beds of the sand plain. The scarp then the riverbank exposures. Farther west, on the flood- drainage divide with a relief, both topographic and time. The nearly universal coincidence of new and formed was planed off by river erosion us the ter- plain deposits that are higher and hence drier, a zone structural, of 10 to 15 feet. A meander loop of the old scarps. their occurrence in linear zones consistent race was cut, and the 5-foot thickness of terrace gravel of fissures formed during the earthquake. This zone dry channel, cut off on the north side of the rise. is with known bedrock faulting peripheral to the basin. was deposited unconformably above the sand plain continues westward and emerges from the alluvium now a flat-floored dosed depression 20.0 feet wide and numerous other features described in both the sediments. Preterrece movement on the fault was to nearly coincide with a long straight fault scarp in which occasionally contains a shallow pond. South- preceding and the following sections demonstrate this. more than 2 feet: the foot of silt exposed between Bull Lake till, overlapped from the north by alluvium east of this. the structure again cuts and darns the the river The structures cannot be explained in terms of basin- and the terrace gravel on the downtiropped fig. 351. The exposed height of the fault scarp is channel and has produced another small undrained side of the fault ward slumping, for any slumping on the very gentle is obviously from a different bed 18 feet where moraine appears on both sides of it, depression. Farther southeast the topographic rise than is any of the slope of the sand plain would necessarily have pro- sand exposed beneath the terrace and this indicates the amount of posttill offset. Above is as steep as 30° and probably is a slumped fault gravel on the opposite side of the structure was duced nearly horizontal separations. Some local fault. About 150 the steep scarp the surface of the till rises more gently scarp rather than a monocline. The yards eastward along the monoclinal rise is the of minor structures slumping that did occur on low-angle surfaces along edge for another 30 feet. and this may indicate still older reactivated with the usual suite of the river terrace. and the rise continues up the during the 1959 earthquake. the lakeshore is described on later pages. uplift: if so. total structural relief is nearly 50 feet. river-cut terrace scarp, 20 feet high. onto Cougar The prequake faults the main This structure of the sand plain hos thus undergone The sides of the abandoned channel of and monoclines appear as surface of the sand gentle plain. On the main surface the at least four episodes of displacement during late Creek are steeper than the slope of the younger struc- north-facing rises 1 foot to 15 feet high which rise is 10 feet high, and this represents the total dis- Quaternary time. The moraine was offset at least tural rise which deforms the channel. The original interrupt the smooth, gentle northward slope of the placement of the sand plain, or the sum of that ac- once, and probably twice. before the deposition of the slope of the rise therefore must have been less steep sand plain. Frontal slopes of the rises commonly complished before the formation of the terrace and sand plain. The uext episode occurred after the sand- than that of the cutbank: so the rise could not be a range between 5° and 10° and uncommonly reach that after. The relations at the river cat suggest plain sediments were deposited. and the next after slumped fault scarp. but must be wholly or chiefly 15°, rarely 20°. The rises are readily distinguished that more than half this amount is postterrace defor- the sand was cut down by the South Fork to the a monocline. The height of the rise and of the chan- by their relative straightness, and generally by their mation. terrace level. Any of the above episodes might have nel sides is about the same, so that the shorter life- DEFORMATION A CCOMPANYLNG THE HE BUN LAKE E ARTHQUARE 65 54 SHE ERR GEN LAKE, MONTANA, EARTHQUAKE OF AUGUST 17, 1959

Den mall base 0 eigeodor m ad.. el se.. 0400

Illiuston at waif seta till "/14-1959 sesalieti. .111ril h'10 / 1,411 PieleUt halvah to 0-3 masa the ngal toil., want dm! DO a...fel/m*111 11000.10. *11° souks it 1151 basket and `.fist bent of Smmal 10.1 in. alga/. former mal faltios frisian3 Pe, 0511110 altinp,p17:1r , \

treseri mars .soused !me et di lies mars 0 nil Sthm met. 111/.4' Pea aorta ha peas feodref 1110,1011

111 0 earl 04SGOIlsol einsise

\ Oambed lused Pam 1/1,111t1 WW1. N 1431 11111 oildeli of \ }.anal 110011 t APPROSALSTE SCALE thmnel f )% 1000 2900 3090 4030 FEET

,1 1601t 35.—Sketch MAD of muthweet corner of West Yellow.toue Sara gartbqnale assurea follow earlier fault &DO motioelthe 1.a imam- eolidated Quaternary deppelte. Scruernral relief of old fault—dloplacemeat of Hull Lake moraine--is at leant 18 tees between Denny and Buttermilk Creeks MRZ1111111,11 relief of ammonite In mead plain east of South Folk In 10 feet: atructure diem tilt east of road. No new Mathes formed et eurfeee In eaterated modern food Fiala. Figure 34 show. motthelthe and fault emoted La Deerbauk. Efaerward nagrallon Of river fn doe to tilt.. of surto:el deposits Yap compiled from aerial 500101r/0m photogrammetrie eartertjapa

span of a lower scarp is not a factor. The steepest they represent repeated deformation. The occurrence part of the structural rise—southeast of the channel— of ma4or earthquakes within very late Quaternary and is as steep as the cutbank, however, and may be a Recent time is shown by the Recent but prehistoric slumped fault warp. 20- to 40-foot scarp, which may have formed in a The sand plain is crossed by a number of other single earthquake, along the front of the :Madison rises similar to those described above, although most Range and the important old fault scarps that cut slope more gently than does the one at Cougar Creek. the moraines about the West Yellowstone basin. Prob• 1 All the rises are followed by the same suites of struc- ably the sand-plain structures received offsets during tures formed during the earthquake. River-cut ter- some of these earthquakes: and probably, like the race banks show none of the new structures and are structure at Madison Fork Ranch, they also grew by quite different. The rises are commonly distinctive slow deformation without earthquakes. on aerial photographs. The rise now marked by the The new south-facing fault scarp in Hat-lying un- longest new scarp was first recognized on prequake consolidated material east of Grayling Creek near the aerial photographs—the new structure was found by Blarneystone Ranch is on a fault that has had two photographic recognition of the older feature. All previous episodes of late Quaternary displacement. It the rises were walked out for their entire length, and precisely follows an older warp of about the same their depiction on plates 2 and 5 is based on con- height-10 feet (figs. 36, 37). An older and much tinuous field observation. By analogy with the South higher-60 feet—scarp has been eroded back by Gray- Fork and the Cougar Creek structures. these other hng Creek; possibly several episodes of faulting con- rises are inferred to be also dominantly monoclines tributed to this height. Southeastward along strike Floras 3a—flaactieeted fault seam at Grayling Creek. The 1959 warp. about 10 item blZb, rather than slumped fault scarps. the old fault scarps dwindle and give Wily to a low prerlaely utherimposed upon an older fault *carp of Delght. that .00.1 beet loot *anti, Most of these rises in the sand plain can be. dated monocline that is apparently identical with those of east of Grayling Creek. lie, also kg. 31.1 A mill eider and greater dimplacersekt produced a moat within which 11r•0.l10 Creek flowed for a 1.1.115e: the stream rut bath the seath [den only as younger than the sand plain; but it is an the sand plain south of Hebgen Lake. As the mono- formed to produce the 60-foot erosional escarpment that parallels the younger egruccurel 11.S. sorest obvious inference that. like the South Fork structure, cline becomes lower and more gentle, the 1.959 scarp- warps. Dark areas ate aloud Maslow. Vertical aerial photograph by

693-433 0-84-0 6(I THE HEBGEN LAKE, MONTANA. EARTMOZAXE OF AUGUST 17, 1955 DEFORMATION ACCO3rP.A2,-YECG THE IECEDGEN LAKE EARTHQUAKE 67

the anomalous dip and the mole-track thrusts—an ex- planation in terms of the response of weak unconsoli- dated material to gravitational stresses that result from movement on bedrock faults. Because of its lack of strength, the unconsolidated Unconsoadated material of the sand plain cannot maintain a surface materials so oblique to the applied force—gravity—as that farmed in the bedrock beneath, and the weak material s ould fail along a surface more nearly parallel to the vertical shearing stress. As the break in the surficial sediments is steeper than the fault in the bedrock. the surfcial deposits tend to pull away across the devel- oping break, but a variety of slump effects prevents any wide gap from forming (fig. 38). The diverse structures formed during small dis- placements along the faults and monoclines of the sand plain are due to the combined directs of refrac- -on of faults and of slump. Dips of the major faults ,it surficial material are near 75° (fig. 34), and the A the magnitude of the slump effects indicates that the dips of the same structures in the underlying bedrock are about 10° less, or near 65°. We visualize that where the movement along a bed- rock fault is small, with slip beginning at depth and progressing upward during an appreciable interval of time, the resultant slumping toward the fault of the Unconsolidated loose material of the downdropped block causes a materials sweeping deflection of the developing fault past the vertical to 45° and, right at the surface, almost to novas 37.—Reectorato fault heart at Orayllisg Crest. The new fault scarp la amperlaiposed upon an Oder one of similar height. Behind "overhanging" part of the this eoconobito resin la at nctirpment eroded sack by Grayling Creek from a larger older fruit scarp on the same trace. Vi.. looking horizontal (fig. 39). The east across Blerneyecone Ranch; are nine arum 34. Photograph by John B.. Sur,. footwall is lowered smoothly, and the product at the surface is a monocline with a mole-track thrust along lets superimposed on it lessen underthrast produced correspondingly, and the its modern flood plain, parallel to and just above the t its base. The little thrust is an last of the new fissures and the old monocline die out prequake fault scarp (see figures 36 and 37), where by slumping in response to tension. New scarps may together near the park boundary. it apparently flows along a slump moat of a branch or may not be produced at the upper side of the mono- Features associated with the younger of the two of the fault. Tiny en echelon fissures, formed dur- dine. as scarping and flexing alternate at the surface prequake scarps are clearly shown on pre-1059 aerial ing the earthquake, were traced upstream for more along the trend of the structure. Mole-track thrusts photographs and are strikingly similar to those asso- than a quarter of a mile in the surficial deposits on the converge in plan with the scarps at places, so the ciated with the 1959 scarp. A slump graben, or moat, northwest side of the creek and may mark very slight structures must converge at depth also in some fash- marked by low inward-facing scarps in front of the quake shifting on this branch structure. ion like the one shown. main scarp. is obvious in each case. Along the seg- Moats and slump grabens like those along the high MECHANICS OF STILI.ICTURES ment represented in the central third of figure 36 the new scarp east of Grayling Creek are due to slump Fromm 55—Effect of fault refractloti span unconsolidated material. prequake moat, is an undrained depression. On the Where bedrock faults had large effects along faults refracted to steeper dips in thin Fault is creeper change In dip exaggerated) ha .bronsparInty6 new displacements arterial than In bedrock. A, Gap la aerficial depots, produced north side of Grayling Creek—and on a lower and during the earthquake, they broke through to the stir. sections of unconsolidated material. The thinner the by dibplbeen7.17, on fault; B, fault-prodbeed rip ID surflolal younger stream-graded surface—the extension of the face and formed continuous new surficial fill, the narrower and better defined is the pcoita rinsed by slump effects 1 scarps like the one earlier fault scarp is clearly marked. Although the mist of Graying Creek. The large new scarps, all of graben. the scarp was artificially accentuated by roadway fill be- which are in surficial material, are unusually steep, Where the fill is thick. a broad shallow sag is pro- The width of the sac is presumably related to hind the ranch buildings before the earthquake, the dipping about 70° to 80°. The small new scarps, also duced. instead of a narrow moat, in front of the scarp depth of basin fill. Depressions wider than 100 feet base of the natural feature there can be traced on in surficial material, are steeper still or even dip in the or monocline as the surface expression of the slump of were not recognized in the field, hut a slight sag about figure 37 as ar alinement of aspen trees surmounted "wrong direction: they have dips between vertical the subsiding block.. Structural relief, as measured 700 feet wide is suggested by the pattern of the 6.600- by a the long few feet of sagebrush-covered slope, above which and perpendicular to the monoclinal slopes upon across the surface monocline, is greater than the ver- foot contour where it crosses in front of the quake scarp of the sand plain miles south of rises. which they occur. These smaller structures require a tical displacement of the basement beneath by an northern scarp such a depression, Grayling Creek is deflected southeastward across complex explanation to account for such features as amount about equal to the depth of the slump sag. Madisou Arm. If this is indeed 68 TEl HEBGEN 1..ARE, MONTANA, ACCOMINANTLNG THE FIERCER EARTHQUARE OF AUGUST 17, 1955 DEFORMATION- LAHR EARTHOTAHE 69 blocks slumped northward toward the center of the a t basin. 4 . qA A A t t SLIGHTLY REACTIVATED STRUCTURES OLD CONICAL DEPRESSIONS eaansters Conical depressions in the sand plain occur in two areas in the eastern part of the main zone of mono- clines. north-northwest of West Yellowstone (pl. 2). bud I,imam. we Several of these 1.1 uflow structures were reactivated during the earthquake, and their mechanism of formation was re- vealed. One Iine of five depressions begins 500 feet east of the east end of the 4-mile-long structure in secs. 17 and 18, T. 13 S., R. 5 E. and continues for 800 feet along strike. :fprt The depressions are circular, it.A• ranging resit in diameter from 6 to 80 feet, and in depth rvtetiairco„.4„,— %OW. i•laimiamil from X' to 12 feet; 100 -e• inward slopes vary from 17° 3. az.. the steepest cone being also the smallest. In 1:7177 three of these five depressions, central cylindrical plugs 4 to 7 feet across subsided 1 foot to 2 feet during the earthquake. Two of these plugs dropped by a combination of downwarping and faulting, and the third dropped within a vertical ring fault with 18 inches of displacement. Two more depressions occur near the west end of the monocline three-quarters of a mile southeast of those described above. • ...... One is a conspicuous circular, conical depression (fig. 40) 130 feet in diameter that has a closure of nearly 10 feet and a central plug 18 • feet in diameter that subsided a few inches during the Proems 10.—Old maimi deprenloa in the amid plain. Depression is 130 earthquake. The second depression is smaller, feet among. dearly 10 feet deep. •ad probabir yea formed an a sand by LL. Two belles northbortherear of West Yellowstone. Photograph by dohs A. sang. much shallower, and was not reactivated. These conical depressions are collapse features that pock an undissected sand plain which slopes only very spicuously downdropped central plugs. The southern fragments. Many similar collapse cones formed dur- tad, •41,..-bni firsper RN* gently northward. They are interpreted as the result $ group may be older, be degraded by slow slopewash, ing Alaskan earthquakes in 1958, and ejected sand FPIA.1.1 of collapse into and the pits of giant sand boils formed a have cores that were better compacted before was deposited thinly over broad surrounding areas during past earthquakes and are similar to those pro- I 1959 than those of the northern group. (Davis, 1960). duced during the 1958 earthquake and described hr Both croups of conical depressions occur at the F. A. RING-FAL-MT DEPRESSIONS Swenson (chapter N). The cylindrical central ends of monoclines. at or beyond the visible limits of 3007.1 plugs that subsided during the 1959 earthquake prob- t the structures as surface features. Such depressions Seventeen small depressions, defined by ring faults abiy represent the compaction of the sand in the were seen nowhere else, so a genetic relation to the formed during the earthquake, ere gronped in a belt throats of these short-lived sand-clogged artesian monoclines is indicated. At the ends of the underlying less than 200 feet wide that stretches for 1,150 feet springs. Although no deposits from these postulated bedrock faults, relatively intense warping may occur east-northeastward across the quake-emerged. Sat Flooen 311.—Ifeehnolsm of ". formation of monoelfrm tad mole-frank sand boils have which affects the northeast of Edwards Island (pl. 2). Each depres- thrust mouth of Radium Arm. been recognized, such deposits would entire thickness of surficiaI material See text for explariatron. be recognizable only by topographic form, as they (in contrast to the superficial warping that occurs sion is 6 to 25 feet across, a few inches deep, and would consist simply of in the monoclinal, and this deep warping may occurs in modern lake silt that here veneers the sand the unconsolidated fill detritus reworked from the have beneath it might be about said plain. squeezed plain. These con. 1,200 feet deep. The large sand boils developed during out saturated sands, forming the sand boils. features differ from the piequake the ical Most of the 1959 earthquake produced only thin veneers of Similar features, locally called sinkholes. were depressions south of the lake in that they are new scarps in the sand plain lie parallel to the sand and gravel despite the considerable formed during the New Madrid earthquake and have simply depressed plugs unaccompanied by surround- contours of subsidence along the size of the south edge of pits from which this Leen described by Lyell and others (Fuller. 1912. ing cones of collapse. The depressions are probably the subsided termite, where material came: such veneers. p. warping during the earth- formed of 87-881. the formed quake produced a material identical with that beneath them, Eyewitnesses observed fountains of mud and quake-compacted throats of sand boils that surface convex upwards. The faults may be would not long be recognizable. water thrown up from some of the "sinkholes." At along a line of deformation of some past tectonic epi- shallow products of tension caused by this Of the two least some had the original form of an warping and may have groups of conical depressions. the north- inverted cone. sode. Although no new fissuring developed along formed as their hanging-wall ern or ones are the steeper sided and have the more con- crater, surrounded by a ring of sand and lignite this zone in 1959, two new sand boils were formed 70 ?ail: stkeeear LSF.E, MONTANA. EARTHQUAKE OF AUGE'ST 17, 1059 DEFORMATION ACCOMPANYING THE HEBGEN LAKE EARTHQUAKE 7i within it—the only ones for several miles in any direc- was badly damaged, partly by this flood but largely Five-eighths of a mile southeast from HiIgerd Kirkwood Creek at the time of the earthquake. The tion—and these eugg.eet that a slight shift did indeed by the preceding earthquake (fig. 81, chapter K). Lodge, the surge high-water line was only 6 feet man was knocked down by the quake, and when he occur. The trend of this zone parallels that of two The downstream face of the dam, a fill of densely above the postquake lake level; and three-fourths of a got up he noted a wave "10 feet high" coming north- earthquake scarplets along the south side of Horse compacted. calcareous stony alluvium considerably ce- mile farther southeast, at the south limit of the Kirk- east toward him. This was closely followed by a Butte, 2 miles to the east. mented by lime, was carved by gullies 3 to 4 feet deep wood Creek fan, the high-water line was only 3/12 wave "13 feet high- moving directly down the lake (fig. 79). The crest of the dam showed little evidence feet above the postquake lake level. The height of (northwest) ; and this in turn was followed by a ware EFFECTS OF EARTHQUAKE ON HEDGEN LAKE of the overflow, which suggests that the water welled the highest surge thus decreased away from the darn. coming northward, obliquely down the lake. toward The deformation that accompanied the 1959 earth- over slowly. The concrete core wall was covered dur- . Such evidence of the main surge was for the most /dm. No times were noted, but the total interval was quake seas particularly pronounced in and around the ing construction under about half a foot of rubbly part limited to the lower 7 miles of the north shore. short. basin of Hebgen Lake. The basin subsided differen- fill that was partly cemented by lime: this cover was The southern part of the lake subsided less than the Ervin S. Armstrong, owner of the ranch at Corer tially and the lake bed was thus abruptly warped. eroded away by the overflow in some places. but in rest and hence became emergent; its shoreline is en- Spring, reported thnt 13 to 17 minutes after the earth- This caused extensive slumping of the banks and others only the fines were stripped off its upper sur- tirely in porous surficial deposits, largely those of the quake. the water below the ranch, where the shore- threw the water into violent surges that alternately face. Streamers of grass roots. some with soil clods obsidian sand plain (pl. 5). Although the abundant line is on a gently sloping fan. was about 100 yards drained and flooded parts of the shore and sent large still attached, were locally overturned and alined water weeds growing in the lake must have been our from its prequnke position. Then during the wares over other parts. downstream, proving that the settling of the fill which oriented by the surges, the rapid lowering of the water night—he did not witness the event—the water rose exposed the core wall took place before the last of level resulted 'in such an outflow of water from the well above its prequake level. By morning it had SURGES the overtopping surges. Evidence of water flow was lake banks that the weed orientations, studied after essentially stabilized at its new level, several feet found maximum of above the old. During the earthquake, the main body of Hebgen a 1/12. feet above the top of the the earthquake, recorded only the pattern of this out- George Gunnell reported that he ran out of his Lake subsided more than either the Madison Arm or core walls,,about 1 foot above the original top of the flow. fill on cabin on the south share of Madison Arm during the the dam, and a southeast-trending trough was formed. the core wall, and the preservation down to this Clear evidence of strong longshore currents was limit of earthquake and saw in the bright moonlight that the The water surface was warped so that, just after the delicate soil structures and accumulations of found at only two of many points examined. CM a fragile lake was covered by vertically bouncing waves. These earthquake, water in Madison Arm stood 5 feet plant litter showed clearly that the water could small low hill of glacial till (Edwards Island in the not have gone higher. Horse Butte broke the rope by which his boat was tied to the dock. higher than the dam and 10 to 15 feet higher than Hungerford's estimate that full prequake lake, but connected to "3 or 4 of As the earthquake ended, the water began to move water in the main part of the lake. Water from feet" water poured over the dam during the peninsula by ri dry sand flat in the postquake lake). westward, down the lake, and rapidly withdrew from Madison Arm accordingly rushed northward, and its second surge is certainly too high. long filaments of water weed recorded a strong north- the shore. Like others who witnessed this phenome- momentum carried it on us a great surge that over- westward current that extended half a foot above the SURGES ALONG THE LkETSHORM non. Gunnell assumed that Hebgen Dam had burst. topped the dam. The dam subsided only half as much postqunke static water level. At one place on the By studying features along the Another boat avas broken loose by the earthquake as the deepest part of the subsidence trough. and lakeshore it was south shore of Madison Arm an orientation of water possible to determine the from n dock a mile east of the inlet of the South therefore the new static water level against the dam. height of the highest surge weed produced by an eastward-directed cnrrent (a at a number of places. It is a Fork of the Madison River and was found the fol- instead of being above the dam as it would have been reasonable assumption returning surge) was preserved after the streaming thnt the highest surge at any one place lowing day in the sagebrush near Hebgen Lake Lodge, had the dam subsided as much as the main lake bed. represents the of the water stored in the porous first surge. Along the relatively submerged shore- 6 miles away. was actually three-fourths of a foot lower than the A witness to the behavior of the surges, Col. H. B. prequake level. line, particularly along the northeast or lee shore of Crowell, was in his cabin at the end of the road along PIER/OD OP w SURGES- the northern arm, the The first surre of water rose 10 feet above the post- surges left swnsh marks of the southwest shore 2 miles from Hebgen Dam. "About Different approaches to the problem of determining abundnnt debris quake static water level at the dam. As the crest fell lodged in plants and dumped along 7 to 10 minutes after the quake" he heard the sound the interval between successive surges yield contra- the ground. From away from the dam and withdrew southward, the this sort of evidence it is plain of rushing water and found that the lake had risen dictory results. No firm, unambiguous data are avail- that the surges momentum of the water again carried it beyond the reached their highest level near the shout 20 feet vertically above its prequake position. able. darn. equilibrium position. Like the slashing of water in a About half of this amount represents submergence of The darn foreman, Hungerford. and his assistant bathtub, the lake water continued to oscillate back Dramatic evidence for the highest surge was found the shoreline beneath the postqunke static water level. witnessed two or three surges and know that water and forth with gradually decreasing intensity for at to H4-fled Lodge, on the north shore of Hebgen and about half was the additional height of the surge. flowed over the shun once before that. They did not Lake 1 mile from the dam. minute or two" the water was rushing out least 12' hours after the earthquake. Because of the Subsidence during the Within "a presume any clear estimation of the times involved. earthquake brought the new static lake level up to again. Boras and docks were washed away and were large amplitude of the waves And their general vio- People who had been camped in the Mndison River the doorsteps of two cabins and put the base of a subsequently found on the northeast shore of the lake, lence, they are here called surges to distinguish them canyon above the Mathison Slide left their camps after from the much Douglas-fir tree, for example. under 6 feet of water directly north of their original position. Later that milder but otherwise similar oscilla- the earthquake and apparently began to arrive at the tions of closed water (fig- 9). The surge then rose 9 feet above the new night and early the next morning the water level bodies called seiches. (For dam about the time of the last overtopping, although similar usage. see hike level and left a clear high-water line of debris changed gradually by much smaller amounts as the Prins, 1958.) Hurigerford's statement only this. From his in the sagebrush behind the cabins. As the two cabins surges became more regular. Crowell also quoted re- infers arrival of OvP_IrrOpMrNG Or THE DAM floated off their foundations. they tipped, and the ports by friends at Watkins Creek Beech who said account of the things he did before the and from the Water poured over Hebgen Dam three or four times. high-water marks on the inside walls therefore sloped that shortly after the earthquake. several large waves . the campers turned him to other tasks, as not reported by the dam foreman. Mr. Hungerford conspicuously. One cabin was sec hack on its founda- rolled up on the alluvial fan there. presumption that the first of the campers did (chapter A. L). It flowed first through the spillway tion only a few inches out of place and was little Another report of waves an the lake was obtained linger in the canyon before driving up to the dam, it outlet, and continued to rise until -it poured across damaged. whereas the other was dumped a yard out by engineer Glenn Jones of the Montana Power Co. seems that the total sequence of overflowing,s took the entire width of the dam. The concrete spillway of place and was badly skewed. from a couple who were standing by the road near about half an hour. 72 THE frEacEy LAKE, MONTANA.. EARTHQUAKE OF AUGUST 17, 1939 DEFORMATION ACCOMPANTLYG THE HEBGEN LAKE EARTHQUAKE 73 Overtopping of the dam was recorded by s con- about 10 minutes after the earthquake. By several tinuously operating stage recorder at the river-gag- hours later. a complex or multinocial surge with an is ing station one-fourth mile below the dam. The river effective short period was apparently active. height here varied not only as the water overflowed ic- ____ the dam, but also as the water depth, and hence pres- PROFLLE Or TILE HICIELEST StiRCIS PROFILES______sure and outflow, varied at the intake The approximate MATER tower in the profile of the crest of the highest Pottanakt static water level lake. The time surge - scale on the graph of the river-depth in the lower part of the lake is plotted on recorder is only 0.1 inch = 1 hour. and the instru- figure 41. The warped shoreline, and hence the de- ment is designed so that fast changes in level are formed position of the water surface assuming. that rP- damped out and nor recorded immediately. The no movement of water occurred during the earth- earthquake knocked the time scale out of position, quake. is also shown. and the - first surge made the pen slip to a new depth Maximum surge height increased rapidly toward a YS position. These various factors combine to make the dam, largely due to the funneling effect of the the graph — ambiguous during the first several hours convergent steep valley sides. but the highest point HEBGEN LAKE after the I earthquake. The main earthquake occurred was several hundred yards from the dam.. This is v.ERTICAL EXAGGERATION x 528 at11:37 p.m. local time. The record before. 2 or only in small part due to draining off of the crest by 3 a.m. is ambiguous, but for 6 hours or so after that flow over the dam. Much of this effect was due to the recorder clearly indicated alternate high and low refraction of the surge toward the valley sides as the water levels, with a period (high to high) near 15 surge was.retarded by these sloping shores. Ii minutes. ta ire Howard 31atthai (written communication. The form of the surge was apparently intermediate its 1959) of the Id Geological Survey interprets the graph between a wave sloshing back and forth aloug the lake ta for the few hours immediately after the earthquake. and a seesaw rocking of the water. Modern studies including the period when the surges flowed over the of water wares suggest that there is a continuous dam, to ehr at re.051 Loire rut a ewgnaw indicate that they arrived at the dam about variation in waveform and behavior between the see- 17 minutes apart. saw movement of a low uninodal seiche in deep water (F...m.r.Postouraki04_ ,. The period of uninodal oscillation in a rectangular and turbulent surges where the height is large rela- tank can be calculated from the simple Merian tive to the water depth (Prins. 1256). In both types, Bt l".41.-1,..eitodinal projection or Tiebern Lake. stowing deformation doe to earthquake. formula velocity is proportional to the square root of the , 2/ depth. RANK slumps I/ Ph SLUNEr crescentic; at the intervening headlands the took place within a few minutes after the earthquake disappear under water. but that slow creep continued for at least 18 hours. The banks of Hebgen Lake slumped extensively where t is the period, 7 the length of the basin- p the during and shortly On the south shore of Madison Arm, a floating dock The separate supports, although not connected by any acceleration of after the earthquake. Along the gravity. and ft the water depth (Rutt- steep and its Iakeward anchor posts were shifted basinward structure. retained their relative positions throughout ner, shores of the lake near Hebgen Dam. large land- 1953, p. 43). Using 15 miles for the length, the as the shoreline slumped. The supports were carried the sliding. slides went into the lake, greatly changing. its bor.bot- period indicated by this formula is about out 31 feet yet lost only 2 feet in altitude: this in- The above types of slump resulted from the rapid an hour, tom topography__ These slides and their effects are varying with the depth assumed. As the dicates sliding of sediment on a surface sloping only outward flow of ground water from the bauks as the ratio of described by J. B. Hadley (chapter K.) and by H. length to depth is large, W. 3.5°. A witness reported that most of the movement lake water retreated. Large quantities of water stored this variation is not great. Jackson (chapter H). Hebgen Lake is of course not a rectammlar tank. In the eastern part of the north shore of Madison Jeffreys (as quoted by Hutchinson. 1157. p. 301-6) Arm. a high steep bank of unconsolidated sand emphasizes that long faces narrow lakes, especially those bays formed by the with flooding of deep-cut meanders cuspate ends. hare periods markedly /Modified meander scar shorter than in the sand plain. During the 40-year life of the indicated by the simple Merin], formula. reservoir, a narrow beach was formed at the high- This calculated period is much longer than that water level: lakeward, the indicatedby flooded meander flat was Cnscontinuous line of My all marks either Hungerforris account or Sail of high water head./ the river covered by a mat of waterlogged timber. During or 0051Clutilne position of slumped beach block, gage. It seems that either surmounted Dy ChADDCIDODOle Of Water- the surge was multiraxlal shortly /Prenuake sewn block after the earthquake, this beach slipped away / logged limo, or that the great height of the initial surge caused it from the bank and Pferauake wilier level to move far left behind shallow moats as wide faster than the ideal formula would sug- as 30 gest. feet. The timber mats were josrled to chaotic Postquake water level The lake bed was so warped that immediately jnnibles of logs at all angles (fig,. FEET after the earthquake 42). The slip sur- the water in Madison Arm stood faces are nearly essentially the downslope extensions of the 10 50 /5 feet above the water in Rooted slump the main part of modified meander scars. At the lake only 5 miles several places the slump to the north. The wave result- ing from this escarpment is continuous throughout the full extent warping apparently reached the dam of a concave meander bay. These features are thus Flamm 12—Protlie of slumped 'vane Os earth aide of Madison Arm of Henna Lake. Emerked takkveeter beach slid 'Leeward. filial meat keeled. 74 THE BERGEN LAKE, arOarmarara. EARTHQUAKE OF AUGUST 17, 19 .5 DEFORMATION ACCOMPANYMG THE HEBGEY LAKE EARTHQUAKE 75

HIGICIVA,T PROF/LE DETERMINaTION in the highly permeable and porous banks doubtless ABSOLUTE SUBSIDENCE marks between the 195-1 and 1959 surveys are indi• moved out under differential preesure, thus reducing The large fault scarps and many of the other earth- sated on plate S. Bench marks within the Madison River canyon the cohesion and resistance to flow of the bank mate- quake features around Hebgen Lake are spectacular It is quire possible that a smut] part of the sub- -west of Beaver Creek were buried by Madison Slide rials. The normal slow annual drawdown of the lake but localized effects of the wholesale subsidence of a sidence between the 1934 and 1959 surveys took place and Earthquake Lake and so could nor be recovered has little such clump-causing effect. large area. Subsidence has been proved over much before the earthquake, but no convincing arguments during the Coast and Geodetic Surrey releveling. The far Gravity sliding. different from the slumps noted of an area nearly 15 miles north-south and almost can yet be made. The data are plotted on plate- level line was run along the highway nearly as above. affected the lake bank in the west part of twice as long east-west. About 60 square miles sub- with the assumption that the total measured subsid- as it was exposed within the canyon. however, and Grayling Arm, just east of the Narrows. A sheet sided more than 10 feet, and about 200 square miles OlVe occurred during the earthquake. we determined the highway profile from the indi- nearly a mile wide along the lakeshore pulled away subsided more than 1 foot. Only about half a foot of Llano of the releveled bench marks are in pairs, vidual level-rod measurements. By comparing this from the hillside highway and moved outward more than 6 feet absolute elevation has been proved. 1::,2 feet apart, on opposite sides of the highway. postquake profile with the prequake profile, while dropping only 1 foot. The east shoreline of the The maximum proved subsidence-22 feet—oc- t Northward from West Yellowstone toward Duck we were able to extend the measure of absolute sub- crescentic tip of Horse Butte peninsula is rimmed curred in the West Yellowstone basin, Part of the Creek the subsidence indicated by the bench marks sidence well into the canyon from both ends. The br almost continnous slump scarps that follow the Madison Valley near the latitude of Hebgen Lake increases gradunlly. Where the total subsidence is prequake profile, provided by the Bureau of Public deeply curved shoreline for more than a mile. Pull- also subsided, though not as much: and subsidence • small, subsidence of the paired stations tends to be Roads. was in the form of a profile of turning points apart scarps and open fissures are steep. The posts has been proved through much of the Madison River uniform: but where the total subsidence is greater, of tangents to the road surface. Applying the usual of a barbed-wire fence extending directly into the canyon, which cuts across the intervening Madison a general side tilt toward tile lake is recorded. Of engineering formula. we derived the actual parabolic lake across the slide zone still good vertical landward Range. /0 pairs. 6 were tilted lakeward. 1 was tilted away shapes of the prequake road crests and saddles from of the zone but within the slide were rotated. tops f•arn the lake. and 3 were not tilted. The three these tangents. shoreward. Most posts were pulled out by the wire, .MEASUREMENT OF SUBSIDENCE arthern pairs. where subsidence is the greatest for The old and new profiles in the Madison River but one post that remained partially imbedded moved that increases are shown in plate 3. Subsidence is demon- The northwest side of the West Yellowstone basin, this leg of the line, show a side tilt canyon at least 6 feet horizontally. strated throughout. Bench marks, as releveled by The slope of the lake where subsidence was greatest. is traversed by a pre- northward from 5.2 to 10.1 feet per mile. floor here is less than 1°, and it is probable that the absolute sub- and Geodetic Survey, subsided compatibly vise level line established by the Coast and Geodetic The figures shown on plate 2 for the the Coast water-saturated sediments actually slid out into the between highway as indicated by the profiles: and Survey in 1934. Releveling of this line after the sidence of bench marks are the differences with the lake on glide surfaces of similar gentle slope. Rapid unadjusted 1952 points on or near bedrock subsided compatibly with earthquake revealed the amount of absolute subsid- the adjusted 1934 altitudes and the drawdown played no part: this slumped area was applied the surficial material. The profile determination ence in the West Yellowstone basin and in the Madi- field altitudes, to which however, we those on slightly submerged as a result of the earthquake de- the Coast and extended along Hebgen Lake. and it was found son Valley. We constructed a continuous postquake orthometrie corrections computed by was formation and is located where the surge was small. based everywhere compatible with subsidence indicated by road profile from the individual level-rod readings Geodetic Survey. These 1959 field altitudes are The Grayling Arm slide may hare resulted from a West marks except in the places where slumping of and compared this with the prequake pro5le to meas- an the assumption that bench mark N-33 at the bench transient excess pore pressure caused by local tom- the unchanged in alti- highway was made obvious by landslides and sur- ore absolute subsidence oki„g the parts of the Yellowstone railroad station was the Nation of surficial sediment during the earthquake. by the nearly scnrps. The continuous smooths character of the Madison River canyon not buried by Madison Slide tude, and that this stability is indicated ficial Downdropping of surficial material was followed by N-33 and the next tectonic subsidence is made clear by the profiles. and Earthquake Lake. Displacement of the prequake unchanged relative altitudes of abrupt deceleration, and the momentum briefly ap- surrey Small uncertainties in the profile comparison re- shoreline of Hebgen Lake relative to the new high two bench marks to the east. The reiereling plied a load throughout the material. As emphasized traverse miles to the from differences in height of the shoulders and water line. the absolute altitude of which was deter- was ended at Sappington, 110 sult by Rubey and Hubbell. (1950, p. 183), instantaneous en this assump- of the highway, height of the level instrument. mined during the leveling, made it possible to deter- north, a-here the field altitude based center loading produces the greatest excess fluid pressure change of -6.1 foot: depth to which the level-rod pin is set, and ambiguity mine absolute subsidence all around the lake. Mens- tion indicated a 1934-59 attitude (and hence the tendency toward flotation/ top validates the assumptions as to exact position of ridge crests and valley floors. at the urements of displacement along the new scarps and a the minuteness of this figure of a sedimentary column in compaction equilibrium. good agreement of 1034 uniformity of subsidence shown by comparison fathometer survey of the lake floor provided further within very small limits. The The If the Grayling Arm gravity elide was rendered ef- bench marks from Sapping- the two profiles, however, makes it clear that such information on the changes of level that accompanied and 1959 altitudes of the of fectively frictionless by pore-water pressure, it is of area is further validation. uncertainties generally total lees than 1 foot and are the earthquake. These data are plotted on plate 1 ton south to the subsided the type illustrated by Rubey and Hubbert (1959. Assuming that some other part of the little-changed far smaller than the 6 to 22 feet of absolute sub- fig. 91. REZLEVELING relevel line represents nbsolute stability would change sidence. At the Narrows ramp SHORELINE ME.1.51TREMENT and eastward along the north In 1984 the Coast and Geodetic Survey established the subsidence figures by n maximum of several tenths shore from it. effects of slump and of structurally a line of second-order levels between West Yellow. of a foot. Such a change has been made by the Coast About 160 observations were made around the lake controlled surface cracks are mingled; featares which stone and Sappington. Mont.. following the highway and Geodetic Survey in a provisional adjustment of the vertical difference between the prequake high- in detail are clearly related ro the hydrographic basin along the northeast side of liegben Lake, through based on the assumption that the first three hench water mark and the water level of the moment. The are in a more general way apparently related to struc- the 31adisan River canyon, and down the Madison marks :beginning with Y-138} north of the subsided surrey was conducted from an outboard-powered boat turn] deformation. A continuous single scarp 1 foot Valley. All previous bench marks along the route zone remained stable. This adjustment minimizes the loaned by the V.S. Army Corps of Enguiers. The high and facing the lake thus parallels the gently were tied into the level line at that time. In Septem- changes between the 1034 and 1959 runs outside of the entire shoreline was examined. anti points were chosen curving shoreline at a distance of only 50 to 100 feet her and October 19111 this level line was rerun by subsided area and lowers the 1939 unadjusted altitudes where measurements could be made most reliably. for a mile eastward from the Narrows. This scarp the Coast end Geodetic Survey. this time with first- by as much ns 0.2 foot. The assumed stability of West The lake level is measured daily at Hebgen Dam is parr of a broad zone of scattered surface breaks order accuracy. Along a 32-mile part of the route. Yellowstone that is incorporated in the data plotted by the darn foreman, so that all shoreline measure- that seems front its general pattern to be structurally -1.7 bench marks, all but one of those recovered, sub- plate 2 thus gives maximum values for the uplift ments can be related to the Coast and Geodetic Sur- controlled (pl. 2). on sided more than 0.5 foot. Changes in level of bench outside the subsided area. vey level line and thus define the amounts of absolute 76 THE HEBGEN LAKE, MONTANA. EARTHQUAKE OF AL,GUST IT, 1059 DEFORMATION ACCOSIPANYLNG THE HEBGEN I at-' EARTHQUAKE 77 subsidence. The shoreline measurements were made The entire lake basin subsided in absolute altitude. several weeks after the eurtliquake. by which time but the lake-level changes relative to the postquake the lake level had been allowed to drop about St., feet full water level were such that the prequake shoreline below the high-water mark at the dam. The deforma- of Madison Arm emerged about 5 feet. Grayling Arni the lake tion of bed was such that the relative water was little changed, the shore of the main body of the level at the data after the earthquake was three- lake was submerged 5 to 10 feet, and there was little quarters of a foot lower than it was before the quake. change at the dam. The emergent pert of the lake Shoreline observations were recorded as positive and subsided less than did rite submergent part. The negative readings from the pool level of the moment amount of absolute subsidence wined from about and were corrected subsequently for chances in lake 5 feet in Madison Arm and 12 feet in Grayling Arm level. although during the time of shoreline measure- to 22 feet in the main part of the lake. but only ments and of the futhometer surrey the hike level 9 feet at the dam. varied little more than 0.2 foot. SCARP ELPTG Errs Where the new lake level is lower than the old, the At height of the high-water mark above the present level each new fault scarp, the amount of displace- ment of the was measured. The lake is kept full for several ground surface changes abruptly. The heights months each summer. whereas other levels are incon- used on plate I for tile major scarps north- east stant: the lake is drown down more or less steadily of Hegben Lake are those reported by Witkind (chapter,G) ; in the fall and winter and filled steadily in the spring. he describes methods of determining scarp heights The summer high-water mark is in most places easily and corrected displacements on the ground surface. recognizable by staining of rocks and posts. by Heights of lesser scarps east and south of the grounded driftwood, by the upper limit of beach sand, lake were measured or estimated by the writers and and by nicks cut in unconsolidated material. Re- others. peated independent determinations by different geolo- FATROMETER sukvz-ir gists showed that the uncertainty in specifying the A contour map of the lake floor after the earth- high-water mark was commonly only about 0.1 foot. quake was made by Wayne Jackson and is incorpo- Most of the determinations were made with Abney rated in his companion paper. chapter IL No com- level and stadia board and were reproducible in the parable surrey was made before the earthquake. but immediate vicinity of any observation to within 0.2 deformation during the earthquake is clearly indicated foot. Along parrs of the south shore of Madison by the data. As the fathometer-indicated deformation Arm, old mid new shorelines were widely separated. cannot be measured precisely, it was not specifically and determinations were made with planetable and incorporated in the deformation map, but the isobases telescopic alidade. 1 of plate 2 are consistent with it. FloCI. 43 —.5thbwerred 1.55 ebrubs sod grew in fletrec Lott Nonbeass rhore of lake. pear Kirkwood Creel, 0.5. Perrot Service Measurements were necessarily less accurate where I plunO5raph. the old shoreline was submerged beneath the new lake CHARACTER OF SUBSIDENCE level. Suspended organic matter and clay generally Most of the points whose subsidence was measured ing was generally made obvious by features such as ern base of the Madison Range Was reactivated and limited visibility to less than a foot, so the old. shore- are on unconsolidated material, but some points on landslide scars and fissures and by its local and ir- formed a new scarp a few feet high. line could seldom be seen beneath the water. The sub- both sides of Hebiten Lake are on bedrock (pl.-2). regular character. Critical analysis of releveling data merged shoreline was determined mostly on the basis MADLNON VALLEY of its The subsidence is smoothly continuous from bedrock in wet bottomlands shows that purely surficial sub- submerged land plants and trees (fig. 45). The to surficial deposits, so nbviously the bedrock beneath sidence there cannot have exceeded 1 foot and gen- Releveling of bench marks by the Coast and Geo- depth on the lakeward side of submerged trees, shrubs, the cover sank also. The long uniform new scarps. erally was less than a few tenths of a foot. detic Surrey demonstrated that the floor of the Nadi- and plants RAS probed with a pole marked in feet and notably those northeast of Hebgen Lake. clearly re- son Valley subsided 7 feet (p/. 2) near the mouth of tenths. Willows and herbaceous plants grow DEFORMATION WEST OF MADISON RANGE flect faulting in buried bedrock. In the large the Madison River canyon. The line of levels trends down to the ]sigh-water mark of emergent sectors, so areas where only surficial deposits are seen, the uniform The spectacular effects of the earthquake are near obliquely into the topographic trough of the valley. readings based on them were used without correction. For pattern of subsidence makes it clear that here also Hebzen Lake. in the West Yellowstone basin and then follows the trough northward. Five miles below conifers along the submerged shoreline northwest of the the bedrock beneath shared in the deformation. Slid- adjacent parts of the Madison Range, but substantial the mouth of die canyon the subsidence was little Watkins Creek fan, an arbitrary correction factor of ing acid compaction are quite inadequate to explain deformation extended many miles to the west. Fart more than a foot. and at Kirby's Ranch. about 51:/e 0.f foot, based upon inspection of the emer- more than local details of the changes. Most of of the Madison Valley near the latitude of Hebgen miles farther, it was only a few inches. No subsidence gent shore. was added to the observed the submergence of subsidence was tectonic. Lake also subsided, although not ns much as the lake was recorded at bench marks still farther north the base of the trees. This method can only yield mini- Modification of the basin. Just west of the valley, subsidence and tilting along the line. mum values: in tectonic subsidence pattern by some pieces probing suggested that are both the slumping and possibly compaction of unconsolidated indicated by lake features. East of the The greatest proved subsidence woe thus at the high-water beach was as touch as a foot or even deposits occurred also. Important examples of this valley, a few miles south of the mouth of the Madison southernmost bench mark surveyed. Subsidence of two feet below the reported level. are described by J. B. Hadley (chapter It:). Slump- River canyon. a segment of the fault along the west- the Madison Valley was probably near its maximum 78 THE HEDGES LAKE, MONTANA. EARTHQUAKE OF AUGUST 17, 1939 DEFORMATION ACCOMPAYYLNIG THE HEBGEN LAKE EARTHQUAKE 79 there. as the bench mark is close to the axis of a major In 1959 a new, smaller scarp formed on this pre- warping at the range front. The highway profiles subsidence along the lienzen fault scarp at the lower y-ncline formed by repeated late Quaternary warping historic scarp, Between Little Mile Creek and Sheep from the Stagger Ranch toward the landslide (pl. 2) end of Hebgen Lake might be due to similar moat ( p.89-97). Creek. a distance of 1.5 miles. discontinuous near-ver- show that the mouth of the canyon and the adjacent collapse into a fissure in bedrock at depth.) Other data. relative rather than tical absolute. gathered new scarps facing valleyward with up to 3 feet part of Madison Valley subsided evenly; most values In the western part of the Beaver Creek basin, the both east and west of the Madison Valley indicate a of offset were formed along the old scarp. At the range from 7 to 8 feet. The Madison River runs highway passes less than 100 yards from a bedrock complex pattern of subsidence for the valley. For mouth of the unnamed canyon next north of Little near bedrock at the mouth of the canyon. Although hill that proj4cts through the valley fill: no evidence example, ditches carrying irrigation water around Mile Creek the prehistoric scarp was formed in ex- the exposed part of the highway rests on colluvium, for slump of the fill away from the bedrock was found the alluvial fan at Sheep Creek (pl. 2) were tilted. tremely coarse morainal debris containing abundant no slump scarps or other evidence for sliding of this during a search. The lack of such slump and the and some became useless as a result. blocks many feet across. No new scarp formed across material were seen, and the striking uniformity of evenness of subsidence of the highway contradict the the canyon during the earthquake, apparently subsidence from Stagger Ranch to the landslide is con- hypothesis of Fraser, Witkind. and Nelson (chapter CLTFF LAKE AND WADE L ATE because of the extreme coarseness of this till. Instead, the vincing evidence for tectonic subsidence. J) that the subsidence represents compaction of sur- Cliff Lake. which is more than 3 miles long, within blocks were jostled about. and the old warp became The interpretation of the 1959 deformation pattern ficial sediment rather than tectonic change. the mountains just southwest of the Missouri Flats heightened. This is shown graphically by the stretch- by Fraser, Witkind, and Nelson (chapter J) and by The Bureau of Public Roads releveled the highway par: of the Madison Valley. was tilted during the ing of a wire fence that goes down the scarp. The Witkind 11961) illustrates the hypothesis that the profile in the Madison River canyon in the spring of enrtnquake. The lake was visited by J. B. Hadley and fence was stretched so tight that its supporting Madison Range front was warped during the earth- 1960 (Lynn D. Tingey, oral and written communica- District Forest Ranger Neil J. Howarth 5 weeks after weighted tripods—no posts could be sunk in this mate- qnake and that the subsidence shown at the Stagger tions, 1961). This new highway profile agrees very the earthquake, and by Hadley again in the summer rial—were pulled clear of the ground to make a Ranch was tectonic whereas that within the canyon closely with the Coast and Geodetic Survey level line of 1960. They (written communication. 1960) found straight line from the top of the old scarp to a tripod was due to slump. In the light of the road-level data, run in September 1959, shortly after the earthquake, the pi-quake shoreline of the northern part of the about 25 wards beyond the base of the scarp, thus 'nix hypothesis requires that slump and tectonic down- except that local postquake subsidence between the lake to be submerged beneath the postquake lake sur- illustrating an increase in scarp height of several feet. warping were exactly balanced in such a way that the times of these surreys is shown. Between September face. The rise in water level attained a maximum of A nearly continuous new scarp 1 to 3 feet high ex- thick fill of Madison Valley was unafected by settling 1959 and the spring of 1960, a section of the highway 11,,z-.2 feet at north end of the lake, as shown by such tends northward 0.6 mile from this canyon along whereas the very thin fill of the adjacent canyon only 300 feet long, centered 2.900 feet southwest of features as the partial submergence of a parking area. the old scarp. At Sheep Creek, 1.5 miles farther subsided by the same amount entirely by surficial Beaver Creek, subsided up to 9 feet in addition to the By contrast, in all three arms at the south end of the north. there are new cracks along the old scarp, but settling. We regard the evenness of the subsidence 11 feet of subsidence during the lake. old and new there earthquake, shorelines coincided. The bed of vertical displacement on these was limited to an inch of the highway (pl. 3) across the range front as con- the northern and trees were tilted along the newly dropped section. parr of Cliff Lake was thus tilted north- or two. clusive evidence against this hypothesis. ward; the bed of the entire lake This narrow new sag is coincident with the topo- was not tilted east- The prehistoric scarp makes an abrupt bend at Subsidence of the upper part of the Madison River ward. graphic trough, noted above, which crosses the road Sheep Creek, from which it trends northwestward in canyon is related systematically to the preexisting The water level of Wade Lake, which is obliquely at this point at the northwest edge of the just north one direction. and a little west of south in the other; structural topography. For most of its course between of Cliff Lake, was reported by Howarth to Beaver Creek basin. The new sag is so narrow that a be higher reactivation during the earthquake was limited to Cabin Creek and Beaver Creek the highway runs along after the quake shallow than previously, without obvious dif- the south-trending leg. or near a bedrock slope and subsided as little as 6 cause is indicated. Possibly this cause is in ference in relative level from one end to the other. feet (pl. 3). At Beaver Creek the highway enters a slump of surficial material toward the new Earthquake The lake is dammed by an old landslide, and its DEFORMATION IN MADISON RANGE broad moraine-filled basin for which a structural ori- Lake; but the stability of the coarse till, the distance deepening may have been due to increased inflow or to The Madison River canyon, entrenched directly gin seems required: subsidence is 8 feet at the bench to the lake, and the gentle slopes involved argue tightening of the permeable dam rather than to de- through the Madison Range. subsided fully as much as mark on Beaver Creek bridge and increases smoothly against slump. The coincidence with the probable formation. did the adjacent parts of Madison Valley and the to 14 feet at the point where the highway disappears structural trough leads us to infer a structural cause The isohase map by Fraser, Witkind, and Nelson West Yellowstone basin. This is shown primarily by beneath Earthquake Lake. Deformation during the instead. We suggest that the Beaver Creek basin {chapter J) of subsidence during the earthquake does comparison of the prequake and postquake highway earthquake increased the depth of the basin within was pulled slightly away from the fault bounding it not account for the tilting of Cliff Lake. profiles (pl. 3). All bench marks from Beaver Creek the Madison River canyon at Beaver Creek. on the north during the earthquake, opening a near- MADISON RANGE FAULT westward to the mouth of the canyon were buried The Beaver Creek basin is surfaced largely by a surface fissure in the Precambrian rocks, and that the under Madison Slide and Earthquake Lake, but the crescentic Bull Lake moraine that trends obliquely surficial materials have slowly collapsed into this fis- continuous Recent but prehistoric fault scarp highway-profile level data extended the measure of toward the northwestern canyon wall. The moraine sure since the earthquake. runs along the west foot of the Madison Range from proved subsidence nearly as far as the is broken by a narrow The Bureau of Public Roads survey ran along the north of the Madison River canyon to Mile Creek road was ex- trough at the foot of that wall, posed—eastward to the landslide colluvial north shore of along the (pl. 5). At Mile Creek this scarp leaves the front from the mouth of and on the north side of the trough a remnant of Earthquake Lake. the canyon, and of the range slid trends along a canyon oblique to the westward to Earthquake Lake from what appears from both photo interpretation and field route of a new road since built. A resurvey in Sep- crest of the mountains. The scarp is most impres- the head of the canyon. reconnaissance to be the same moraine occurs 150 tember 1960 showed that between spring and fail of sive: with a height of 20 to 40 feet and a slope of At the mouth of the canyon. the last recovered feet ah

DEFORMATION EAST OF MADISON RANGE imaginary surface, horizontal before. the earthquake, fault reactivation along this slope was found despite moth Hot Springs was by a maximum of several Deformation extended east of the Madison Range and deformed during the earthquake into the pattern intensive search. The isobases of subsidence trend into tenths of a foot. into the northwestern part of the Yellowstone Plateau. shown. Such contours were named "isobases" slope at high angles, and surveys showed that by this RELATION BETWEEN FAULTING AND wa.n.Pnea During September 1060. the Coast and Geodetic Sur- DeGeer (1893). Hebgen Darn was not tilted crossways. The relevel- of warping and vey releveled bench marks along the highway from Releveling showed much absolute subsidence but ing data prove continuous subsidence between Hebgen Various combinations faulting by West Yellowstone to Livingston, Mont., via Madison has proved a maximum of only 1.7 feet of absolute Lake and the Beaver Creek basin. which the subsidence during the earthquake was ac- Junction. Norris Junction. and Mammoth Hot Springs. elevation. The subsidence proved by this surveying The greatest subsidence of the West Yellowstone complished are illustrated by the profiles in figure 44. Data from this releveling have not been rigorously is quantitatively adequate to account for all of the basin occurred well to the South of the Madison Rirer Profile A-A' illustrates the situation in which the adjusted and are not shown on plate 3, but applica- observed features, and in preparing the map it was cenyon, and most of the isobases trend westward at near-maximum subsidence of 18 feet took place by single fault. the Hebgen fault. tion of approximate orthometric corrections to the assumed, accordingly, that practically all the deforma- high angles into the Madison Range. The basin of displacement along a surface by the preliminary figures indicates that the sector between tion during the earthquake was by subsidence. It was 1959 subsidence is continuous via the Madison River The net displacement of the ground 4 miles east of Madison Junction and Norris Junction further assumed that the change in altitude across canyon with the depression of Madison Valley. Ab- fault (the height of the new warp, 20 feet, less a es- subsided 0.3 to 0.7 foot. the maximum having occurred new warps continued far beyond the faults—that is, solute subsidence of Madison Valley was determined 2-foot correction for small backfaults) is 18 feet, in the vicinity of Gibbon River Rapids at the axis of that large areas moved relative to one another across only northwest of the mouth of the cnnyon, but the sentially equal to the absolute subsidence proved by a youthful syncline that trends N. BO' W. to merge the faults. This assumption was proved valid wher- northern part of Cliff Lake is known to have been the highway profiles and the warping of the bed of with the West Yellowstone basis. Subsidence con- ever it could be tested by level or shoreline data. tilted northward. 'We infer that Cliff Lake marks Hebgen Lake immediately to the south. The down- tinued with decreasing amount for 9 miles north of The isobase map is tightly controlled in the vicin- the southwest limit of subsidence, that Wade Lake warped area was not simply tilted, as its profile of Norris Junction. beyond which there was uplift—the ity of Hebgen Lake and fairly well controlled Iles along an isobase, and that the axis of 1959 sub- subsidence is concave upward. Northwestward along peripheral upbowing, recognized north of the sub- throughdut the West Yellowstone basin and in the Eidenee follows the general axis of late Quaternary the Hebgen fault the height of the 1959 scarp de- sided terrane of the West Yellowstone basin and Madi- Madison River canyon: only minor adjustments could warping across Madison Valley. Whether the subsi- creases and the subsidence measured by displacement son Valley also--by an apparent maximum of about reasonably be considered in these places. Elsewhere, dence basins of Madison Valley and West Yellow- of nearby bench marks on the downdropped block 0.3 foot. decreasing to zero at Mammoth Hot Springs. several interpretations could be made. The greatest stone basin are connected by a broad zone of de- decreases correspondingly. The northeast margin of Madison Junction, south of the newly deepened sem- uncertainties lie in the possible extensions of the zone pression of the intervening Madison Range or only the deformed area is thus the warp of the Hebgen cline, rose 0.1 or 0.2 loot to define the southern mar- of deformation to the east and to the west, and the by the proved depression along the Madison River fault northwestward from profile A-A', and in this ginal bulge. No consistent change was detected be- biggest problem is in the behavior of the Madison canyon through the range cannot now be demon- short interval the total subsidence is effectively repre- tween Madison Junction and West Yellowstone. Range southwest of Hebgen Lake and south of the strated. In either instance, a continuous basin of new sented by the height of the new warp. By far the greatest damage to roads within Yellow- Madison River canyon. subsidence formed that was elongate east-west rather The pattern is quite different southeastward from stone National Park occurred from rockslides along The West Yellowstone basin and Madison Valley than parallel to the major new northwest-trending profile A-A'. The new scarp on the Hebgen fault the Madison and Gibbon Rivers, within and just depressions of new subsidence are connected via the fault warps. becomes lower so rapidly that it is only 2 feet high south of the subsided tract. continuous zone of subsidence along the Madison Synthesis of the data (ph 3) suggests a broad com- 0.6 mile from A-A'. As the subsidence of the high- The fire lookout on Mount Holmes, in the Gallatin River canyon. As the canyon cuts directly across the pound basin of subsidence that plunges gently east- way, which is only a few hundred yards southwest of Range between Hebgen Lake and the Gibbon River, Madison Range. the nearby part of the range must ward across the upper Madison Valley, the southern the fault, remains constant at 22 feet in this distance, reported that a crack opened across the ridge between have subsided also during the earthquake. Whether part of the Madison Range. and the West Yellow- the dwindling of the scarp must record e downwarp Mount Holmes and Trilobite Point. Whether this was the Madison Range some miles to the south of the stone basin. It is markedly asymmetric, with the toward the south of the high-standing block northeast a fault scarp of small vertical displacement or merely canyon subsided more or less than did the nearby north flank dipping much more steeply than the of the fault. The Hebgen fault, which forms the a slump feature in unconsolidated materials was not parts of West Yellowstone basin, Madison Valley, and south. In the West Yellowstone basin the south flank margin of the subsidence basin at profile A-A', determined. the Madison River canyon cannot be proved: but for is a gently dipping platform whereas the north flank changes southeastward to become an element of rela- reasons outlined below, we infer (pl. 3) that is terminated obliquely and abruptly by the large new tively minor displacement that extends for 7 miles GEOM ETRY OF DEFORMATI ON the crest of the range did subside. although less than did scarps of the Hebgen. Red Canyon. and other faults. within the deeply subsided tract. The northeast mar- SUBSIDENCE PATTERN the basin. Decreasing subsidence continues eastward into the gin of the snbsided basin has shifted northeastward The absolute deformation during the earthquake The existence of the Madison River canyon might Yellowstone Plateau at least ns far as Gibbon River to the warps and associated warps of the West Fork can be defined with precision along the line of repeated itself be taken es evidence for a structural depression Rapids. The basin of proved subsidence is 43 miles and Red Canyon faults. levels and new highway-profile determination and across the Madison Range. and the structural basin at long, from Norris Junction to Cliff Lake, and trends The orientation of the downwarp that diverges around Hebgen Lake. Other data. relative rather the mouth of Beaver Creek may he only a part of a N. 78° W. from the Hebgen fault cannot be firmly established, than absolute, are provided by scarp heights and by bigger Madison River canyon depression. The general RelereIing by the Coast and Geodetic Survey dem- for the only control by which deformation can be such features as the tilting of Cliff Lake. From axis of late Quaternary depression of Madison Valley onstrates that a belt 9 to 19 miles wide flanking the determined northeast of the Hebgen fault is the these diverse data, a deformation-contour map was trends eastward toward the mouth of the canyon. as subsided terrane on the north was elevated, mostly height of fault warps. However, as the new warp prepared (pl. 3l, the contours representing the chang.es is discussed on subsequent pages. The southwest only 0.1 to 0.3 feet, above previous altitudes between of the West Fork fault trends southwest in the foot- in altitude that accompanied the earthquake. The shore of the lower part of Hebgen Lake is a straight. surveys, presumably during the earthquake. Meas- wall of the Hebgen fault, directly toward the segment map is thus a special type of streeturel-contour map steep slope for which a fault origin is probable, sug- ured uplift reached 1.7 feet alone Grayling Creek, of the Hebgen fault where the abrupt downwarp is in which the datum is the irregular prequzike ground gesting that the Madison Range has there tended to about 1.5 miles north of the major scaup near Duck recorded. it is a fair nssnmption, that the downwarp surface rather than a striwiterephic horizon. It may stand while the Beaver Creek and West Yellowstone Creek. Elevation in Madison Valley and in Yellow- and the West Fork fault are parallel, interdependent also be pictured as a mirth:rural-contour limp of an basins have subsided in the past: but no evidence for stone National Park between Obsidian Cliff sad Ham- features. This interpretation is shown on plate 2. 693-455 0-64-9 82 THE ILE BGEN LAKE. MON-TANA, E AR TH Q CAKE OF AUGUST 17, 1559 DEFORMATION A CCOMF/WYLNG THE HFBGEN LASE EARTHQUAKE 83

height along its main southeastward trend. By ana- prove more than 15 feet of subsidence. This subsi- up against the fault zone by a rela- a logy with the known behavior of the subsided Hebgen dence is taken Lake block and with the proved abrupt warping below tively steep down warp. At Corey Spring, the lake- the faults at Corey Spring (discussed later), it is shore approaches to within a third of a mile of the presumed that the irregularities record variations in fault zone in a deeply indented bay. The tilting of amount of warping accompanying the faulting. the shoreline of this bay and the changes in altitude SOUTH A A' NORTH the cr Maximum net displacement across the nets- scarp is of nearby bench marks prove that warping here, recognized faulting, accomplished a change in 20 15' feet, a value found at two localities. These two without AR-1 ;laces are interpreted as points where the total dis- altitude of 9 feet in a horizontal distance of about placement of the evenly subsided Red Canyon fault 650 feet; this is equivalent to a slope of 3/4 of 1° (fig. block was by movement on a single fault. We assume 45). that elsewhere along the fault, where the scarp varies The prequake shoreline of the bay at Corey Spring in net height to as little as 8 feet, the sum of fault was warped smoothly downward, and the least sub- displacement plus warping is from 13 to 15 feet. sidence was at the bead of the bay. The strike of Red Canyon fault coutinues with its irregular the warp is N. 73° W., a direction clearly defined by REMO, Lou The Red Canyon Creek; then, a the intersection of the postqnake water surface and B' southeast trend across mile and a half north of Grayling, it curves south- the deformed prequake shoreline on opposite sides of ward, rapidly dwindles in height, and disappears as a the bay. From this line, at the time of measurement, I- -ontinuous feature but gives way to a curving com- the old shoreline rose smoothly northward to a height plex of many small scarps distributed irregularly over of 5.3 feet above the water level at the head of the a broad zone. Net displacement on each small fault bay, 350 feet distant. Several hundred feet northeast is no more than .2 or 3 feet. (in chapter G, Witkind of the head of the bay, a pair of bench marks on op- retains the name 'Zed Canyon fault scarp" for this posite sides of the highway record the continuation of Grayling the warp. The southward slope of the downwarp is NERCEN LAKE zone and for the new scarp southeast of , • Creek into which it trends.) slightly convex upward. The broad zone of disconnected fissures and small These data prove that most of the snbsidence in the 27 scarps curves from a southward trend along the east vicinity of Corey Spring was accomplished by rela- side of the valley of Red Canyon Creek to an east- tively abrupt warping (profile 0--C' fig. 44). The ward trend along the hillside north of Hebgen Lake. amount of snlisidence accounted for by this warping faulting alone feLLS One scarp swings southwestward instead of southeast- is essentially the same as that due to in some places (profile A—A') or to the combination VERTICAL ElAGGEILATIRY 2 :CO ward as this zone of general curvature is entered. The main zone of small faults, here referred to as the of faulting and warping in others (B—B'). The nartr 44.—ProAle• of 44041denee of horizontal plant during 242L11244.114 or AvrY4t 17. MEL 1-1.1.1 of ,1 42.141 shown on 21414 2. Corey Spring fault zone, continues eastward north of northeast boundary of the subsided basin at Corey the lake and past Corey Spring to another area of Spring is an abrupt downwarp, broken near its upper 1 The new Red Canyon fault scarp, the longest warp hut suggests a low basinward rise as well. The curvature where it swings again to a southeast trend. limit by a zone of minor discontinuous faults. formed during the 1959 earthquake, begins a mile east scarp maxima that suggest this rise are at the appar- The Corey Spring fault zone continues as a broad Discontinuous, low scarps continue eastward from of the point at which the West Fork downwarp di- ent intersection of the Hebgen fault and the axis of the zone of disconnected, snbparallel faults almost to Corey Spring almost to Grayling Creek. where a verges from the Hebgen fault. For the first 2/12 arcuate anticlinal fold which controls the Red Canyon Grayling Creek, where it gives way to a single high, single high scarp appears abruptly. The scarp is 1 miles of ics course the fault trends northeastward fault (pl. 5). The isobases incorporate the implica- impressive new scarp. or 3 feet high 50 yards uphill from the edge of the parallel to the scarp of the West Fork fault, but it tion that the old fold influenced the subsidence. The displacement on many of the small fanits of valley fill, yet 8 feet high at the base of the hill and then curves through 90° and trends southeastward. Subsidence profile E—B' illustrates the effect of the the Corey Spring zone is down on the uphill side 34 feet high a short distance southeast of Grayling New displacement on the Red Canyon fault increases displacements of the Red Canyon and Hebgen faults and is thus directed opposite to the general sense of Creek (pl. 2, and figs. 36 and 37). This abrupt in- to a maximum of 1:? feet in the big curve. This in- and of the West Fork downrearp upon the pattern of displacement. Some of the scarps must represent crease in the scarp height represents the change in creasing displacement records the eastward transfer of deformation. The basin of subsidence is broad and movement, of purely superficial material and be part the boundary of the basin from a downwarp, as at subsidence from the West Fork structures to the larger asymmetric. The deepest point lies more than a mile of the pattern of downhill mass movement of talus Corey Spring, to a single fault scarp. The monocline scarp. south of the Hebgen fault, not against the scarp. The and glacial till, some results of which are described of the warp must trend into and terminate against Followed southeastward from the 18-foot maximum remarkably smooth profile of subsidence south of the by Hamilton in a section of the companion paper by this fault. height at profile A—A', the Hebgen scarp dwindles scarp is controlled by 3.1 instrumentally determined Hadley (chapter K). The fault trends southeastward from Grayling rapidly to feet as described above, then increases points. most of which are measurements of the warp- c. The low and discontinuous character of the scarps Creek, although the single scarp gives way to a series to 9 feet in about half a mile, and diminishes again ing of the prequake shoreline: all are within 1,800 of the Corey Spring fault zone might suggest that of discontinuous scarps within a narrow zone. Aggre- to 5 feet in a like distance. An essentially constant feet of the line of profile and within 54/2 miles south the basin to the south subsided relatively little along gate displacement on these scarps decreases irregu- subsidence of the Lakeshore over this distance indi- of the Hebgen fault. this interval, but this is not so. Displaced bench marks larly. and where the fault curves to the east displace- cates not only ;be presence of the West Fork down- The sinuous Red Canyon fault varies irregularly in along the highway immediately south of the zone ment is only about a foot. These low scarps dwindle 84 THE BERGEN LAKE, IMONTeLNA, EARUELQUAKE OF AUGUST 17, 1959 DEFORMATION ACCOMFANYU,'S TILE ILEBGEN LAKE EARTHQUAKE 83

in turn and give war to a zone of email disconnected The Red Canyon warp is continuous and impressive scarplets with only an inch or two of vertical dis- where only surficial deposits are on the downhill and placement: at the park boundary only a few small downdropped side; but near its west end, where bed- cracks without vertical offset appear. This decrease rock is encountered in both blocks, it disintegrates Moraine in warp height is due to the decrease of absolute sub- into small, irregular, discontinuous warps within sidence southward, as is shown clearly by the relevel- single formation. A -6.5 ft Ofire 45 0 97) ing of the bench marks. Near the east end of the new scarp of the Hebgen fault north-northwest of Lakeview, Cambrian carbon- RELATION TO OLDER STRUCTURES -7.5 ft ates crop out a few yards above a single new warp that STRUCTURES (DM e5 D 96) NORTIECEAST OF EEEGEN LAKE is about 4 feet high, and large outcrops of rhyolite = p The basin of new subsidence ends obliquely and welded tuff lie only 30 yards below the scarp. The Issmungsr Raba abruptly northeast of Hebeen Lake against. the Red linear bedrock contact can reasonably be assumed to PreouiCe shoreline Canyon and Hebgen fault scarps. That these large be a preexisting fault, and the clean new break between -15.5 new scarps and the major warps associated with the two rock types reinforces the assumption. A nor- Postouake shoreline them are related systematically to older structures mal fault here has probably dropped the rhyolite of Aug. 22. 1959 can be seen readily by comparing the geologic map down against the limestone, and was reactivated dur- (pl. 5) with the -earthquake- map jpl. 2). ing the 1959 earthquake. It is significant that here, The broad pattern of structures produced during the only place where an old bedrock fault is clearly the Laramide orogeny, culminated in very early demonstrable, the new scarp is a continuous feature. Alluvial tan which ADOoNsinlete position of of northwest-trending folds, Hebgen fault southeast of Kirk-wood Creek is (------prequake shoreline fertiary time. is one The many of them overturned northeastward. Dips in the represented by a continuous single new scarp on each upright limbs are low, commonly less than 20°. and side of a limestone knnb. As these scarps approach in various directions, whereas the overturned limbs the large but discontinuous limestone outcrops from x-15.75 generally dip steeply to moderately southwestward. both sides. they break up into a number of small BERGEN LAKE The folds are broken by southwest-dipping thrust warps. The bedrock of the ridge is shattered by a faults that are roughly parallel to the fold trends. complex of fissures and small extremely irregular These Larnmide structures were deformed during later faults, dropped down on either side; where these Cenozoic time by normal faulting and associated breaks cut nearly continuous outcrops, there is no FEIT warping. clear evidence for older displacements. The new contacts. and there does RELATION OF MODERN FAULTING TO CENOZOIC breaks represent no lithologic FAULTING not appear to have been a previous topographic warp. As the 1059 scarps northeast of Hebgen Lake are A limestone outcrop west of the new breaks, however. commonly high on hillsides but below the lowest out- dips eastward moderately, but east of the quake scarp- Shoreline Change Shoreline change Bench shirks dips are overturned. The divergence of dips -11.5n -10-2 -7.55 -8.6 crops, with only talus slopes beneath, there is no lets the direct way to prove the interpretation, shown by may indicate an old fault, but the relations suggest plates 2 and 5. that these new scarps are coextensive instead that the new scarps advanced into previously .4 preopace horizontal plane - I I A' with older major normal faults in the bedrock. The unfaulted rock ao this locality. Postouake position. true Scale existence of high ridges above the new warps. how- Despite this general lack of conclusive evidence for ever, is presumptive evidence that much of the relief faulting of exposed bedrock. there is much evidence in is indeed structural. It is instructive to examine the surficial material for previous faulting during some of these same trends. The PcisfOLalci position x 10 ..eittical ecitS9e'ali°P the only four localities at which bedrock was found Quaternary time on to crop out near the scarp on both flanks. northernmost fissures of the Corey Spring fault zone. The sew fault scarp east of Grayling Creek is high for example. lie along the base of a steep bedrock slope. and impressive where it cuts surficial deposits but and follow closely an old lineament developed in both abruptly dwindles to small disconnected and highly bedrock and surficial deposits. Just west of the promi- 100 am loo 7 FEET 4°Di irregular scarplets ns it enters the bedrock Hi formed nent gully north of Corey Spring, the prequake line- of Pliocene( ?) welded tuffs immediately west of ament crossed a colluvial terrace as a narrow straight II/ntran 45.—Mau and °reale allowing 11/11,13filf mouth of Corey apnea fault In I:anther. Carnet. auboldoone, In fee/. det.r...11 outlined by relevellog of bends marks, bY rclationa of old and non sbonsiiaeis. and by aubmervenee of laud Pialna. Slap enlarynt from tenni Blarneystone Ranch. The rhyolite shows no evidence welt: during the 1959 earthquake this was ybotorrapb. Wit- of older faulting, whereas the sot-neinl deposits were by new opposed scarplets. largely mapped by broken in previous late Quaternary time by two gem kind, and thus was made recognizable as a reactivated lineament motions of major warps. This suggests that there is horst. Half a mile to the west, the old allu- no older fault along the course of the new breaks in splits: one arm continues westward toward the curves the bedrock area. vial fan of Red Canyon Creek. and the other so THE HERGEN LAKE, MONTANA, EARTHQUAKE OF AUGUST 17, 14.59. DEFORM-n-10N ACCOMPANYING THE UESGEN LAKE EARTHQUAEE 87 northward along the hillside on the east side of the southeast. alinement of three springs now within the a broad arc which has a curvature of 90°. The Red Madison River canyon. although no such fault is shown valley: this latter trend is emphasized by a new lake but shown on an unpublished map of the prelake Canyon scarp closely follows this arc, and for much on the geologic map (pI. 51—and remains in them to fissure. The an-ring features roughly parallel the basin, marks the continuation of this apparent fault. of its length diverges very little from a stratigraphic its oddly curved terminus one-half mile to the north- beds of the Paleozoic sequence higher up the hill. The small alluvial cone a mile southeast of Kirkwood horizon on which slipping is expectable—the contact west. Despite the apparent complexity of these de- which swing northward into the west flank of the Creek is discordantly below the dry gully at its head between the massive Mission Canyon limestone and tailed relations, the recently active fault is broadly steeply overturned northward-plunging syncline. The and seems to hare slipped down along an upsiope re- the thin-bedded shely Amsden formation. Displace- parallel to the overall attitude of the Paleozoic strata old lineament that continues westward apparently fraction of the fault—much as the fan at Hebgen Dam ment on the fault has been down the dip of the beds and to thrust faults within the sequence. marks a line of earlier movement on the Corey Spring has done on the refracted Hebgen fault during this and clearly has occurred there because of this favor- The deformation map (pl. f1) shows that faulting fault zone. A thousand feet north of the divergence, earthquake. utile attitude for slipping. Above the scarp, the high and warping went on together along the 1959 scarps. new fissures of the two trends cross: this emphasizes The position of the new Hebgen and Red Canyon bedrock ridge, presumably itself a fault warp, is coex- Major scarps of both the Hebgen and the Red Canyon the contrast in the structures that control snitsidence fault scarps at the base of high bedrock ridges that tensive with strata with this general attitude, faults lie updip along panels of hasinward-dipping here. Half a mile north of the divergence, the short stand above long talus slopes is convincing presump- The warp abruptly disintegrates into crude zones planes of easy gliding formed by the overturned limbs irregular fissures of the northward trend give way to tive evidence that these ridges are in part due to of small warps and fissures at both ends where the of folds. The height of new warps tends to vary the new scarp of the Red Canyon fault, which rises previous normal faulting. Topographic evidence for overturned Paleozoic beds roll to gentle upright atti- systematically with the dip of the strata. Deforma- gradually northward as a single scarp to attain an such faulting would be ephemeral, as the warps tudes and the ridge loses its commanding height. tion has occurred by faulting where slippage was easy impressive height. formed would greatly oversteepeu unconsolidated talus Where the beds are oriented unfavorably for slippage and by warping where it was more difficult. Pre- The num scarp east of Grayling Creek is on a fault which lay already at its angle of repose. and debris and must be crosscut by any normal fault developed, sumably this was true in the past also. that has had at least two previous episodes of late shed from cliffs above would hasten the burial or them has been little or no precious modern faulting. Of the three major fault scarps developed during Quaternary displacement (figs. 36 and 37). Old and destructioneof the warp. The talus would also bury The Red Canyon fault is now being extended into the 1959 earthquake, the two that are in areas of ex- new structures die out southeastward via a monociine any outcrops on the downhill side. The bedrock those areas. posed bedrock are rigorously controlled by structures— of decreasing height: but in the surficial deposits 3 ridges rising above the impressive new fault warps The situation along the Hebgen fault warp is simi- bedding planes—which can hardly extend to a depth miles east of the end of the structure. another old must be in substantial part products of previous off- lar. The trend of the fault is straight so chat no of more than I or 2 miles. The surface fault pat- fault scarp rises gradually in the sand plain to be- sets along the same faults. concentricity of fault and strata can be demonstrated, tern must accordingly differ from the pattern of come a prominent structure cutting the east margin It is conversely evident that where the new scarps but fault and beds are essentially parallel in strike deeper, and more fundamental, displacements. of the West Yellowstone basin. Whether trend into bedrock areas not marked by prequake length of the fault. Southeast of or not this for most of the STRUCTURES SOUTH OF ECEROEN LAKE old scarp was reactivated in 1955 was not determined. scarps, as in cases described above, there had not been Hebgen Lake Lodge, 1959 displacement in the fault The bedrock fault between Cambrian limestone and appreciable late Quaternary movement on faults in zone is represented by small discontinuous scarps and The continuity of the many faults and monoclines, Pliocenef!) rhyolite. reactivated in part as described that bedrock. As fault scarps are widely preserved in fissures that extend for 11/2 miles in rocks that have a both of 1959 and older, that displace the surfaces of above, separated varied pre-Tertiary rocks from rhyo- unconsolidated mateeial (other than talus) of varied gentle to moderate northeast dip. The beginning of a the sand plain, moraines, and other deposits of un- due to lite over a considerable distance where it was not re- late Quaternary ages in the region, it is reasonable to single continuous warp is coincident with the abrupt consolidated material, indicates that they are although activated in 1959. East of the newly active part. this expect that bedrock warps of similar age should be overturning,of the beds to a steep southeast dip just displacements of the underlying bedrock, basin. fault has probably offset the base of the Paleozoic sec- still better preserved: their absence is clear evidence east of the lodge. Farther northwestward, the pre- the bedrock is not exposed within most of the 1959 tion by about 400 feet. and all or much of this may of the lack of late Quaternary faulting of those out- vailing dip of the beds exposed above the scarp rolls No detailed analysis of the relation between the struc- have been accomplished since the deposition of the crops. back to an eastward position, and the scarp is parallel faults and monoclines and the older bedrock Yellowstone basin is rhyolite,s. No Quaternary surficial units are recog- The new fault warps north of Hebgen Lake thus to the beds in strike but cuts directly across them in tures can be made. The West south sides, nizable offset along the fault trece, in contrast to the seem to be partly on faults which had previous, and dip. blocked out by faults on both north and relief between abundant scarps in unconsolidated materials to the perhaps major, offset during late Quaternary time. Throughout a length of more than 4 miles, the although to a large extent the structural warping south and east. At Kirkwood Creek, however. the and partly on extensions of those faults into areas Hebgen fault scarp is but a few hundred feet west of basin floor and adjacent highlands is due to have been repeatedly surface of the Pinedale fan is furrowed by a prominent which had little or no such offset. The old faults are a west-dipping thrust fault. Beneath this thrust the rather than faulting. The faults brought out old drainage course, now dry, that heads just below increasing in both total offset and lateral extent. beds are overturned almost everywhere and are nearly active during late Quaternary time. As is those faults and the new Hebgen fault scarplets and appears to mark parallel to the thrust: farther northwest this thrust in a later section, we believe that CONTROL OF MooBRN FAULTTNO BY L.RRMITOS most of which the site of a group of springs that were created by RTRUCT4RE becomes nearly a bedding-plane feature. It seems that monoclines which have easterly trends, slight Madison Arm, have a late origin and prehistoric movement on the fault. Older Quaternary displacements of the faults north at shallow depth, the Hebgen fault which at the sur- are south of the The toe features. of the Kirkwood fan is cat by a proud. of Hebeen Lake seem to have taken place in the zones face is steeper than the thrust, is either coextensive are not Laramide neat the southwestern part of the transverse warp that is interpreted (pl. 5) as of easiest slippage, ns determined by the attitudes of with the Laramide thrust fault or else lies closely Three fault zones in an converge toward Targhee erosionally modified fault warp of interglacial age the pre-Tertiary rocks. whereas what appear to be new below in the nearly parallel beds of the footwall. West Yellowstone basin which Pass, the low route across the Madison Range west of parallels the Hebeen fault. sunlit 350 yerds to extensions of the faults tend to be in zones where North of Hilgard Lodge the scarp ruts obliquely across the One zone trends northwestward east. Pinedale channel of Kirkwood Creek attitudes are less favorable for slippage. This is the overturned sedimentary contact of the Cambrian West Yellowstone. was deflected pass along the topographic margin of the along the base of the scarp. at right readily seen upon comparison of plates 2 and 5. beds on the Precambrian metasedimentary rocks and toward the angles to its curve westward and southward. upstream course. by a slump graben or The new Red Canyon fault scarp is parallel to the remains within the latter as far as Cabin Creek, 11/s basin: the other two moat. The three zones arc topographic offset across the scarp ranges bedding of the limestones of the high-standing block. miles farther north. At the creek the warp enters respectively, toward the pass. All from 80 to 100 converge. No feet and appears to reflect fault dis- The strata dip moderately to steeply lakeward in the upper Paleozoic limestones—presumably by crossing together: they do not intersect, they placement of a surface of material of Bull Lake age. To the strongly overturned limb of a fold and strike through n normal fault that bounds the north wall of the warps were recognized to cut the surficial 88 THE BERGEN LAKE, MONTANA. EA RTP1Q VASE OF AUGUST IT, MS DEFORMATION A CCOM.PANYriG THE HEDEEN 1 •NF EARTHQUAKE S9 the pass. and the bedrock there was not mapped; but Madison Arm and suggests downwarping of at welded tuff, whose axis strikes about S. BO° E. and In the southwest corner of the basin, progressive this convergence least suggests that major structures cross 80 feet. lies about 5 miles north of Madison River from the eastward shifting of meander scars of the South Fork the mountains through the pass. Two of these three If the bedrock floor of Park boundary to Madison Junction. The youth of of the Madison Rirer is apparently due to basinal fault the begin once drained to the zones in the basin have displacements that in- Hebgen damsite. a greater downwarping is indicated. I this syncline is indicated by the drainage reversals sagging (fig. 35). crease the relief of the basin. but the third has the The obsidian sand has been penetrated only in two f. that occur along its northern Mace line. Straight The Grayling and Red Canyon fans are near the opposite sense of offset: along the zone trending oil-test wells, 2 miles northwest of West Yellowstone; Creek, 17 miles east-northeast of West Yellowstone axis of subsidence of the recent earthquake (pl. 2). southward along the west side of the basin, the west, there the sand is about 200 feet thick, and its contact and just north of the area shown on plate 5, is an The surface of these features is considered by Rich- or mountain-facing, side is the r downdropped one. with rhyolite bedrock is essentially at the altitude underfit stream that flows northward to the Yellow- mond (p1. 5) to be Pinedale (latest Pleistocene) in Viewed from the framework of the "Madison Range, of the bedrock rim of the basin at Hebgen Dam. A sione River. It heads just north of an imperceptible age, but their size suggests a considerably greater age however, all three fault zones have the same sense of water well on the south shore of Madison Arm,6 miles drainage divide in a conspicuous old channel that is for their cores. The lack of dissection of the fans displacement. as though the Madison Range north of closer to the dam. did not reach bedrock at a depth continuous southward to the Gibbon Geyser Basin. implies that the axis of sagging has been in about the Targhee Pass has been downfaulted along varying of 240 feet and thus reached sand about 130 feet lower At the northwest edge of this basin the old channel, same position—that is, near the point of entry of these combinations of these three fault zones. in altitude than the bedrock rim at the dam. The bed- little changed at this point. turns east, and although streams into the basin—since the inception of the EARLIER WARPING rock floor of the basin has a minimum closure of modified strongly by later erosion and deposition, it fens. OF THE WEST YELLOWSTONE 135 i 21A.S/N feet. Down warping, with a low between West Yel- appears to continue eastward beneath the Recent At some time before the late Quaternary, the basin lowstone and Horse Butte and a closure of several rhyalite dome of Gibbon Hill. east of Gibbon Meadows probably emptied southward into the upper Snake The distribution of surficial deposits is evidence that hundred feet, is indicated. The axis of 1959 sub- and about 6 miles from its present divide. The south- River Plain. The basin is now rimmed to the south the West Yellowstone basin has been warped during sidence was about 5 miles north of the apparent low ern limit of the essentially unmodified pan of the by huge flows, each hundreds of feet thick, of late late Quaternary 1.• time, before the :959 earthquake. indicated by the well logs. channel is 3 miles from the modern divide at an alti- Quaternary rhyolite: the flow immediately south of Major changes in the drainage pattern of the sur- tude nearly 150 feet lower than the windgsp. Assum- West Yellowstone, is younger than Bull Lake moraines. rounding CELLICCES Iti DR-LINAGE area are further evidence of long-continued ing that the now reversed drainage had the same The Madison Range dwindles southward, and at Rees sagging. Long-continued sagging of the West Yellowstone northward gradient as does the present upper part Pass (altitude 6.932 ft.) the crest is less than 400 feet basin is responsible for some major changes in the SITILFICIAL DEP0917-3 of Straight Creek, a backtilting of about 275 feet is above Hebgen Lake. Before the eruptions the basin drainage pattern of the surrounding area. At least Hebgen Dam is at the retarded in this 3-mile stretch of channel. Three probably drained southward around the end of the entrance to the narrow Madi- part of the area of the present basin probably once son River canyon, where the miles to the east, the ripper part of the former channel Madison Range. The ancestral north fork of the river profile steepens drained northeastward to the Gallatin River. The abruptly (fig. 46). A detailed of Obsidian Creek. which parallels that of Straight Snake River probably headed where Beaver Creek geologic section made deep canyon of the Gallatin River along the State of the dam corewall excavation Creek, is similarly backtilted. The axis of drainage now flows southward toward Hebgen Dam. Low- by Montana Power boundary is continuous with that of Grayling Creek, Company engineers shows that the reversals trends eastward. The syncline defined by ering of the site of the present Madison River canyon transverse bedrock and so inconspicuous is the low divide separating the surface is nearly horizontal and is these reversals and by the deformation of the welded is thought to have taken place by downdropping of lower than the oppositely flowing streams as to give the illusion that modern gravel-filled channel tuff was deepened as much as 0.7 feet between level the block south of a fault postulated in the canyon, across the entire length Grayling Creek even now is simply the upper Gallatin. of the dam. At the northeast surveys, presumably during the earthquake. with consequent reversal of drainage. end of the dam. MU, The upper part of present Grayling Creek flows north- ficial material is 115 feet thick. and Within the West Yellowstone basin, continued sag- bedrock lies 25 feet westward, an orientation which suggests that this part A. below the modern channel. The ging has clearly controlled stream processes. South RELATION TO REGIONAL STRUCTURE unconsolidated mate- formed as a tributary to the north-Rowing Gallatin. rial. as interpreted by Richmond of the present axis of subsidence and west of U.S. .A. comparison of the overall pattern of deforma- (chapter T), includes. Numerous other drainage details, among them the f• from the base upwards, Highway 191. the sand plain on the south side of the Lake obsidian sand; a persistent barbed tributaries of the now southward flowing part tion and seismicity accompanying the Hebgen layer of red clay inferred to be a Madison River is marked by the traces of old meander deformation in the fossil soil developed of Grayling Creek, reinforce this suggestion. Drain- earthquake with evidence of recent on pre-Bell Lake deposits; more scars over a belt locally more than a mile wide. These that the deformation of obsidian sand: and, age of the West Yellowstone basin toward the Gal- surrounding region suggests intertonguing with and overlying scars show a consistent northward shift of the former of a complex of sagging that extends east- the upper sand, a latin seems to have been interrupted by backtilting. 1959 is parr steep alluvial fan of Bull Lake channels. The north bank of the river is almost Lake are. Growth of the Along the northeast edge of the basin, the present ward from the Centennial Valley and Henrys fan narrowed the canyon everywhere a single prominent escarpment cut by the, structures and deflected the river drainage pattern suggests much modification of the depression across the older north-trending against the modern prereservoir) channel of the river. North- and that southwestern bedrock wall, but there has earlier north-draining streams by headward erosion of of the Madison Range and Madison Valley been no ward tilting during the channel cutting is indicated. Plain to the south. lowering of the bedrock lip of the West Yel- tributaries from the sinking basin. Farther within may he related to the Snake River lowstone East of the highway the headward trend of the river basin during late Quaternary time. Sources the basin the drainage is entirely centripetal; a smooth for the bowl is more to the south, and the stream flows down the MADISON VALLEY obsidian sand are all to the southeast. and -shaped basin is cut radially by drains which suggests that slope of the basinal sag; the meander belt narrows Madison structural drainage of the West Yellowstone basin through the the streams are consequent on a surface The Valley is a north-trending Madison shaped by basinal sagging. upstream, and no shift of the stream toward the right trough that lies near the crest of en older Laramide River canyon is thus established for a con- This relation is thought to hold for bank is evident. The gradient of the stream in this by the siderable part of the late Quaternary. the slopes south and east from Grayling uplift. The valley is bounded on the east side Creek. past the downslope section is anomalously high, as would be impressive but much eroded fault scarp of the Madi- The base of the Ball Lake till beneath the entrenched Madison River, and south- lake basin westward expected (fig. 46). end has been located to the east margin of the great lobate flow of son Range for most of its length. At the south uncertainly in well logs. Even the Tilting is also suggested by the apparent post-Bull Lake rhyolite south of the town northward of the Madison Range, in the Snake River drainage most conservative interpretation indicates of West a reversed Yellowstone (pl. 5). migration of Cougar Creek and its abandoned chan- south of the Madison Valley, faulting is of miner slope of more than 50 feet for the base of these de- nel to the south. The present southern (oral com- posits The extension of this downwarp eastward into the distributary importance. According to J. B. Hadley between the dam and the south shore of the Yellowstone Plateau is shown by a broad syncline, in of the creek probably uses part of the old course. munication. 1960), the west side of the Madison Val- 90 THE FIE/IGEN LAKE, MONTANA, EARTHQUAKE OP AUGUST 17, 1559 DEFORMATION ACCOMPANYING THE 13RBOEN LAKE EARTHQUAKE 91 ley is bounded against the Gravelly Range by a series deposits dates Henrys Lake. These young faults trend i - PTV ....9Y3 ..\ of faults and warps. and the valley is largely younger northeastward toward the low transverse valley of :...-'' uonsunr :rammer—i\ than the volcanic rocks of Oligocene age and may be Targhee Pass, but do not enter it; at the front of the of very late Cenozoic age. Madison Range the structures turn abruptly to form A. study of aerial photographs of former courses low discontinuous scarps that trend northwestward. ONTIM0.3,15 of the Madison River southwest of the mouth of the The latest episode of faulting has thus been limited Madison River canyon indicates en impressive amount to the depression of a block at the intersection of the AsivGNiloe 11155 of geologically recent tilting in the part of the valley Centennial Valley and Madison Valley subsidence affected by the 1950 deformation (figs. 47, 48). The trends. As other young scarps. downfaulted on their ulilq 4 younger flood plains are cut in unconsolidated fan north sides during late Quaternary time, trend toward 1.03..0101,1, I t gravel that is at least in large part of late Pleistocene Targhee Pass from the West Yellowstone basin on the isam Avec N39e2H— age; the older flood plains cut rhyolites which are east, it seems likely that the underlying faults are con, >Os.° ,txina assumed to be correlative with those dated as of Oligo- tinuous across Targbee Pass, and that the low divide cene age in the southern part of the Madison Range. is a structural depression much modified by erosion. Ouse ...speed to esess tr u"= ti The oldest flood plains are far southwest of the present river and enter the trough of Cliff Lake, part of a CENTENNLAL MOUNTAINS AND CENTENNIAL VALLEY srxe an., 6i61-a4:11 north-trending canyon whose present stream now The Centennial Mountains are a compound tilted ., 01! P.AIL\ joins the Madison. River a few miles northwest of the fault block that dips smoothly southward beneath the area in the photographs. Maximum structural relief lavas and tuffs of the Snake River Plain, and thus on these flood plains is about 500 feet. The several they form the structural border of the plain; the north younger plains lie successively closer to the modern slope, facing Centennial Valley, is a series of arcuate river. Tilting in the younger plains is difficult to fault scarps that rise more than 3,000 feet above the define precisely as they were cut successively deeper valley. Fault scarps young enough to offset glacial in the fill and as they start and end near the present deposits, alluvium, and lake sediments of latest Pleis- river; but the southernmost and oldest has been tilted tocene and Recent ages lie in many places in the val- northward about 100 feet, and an intermediate one ley and along the mountain front and are as high as has been tilted about 60 feet. • 50 feet (fig. 49). Viewed broadly, the valley is a long 500155 NOSA7VAI-- Until late Pleistocene time, the Madison River thus east-trending sag. bounded on the south by the faulted flowed southwest of its present course to Cliff Lake face of the Centennial Mountains. The eastern faults 0.11.03 p Sea Jaaa and thence northward along the course that is now a are en echelon, stepping to the left. The valley PIVO Nos!orw-- deeply entrenched canyon. Tilting toward the east- formed by collapse of an area once continuons between northeast diverted the river closer to its present the Centennial Mountains and the ranges to the north. course. but Inter the direction of tilting changed to the Madison fronts are about 3ao.13 moods Both the Centennial and north-northeast. The modern river flows slightly 3,000 feet high, but the Madison scarp has been very north of west for nearly 5 miles after leaving the much more eroded than has the Centennial scarp. The Madison River canyon. and in this distance it flows Madison Range crest lies 1.5 to 5 miles from the fault along a structural axis of late Quaternary subsidence. trace at the foot of the scarp, and permanent streams have cut deeply into the block; Sheep Creek, for ex- HENRYS LA.KE BASIN ample, falls only 400 feet in the mile above the fault The Madison Valley is separated from the drainage trace. The crest of the arcuate Centennial fault block of the Snake River by low Reynolds Pass. Just south- south of Upper Red Rock Lake, by contrast, is only east of the puss is a broad valley, clearly the southern 0.8 mile to 1.5 miles horizontally distant from the fault extension of the Madison Valley structural depression, trace 3,000 feet below, and even the stream most deeply which trends south-southeastward for about 20 miles block falls 1.200 feet in its lower and merges with the northeast corner of the Snake entrenched into the faulting along the Centennial and River depression. Henrys Lake lies in a relatively mile. Although Centennial depressed part of this valley directly east of the east- Madison fronts has overlapped in time, the trending structural basin of Centennial Valley, which structures are, viewed broadly, much younger than is described neat (fig. 49). the Madison ones. Perhaps the Centennial structures The Centennial Mountains jut eastward into the formed chiefly in Quaternary time and the Madison valley south of Henrys Lake and are bounded on the ones largely in the Pliocene. north by faults which lose structnral relief eastward. Prominent shorelines clearly recognizable on the 1131 Nl iarouiv One of the prominent north-facing scarps in surficial ground and in aerial photographs show that a large 92 THE BERGEN LAKE, MONTANA. EARTHQUAKE OF AUGUST 17, 1929 DEFORMATION ACCOMPANTMG Lux. JUERGEN 7 AFT EARTHQUAKE 93

Rtght photograph belt photograph

0.• s N ces ,/

tor BevrIce. ' 0

org.4.:1 ‘14 alb Hap

P!nrioi - Army Recent tombstone) SCIt. 20..42 ft high ll.S.

b, 6Pa 0 MISSOURI toa tzik' •-• FLATS

Photorrsph. 6ztro lanais of IMadison Rrnsr - ...di. ‘441:.*---7- 4.22V *0400 7, te'diV.-..-,:. .T...... 4,1 .i V .-.;4-1!=!0%tie , Z,".. 00 \ '..."•r"..Z.1-r..0 4, .--.* 4.-`-- -t.,„, . ''''.. -::a ijoto`,..„ ..•-•- -,-'''''. ‘$ 4, ,„-- 4•,,,.„:-._1'7.77> -f,•11.7..,-,p4!,.;..--._.,-- - ,- ,•1 7 ,,-i , skt.", -,-.4.. ::.„..1:-P...7"-:„._7 ' ! Neetya Camp ,,...441 „.. X13'-. '. ...att'N.4,14A '--11'-', ..iT;:_--;- z6o /

.otaV..!MVS1.. J VOeNr'gc..tizp.' re_o ° 2166 • . 1**, • Tgatng-tea''s LIZ" APPligtXMAft SULLt 1 2 3 3 MILLS

fastness in Photuarsib. of agora 7101M.S 46.—Mep of rdhaouti Flats area, south end of Madison 7.1Ie7. ebovnig nodal within 2 miles of Cli2 Lake have been 47. L1thre (dashed where uncertain, stark former nurses of iiatlison IUrer. Those show 'Sat the rice, formerly flowed through Lake Creek muck modlied by SoOsequent warping. fasEtiug. and erosion but relief of 400 or 500 feet. Younger caaaTon. These old flood plains have been tilted to the salt.northeam and have a structural •ILkaresset.,.., nett-northeastward with a maximum arrnetaral colleen Cr. noised to the pressor valley of the Nathan and have been tilted from 1:24.01)0 yeozomatamealc mantwarlot maps toilet of about 100 feet_ The altitudes (marked by L'at were drteradord produced by deformation of Initially con [Joao's. contour intervals; underlined altitudes mark drainage divides. " with 20-toot t Ve.:1P.t. et.6.6.1a.

Record time At the southwest corner of present Upper Red Rock _;, -. .4., •,," lake occupied Centennial Valley during ....: ' I ' shoreline has a minimum altitude of 6,630 ,..p...... ew-...... -•Tti„. f. .4„ (fig. 49). Just south of Upper Red Rock Lake the Lake, the east; it cones de- feet. but it rises to 6.690 feet in 3 miles to the -.,...., ,• `:...,_-.,,,—,E, . shoreline is cut into alluvial fans and talus shoreline descends 20 feet ..r.,L-73"..,&- ' -- _...:V.,..,' ! 1 rived from the youngest Pinedale moraines at the northward from here, the .:...":::' r-W .'"-,,;:s.---'-:- '. E front. This indicates and then disappears where it has been cut by the flood .,„.. ,t r .1. foot of the imposing mountain I . ...., .2,...1.5,.„;.. Creek. North of the stream t that the lake was full in post-Pleistocene time, per- plain of modern Red Rock Rock Lake. the shoreline reappears haps during the postaltitherrnal stage of minor glacia- llat and Upper Red feet. Farther tion known elsewhere in the region and dated at about and has an even altitude near 6.650 is lacking: but 1800 B.C. Despite its extreme youth, this lakeshore west, close control of shoreline altitude floor is oCset about has been waived. at least 60 feet and broken by faults. north of Lima Reservoir the lake irregular fault scarps. Faulting The closed basin filled by the lake has been nearly 20 feet along highly occurred simul- emptied—only the shallow Upper and Lower Red Rock and much of the warping presumably and probably Lakes remain—by this deformation and. to a lesser taneously during a major earthquake 44. extent, by erosion of en outlet channel to the west. emptied the lake. •

94 THE HEBGEN LAKE, MONTANA, EARTHQUAKE OF AUGUST 17, 1039 DEFORMATION ACCOMPANYING THE FEEBGEN tackr EARTHQUAKE 95 Bull Lake moraines are well developed in the eastern Valleys truncate a fault basin too old to have much part of Centennial Valley and indicate glaciation more inffnence on present topography: a synclinal fault extensive than that of the subsequent Pmedale stage trough trends northeastward in the central part of whose moraines are nested inside them. In the cen- the Centennial Mountains, disappears beneath Cen- tral part of the valley, however, no Bull Lake moraines tennial Valley, reappears in the Gravelly Range. and are exposed: although such moraines must have is again cut off at an angle of 45° by Madison Val- formed in the vicinity of Upper and Lower Red Rock ley. This old trough cuts sharply across Laramide 4 Lakes about a piedmont glacier more extensive than structures and drops down volcanic rocks of probable the small Pinedale glaciers. the old moraines have Oligocene age. The deformation of the southern been buried by lake and other sediments. The exposed part of Madison Valley by structures trending east- alluvial fans along the mountain front are small and ward from the Centennial area was described previ- extend only a short distance to the edge of the lake ously. sediments; the minimum-altitude valley flats begin In the Madison-Centennial region, Laramide struc- I as close as 0.1 tulle from the trace of the mountain- tures that trend broadly northward thus were trun- front fault. These features indicate that subsidence cated in Miocene(?) time by a northeast-trending here has been very rapid. graben. The graben was deformed by north-northwest Until late Pleistocene or Recent time. Centennial trending Madison Valley (largely during the Plio- Valley drained northward from its east end, through cene?). and both the old graben and Madison Valley the canyon occupied by Cliff Lake. to the Madison have been deformed (mostly during the Quaternary?) River. The once continuous canyon has been seg- by east-trending Centennial structures. mented by large landslides, perhaps quake triggered. SEISMIC EVIDENCE OF MOVEMENT behind which stand the various lakes (Mansfield, PATTERN 1911) ; the rock floor beneath the slides probably A plotting of recorded earthquake epicenters in the os slopes continuously northward from Elk Lake, which northern Rocky Mountain area (Ross and Nelson, is less than a mile north of Centennial Valley. al- chapter E, fig. 14) emphasizes a major seismic zone though Elk Lake now drains into Centennial Valley that trends northward through Yellowstone National because of a large landslide dam at its north end. Park. A striking detail of the pattern is the lack of The lake surface is only 25 or 30 feet above the nearest recorded activity in the upper Snake River Plain. a part of the Recent Centennial Valley shoreline: this major structural feature that was active in late Ter- height may represent either building up of the chan- tiary and Quaternary time. nel by sedimentary or slide processes or structural North of the Snake River Plain the plotted epi- tilting. The Centennial Valley lake recorded by the center of a major earthquake on November 23, 1947, shoreline may have drained northward through the lies about 40 miles west of the north-trending seismic canyon, or this drainage course may have been aban- zone throngh Yellowstone and 10 miles north of the doned somewhat earlier in favor of the westward one. axis of Centennial Valley. This epicenter (44°47' N., 112'92' IV.), at the drainage divide between the Cen- SUPERPOSITION OF STRUCTURES IN THE MADISON- tennial basin and the region to the north, may record CENTENNIAL REGION activity along the east-trending Centennial subsidence Although structures of different types are super- zone. No other recorded epicenters plotted by Ross imposed in parallel array in many regions, successive and Nelson seem to be related to this trend. structures in the Madison-Centennial region have cut The epicenters of the Hebgen earthquake and the markedly across the older structures. This is in large aftershocks that recurred over a period of 7 weeks part obvious even on the geologic map of Montana. (Murphy and Brame, chapter C. fig. 11) constitute Madison Valley lies along the core of Precambrian an east-trending array that is twice as long as it is crystalline rocks of a broad Larnmide uplift: although wide. Omitting the uncomputed ones. the computed oblique to the Laramide structures nt only a low angle. epicenters form an east-trending group that is four L the modern volley is where the Laramide range crest times as long ns it is wide and which lies generally 3 t once lay. The Centennial Mountains and Valley cut north of the West Yellowstone basin. just as the epi- sharply across a northeast-trending Lnramide basin center of the 1947 quake is north of the Centennial \\\ • )! in which are preserved high Upper Cretaceous rocks basin. • 5 in the west and the Precambrian core of a Laramide A different pattern of epicenters was determined uplift—the same one that is dropped down by Madi- for a series of aftershocks recorded within the space son Valley—in the east. Both Madison and Centennial of a few days by Stewart, Hoffman. and Dimenr a

96 TELE PEERGEN LAME, MONTANA, EARTEQUAME 01' AUGDST 17, 1959 DEFORMATION ACCOMPANYING TIIE 'PLEDGER LARD EARTHODA,KE 97 (chapter D, fig. 12). The epicenters fall in a broad largely younger than the north-northwest striking a The known deformation accompanying the earth- tionship between that volcanism and the concurrent arc whose oblique inflection lies a few miles northeast Madison structures. quake consisted largely of downward movement. From deformation of the ad)acent Madison-Centennial re- of West Yellowstone. The northern arm of the arc The broad aspects of deformation during the earth- this we conclude that the immediate cause of the gion. The geometric ties between the volcanic prov- trends about N. 80° W. across the zone of main north- quake. considered with the pattern of Quaternary de- earthquake was collapse in response to gravity, re- ince and the flanking highlands are so numerous that west-trending reactivated fault scarps: the other arm formation of the region, suggest that the Centennial gardless of the ultimate tectonic, magmatic, or other the stress systems controlling them must be closely re- trends southward along the Wyoming State line. The structural system is now being extended eastward I reasons for that collapse. The surface effects can be lated. The northeastern part of the Snake River lack of agreement between this pattern and that found across the Madison structural saistem and into the viewed as due to slumps of various scales and at Plain—which includes the Yellowstone Plateau as its ' by Murphy and Brazes may be due to the shortness of Yellowstone Plateau. Although the large new scarps different depths, variously interacting and interfer- high lava-filled end—is rimmed almost continuously by the Geological Survey record. northeast of Hebgen Lake utilize older Madison ing, related partly to near-surface directions of easy high mountain blocks that dip inward beneath it. The A first-motion study, based on P waves as recorded structures oriented northwestward, the subsidence slipping and partly to the motion of deeper material. depression trends northeastward at a high angle across at it stations, was made for the Hebgen Lake earth- breaks across such individual structures to form a corn. The tight control of the positions of two of the both Laramide structures and most of the Tertiary quake by Ryall (1960, who used the method of posits basin with an eastward trend. It seems that major 1959 faults by the orientation of strata in old block-fault structures, including the Madison Range, Byerly. Ryars calculations iudicate the initial slip- the structural pattern now in early stages of develop- structures of shallow extent shows that these faults vet is not itself deformed by structures of such trends ping to have been on a fault plane striking N. SO° aa... ment in the northwestern part of the Yellowstone are in effect giant slump scarps. The utilization of (Hamilton, 1960). The Centennial Mountains are 10° W., dipping 54° 8° S.W., and with dip-slip region will result ultimately in the formation of any specific preexisting structure is therefore inci• s-ubparallei to the depression at its edge, and on the movement. This calculated strike lies far to the west major east-trending structures, similar to but east of dental and dependent ou a more fundamental control opposite side of the depression the Teton Range of the dominant northwestward trend of the major the present Centennial Mountains and Centennial Val• of the zone of subsidence. As these near-surface faults emerges to complete a pattern of remarkable dynamic faults reactivated to the surface and is consistent with ley, and bathe progressive obscuration of the Madison bound a very large basin of new subsidence, we regard symmetry. Jackson Hole, the structural counterpart the interpretations made in this paper. Valley and Madison Range. Reactivation of preexist- them as passive products of the general collapse rather of Centennial Valley, has also undergone extraordin- ing structures with other trends is to be expected in than as direct products of deep tectonic stresses. The ary tectonic activity during the Quaternary and also GEODETIC EVIDENCE OF ACTIVE ZONES this process. abruptness of the warps and scarps that bound the cnts sharply across older structures. The intervening The level line run by the Coast and Geodetic Sur- The subsidence during the earthquake of 1959 in- 1059 basin of subsidence on the north and the termina- Snake River-Yellowstone province has been the site vey from West Yellowstone to Bozeman in 1934 and creased the relief of structures clearly belonging to tion northward of these northern structures along a of intense basaltic and rhyolitic volcanism throughout rerun after the 1959 earthquake is but a part of a the Madison Range system. Both Madison Valley vague line trending near N. 80° -W., in contrast to Pliocene, Pleistocene, and Recent time. The regional more extensive network of levels. Third-order level and the West Yellowstone basin were depressed rela- the broad gentle south flank of the basin, lead us to pattern suggests that southwestern Montana is now lines were run by the Topographic Division of the tive to the Madison Range, but the intervening part infer that the deep structure controlling the surficial drifting relatively northwestward: that the Snake Geological Survey in 1943 between the older first- and of the range subsided also, at least near Hebgen Lake subsidence is either a normal fault striking about N. River-Yellowstone depression is a rift zone with a ten- second-order levels. There are systematic large clos- and in the Madison River canyon. The long axis of 80° W. and dipping southward or an abrupt mono- sionalle thinned crust and an abnormally high tempera- ing errors in these third-order lines across the Cen- the compound basin of new subsidence is thus defined, cline with the same orientation. ture gradient; and that the Centennial family of tennial basin which strongly suggest that the basin independent of the incompletely known behavior of The earthquake might have been due to faalt move- structures (and hence the earthquake of 1959) is a subsided about half a foot between 1934 and 1948, the Madison Range. by the proved subsidence of Madi- ment regardless of whether such movement was funda- product of this horizontal extension. perhaps mostly during the earthquake of November 23, son Valley, the Madison River canyon, and West Yel- mental to the subsidence or not; but a minority of 1945. lowstone basin, and the Gibbon River valley to have American geophysicists, and a larger minority in REFERENCES CITED By contrast, Coast and Geodetic Survey line 48. in a general easterly trend across the Madison structures. other countries, no longer support the common as- Dania, T. N., 1960, A Bela report on the Alaska earthquakes Madison Valley north of the area that subsided dur- The general high regional altitude suggests that sumption that simple fault motion causes all major Of April 'T, 1958: Seismal. Sow America Bull., v. 50, p. 489-510. ing the 1959 earthquake, remained almost stable be- the Madison-Centennial region is being broadly ele- earthquakes. and it does appear that the kinetic vated. DeCeer. Baron Gerald, ISM, Quaternary cluiniee. Of level Ia tween 1934 and 1959. Upon this presumed regional elevation is subsiding Hebgen Lake energy of the masses about Scandinavia: Geol. Soc. America Bull., v. 3, p. 65-68. superimposed the more local subsidence to which are is adequate to explain the earthquake in other terms. Fuller, Myron L. 1912, The New Madrid earthquake,: U.S. CONCLUSIONS due the downbowed and downfaulted basins. The The 1919 collapse has the shape of a broad but • Geol. Survey Bull. 494, 119 p. The character of the late Cenozoic volcanism 1953 basin of new subsidence has an overall structures—normal markedly nsymmetric basin warped gently downward Hamilton. Warren, 1960. late Cenozoic tectonics end trend of about faults and gentle warps—suggests regional extension . of the 'Yellowstone region, Wyoming, Montana., and Idaho, N. 80° W. and is superimposed at a abrupt structures on the south flank and bounded by Field Conf.: high angle across the major structures of the Madison that permits the slumping of keystone blocks. Prob- in Billings Geol. Soc. Guidebook 11th Ann. on the north. We presume that the fundamental p. 92-105. Range, which trend north or northwest. ably the region is being slowly elevated, perhaps with- A similar deep structure is similarly one sided. The collapse Hutchinson, G. E., 11157, A treatise on Brandon, New York, crossing of old out earthquakes, at the same time that it is being trends by new is shown by the pattern could have been caused by withdrawal of either Wiley, v. 1, 1015 p. of the intense late stretched: the great earthquakes may accompany the earthquake in central Quaternary deformation of the horizontal or vertical support from beneath the basin; Rota. B., 1693, On the cause of the great region. settling of local blocks rather than the uplift of moun- Japan, Isol: Imperial linty. Japan. Jour College Science. Madison Valley south of the mouth of the regardless, presumably slow motion over a long period tains. The warping and faulting of the Recent lake v. 5, pt. 0. p. Madison River canyon is being deformed by stric- basin in Centennial Valley. the formation of the high produced stresses which were released suddenly dur- Mansfield. G. R., 5911. The origin of Cliff Lake, Montana: tures that strike east from the Gravelly Ranee. Cen- Recent scarps along the fronts of the Tendcrr and ing the earthquake. Geog_ Soc. Philadelphia BUIL. a. 9, p. 10-19. tennial Valley, and Centennial Deformation ee- Mountains; and other Madison Ranges, and the landslides damming Cliff The intensity of late Quaternary volcanism in die 1Iyers, W. B.. and Hamilton, Warren, 1961. active structures strike east from the West Yellow- companying the Hehgen Lake. Montana, earthquake of Lake and the other lakes in its chain may all have Yellowstone Plateau and upper Snake River Plain. stone basin tar August It. 1959. in Short papers in the geologic and hydro- into the Yellowstone Plateau. The occurred simultaneously in an extremely severe earth- where individual eruptions of rhyolite hate volumes as logic sciences: U.S. Geol. Survey Prof. Paper .514-0, Centennial structures. which strike eastward. are quake. large as 5 cubic miles, suggests a possible causal rela- p. sau--43s 0-64-10 •

98 THE ETEEIGEN LAE-E. lir102.7A2.1A, EARTHQUAKE OF AUGUST 17, 1950

Oakeshott, G. H., ed., 1955, Earthquakes in Kern Column Rattner, Franz, ISM, Fundamentals of limnology: Toronto California, cloring 1852: California Jour. Mires and Geol- Unit, Press. 242 p. ogy Bull 171, 203 p. RyalL Alan. 10EC. The earthquake at Hebgen Lake. Montana, Prins. J. 1958. Water wares due to a local disturbance: on August 18, 1089 (GCT). P Wares; Seismol. Sue. Coastal Engineering, tith Cont. Proc., Wave Research. America Bull_ v. 52. no. 2. p. M5-271. Council, Eng. Foundation. p. 1.17-1E2 Witkiad, L. J., 1961, Deformation of the epicentral area. Pleb- Rubey, W, and Hubbert. II. K., 1559. Role of fluid pressure gen Lake, Moors:1s, earthquake of August 17, 1959—Dual. IS mechanics of overthrust faulting. pt. 2, Overthrust bait basin concept, to Short papers In the geologic aud hydro- la genizynelinal area of western Wyoming In light of Sold. logic sciences: U.S. GeoL Surrey Prof. Paper 421--D, pressure hypothesis: Geol. Soc. America Bull_ r. 70, au. 2, p. p. 167-205.