University of Washington Publications in Geology The Geomorphology and Volcanic Sequence of Steens Mountain in Southeastern Oregon

STEENS MOUNTAIN ANDESITIC SERIES
(continued)

UPPER ANDESITIC SERIES

GENERAL FEATURES

Following the extrusion of the great flow, there is no evidence of an erosional interval prior to its submergence beneath a number of thin andesitic flows, which have been locally complicated by the close association with their vents. At a few horizons, accumulations of ejectamenta testify to explosive activity, which has given rise to several small, well-defined cinder cones. The lower part of this upper series consists predominantly of flows which vary rapidly in their physical characteristics, and yet are so uniform in their variations that they can be described as a unit. Some of the more distinctive local peculiarities observed in the series will, however, be subsequently mentioned in a brief description of the andesite in each valley.

Most of this upper series consists of a horizontal alternation of vesicular andesitic breccias and dark gray aphanitic platy lava, which as a rule follows a slightly sinuous course (fig. 48). At a few localities, interbedded stratified tuffs indicate occasional intervals in the extrusion of the series and testify to the fact that it has been formed by the advance of successive flows. The individual members, however, are poorly defined, for the basal breccia of one merges with the irregular surface breccia of that which locally preceded it. The contact as a rule is therefore imperceptible if it is not defined by the accumulation of ejectamenta.

Fig. 48. View of the andesitic flows exposed at the base of the scarp south of Little Dry Creek. The resistant horizons are formed principally of extensive flow breccias.

As in the case of all blocky lavas, the margin of each flow presumably ended in a relatively steep wall with a slope depending on the angle of repose of the fragmental material. Since the successive sheets were apparently derived from a series of adjacent vents and not from a central one, it is obvious that the steep margin of one might have locally curtailed the advance of the succeeding flow. This possible merging of successive marginal flow breccias may explain the vertical zones of brecciation that are both underlain and overlain by the normal horizontal alternation of breccia and platy lava (fig. 49). Some more extensive vertical zones of breeciation however, mark loci of extrusion.

Fig. 49. View of the northern wall of the valley of Dry Creek, showing the local merging of breccias presumably at the conflicting steep margins of flows.

The breccias almost invariably, form the bulk of the series. Their predominance is emphasized, however, by the fact that their consolidation renders them sufficiently resistant to erosion to cause them to stand out prominently, while the platy horizons, which are usually less than a third as thick, are deeply eroded (fig. 50). The resistant members, consisting largely of roughly rounded vesicular fragments that range from 2 to 10 inches in diameter, show as a rule a thickness of from 10 to 40 feet. Many of the individual lines of platy lava can not be traced more than a few hundred yards before they are found either to lense out, or to blend into the breccia, leaving no obvious break to separate the upper and lower members.

Fig. 50. View of the valley of Dry Creek, showing the typical alternation of breccia and platy lava formed by the merging of successive flows. The slight vertical structure of the breccia is caused by erosion. The area shown has an estimated height of about 100 feet.

Locally, however, far thicker horizontal masses of lava occur in the breccia. Some of these are very irregular, both in outline and jointing. Others show curving jointing roughly parallel to an elliptical outline, which may he 20 by 40 feet or more across. Some minor injections can be traced from these larger masses. Some of these curve into a roughly horizontal sinuous course, usually showing a thickness of from one to three feet and a lateral extent of from 10 to 50 feet. Although their orientation probably depended on the direction of flowage, no uniformity was observed in their inclination. In fact, in a number of exposures, the injections were observed to spread laterally in both directions and to develop a pronounced fan shape, which flares upwards (fig. 51). Many of these small injections show well-defined platy structure, but as a rule the more glassy ones exhibit an irregular spinelike jointing. These minor extensions of the larger masses are considered to have been formed as auto-injections, caused by tongues of the still fluid lava intruding and squeezing upwards or laterally into vesicular breccias that either capped or defined the margin of an advancing flow.

Fig. 51. Fan-shaped injection intruding its own flow breccia directly above the shoulder at the top of the great flow north of Alvord Creek.

The contact of the upper part of these auto-intrusions appears invariably to be clean cut, but traced downward their margins as a rule become gradually vesicular and develop more irregular and coarser jointing. At these lower contacts, the injections even exhibit a rough reddish surface identical to that of the adjacent fragments. Locally either a surface of this type or fractured vesicular lava appears to truncate the injections actually normal to the direction of their flowage, as indicated both by the jointing and the elongation of the vesicles.

The lava advancing beneath its breccia would obviously have been more chilled and viscous at its margin. If hydrostatic pressure caused the local advance of auto-intrusions into its own breccia, the viscous marginal lava would have been forced outward in the front of the injection, but it would have been followed by lava that became more and more fluid as it was derived from the central portion of the flow. The porosity of the breccia would have permitted a relief of pressure and a consequent development of vesicles, which would have become progressively more marked as the gradual downward increase in fluidity permitted the volatiles to expand. The chilling induced by this expansion would have caused the vesicular lava to solidify and the injection thus to be truncated. With the further advance of the flow the injection might have been completely brecciated. In the flow breccias the linear distribution of fragments of platy rock locally testifies to the former presence of injections.

These flows resemble the aa variety of basalt more closely than they do the slightly more acidic type of blocky lava observed by Washington at Fouqué Kameni.¹² In these Santorini flows, the upper breccias are formed of loose angular glassy blocks, which, as well as the underlying lava, are practically free from vesicularity. In marked contrast to this type, the breccias of the upper andesitic flows on Steens Mountain are highly vesicular, although far more minutely so than the typical basaltic ones. The primary surface of some of the vesicular blocks is very rough and jagged and closely similar to that frequently developed by basalt. Unlike typical aa basalt, however, the andesite is always aphanitic and as a rule relatively glassy, rather than showing the characteristic coarse crystallinity of the more basic lava. Although the breccias are not very compact, they are for the most part firmly consolidated. In contrast to the aa type, this agglutination appears to be due to later agencies and not to the adhesion of still viscous fragments of slag-like lava.

---

¹²Henry S. Washington, "Santorini Eruption of 1925," Bull, Geol. Soc. Am., vol 37, p. 367, 1926.

---

DISTRIBUTION

Although these flows appear to have been very viscous and as a rule to have advanced predominantly as unconsolidated flow breccias, they accumulated with sufficient uniformity to give the andesite a relatively level surface throughout its entire extent from Little Alvord Creek northward to Mann Creek, where it is truncated by a transverse fault. For this distance of about 11 miles the uppermost member of the andesitic series is, as a rule, 2,000 to 2,500 feet above the base of the scarp. Owing to the irregularity of the surface on which they were extruded, however, the thickness of this upper series of andesitic flows is very variable.

These upper flows are not present above the vent in the valley of Little Alvord Creek. Here the later activity is merely suggested by isolated blocks of hornblende-free andesite scattered throughout the capping zone of soft altered glass with its obvious content of hornblende. It was not determined if the blocks were derived from pyroclastics. It is possible, however, that some of the previously mentioned explosive characteristics at this locality may indicate a transition to the later phase of activity. A little farther to the north in the southern part of the valley of the south fork of Alvord Creek, a number of poorly defined flows of andesite fill the local depression in the surface of the great flow to a depth of about 400 feet. Although their jointing is predominantly horizontal, it locally is very irregular, and in a few horizons is associated with vesicular breccias which here form a minor constituent.

Northward this upper series of flows decreases in thickness as the surface of the great andesitic flow increases in height above the vent south of Alvord Creek. In this region, vesicular andesitic breccias may be seen to be locally cut by injections, but the outcrops are not of sufficient magnitude to permit an interpretation. It is only in this vicinity, however, that the basal contact of the overlying basalt was observed. The basaltic flows were invariably found to have lapped against a slightly irregular surface that sloped gently westward. The andesite is covered with a thin layer of tuff that shows only a slight tendency to stratification. There is no indication of erosional agencies. Immediately below this upper contact, blocks of glassy andesite outcrop. Some of these appear to represent the surface of an exceptionally vitreous flow, while others may have been distributed as coarse pyroclastics.

On the northern side of the valley of Alvord Creek, the upper series of andesitic flows is excellently exposed above the great flow (fig. 52). At this locality the upper andesite shows a minimum thickness of about 600 feet. A succession of flows accompanied by extensive breccias form the lower half (fig. 53). These rapidly increase in number to the west with the thinning of the underlying extrusive mass. In the upper part of the exposure, there are several thick irregular masses of platy lava that definitely crosscut and indurate the breccia. These appear to have welled into a thick flow. Farther to the west along the divide between the Alvord Creek and the Cottonwood Creek drainage, there are additional indications of small loci of extrusion. The surface of these is characteristically formed of a dark dull glass which breaks into spinelike fragments. This vitreous lava is extensive throughout the upper part of the series. It appears to indicate an increase in the viscosity of the magma, for the composition remains approximately uniform (table VII).

Fig. 52. View of the andesite exposed on the northern wall of Alvord Creek showing the upper portion of the great flow overlain by the later series. The lava forming the massive exposure in the upper center appears in part to overlie the minor vent or vents from which it was derived. Here the upper andesitic series shows a minimum thickness of about 600 feet but it increases rapidly to the west.

Fig. 53. View of the upper andesitic series capping the great flow north of Alvord Creek (cf. fig. 52). The resistant horizons in the foreground are formed predominantly of flow breccias. Above them is a near-vent accumulation of lava. The exposures on the skyline show a total height of about 500 feet.

Both here and to the north there is evidence of a discontinuous series of vents, which as a rule are a half mile or more to the west of the base of the scarp. To the east, the upper andesite appears like a fairly normal series of blocky flows, but when the exposures are followed westward, the local variations accompanying the near vent facies render them far more irregular. The horizontal alternation, which usually defines the series, becomes locally no longer apparent. The actual mechanism can seldom be conclusively interpreted, for as a rule such criteria as the orientation of the platy jointing are too variable to be relied upon.

Near the head of the south fork of Cottonwood Creek, there is additional evidence of a vent, which was active at an horizon several hundred feet below the uppermost members. To the east on the northern wall of the north fork, a homogeneous breccia of dark vesicular glass forms a vertical zone that is over 500 feet in height. Immediately adjacent to this zone, the platy lava forming the horizontal flows is far more extensive than usual. The inclination of their jointing suggests that several consecutive flows have been derived from this vent, which subsequently gave rise to explosive activity.

The other noteworthy feature observed in the valley of the north fork is the occurrence of a large inclined block of acidic stratified tuffs at the bottom of the canyon to the south of the vertical zone of brecciation, and only about 100 feet above the top of the great flow. This block, which shows dimensions of about 70 by 30 feet, rests directly on andesitic breccias. Locally it is indurated at its contact with the lava. The content of quartz and biotite in these sediments proves the block to be of exotic origin. An explanation of their presence appears to demand that it has been carried upwards in a vent from the Alvord Creek Beds, the uppermost horizon of which is 600 or 700 feet below.

Northward in the valley of Willow Creek, the andesitic flows are continuously exposed. Although the breccias are very marked, they are locally less predominant than usual. In the eastern portion of the valley, they appear to form a normal series, but to the west, at the head of the lower cirques, there are indications of a vent of indeterminable dimensions midway between the two forks. Here dark vitreous injections protrude from the summit of the divide. The vertical structures may be traced downward on either side for several hundred feet. At this locality, the discordant relation of the basalt is more marked than observed elsewhere and the contact has an inclination of at least 30°.

Mosquito Creek forms the next valley to the north. Its walls are less precipitous and its exposures therefore less distinct, but isolated outcrops permit the andesitic series to be traced continuously as a normal succession of flows. Immediately to the north, the narrow gorges of four closely spaced valleys show magnificent exposures, which permit more detailed observation. Although each of these valleys exhibit distinctive features, they disclose principally a repetition of flows of the blocky type, which have been previously described. At several horizons, these valleys show evidence of explosive activity, either as interbedded tuffs or as well-defined cinder cones. The center of activity appears to have moved progressively westward. Higher in the valleys, the series is cut by elongate intrusions from which at least some of the uppermost members have been derived.

The most southerly of these four creeks is known as Dry Creek, although it is usually well supplied with water. Near its mouth, several large flows are exceptionally distinct in their relation. They are separated by typical flow breccias, which, however, form a relatively minor constituent. At two of the lower horizons an otherwise imperceptible contact of a basal and a surface breccia is rendered apparent by an accumulation of tuffs. In one of these localities, on the northern wall at about 300 feet above the lowermost exposures, a small bed of well-stratified tuffs testifies to the water deposition of tuffaceous sediments in a minor depression on the surface of a flow.

To the south, a dissected cinder cone, at least 400 feet in height, caps the divide between Dry Creek and Mosquito Creek (fig. 54). On both its western and northern sides, the pyroclastics are excellently exposed and show a distinct slope of about 30°. To the south and east of these main tuffaceous deposits, the center of the crater has been removed by erosion, but the position of the neck is indicated both by vertical platy lava and by extensive agglomerates, which are so highly indurated that their structure is apparent only on the weathered surface.

Fig. 54. The western part of the southern cinder cone viewed from the intrusion which cuts its western slope presumably near the crater.

The cone is also cut by several extensive intrusions, which have resulted in similar induration. To the east at its base, one of the largest of these has altered the flow breccias immediately beneath the cone. Adjacent to the major intrusion, the series shows an alternation of horizontal platy lava and irregularly jointed breccias, which are completely indurated. Locally it is impossible to distinguish the platy lava of the flows from lateral extensions of the intrusion. To the south, the transition to the unaltered rock, however, renders the original relation apparent (fig. 55). The uppermost exposure of the western limb of the cone is also cut by an elongate intrusion, which trends a little west of north (fig. 56). This dike-like intrusion has also resulted in extensive alteration. The northern limb of this cone is cut by at least two additional minor andesitic dikes.

Fig. 55. The margin of the zone of induration in the flow breccias beneath the northeastern margin of the southern cinder cone. To the left, away from the intrusion, the agglomerates grade into a poorly consolidated breccia.

Fig. 56. View of the main elongate intrusion cutting the southern cinder cone. The jointing of the platy intrusive and the associated indurated agglomerates are both predominantly vertical.

North of Dry Creek, another cinder cone is so close to the one just described that their pyroclastics almost coincide at the bottom of the valley. This northern one, which is at approximately the same horizon, may have originally been more extensive, but now its remnants are widely scattered. Its southern limb, which is formed of buff colored tuffs and angular glassy fragments, is exposed on the northern wall of Dry Creek. To the south, near the summit of the divide separating it from Little Dry Creek, a tuffaceous pinnacle has been sufficiently indurated by an andesitic intrusion to survive erosion (fig. 57). This pinnacle exhibits irregular masses of lava that are inclined northward and suggest an accumulation of viscous splashes of lava on the inner slope of a crater. To the east, remnants of the eastern limb survive, but to the north the pyroclastics have been almost completely stripped from a continuous surface that appears to be formed by a flow breccia that slopes northward gently. On the northern wall of Little Dry Creek, the northern limb of a cinder cone shows a thickness of over 100 feet. Presumably it is part of the same cone, but if such is the case, the accumulation of ejectamenta must have been very assymetrical, and the cone must have had a far longer and more gentle slope to the north. In the valley of Dry Creek, thin andesitic flows are exposed between these two cones, indicating that they were subsequently just submerged. These flows appear identical to the lower series, except that the lava forming them is slightly more vitreous.

Fig. 57. The northern fork of Dry Creek is on the right. To the left of it, the inclined beds of the northern cinder cone are visible above the flow breccias. As a result of the induration of an elongate andesitic dike, a remnant of the inner lip of the crater forms a red pinnacle on the divide to the north.

A little to the north of this locality, there is also evidence of explosive activity in the lower part of the series. At the mouth of Little Mann Creek, part of a small andesitic cone forms the lowermost exposures (fig. 58). The only other break observed in the extensive flow breccias is near the top of the main series. The thick flow, on the surface of which the northern cone accumulated, shows at its base a persistent thin bed of tuffs, which varies from 2 to 5 feet in thickness (fig. 59). The ejectamenta apparently settled as a relatively uniform coating on an irregular surface. The sharper projections locally failed to be covered. The bed at present follows a sinuous course, which in places may be due to later slumping.

Fig. 58. View of the base of the scarp at Little Mann Creek showing the inclined tuffaceous deposits that underlie the thin flows. To the right, just beyond the margin of the picture, the beds dip in the opposite direction, presumably because of their deposition on the inner slope of a crater.

Fig. 59. An exposure on the scarp north of Dry Creek showing a persistent bed of stratified tuffs, which demarks the contact of two flows in the upper part of the series. To the right of the center, a typical auto-injection projects above the surface breccia of the flow.

To the north in the valley of Mann Creek, at several hundred feet higher than the top of the great flow, exposures of tuffs appear to indicate remnants of another smaller cone. Here an outcrop showing platy vertical jointing suggests the position of the neck. Higher in this valley, elongate intrusions form pronounced exposures with a general north-south trend. Like the intrusions cutting the southern cinder cone, they are associated with highly indurated agglomerates.

PETROLOGY OF THE UPPER ANDESITIC SERIES

The vesicular flow breccias, which form such a large bulk of the upper andesitic series, have already been briefly described. As previously stated, most of the blocks range from 2 to 10 inches in diameter. Many of these have been roughly rounded by the breaking of their fragile margins during the advance of the flow, but where the primary surface may be observed, it is found either to be smoothly fractured or extremely jagged and so roughly pitted by the disruption of its vesicles, while still viscous, that it appears almost granular. These vesicles are seldom over 1 mm. in diameter, but they are usually greatly elongated and distorted by flowage.

During the advance of a flow, the grinding of the fragile margins of the blocks has resulted in the interstitial accumulation of tuffaceous material. Evidence of this disintegration may be observed at numerous localities. The disrupted particles, together with the margins of the dark gray blocks, are locally more or less altered to a dull red, or more rarely to a pale buff. Probably owing to fluxion, the finest of this material locally exhibits irregular stratification, which roughly parallels the margins of the enclosing blocks. Somewhat similar results are also attained by the deposition of ejectamenta on the surface of a flow breccia. Beneath the thin interstratified beds of tuff, the fine particles fill the spaces between the upper blocks.

The rock forming the breccia as a rule is completely aphanitic. Minute feldspathic laths, however, are visible megascopically in some of the platy lava, which for the most part is of a slightly lighter shade of gray. The planes of flowage, along which it splits, occur usually at intervals of from 2 to 7 mm. Some of the more vitreous auto-injections are almost black, and break into irregular splinters of various sizes. The major axes of these are parallel to the flowage. Locally this spinous jointing results in the formation of cigar-shaped fragments.

The larger fragments in the ejectamenta show a dull vitreous surface, which also is nearly black. As a rule, they are completely free from vesicles. Some of the blocks have a conchoidal fracture. In others, planes of weakness, which were probably formed by contraction during cooling, cause them to break roughly parallel to the surface and thus to form rounded residuals. Many of the smaller fragments, which compose the light buff or gray matrix, may be seen to have been pumiceous. Although most of the blocks and lapilli exhibit fractured surfaces, the rounded chilled margins of some testify to their extrusion while still plastic.

Petrographically these various facies are so uniform that they may be treated as a unit. The most vitreous types show scattered minute feldspathic microlites averaging about .05 mm. in length in a dark opaque ground. In other very aphanitic types, an ill-defined feldspathic crystallization gives the rock in thin section a slightly mottled appearance. The more crystalline specimens, however, usually exhibit a seriate development of clean cut andesine laths, which show no zoning. The largest of these crystals seldom exceeds 1 mm. in length, while in the main mass of the ground they range from .05 to .2 mm. As a rule these show an irregular alignment. In these more crystalline types, the ground may be formed largely of a fairly light colored glass or it may he cryptocrystalline. It usually has a high content of black dust, which presumably consists of magnetite. In some specimens the ground is clouded with the opaque white dust. Indeterminable mafic grains less than .02 mm. in diameter locally form a marked constituent. Aside from these, a few sections exhibit scattered small crystals of hypersthene and of pale greenish augite. Neither of these minerals exceeds .5 mm. in length.

In no specimen was observed either the coarser andesine crystals or the hornblende phenocrysts, which locally form a marked constituent of the great flow. Some types, however, are very similar to its altered facies both in their indistinct feldspathic crystallization and in their mafic content. A chemical analysis of one of the uppermost vitreous injections cutting the breccia north of Alvord Creek emphasizes the resemblance. It proved the composition of the rock to be practically identical to that of both the upper and the lower portion of the great flow in that locality (table VII). The extreme monotony in both the megascopic and the petrographic characteristics of the rock, forming the upper andesitic series, suggests that it is relatively uniform in composition.

Only slight variations in appearance are shown even by the indurated agglomerates, adjacent to the later andesitic intrusions in the northern valleys. Megascopically their structure is hardly visible except on the weathered surface, but as a rule their irregular jointing contrasts strongly with that of the platy intrusions. Most of the fragments are angular, but some are distinctly distorted. Petrographically their fragmental origin is distinct. The finer particles and usu ally at least the margins of the larger fragments have been rendered almost opaque by kaolinitic alteration, which locally is very blotchy. The minute feldspathic laths, however, appear to be unaltered. The alignment of some of the feldspar of the interstitial material suggests subsequent flowage. The refusion of the breccia is potentially possible, since the heat of crystallization would not have had to be overcome in the alteration of the glassy fragments.