Geological Survey Professional Paper 1356 Neogene Tectonics and Geomorphology of the Eastern Uinta Mountains in Utah, Colorado, and Wyoming

Drainage off the eastern Uinta Mountains formed concomitantly with Laramide uplift, and the gradual degradation of the mountains in Paleocene and Eocene time produced immensely thick deposits of clastic sediment and carbonates in the adjoining basins. Drainage at that time was centrifugal, away from the mountains in all directions, unlike the modern drainage, and it remained largely so until after the deposition of the Bishop Conglomerate (fig. 35). One exception was the structurally controlled drainage of Blue Mountain, where the Wolf Creek—Mud Springs Draw drainage was controlled by the east-west folds and faults on Blue Mountain. Even that drainage was consequent to the structural trough between the subsidiary folds of the Blue Mountain highland. The distribution of Bishop Conglomerate shows that Wolf Creek was constrained, then as now, by the bounding upwarps of Blue Mountain. Modern Wolf Creek and Mud Springs Draw, flowing across Bishop Conglomerate, are re-incising themselves into the sub-Bishop bedrock.

FIGURE 35.—Steps in the development of the Green River drainage system, telescoped in time and generalized. Details in text. Top—Late Oligocene time. Diagrammatic drainage off the Gilbert Peak erosion surface and across the Bishop Conglomerate into the adjacent basins. At this stage the Bishop formed a nearly continuous bajada around the flanks of the Uinta Mountains. Center.—Latest Oligocene to early Pleistocene time. Subsidence and northeasterly tilting have redirected the drainage in the eastern part of the range. Note new drainage through Browns Park (B) along the former crestline, leading to the downwarped area (D) at the east end of the range. Note reversals of drainage direction on the northeast flank and the new southeasterly drain age at Diamond Mountain (DM). Vigorous headward erosion into the Green River Basin from the south is threatening the east-flowing ancestral Green and its northeast-flowing tributaries, Blacks Fork (BF) and Henrys Fork (HF). Hells Canyon (H) has captured the headwaters of Wolf Creek (W). Yampa Canyon (Y) is eroding headward toward D. Bottom—Present drainage. The southeasterly drainage along the Uinta axis spilled across the range, in late Miocene or early Pliocene time, at Lodore Canyon (L). The Green River was captured at Green River, Wyo. (GR), in the early Pleistocene. Black Fork and Henrys Fork have been redirected southward. Note the barbed confluence of Bitter Creek with the Green River.

As drainage from the mountains flowed out toward the basins, it deposited the Bishop Conglomerate in broad, sheetlike fans on the subjacent Gilbert Peak erosion surface. The present major streams of the area obviously did not yet exist; their courses are younger then the pedimentation and the alluviation that produced the Bishop Conglomerate.

Deposition of the Bishop Conglomerate ended when renewed deformation terminated the long period of crustal stability that followed Laramide uplift. One consequence was a drastic overhaul of the drainage regime. Gravitative movements on the zone of faults along Red Canyon and Browns Park tilted the whole crestal region and south flank of the range northward and eastward, deforming the Gilbert Peak erosion surface and lowering the crestline as much as one and a half kilometers. These movements began in late Oligocene time and lasted intermittently through the Miocene and, perhaps, the Pliocene. Some faults predated the Browns Park Formation; others displaced and dragged it to steep angles. Some faults have had Quaternary movements. At the east end of the range the lowering of the Bishop below present drainage formed a broad downwarp that became a catchment basin for the Browns Park Formation. Between Browns Park and the Uinta fault, the north flank of the range was tilted unevenly southward.

All these crustal movements helped set the stage for the drainage changes that followed. Much of the runoff that had flowed from both flanks of the range toward the Tertiary basins was reversed and became tributary to the newly forming valley of Browns Park. The new drainage lines in Browns Park then carried the combined runoff eastward to the downwarp in the Lay—Maybell—Little Snake area. The former crest of the range was now a deepening valley that ultimately would carry the full flow of the Green River. As discussed later, the Green was eventually diverted south from Browns Park into Lodore Canyon.

ADJUSTMENTS ON THE SOUTH FLANK—IN THE POT CREEK AREA

Former southward drainage south of Browns Park and Red Canyon is indicated by the barbed tributary patterns of all the major valleys between Cart Creek (fig. 36) and Lodore Canyon (Kinney and others, 1959; Ritzma, 1959, p. 87; Hansen and others, 1960, p. B258; Hansen, 1965, p. 174; 1969a, p. 34; Hunt, 1969, p. 97). During Bishop time all these valleys drained into the Uinta Basin and contributed detritus to the Bishop Conglomerate (fig. 37). All still contain Bishop Conglomerate, consisting mostly of coarse, nearly white, friable pebbly sandstone and conglomerate. As these fills merged southward they formed an extensive bajada on the south flank of the range (Hansen and others, 1981).

FIGURE 36.—Drainage map of the Cart Creek area, showing barbed headwater tributaries that formerly flowed south to Reader Creek. These tributaries meander sluggishly through McKee Draw and East McKee Draw before plunging into the Cart Creek canyon, shown by hachures (hachures coincide with canyon rims).

FIGURE 37.—Distribution of Bishop Conglomerate (shaded), anomalous present drainage pattern, and barbed tributaries of north-flowing streams in Pot Creek area. Hachures indicate canyons. Note how close heads of Jackson Creek and Crouse Creek are to Pot Creek. Pot Creek is held at grade by the numerous crossings of the Uinta Mountain Group (unshaded), and its capture by Crouse Creek is imminent at point shown by arrow. (click on image for a PDF version)

With northerly and easterly tilting, however, all these streams stagnated and partly reversed their flow directions, leaving barbed tributaries. In part these reversals were caused simply by capture due to tilting. For example, Cart Creek, Gorge Creek, and Jackson Creek were captured by vigorous north-flowing streams emptying into Red Canyon through deep, narrow tributary canyons. All the reversed streams, in fact, now pass through such canyons before joining the Green. Most of the valleys at some midpoint contain low divides in Bishop Conglomerate, the northerly reaches flowing to Red Canyon or Browns Park and the southerly reaches to Pot Creek. A person can cross such a divide by automobile without realizing that the drainage direction has changed 180 degrees. Similar low divides separate the drainages of Pot Creek and Diamond Gulch.

Pot Creek and Diamond Gulch are post-Bishop drainages that formed by the diversion of southward runoff into east-southeasterly courses after regional tilting. Because of the concavity of the original Bishop profile, tilting created new low points along the formerly south-flowing streams, as shown diagrammatically by figure 38. Pot Creek and Diamond Gulch sought out these low points and consequently funneled the runoff toward the east-southeast. The near coincidence of Pot Creek and Diamond Gulch with fault lines, however, suggests that their drainage direction was influenced partly by faulting. I visualize south-building alluvial fans gradually rising toward and locally overtopping low places at fault lines in the interfluves. With tilting, drainage found its way across these partly buried interfluves.

FIGURE 38.—Effect of tilt on the low point of a concave-upward profile. New low points are indicated by arrows on tilted profile. The local flattening of the original profile at the future site of Pot Creek is attributed to a constriction in the valley upstream from a belt of resistant Precambrian rock across the drainage line.

Drainage adjustments are still taking place and, on a geomorphic time scale, Pot Creek itself is threatened with imminent capture by Crouse Creek (Hansen and others, 1981). Crouse Creek heads within a few hundred feet of Pot Creek, has the advantage of a steeper gradient than Pot Creek, and is cutting rapidly downward into the intervening divide of soft Bishop Conglomerate (mostly friable sandstone at that locality). Pot Creek, on the other hand, flows alternately across the Bishop Conglomerate and the Uinta Mountain Group and is held at grade by numerous thresholds of hard red quartzite. If left to itself, the entire discharge of Pot Creek upstream from the divide will eventually flow north into Crouse Creek. The capture of Pot Creek, in fact, has been partly fulfilled artificially by a diversion ditch that directs water from Pot Creek across the low divide into Crouse Creek to irrigate hay fields in Browns Park. A narrow trench 10-15 m deep has already been eroded into the divide.

ADJUSTMENTS ON BLUE MOUNTAIN

Middle Tertiary drainage adjustments on Blue Mountain parallel those of the Pot Creek area in time and style. The pre-tilt drainage flowed east down Wolf Creek in a valley floored with Bishop Conglomerate. Hells Canyon, which drains north to the Yampa River, has captured most of the headwaters of Wolf Creek and in so doing has greatly expanded its watershed at the expense of Wolf Creek (fig. 39). Hells Canyon is a steep, narrow gorge that drops nearly 600 m in about 11 km at an overall rate of 38-57 m/km; the gradient steepens locally to as much as 75 m/km. Wolf Creek has a gradient of only about 15-19 m/km. Mud Springs Draw, Spike Hollow Draw, and Bobcat Draw are sharply barbed former tributaries of Wolf Creek. Turner Creek, a former reach of Wolf Creek, has been reversed and is now separated from Wolf Creek by a low divide. Northerly tilt probably shifted the hydraulic advantage to Hells Canyon. Farther upstream, north-flowing Meadow Creek captured the headwaters of K Creek, probably at about the same time.

FIGURE 39.—Drainage map of the Blue Mountain area. Stippling indicates Bishop Conglomerate. Note how former drainage of Wolf Creek has been diverted into Hells Canyon, reversing the direction of Turner Creek. Meadow Creek formerly flowed into K Creek. Note barbed tributaries of Johnson Draw. (click on image for a PDF version)

Johnson Draw, which once drained south independently into Wolf Creek, now drains north to the Yampa River via Johnson Canyon, a gorge that resembles Hells Canyon but is smaller. Johnson Draw heads only 1-1/2 km from Wolf Creek at a wide low pass in Bishop Conglomerate, then flows 11 km north to the Yampa. A few kilometers to the east, Bear Valley Draw, shown on figure 26, still drains into Wolf Creek, but part of its headwaters has been diverted to the Yampa via Bear Draw through a deep gorge cut in the Madison Limestone. Bear Valley is being threatened by Thanksgiving Gorge at East Bear Valley Draw, but capture is not yet imminent.

ADJUSTMENTS ON THE NORTH FLANK

Drainage adjustments on the north flank of the Eastern Uinta Mountains, north of the anticlinal axis (fig. 40), have had far-reaching geomorphic consequences. Many adjustments were imposed by the deformation that ended the Bishop depositional episode. Erosion in the southern Green River Basin since the end of the Bishop episode in late Oligocene time has been extensive, and evidence for ensuing drainage changes has been partly lost to erosion. In brief, a reversal of the regional gradient, accompanied by a large-scale redirection of the drainage, was caused by the foundering of the Uinta arch in late Oligocene and early Miocene time.

FIGURE 40.—Principal hills and drainages of northeast flank of Uinta Mountains and vicinity and their relationship to the Uinta anticline and the Uinta-Sparks fault zone. Bar-and-ball symbol indicates Laramide downthrow; (D) indicates post-Bishop downthrow; hachures outline mountains and escarpments. Notice at upper center of the map how Salt Wells Creek has captured the headwaters of Vermillion Creek. (click on image for a PDF version)

From the time Lake Gosiute disappeared (middle Eocene) until the foundering of the Uinta arch (late Oligocene), the drainage of the southern Green River Basin was northward, away from the Uinta Mountains. The northerly gradients on Bishop-capped remnants of the Gilbert Peak erosion surface and the northward fining of the gravels leave no doubt about the direction. The northward drainage must have arisen south of the exposure area of the Red Creek Quartzite, as cobbles and boulders from the Red Creek and from the Uinta Mountain Group are abundant in the Bishop Conglomerate on Little, Miller, and Pine Mountains and as far north as Aspen Mountain. The drainage divide was well south of the Uinta fault, for the Red Creek Quartzite is truncated south of the fault by the Gilbert Peak erosion surface. The crestline must have been as far south as Browns Park, along the axis of the Uinta anticline, because the reconstructed profile of the Gilbert Peak surface projects into the air above Browns Park. The coarseness of the Bishop, furthermore, indicates that its source to the south had strong relief. Bradley (1936, p. 177) suggested that the master drainage was eastward or northeastward out of the Green River Basin, possibly toward the ancestral Platte River, a view endorsed by Ritzma (1959, p. 87) and elaborated by Hansen (1969b). Sears (1924a) had hinted at it decades earlier.

As pointed out previously (p. 32), the shift in drainage direction along the mountains followed the deposition of the Bishop Conglomerate and preceded the deposition of the Browns Park Formation, because the Browns Park Formation contains abundant clasts derived from sources to the north. New south-flowing drainage lines must have begun to form all along the north flank of the range concomitantly with tilting. With modifications, these lines persist in the modern drainage. The master drainage was diverted much later (p. 31).

FLAMING GORGE AREA

The foundering of the Uinta arch almost surely played a critical part in redirecting the drainage of the Green River Basin toward the south and in establishing the present Green River as the master stream of the basin. Some idea of the extent of the subsidence at Flaming Gorge, where the Green enters the mountains can be had by comparing the present height of Bear Mountain with Cedar and Little Mountains in Wyoming (fig. 41); all three of these are remnants of the Gilbert Peak erosion surface. (Cedar and Little Mountains are also capped by Bishop Conglomerate; Bear Mountain is not). In longitude, Bear Mountain is about halfway between Cedar and Little Mountains, and it stands about 18 km to the south. The elevation of Cedar Mountain approaches 2,620 m, and Little Mountain exceeds 2,774 m. A point on the restored Gilbert Peak surface halfway between at their latitude would be about 2,700 m, or about 2,670 m if 30 m is subtracted for an assumed capping of Bishop Conglomerate, which is about 30 m thick on Twin Buttes, 4.8 km to the north. Bear Mountain, however, reaches only 2,356 m, about 300 m lower than the projected surface midway between Cedar and Little Mountains. Originally, though, Bear Mountain was higher than the restored midpoint. If the restored pre-tilt gradient of the Gilbert Peak erosion surface (there assumed to be about 38 m/km) were projected south to Bear Mountain, it would reach about 3,350 m. Thus, the Gilbert Peak surface at Bear Mountain probably has been lowered at least 975 m. That amount of lowering requires a back tilting toward the mountains of about 17 m/km, and with that amount of tilt, a reversal of the drainage should not be surprising.

FIGURE 41.—Diagram showing subsidence of Bear Mountain and reversal in gradient of Gilbert Peak erosion surface north from Bear Mountain. The restored gradient is assumed to have been a uniform 38 m/km, although it probably was actually concave upward. In any event, subsidence of the magnitude indicated would have drastically altered the drainage directions across the area. Vertical scale is 10 times horizontal scale.

In fact, the new southward and eastward drainage should have been vigorous and should have agressively enlarged its watershed at the expense of disadvantaged north-flowing streams nearby. The flow through Red Canyon toward Browns Park should have rapidly extended its watershed north into the Green River Basin, capturing such streams as Henrys Fork and Sage Creek in the process. Just when it captured the master drainage of the basin, thereby greatly shifting the position of the Continental Divide, is uncertain, but it probably did so in Pleistocene time (Hansen, 1969b; Hansen and others, 1983). Such a momentous capture, diverting and acquiring the headwaters of the Platte River, would have vastly enlarged the competence and erosive capacity of the system. The effect on the Platte should have been the opposite—aggradation and reduced competence.

Drainage changes in the various tributaries of the Green River on the north slope of the Eastern Uinta Mountains since Bishop time have been outlined by several investigators (Sears, 1924a; Bradley, 1936; Ritzma, 1959; Hansen and others, 1960; Hansen, 1965, 1969a, 1969b, 1984). In brief, drainage quickly adjusted to the topographic changes brought on by the foundering of the Uinta arch, and the earliest new stream courses were largely consequent down the new slopes.

Subsidence and topographic distortion caused by the foundering diminished gradually westward to a minimum. Drainage west of Flaming Gorge continued to flow northward, but all of it as far west as Henrys Fork was eventually captured by structurally controlled subsequent drainage that flowed directly to the Green River in response to the new gradients (Hansen, 1965, p. 175-180).

East of Flaming Gorge the topographic dislocations were severe, and the resultant drainage changes were correspondingly drastic. New south-flowing consequent streams included Spring, Goslin, Red, Willow, Beaver, and Vermillion Creeks, the Little Snake River, and many lesser streams. (Incidentally, remnants of the Browns Park Formation deposited within the valleys of Vermillion Creek and the Little Snake River are confirmatory evidence that southerly directed drainage and hence southerly tilting predated the Browns Park Formation). Further realinements took place along some of these streams as they adjusted to the structure of the underlying bedrock and as subsequent drainage developed along belts of nonresistant rock. Spring Creek, for example, was later diverted west into a strike valley in the Hilliard Shale (Cretaceous), leaving its initial course high and dry at Dutch John Gap (Ritzma, 1959, fig. 8C; Hansen, 1965, fig. 67). (Dutch John Gap is a conspicuous wind gap just northwest of Dutch John, Utah).

VERMILLION CREEK AND IRISH CANYON

Changes in the course of Vermillion Creek at the southeast end of Cold Spring Mountain have been described briefly by Sears (1924a, p. 298). With a drainage area of more than 2,600 km², Vermillion Creek has by far the largest basin of the south-flowing tributaries of the Green between Flaming Gorge and the Little Snake River. Vermillion Creek crosses the flank of the Uinta arch in a narrow canyon about 215 m deep. Sears attributed this canyon, correctly I think, to superimposition down through a former cover of the Browns Park Formation. However, the broad basin of Vermillion Creek upstream from the canyon plainly was excavated before the Browns Park Formation was deposited in it. The present valley may have evolved through a process of cutting, back filling, and renewed cutting, perhaps by the mechanism of anteposition, as postulated by Hunt (1956, 1967, 1969) for many stream valleys of the West. Sears also pointed out the probability that Vermillion Creek formerly flowed through nearby Irish Canyon (fig. 42) but was captured by a tributary of the "East Fork of Vermillion Creek." (The East Fork is now called Dry Creek.) With that capture, Irish Canyon was abandoned by perennial drainage and is now a dry but impressive beheaded valley, 6.5 km long and nearly 300 m deep.

FIGURE 42.—Geologic map of Vermillion Creek area affected by the beheading of Irish Canyon. Heavy dashes mark wind gap at head of Irish Canyon. Geology modified from Rowley and others (1979). (click on image for a PDF version)

Sears did not elaborate on the origin of Irish Canyon, but he noted that Irish Canyon's downstream extension, Bull Canyon, though incised into the Uinta Mountain Group, has undoubtedly been superimposed. The Browns Park Formation still exists at the head and mouth of Bull Canyon and on both rims. Whether or not the Browns Park Formation also once covered the rims of Irish Canyon, however, is doubtful. The head of the canyon truncates a bermlike remnant of a pediment that flanks Cold Spring Mountain on the northeast. Just 2.4 km farther north the pediment passes beneath a basal conglomerate of the Browns Park Formation that, in turn, underlies typical Browns Park sandstones. Soon after the pediment formed, the Browns Park Formation probably was deposited on it. Irish Canyon, which cuts the pediment, probably is younger, therefore, than the nearby Browns Park Formation. The slope of the pediment, moreover, is wrong for drainage into Irish Canyon. A hypothesis that would not require superposition of Irish Canyon down through the Browns Park Formation is as follows:

Irish Canyon, like other north-flank drainageways, began to form with the onset of tilting and collapse of the north flank. Eroding headward across the Uinta Mountain Group and the lower Paleozoic rocks, Irish Canyon then uncovered the soft Morgan Formation and started to excavate a strike valley on the northeast side of Cold Spring Mountain over the present upper reach of its canyon. Meanwhile, another south-flowing drainage was excavating Vermillion Creek basin. On the southwest side of the basin—on the northeast flank of Cold Spring Mountain—the pediment that is preserved today as the aforementioned berm was forming at that time also, graded to the floor of Vermillion Creek basin. The Browns Park Formation soon began to accumulate on the pediment, on the floor of Vermillion Creek basin, and in the valley of Browns Park below Irish Canyon. Irish Canyon continued to erode headward, finally truncating the pediment along the strike of the Morgan Formation, leaving the stranded berm. Perhaps it then extended its headwaters north to become the upper reach of ancestral Vermillion Creek.

At that time, the level of Irish Canyon was about 425 m above modern drainage in the Vermillion Creek basin, as indicated by the altitude of its rim. Ancestral Vermillion Creek would have continued to flow through Irish Canyon until a tributary of Dry Creek beheaded the canyon and set the present drainage pattern. Dry Creek, which probably was flowing on the Browns Park Formation at that time, would have had an erosive advantage over Irish Canyon, which had to cross the quartzites of the Uinta Mountain Group and the resistant Paleozoic limestones.

Downstream from the point of capture, the southeast-trending reach of Vermillion Creek meanders in a trench that parallels, crosses, and recrosses the Sparks fault in a way that suggests a causative relationship between faulting and drainage. The fault is downthrown to the northeast, and the Browns Park Formation is tilted northeastward by drag in the upthrown block. Late movement on the fault thus postdates the Browns Park Formation. This movement moreover was directly athwart ancestral Vermillion Creek and was parallel to the capturing tributary of Dry Creek. It therefore would have inhibited the flow of ancestral Vermillion Creek while encouraging headward erosion by the capturing tributary. Capture, therefore, probably shortly followed movement.

The capture of Irish Canyon must postdate the diversion of the Green River into Lodore Canyon, because the Green River is the base level of Vermillion Creek and because the wind gap at the head of Irish Canyon, at an altitude of about 2,035 m, is nearly 300 m lower than the rims of Lodore Canyon at the Gates of Lodore. The diversion of the Green River into Lodore Canyon should have rejuvenated all upstream drainage. The capture of Irish Canyon probably was an early Pleistocene event, inasmuch as it significantly postdated the deposition of the Browns Park Formation and inasmuch as the wind gap at its head is only about 180 m higher than nearby modern Vermillion Creek. The latest ground-breaking movement on the Sparks fault, therefore, may be as recent as early Pleistocene. An earthquake on September 24, 1983, strengthens the case for Quaternary movement. This earthquake, of magnitude 4.2, was centered beneath Cold Spring Mountain at lat 40.8° N., long 108.8° W. (John Minsch, U.S. Geological Survey, oral commun., October 3, 1983), and it may have been caused by a movement on the Sparks fault, which dips southwestward beneath Cold Spring Mountain.

Bull Canyon, like Irish Canyon, also could not have been eroded until after the diversion of the Green River into Lodore Canyon, inasmuch as the rims and floor of Bull Canyon are much lower than the rims of Lodore. Bull Canyon has an average depth of about 75 m, and at its mouth its floor is only about 75 m above its base level downstream at nearby Vermillion Creek. Bull Canyon, therefore, must have been incised largely in later Pleistocene (early Wisconsin?) time, well after the beheading of Irish Canyon. Despite the loss of its Irish Canyon headwaters, Bull Canyon still drains about 30 km² of steep terrain on Cold Spring Mountain, much of it possibly gained at the expense of the drainage of Little Joe Basin.

PHYSIOGRAPHIC SETTING AND DEVELOPMENT OF LODORE CANYON

For more than 100 years geologists have pondered the course of the Green River through the Uinta Mountains. Most rivers flow away from mountains, not toward them. But the Green, draining the mountains and plains of southern Wyoming, cuts sharply into the Uinta Mountains at the Wyoming-Utah State line, then flows 175 km east and south across the range through Utah and Colorado without regard for topographic relief or geologic structure. Why, at the Gates of Lodore, does the river turn south from the wide valley of Browns Park, where logical egress is eastward, reenter the mountains to the south, and for the next 80 km drain one canyon after another before reaching the Uinta Basin? The hypothesis I've advanced in earlier publications (Hansen, 1969a, p. 42-44, 58; 1969b; Hansen and others, 1982) is briefly recapitulated here:

The Green River was flowing southeast through the valley of Browns Park in middle Miocene time on a thickening fill (the Browns Park Formation) toward a possible junction with the Yampa River and ultimately, perhaps, the White River. The fill eventually overtopped the valley rim to the south, and the river, turning southward at the site of the present Lodore Canyon, found a new course across the Uinta Mountains toward the Uinta Basin. From the Gates of Lodore southward, the river established itself on the old upland surface of the Bishop Conglomerate, ultimately eroding its way down through the underlying rocks to the present canyon bottom. Then, as now, the Bishop Conglomerate extended deep into the mountains as a network of coalesced alluvial fills (figs. 5, 8, 13, and 37). On the south slope of the range the Bishop is still so widely disposed along and near both rims of the canyon that there can be no doubt about the superposition of the river through it. Quite likely, a shallow south-flowing drainage had already been established over the site of Lodore Canyon on or through the Bishop Conglomerate at the time of spillover. If so, the Green simply took advantage of the already available course.

This concept calls for a much-thickened valley fill in Browns Park and for the subsequent removal of much of it by the Green River, but the evidence seem irrefutable. East of Lodore Canyon the Browns Park Formation still reaches altitudes comparable to the rims of the canyon, which are about 2,285-2,315 m. The highest little-deformed remnant is at John Weller Mesa (shown on fig. 13), on the drainage divide between the Green River and the Little Snake River, 30 km southeast of the Gates of Lodore. There, the eroded top of the Browns Park Formation is about 2,195 m above sea level. John Weller Mesa slopes gently northeastward and probably has been gently tilted in that direction. Beds stratigraphically higher in the Browns Park Formation crop out to the northeast but, because of dip, at lower present altitudes. Twenty kilometers north of John Weller Mesa at Dry Mountain, remnants of the Browns Park Formation stand even higher—2,340 m—in an area that has been involved in strong late Tertiary warping. Their altitude may formerly have been higher than now.

In any event, a vast amount of material has been removed from the valley since the onset of canyon cutting. The highest part of John Weller Mesa is 540 m above the level of the Green River at the Gates of Lodore, a height only about 122 m lower than the canyon rim, or about 82 percent of the height of the canyon wall. A rough guess would place the volume of material removed from the old valley of Browns Park below the level of John Weller Mesa at 210 km³. (Here, the "old valley of Browns Park" is considered to be the area between the head of modern Browns Park, to the northwest in Utah, and Lone Mountain, to the southeast in Colorado—a total length of about 72 km.) If the fill reached to the rim of Lodore Canyon, as postulated here, about 80 km³ of additional fill must have been removed. An even larger volume of Browns Park material has been eroded from areas to the east in Moffat County, and much more still remains.

The above hypothesis varies from views of earlier workers. Major Powell (1876), using the Uinta canyons as his type example, advanced the then-new concept of antecedence, which assumed erroneously that the Green River predated Laramide uplift of the Uinta Mountains and that as uplift progressed the river kept pace in eroding its canyons. Powell alluded to a stationary saw cutting a moving log: "the river was running ere the mountains were formed * * * before the rocks were folded" (Powell, 1876, p. 152). Antecedence is an elegant concept, but cogent evidence rules it out for the Eastern Uinta Mountains, as S. F. Emmons (in Hague and Emmons, 1877, p. 194) was the first to point out.

Noting the abundance of high-level Tertiary deposits (Bishop) in the mountains, Emmons advocated superposition, a concept that nearly coincides with the view expressed here. Bradley (1936, p. 189) agreed that the Green was superimposed upstream and down from Lodore Canyon, but he doubted the supposed depth and extent of the old Tertiary fill and, elaborating on a suggestion of Sears (1924a), proposed instead that an ancestral east-flowing Green River was captured by a small but vigorous stream in Lodore Canyon, "Lodore Branch of Cascade Creek," and was thus diverted southward. (Cascade Creek is now Pot Creek.) This hypothesis disavows any extraordinary thickness of Browns Park fill and calls for headward erosion by "Lodore Branch," rather than a spillover by the Green, as postulated here. Bradley did not address the difficult problem as to how a small stream with a very small watershed in hard rock would erode across a divide to capture a much larger stream—a problem obviated by spillover. Hunt (1969, p. 90-99), in his brilliant analysis of the history of the Colorado River and its tributaries, proposed a compromise position between Powell and Emmons, a combination of antecedence and superposition, which he called "anteposition." Also questioning the extent of the fill, Hunt suggested that much of the depth of Lodore Canyon is an erosive response to renewed uplift of the range since the drainage was established. Hunt's surmise probably is true, but there is no known demonstrable evidence that the canyon area has been uplifted differentially relative to the valley of Browns Park in a way that would account for the depth of Lodore Canyon. The evidence suggests otherwise: The Green has eroded hundreds of meters downward in Browns Park also. The small difference in altitude owing to post-Bishop tilting between the head of Lodore Canyon and Browns Park is just a few meters per kilometer and, in that short a distance, is negligible compared to the depth of canyon erosion. Browns Park and Lodore Canyon are both parts of the same crustal block, and if renewed uplift has occurred, it probably was regionwide—epeirogenic—and included Browns Park as well as Lodore Canyon.

YAMPA RIVER

The physiographic evolution of the Yampa River in the Uinta region is more obscure than that of the Green, and its complete history may never be fully resolved. Some episodes in its development, however, are reasonably clear. Like the course of the Green River, the course of the Yampa through the Uinta Mountains is younger than the Bishop Conglomerate; the conglomerate is well preserved south of the Yampa River, and its transport direction was plainly from the north, from the crestal part of the range toward the flanks at right angles to the present course of the river. Neither the river nor the canyon could have existed at the time the Bishop was deposited. Upstream from the Uinta Mountains, however, before the main body of the Browns Park Formation was deposited, an ancestral stream—perhaps the Yampa—was delivering pebbly gravel from the Park Range to the basin in the Lay—Maybell area. This basin was formed after the deposition of the Bishop Conglomerate; tilted Bishop crops out on the distal side of the basin, dipping toward its source.

With the tectonic subsidence of the Eastern Uinta Mountains, the broad downwarp in the Lay—Maybell area became an accumulation basin for the dominantly lacustrine and eolian deposits of the Browns Park Formation. Fluvial deposits coarser than medium-grained sand are rare in the Browns Park of this area (Luft and Thoen, 1981), and currents were seldom strong enough to transport pebble-size debris (except at the onset of Browns Park deposition, when the basal conglomerate derived from the Park Range, described on previous pages, was laid down).

Through drainage, therefore, was weak during deposition of the main body of the Browns Park Formation. No outlet from the basin at that time has yet been identified, but trunk drainage probably skirted the north end of Cross Mountain, essentially along the present course of the Little Snake River between Cross Mountain and the easternmost Uintas. It then probably turned west along the present Yampa Canyon. (See discussion of Deerlodge monocline, p. 53.) The Yampa River as such did not yet exist through Cross Mountain Canyon or Juniper Canyon, because both are superimposed canyons that postdate the deposition of the Browns Park Formation. In any event, drainage probably was overwhelmed by sediment. The abundance of eolian bedding in the Browns Park Formation in the Lay—Maybell area suggests aridity, and the meager runoff into the rather capacious basin may have been incapable of overtopping the basin rim. Basically, the Lay—Maybell area was an aggrading sink throughout Browns Park time. Dyni (1980) noted sandy calcitic concretions within the eolian sequence and thin horizontal beds of sandy limestone that he regarded as deposits of interdunal ponds. Continuing subsidence of the basin floor, moreover, may have prolonged its filling.

Downwarping of the magnitude indicated by the displacement of the Gilbert Peak erosion surface and the Bishop Conglomerate must have required a long period of time. Guided in part by faulting, the ensuing basin over the subsided Uinta arch was partly structural and partly erosional, and concomitant runoff from the higher mountains to the west must have followed available low ground and weakened rock. Erosional breaching of anticlines is common in a first cycle of erosion following uplift. A prominent set of joints parallel in trend with the axis of the Uinta anticline, moreover, suggests lateral extension normal to the axis. Such fracturing along the crest of the fold would abet breaching. Thus, the depositional basin in the Browns Park—Lay syncline nearly coincides with the subsided Uinta arch and Axial Basin anticline. Inasmuch as the subsidence and valley cutting postdated the deposition of the Bishop Conglomerate and predated the deposition of the Browns Park Formation, a significant time break must have intervened between the deposition of the two formations. Radiometric dating indicates a break of perhaps 4 m.y. (See paragraph on age of the Browns Park Formation, p. 31.)

Meanwhile, drainage must have begun to develop along the present course of the Yampa River in Dinosaur National Monument soon after the Gilbert Peak surface was formed and the Bishop Conglomerate was deposited. That area is an asymmetrical, partly faulted structural trough between Douglas Mountain to the north and Blue Mountain to the south (Hansen and others, 1980, cross sections), and the early drainage should have been guided by the trough into a roughly east-west trend. The exact course is unclear; I view it as flowing west but extending itself east by headward erosion, possibly as far as the east boundary of the monument. Structurally, the trough is not nearly as deep as the Browns Park—Lay syncline, but it is deepest along its southern border next to the Yampa fault. Bishop Conglomerate is preserved at the east and west ends but not in the center; the Gilbert Peak erosion surface is sparingly preserved as accordant ridgelines and summits and, as noted previously, it slopes gently eastward and northward in response to post-Bishop tilting.

Thus, the westerly flow of the Yampa River in Dinosaur National Monument is against the direction of the present slope of the Gilbert Peak erosion surface. The rim at Harpers Corner, near the confluence with the Green, is more than 300 m higher than East Cactus Flat; westward drainage, therefore, must predate eastward tilting. Otherwise, the Yampa would have flowed east and would still do so.

Before tilting, drainage wandered rather widely between Blue Mountain and Douglas Mountain, and in so doing, it planed off the smooth platform still well preserved 300 m or so above present drainage on East and West Cactus Flats (and less well preserved at Schoonover Pasture). This platform truncates bedding at an angle of 8°-10° The northerly component of tilt, then, would have encouraged the river to shift gradually northward, away from Blue Mountain and the Yampa fault and toward the flank of Douglas Mountain where it is today.

The incised gooseneck meanders between Harding Hole and the Yampa's confluence with the Green at Echo Park may have formed as a hydraulic response to a rising base level induced by tilting up to the west and south against the direction of drainage. Canyon cutting followed, and it faithfully preserved the meander pattern. As entrenchment progressed, however, the meanders gradually shifted, undercutting the outsides of bends, slipping off the insides, and forming the dramatic overhangs for which the canyon is famous. The largest of these is more than a thousand feet high. Meander shifting was influenced by dip, inasmuch as most of the overhangs are on the downdip sides of meanders. The time of entrenchment can only be surmised, but downcutting would have been greatly enhanced by the spillover of the Green into Lodore Canyon and the resultant rejuvenation of the drainage system. Further entrenchment surely would have been enhanced by the postulated diversion of the Upper Green River near Green River, Wyo. (See next section.)

When drainage was flowing across East and West Cactus Flats, it also was flowing between Cross Mountain and the east end of the Uintas, now the valley of the Little Snake River. At that time it was cutting a broad terrace into the Bishop Conglomerate at Klauson Pasture (fig. 43)—present altitude about 1,980 m. This truncation predated the northerly tilt of the Eastern Uinta Mountains. Whether or not that stream should properly be called the Little Snake River is a moot question: It was the outlet for the growing playa or sedimentary basin in the Lay—Maybell area, it probably carried the overflow from the Browns Park area as well as the Little Snake and Yampa headwaters, and the Little Snake as now constituted did not yet exist. Drainage downstream along the present Yampa Canyon appears to be anteposed—established before tilting but maintained by entrenchment as tilting progressed and as aggradation upstream raised its bed on the thickening fill of the Lay—Maybell area.

FIGURE 43.—View north toward Klauson Pasture (flat surface in distance at foot of easternmost Uinta Mountains), Moffat County, Colo. Klauson Pasture is capped by Bishop Conglomerate and has been terraced here by a second, lower erosion surface. Such incision of the Bishop Conglomerate is fairly common along the Uinta Mountains. Bradley (1936, p. 198) noted a similar incision of the Bishop Conglomerate on Cedar Mountain in Wyoming.

The present course of the Yampa through Juniper Canyon and Cross Mountain Canyon did not exist until the end of Browns Park deposition (Hancock, 1915). Both canyons are accidents of superposition that formed after the level of fill had risen high enough to superimpose the drainage. Farther upstream, what is now the Yampa merely emptied into the playa and was lost in a sink of shifting sand. When the drainage changed from a depositional to an erosional mode, the Yampa became organized as a through-flowing river across the buried spurs of Cross and Juniper Mountains. This change may have been triggered by the diversion of the Green into Lodore Canyon.

Hunt (1969, p. 89) has suggested that Cross Mountain has risen 150 m since the superposition of the Yampa and, hence, that the lowest 150 m of canyon is due to antecedence after superposition (fig. 44). Such a rise should be expressed by upturning or faulting in the flanking Browns Park Formation. Exposures are inconclusive on the east side of Cross Mountain—upturning, if present, is minimal—and the Browns Park Formation is not preserved on the west, but just south of the Yampa River on the west side, the Bishop Conglomerate (Dyni's lower conglomerate unit of the Browns Park Formation) has been lowered as much as 150 m in a narrow graben along the Cross Mountain fault, well shown on Dyni's map (1968). The time of faulting is not precisely known, but Dyni (1980) noted post-Browns Park monoclinal folding and inferred growth faulting in adjacent areas. Farther upstream, the rather tortuous meandering of the Yampa River above Juniper Canyon suggests that the river there has been base leveled by late resurgence of the Axial Basin anticline. These meanders are incised to a depth of about 150 m.

FIGURE 44.—Yampa River emerging from Cross Mountain Canyon, incised into the Madison Limestone. The course of the river here is superimposed, but Hunt (1969) has suggested that the full depth of the canyon in this view, about 150 m, is antecedent to Pliocene-Pleistocene renewed uplift of Cross Mountain. The Cross Mountain fault at the base of the slope has had post-Bishop movement (Dyni, 1968).

Large horseshoe-shaped or amphitheatrical basins in the north wall of Yampa Canyon at various places are not "meander—migration scars" cut by the Yampa River at an early time, despite the belief of Sears (1962, p. 9), and they thus have no bearing on the development of canyon drainage. Rather, they are products of retrogressive landsliding down a dip slope (figs. 45 and 46). The hummocky but essentially translatory slides are restricted to places where the downcutting river has uncovered the incompetent lower member of the Morgan Formation dipping into the canyon. Failure within the Morgan removed support from the canyon walls above, and failure then retrogressed updip and radially outward from the points of origin. Subsequent gullying through some slides has exposed their rubbly interiors as well as the underlying Round Valley Limestone, which provides the resistant base of failure. The most accessible slide of this kind is not in Yampa Canyon but in Diamond Gulch along the road to Jones Hole, where ongoing failure is indicated by road damage, open ground cracks, and toppled blocks of sandstone tens of meters on a side.

FIGURE 45.—Geologic map and section of part of the Haystack Rock quadrangle (Hansen and others, 1980) showing relationship of large, retrogressive landslides (Ql) in Yampa Canyon to amphitheatrical headwalls and to the incompetent lower member of the Morgan Formation. (click on image for a PDF version)

FIGURE 46.—Block diagram showing relationship of a retrogressive landslide to Yampa Canyon and the amphitheatrical form taken by the headwall scarp. Sliding begins when the lower member of the Morgan Formation is undercut by the river. Where such slides encroach on the river, they cause rapids.

POSTULATED CAPTURE OF THE MASTER DRAINAGE OF THE GREEN RIVER BASIN

The most significant major drainage change, in terms of the total budget of the streams involved, and probably also the most recent, was the postulated diversion of the Upper Green River³ from its ancestral easterly course. As a first step, by late Eocene or early Oligocene time, the Rock Springs uplift had been truncated by the Gilbert Peak erosion surface. Aspen Mountain remained as a monadnock, with a mantle of Bishop Conglomerate at its flanks. Drainage out of the Green River Basin was presumably eastward, across the present Continental Divide, for reasons discussed below.

---

³As the following discussion postulates, the part of the Green River above the mouth of Bitter Creek, near the town of Green River, Wyo., was once a separate stream and has a significantly different history from the downstream reaches of the Green. For simplicity and clarity, this part of the stream is here referred to as the "Upper Green River," even though, in more common usage, the Upper Green is regarded as the reach above the confluence of the Yampa River, the Green's chief tributary.

---

As postulated previously (Hansen, 1969b, p. 99), the slow rise of the incipient Continental Divide in post-Bishop time stagnated easterly drainage. The Great Divide Basin rose so much in late Pliocene time that it was largely stripped of post-Eocene rocks (Love, 1971, p. 79). As the eastern Uinta Mountains subsided, moreover, the new drainage along and across the range gradually extended its watershed, and the Upper Green River finally was captured near the present mouth of Bitter Creek at the town of Green River, Wyo. At that instant in geologic time the entire Green River system was turned south toward the Uinta Basin. Ritzma (1959, p. 88) has suggested that a rise of the Rock Springs uplift in Pliocene time and the extrusion of lavas over the north part of the uplift may have blocked easterly drainage. These lavas crop out in the Leucite Hills 30-45 km northeast of Rock Springs. They are now deeply dissected but were once more extensive, and they surely would have deflected any drainage in their path. Pilot Butte, 13 km north west of Rock Springs, is an outlier, base altitude about 2,380 m, resting on soft gray sandstone that resembles a fine facies of the Bishop Conglomerate. The base of the sandstone, altitude about 2,315 m, rests on the Laney Member of the Green River Formation. Pilot Butte itself is about 520 m above the nearby Green River. The Leucite Hills lavas have been dated at about 1.25 m.y. (Bradley, 1964, p. A58) and more recently at 1.1 m.y. (McDowell, 1971). Their age, therefore, is compatable with Pleistocene diversion of the Upper Green River, but they seem too high and too old, as explained later, to have played a direct part. They could, however, have shifted drainage southward away from an earlier hypothetical course toward the Sweetwater River.

As a topographic feature, the present Continental Divide is very subdued all the way from the Wind River Range to the Sierra Madre. The divide splits and encircles the Great Divide Basin, which has interior drainage and contains many ill-defined watersheds and closed depressions, some of which contain playas. If the Great Divide Basin were filled to overflowing, it would spill to the North Platte River across a broad threshold about 2,000 m above sea level. The low point in the west rim of Great Divide Basin is just west of Tipton, Wyo., at an altitude of about 2,070 m. This is the true position of the Continental Divide. It is about 245 m lower than the floor of the Gilbert Peak erosion surface near the center of the Green River Basin and about the same amount lower than the rims of Lodore Canyon at the Gates of Lodore. Just north of the Great Divide Basin the Sweetwater River, arising on the southwest side of the Wind River Range in an area that is structurally part of the Green River Basin, turns south, then east, and flows to the North Platte. Thus, even now, part of the drainage from the Green River Basin flows to the Gulf of Mexico, and has, according to Denson and Chisholm (1971, p. C125), since the end of the Eocene.

Once drainage turned across the Uinta Mountains at Lodore Canyon, and was thus flowing to the Uinta Basin, it was rapidly rejuvenated. The Uinta Basin is not only the deepest basin of the Colorado Plateau structurally (Hunt, 1956, p. 2), it is also some 300 m lower topographically than the Green River Basin. In late Pliocene time, as thus visualized, the main drainage of the Green River Basin (the Upper Green River) was still flowing eastward at a hydraulic disadvantage across the rising Rock Springs uplift and the rising incipient Continental Divide. The invigorated trans-Uinta drainage, however, with its new, lower base level, was incising itself into the Uinta Mountains, carving out the spectacular meander loops of Flaming Gorge, Horseshoe Canyon, and Red Canyon. Near Flaming Gorge it captured drainage that once flowed north into the Green River Basin, including Henrys Fork, Burnt Fork, and Birch Creek, as noted previously on page 59. These streams flow north off the Uinta Mountains, then turn southeast to join the Green. Other, lesser streams do the same (Hansen, 1965, p. 176). Some of the details can be gleaned from topographic maps of the area. Henrys Fork probably once flowed north through a gap between Hickey Mountain and Sage Creek Mountain ("Sage Creek Butte" in old reports; Bradley, 1936, p. 199). Burnt Fork and Birch Creek, now flowing independently into Henrys Fork, may have merged and flowed north through a gravel-covered gap between Cedar Mountain and Black Mountain. Working headward against the prevailing regional slope, the expanding new river system then captured the Upper Green.

This capture may have been a middle Pleistocene event, although modern indigenous fish species in the Upper Green River drainage differ markedly from those in the North Platte and the Sweetwater. Only the mountain sucker and two closely related daces are common to the two drainage systems, according to R. S. Behinke (Colorado State Univ., Dept. of Fisheries and Wildlife Biology, written commun., 1983). Any postulated connection, therefore, must take into account these differences.

The present Green River at Green River, Wyo., is flowing at an altitude of about 1,850 m, only about 215 m lower than the Continental Divide at Tipton, 90 km to the east. Extending up the Green from a point a few kilometers west of the town of Green River is a series of high gravel-capped terrace remnants. These remnants are about 120-180 m above modern drainage. Richmond (1948) correlated them with the Buffalo glaciation of the Wind River Range, though he later abandoned the name "Buffalo glaciation" because it represents three separate glaciations (Richmond, 1965, p. 218). The highest terrace, which will be mentioned further, is referred to as the "Peru bench" for its occurrence north of Peru station on the Union Pacific Railroad 10 km west of Green River (Hansen, 1969b, p. 99), where it has an altitude of about 2,040 m and is about 195 m above the river.

The gravel on the Peru bench contains an assortment of resistant rock types-particularly light-colored to nearly black quartzite and pink granite—derived mainly from the Wind River and Wyoming Ranges, including many pebbles that may have been reworked out of the Pinyon Conglomerate and (or) Pass Peak Formation. It contains only subordinate amounts of red quartzite from the Uinta Mountains, because little Uinta Mountain drainage ever joined the Green River above that point. (Blacks Fork once joined the Green just above or about at the Peru bench and shared a common flood plain with the Green at that level.)

Significantly, the Peru bench does not extend downstream from Green River. However, a high gravel-capped bench 50 km east at the Rock Springs airport, referred to here as the "Rock Springs bench," may be its easterly correlative and may thus mark the former course of the Upper Green River, roughly along the present course of Bitter Creek, which now drains that area in the opposite direction. Other gravel benches stand at intermediate points between the Peru bench and the Rock Springs bench. The gravel on the Rock Springs bench is very similar to that on the Peru bench, although it is not as coarse and contains fewer granite pebbles and more red quartzite. Perhaps the granite was worn away by attrition in its transit downstream from the Peru bench area; the red quartzite may have been augmented by material carried out of the Uinta Mountains by Blacks Fork and by material reworked out of the Bishop Conglomerate south of Rock Springs.

The Rock Springs bench is about 130 m above the level of Bitter Creek. There is little likelihood, however, that its gravel was deposited by a diminutive stream like the present Bitter Creek, which is manifestly underfit (Drury, 1964) and carries no gravel. A local source for the gravel, moreover, is improbable, because the predominant rock types are derived from Precambrian sources and are much the same as the rocks in the Peru bench. No source for such rock types exists within the drainage basin of Bitter Creek. A more likely source, therefore, was the Upper Green River, before its diversion to the south.

The most easterly deposit carrying these clasts known to me is at Creston Junction, east of the Continental Divide on U.S. Interstate 80, about 130 km east of Rock Springs, Wyoming. This deposit overlies a bed of volcanic ash (Sanders, 1975) sampled by J. D. Love and identified by Izett and Wilcox (1982) as Lava Creek B ash from the Yellowstone caldera complex. The 0.62-m.y. age of the Lava Creek B ash (Izett and Wilcox, 1982) indicates that the Upper Green River was still in its east-flowing course 600,000 years ago.

A continued late Tertiary and Quaternary uplift of the Continental Divide across southern Wyoming (Love, 1971, p. 79) probably was partly responsible for the Upper Green's capture and diversion. The gravel deposits rise toward the east, and their rise is a measure of the uplift of the Continental Divide. Thus, from a point near the town of Green River east to the Continental Divide near Creston, a distance of about 145 km, the gravel deposit rises about 155 m, at an average rate of about 0.94 m/km. Assuming that the gradient of the river was about the same before capture as now, about 0.9 m/km, the rise has been about 285 m in the 600,000 years since capture, or about 475 mm per thousand years. The tilting up toward the east has been about 1.84 m/km, or roughly 3.1 mm/km per thousand years.

RELATIONSHIP OF UPPER GREEN RIVER CAPTURE TO MODERN FISH FAUNAS

Any hypothesis proposing that the Upper Green River was captured as recently as middle Pleistocene time must account for marked differences between the indigenous fish fauna of the present upper Colorado-Green River drainage and that of the North Platte and its tributaries. A mixing of these faunas might be expected, but they are distinctly different. The Platte River is tributary to the large Mississippi-Missouri system and shares its faunal diversity, which includes many different families and species adapted to seasonally warm waters. It also includes a few cold-water taxa such as salmonids (trouts and whitefish) that are restricted to mountain headwaters in areas of geologically recent transfers across the Continental Divide from the Columbia River basin (Miller, 1958, 1965; Uyeno and Miller, 1963; Behinke and Benson, 1983). The North Platte, however, contains no native salmonids, for reasons mentioned below.

In contrast with the Missouri, the upper Colorado-Green drainage has a narrowly limited fauna of only four families and 14 species adapted to two separate habitats: a lower altitude, seasonally warm-water habitat and a higher altitude, cold-water habitat. The warm-water species are endemic: that is, they are restricted to the Colorado drainage system and have evolved adaptations unique to that drainage, owing to long isolation from other river faunas (Behnke and Benson, 1983). These species are now greatly depleted in range and numbers, owing to habitat changes caused by river impoundments and irrigation, artificial population controls, and the introduction of highly competitive exotic species. Some of the endemic, warm-water species are endangered, such as the Colorado River squawflsh (Ptychocheilus lucius), the humpback chub (Gila cypha), the bonytail chub (Gila elegans), and the razorback sucker (Xyrauchen texanus), which were abundant in the Green as far north as Green River, Wyo., before the construction of Flaming Gorge Dam (Behinke and Benson, 1983). They are adapted to turbid, seasonally warm water, and their former ranges were mostly restricted to the main stem of the Colorado River and the downstream reaches of its larger tributaries, such as the Green and the Yampa, where the water is suitably warm in summertime to sustain the habitat (Behinke and Benson, 1983).

Nonendemic fish native to the upper Colorado-Green drainage include cold-water, clear-water species transferred from the Columbia River basin in geologically recent time, such as the cutthroat trout (Salmo clarki), mountain whitefish (Prosopium williamsoni), speckled dace (Rhinichthys osculus yarrowi), and sculpins of the genus Cottus. None of these fish is indigenous to the North Platte drainage, although they thrive elsewhere in the headwaters of the Missouri River basin. All have close ties to the Columbia River basin.

A mountain sucker, Catostomus platyrhynchus, is native to both the Green and the North Platte drainage (R. J. Behnke, Colorado State University, Department of Fisheries and Wildlife Biology, written commun., 1983; Behinke and Benson, 1983) and poses a perplexing question, therefore, of interbasin fish dispersal. This fish also formerly lived in Picket Lake—a small lake without outlet in the Great Divide Basin—along with the lake chub, Couesius plumbia, which was native to the Sweetwater River but not the Green. Both species have been extirpated from Picket Lake by introduction of exotic species (G. T. Baxter, University of Wyoming, written commun., 1983).

Because of the marked differences in the native fish faunas of the Green and the North Platte, the Upper Green is postulated to have been captured at a time and under circumstances unfavorable to the interbasin transfer of fish populations. The least favorable time would have been at the height of a glacial stage, when the riverine environment was cold and hostile and the headwaters were locked in ice. At such a time the Green River Basin was a frigid semidesert, cold in summer and bitterly cold in winter. Widespread relict frost wedges, frost polygons, and remains of cold-climate terrestrial fauna suggest the existence of permafrost and a periglacial tundra throughout the intermontane basins of Wyoming during late Pleistocene time (Mears, 1981). Malde (1961) noted similar evidence in western Idaho and adjacent areas. Mears calculated that the average Wyoming temperature in late Wisconsin time was 10°-13°C colder than now. Similar conditions probably prevailed during pre-Wisconsin glaciations.

The Green River Basin today remains one of the coldest areas of Wyoming, winter and summer (Lowers, 1978). Hardly in jest, the local people proclaim Big Piney, Wyo., the nation's icebox. The prediversion fish of the Green River Basin—linked to the North Platte and adapted to more moderate climates, therefore—would have been forced far downstream to warmer waters during the climatic deterioration of a glacial stage, leaving the basin void of fish life. East of Wyoming, the midcontinent was ice covered as far south as northeastern Nebraska during the Wisconsin stage, and even farther south during earlier glaciations (Flint, 1955, fig. 27; Mickelson and others, 1983).

Eroding northward from Flaming Gorge toward Green River, Wyo., the stream that captured the Upper Green River probably was ephemeral and contained no viable fish population. Inasmuch as its watershed was the dry, central part of the Green River Basin—even drier in late Wisconsin time than now (Gates, 1976)—it probably held little if any water except during prolonged wet periods or after heavy showers, when its flow would have been large and energetic. Moreover, because it was working headward through the saline Wilkins Peak Member of the Green River Formation, it would have contained high concentrations of alkaline salts. I, therefore, postulate a vigorous but ephemeral stream, devoid of fish life, capturing a much larger but also fish-free Upper Green River whose indigenous North Platte fish population had been forced far to the east, out of the Green River Basin, by the harsh Pleistocene climate and, perhaps, by the turbidity of the glacial meltwater. At that instant in geologic time, the entire drainage of the Upper Green was diverted south through the Uinta Mountains, and the diversion was without any transfer of fish populations. Then, with the warming of the interglacial climate, the endemic warm-water species of the Lower Green began to colonize the Green River Basin.

The native but nonendemic cold-water fishes of the headwaters and mountain tributaries of the Colorado-Green River system—the cutthroat trout, mountain whitefish, speckled dace, mountain suckers, and sculpins—are all regarded as geologically recent entrants into the drainage system, inasmuch as they have evolved but little since their interbasin transfer (Behinke and Benson, 1983, p. 9). Because they are absent from the North Platte drainage (except the mountain sucker, Catostomus platyrhynchus), they must have migrated into the Green from the Columbia River basin by way of the Snake River, rather than by way of the Missouri, after the Upper Green was captured. Though absent from the North Platte, all these fish are native to the Upper Missouri; Marias Pass in northwestern Montana was an obvious transfer point across the Continental Divide, according to R. S. Behnke (written commun., 1983), until it was blocked by railroad construction early in the 20th century. The cutthroat trout, the species most adapted to cold water and least tolerant of warmth, never became established much below Great Falls, Mont. The whitefish reached the mouth of the Yellowstone but no farther down the Missouri, and the mountain sucker, the most warm-tolerant species, became widespread throughout the upper Missouri, including the Sweetwater (R. S. Behnke, written commun., 1983).

The mountain sucker may have entered the Platte drainage (the Sweetwater River) from the Wind River, according to G. T. Baxter of the University of Wyoming, Department of Zoology (written commun., 1983). If so, it should have transferred from the Columbia by way of the Flathead, the Missouri, and the Yellowstone. This fish tolerates the warm water of the Missouri and the lower Yellowstone, but the salmonids and sculpins do not, hence they were unable to reach the Sweetwater. The mountain sucker might also have traveled all the way down the Missouri to the mouth of the Platte, though the turbidity of the lower Platte would have been a deterrent to its further migration.

In any event, the cold-water fauna of the Green River must be high-altitude transfers from the Snake, since they are not natives of the Platte. The mountain sucker probably entered the Green along with the rest of the cold-water fauna, but being native to the Missouri and the Snake, as well as the Green, it theoretically could have made the transfer in either direction. The sucker family Catostomidae has a nearly continent-wide Nearctic distribution, from Arctic Canada to Guatemala (Miller, 1958, fig. 10; Patterson, 1981, p. 274).

The time (or times) of fish transfer to the Green River is uncertain. At the time of any pre-late Wisconsin fish transfer, the topography of the interbasin divide would have been somewhat different from now, and possible transfer points accordingly would have been different also. A likely Holocene or very late Wisconsin transfer point is a broad, marshy morainal area straddling the Snake-Green divide in the extreme northwest corner of Sublette County, Wyo., drained by Raspberry Creek on the Snake River side and Wagon Creek on the Green River side (Mosquito Lake 7-1/2—minute quadrangle). The headwaters of these streams intermingle in a maze of ponds and bogs left behind by the withdrawal of the large late-Wisconsin (Pinedale) Green River glacier (G. M. Richmond, oral commun., 1984). Raspberry Creek is a first-order tributary of the Gros Ventre River, which in turn joins the Snake in Jackson Hole, Wyo. Wagon Creek flows directly to the Green. A few kilometers to the southwest, at another possible transfer point, Tepee Creek (Green River drainage) and Kinky Creek (Gros Ventre drainage) are separated by a chain of lakelets along another obscure divide in the same Pinedale moraine complex.

In one late Wisconsin transfer scenario, partial withdrawal of the Green River glacier from its terminus formed small proglacial lakes on the Wagon Creek side of the divide. These lakes then spilled north across the low divide into Raspberry Creek, allowing fish to migrate across to the Wagon Creek (Green River) side. When the ice melted back far enough to drain the lakes, the divide closed to further transmigration, and the fish moved into the Green River drainage. Whether or not a viable fish population could survive that close to the glacier terminus, however, is uncertain. The Colorado River subspecies of the cutthroat trout (Salmo clarki pleuriticus) requires a water temperature of about 7.2°C for spawning (Behnke and Benson, 1983, p. 29), and the meltwater temperature would have been close to freezing. Moreover, the turbidity of the meltwater would tend to bury and suffocate the eggs and suppress the growth of aquatic plants and small invertebrates needed to sustain a food chain. More likely, therefore, well after deglaciation, a physical connection through the bogs and ponds allowed fish to move freely from one drainage basin to the other. Such dual drainage is not uncommon in glaciated areas. Sedimentation in the ponds and gradual expansion of the bogs at the expense of the ponds eventually closed off the transfer routes.

Various low passes along the present Snake-Green River divide, including points along the headwaters of the Hoback River south of Jackson Hole, Wyo., suggest other possible earlier Pleistocene transfer points, some probably involving minor headwater stream captures. Transfers might have happened at more than one place and time. Twin Creek, near Kemmerer, Wyoming, is another early possibility, which could have linked the Bear River and Hams Fork of the Green River, at a time when Hams Fork was flowing at a terrace level about 100 m above present grade (Rubey and others, 1975). At that time the Bear must still have flowed into the Snake. Two Ocean Pass, on the Continental Divide just south of Yellowstone National Park, is a modern transfer point from the Snake to the upper Yellowstone River and perhaps also to the lower Yellowstone and the Missouri (Evermann, 1892, p. 28; Jordan and others, 1930), but the Upper and Lower Falls of the Yellowstone prevent counter migration from the Missouri to the Snake. Few fish, moreover, probably survive the descent of the Lower Falls, which plunge 94 m to a rocky bed.

Fish transfers from the Snake River to the Green most likely took place during one or more interglacial stages, inasmuch as the Snake-Green divide was partly ice clad during glacial maximums, as already noted, and the streams that remained open probably were heavily laden with sediment downstream from the glaciers. These fish prefer clear water and have low tolerance of turbidity. The subsequent dispersal of the cold-water fauna throughout the upper Colorado River basin, however, probably followed during a later glacial stage (Pinedale?), when the downstream water temperature throughout the basin was lowered enough to be tolerated. Meanwhile, the warm-water fauna would have been forced farther downstream. Then, as the water temperature slowly rose again during a succeeding interglacial, or the postglacial, the cold-water fauna would have abandoned the trunk drainages and moved gradually into the high-altitude, cold-water tributaries throughout the upper Colorado River basin. This event may have been Holocene.

QUATERNARY REENTRENCHMENT OF THE GREEN RIVER

A postscript to the diversion of the Green River out of the Green River Basin and across the Uinta Mountains is the relatively brief history of the canyons since then. Field relations suggest reentrenchment of the Green River since the initial cutting of the canyons, especially of Lodore. Despite the steepness of the walls of Lodore Canyon, cross-canyon topographic profiles show a distinct slope break, steepening downward, about 245-300 m above the canyon floor (frontispiece). Profiles in Red Canyon are comparable to those in Lodore. The compound nature of the profile is most pronounced down spurs. Unrelated to lithology or stratigraphic shelving, the break in the profile is regarded as a result of accelerated dowincutting and reentrenchment by the river, probably in early or middle Pleistocene time, after the canyon had been partly excavated (Hansen, 1969a, p. 48; Hansen and others, 1982). Thus, after Lodore Canyon had been eroded to a depth of about 460 m, downcutting slowed and the river began to widen the valley bottom, but the river was then rejuvenated, and it cut down an additional 245 m or so to its present level. This figure is comparable to the amount of canyon deepening along the Colorado River in the Rocky Mountains in Quaternary time—230 m, according to Hunt (1969, p. 72).

Tributaries of Lodore Canyon responded to rejuvenation by renewed downcutting also, which led to breaks in their gradients at heights above the Green River in accordance with the breaks in the cross-canyon profiles (Hansen and others, 1982).

The cause and timing of renewed downcutting are uncertain. Regional uplift in response to unloading or some other factor would have rejuvenated the entire drainage system through the mountains. The capture of the Upper Green River could also have caused reentrenchment by greatly increasing the discharge and competence of the system. Capture and uplift may both have been involved. If initial cutting began in latest Miocene time after the Uinta Mountains were overtopped at Lodore Canyon, most of Pliocene time may have elapsed during the erosion of the upper 460 m of canyon. Reentrenchment, then, may have begun in early Pleistocene time.

The rate of canyon cutting can be inferred. If 760 m is taken as the mean depth of Lodore Canyon and if cutting took all of Pliocene and Quaternary time—about 5 million years—then the average rate was about 15 cm per thousand years. This rate compares favorably with that of the Black Canyon of the Gunnison River in central western Colorado, where the estimated rate was about 30 cm per thousand years (Hansen, 1965) but where the river, though smaller, has a much steeper gradient.

The actual cutting surely involved times of faster and slower erosion. Differential rates of Quaternary cutting are thus expressed by terracing upstream and downstream from the canyons. Terrace remnants are uncommon in the canyons but are abundant and well preserved in the softer rocks outside the mountains and in Browns Park (fig. 47). Climatic fluctuations contributed to Quaternary terracing, inasmuch as the terrace deposits correlate with episodes of mountain glaciation, but regional uplift must also have contributed to downcutting and terracing. The highest gravel-capped surfaces stand more than 150 m above river level, and downcutting of that magnitude can hardly be ascribed solely to climatic change. I favor the view that the terraces represent episodic aggradation in an otherwise degradational regime, when downcutting was temporarily overwhelmed by increased sedimentation.

FIGURE 47.—Quaternary terraces in the Eastern Uinta Mountain area. Top.—Quaternary terraces along Brush Creek northeast of Vernal, Utah. (See Kinney, 1955, p. 128). Remnant at upper right is about 120 m above Brush Creek. Diamond Mountain Plateau, on the skyline, is capped with Bishop Conglomerate more than 760 m above Brush Creek; its top appears uneven here only because the irregular rim of the plateau is viewed from below. Bottom.—Gravel-capped Quaternary pediment remnants—rock fans—along northeast side of Browns Park at Utah-Colorado State line. Canyon of Beaver Creek, upper left. Cold Spring Mountain on the skyline. The highest Quaternary pediment remnant in the background is about 90 m above the level of Beaver Creek and about 150 m above the Green River, out of view at right (Hansen, 1965, p. 132, plate 1).