Quaternary incision rates and drainage evolution of …eps et al_RockyMtnGeo... · Quaternary incision rates and drainage evolution of the Uncompahgre and Gunnison Rivers, western

Quaternary incision rates and drainage evolution of the Uncompahgre and Gunnison Rivers, western Colorado, as calibrated by the Lava Creek B ash

Andrew L. Darling1*, Karl E. Karlstrom1, Andres Aslan2, Rex Cole2, Charles Betton3, and Elmira Wan4

1Department of Earth and Planetary Sciences, Northrop Hall, University of New Mexico, Albuquerque, New Mexico 87131, U.S.A.2Department of Physical and Environmental Science, Mesa State College, Grand Junction, Colorado 81501, U.S.A. 3Grand Junction Geological Society, Grand Junction, Colorado 81501, U.S.A.4U.S. Geological Survey, MS 975, Menlo Park, California 94025, U.S.A.

*Correspondence should be addressed to: [email protected]

ABSTRACT

The Quaternary erosional history of western Colorado is documented in terraces of the Colorado, Gunnison, and Uncompahgre Rivers that contain the Lava Creek B ash (0.64 Ma). This paper reports an important new ash locality that dates ca. 100-m-high river gravels associated with the paleo-confluence of the Gunnison and Uncompahgre Rivers upstream from Grand Junction. Provenance analysis reveals paleo-Gunnison River gravels (containing granite and gneiss clasts) and paleo-Uncompahgre River gravels (containing Uncompahgre Group quartzite and San Juan volcanic field rocks). The paleo-Uncompahgre River gravels are 3 m directly beneath Lava Creek B ash, and the areal distribution of terraces indicates that this area was the paleo-confluence between the Gunnison and Uncompahgre Rivers. This confluence has shifted 11 km to the east since 0.64 Ma due to events related to stream piracy and drainage reorganization. Gunnison terrace straths near the paleo-confluence are estimated to be 106 m above the modern strath, giving an estimated incision rate of 165 m/Ma.

Because of excellent age and geologic control, this is one of the best incision-rate data points in the upper Colorado River system. It is similar to previously reported regional rates, but is substantially lower than upstream incision rates in the Black Canyon of the Gunnison River. This dated Gunnison River terrace anchors the projection of Lava Creek B-bearing Grand Mesa pediment surfaces (e.g., Petrie Mesa) to regional base level and helps constrain a regional reconstruction of the 0.64-Ma profile of the paleo-Gunnison River. This reconstruction shows dramatic differences in incision rate in the Gunnison River system since 0.64 Ma, and that a transient knickpoint migrated past Sawmill Mesa prior to 0.64 Ma. This incision data point has important implications for evaluating major Quaternary changes in the configuration of this part of the Rocky Mountain drainage system. It also provides evidence for a young, disequilibrium drainage system that is responding to base-level changes downstream driven by a stream capture event, which in turn may have been driven by tectonic or climatic perturbations.

KEY WORDS: Colorado, Gunnison River, incision rates, Lava Creek B ash, Quaternary, river gravels, river ter-races, Uncompahgre River.

INTRODUCTION

The Gunnison River is the largest tributary to the upper Colorado River system, which drains the western slope of some of Colorado’s highest topog-raphy in the Rocky Mountains (inset to Fig. 1). This system has been the focus of classic studies of west-

ern geomorphic history of the United States (e.g., Hunt, 1969; Lohman, 1981). Widespread occur-rences of the Yellowstone Lava Creek B ash in river terraces have been used to infer regional variations in incision rate in this system (Dethier, 2001). These, plus more recent studies, have emphasized that the Colorado River system, as well as the entire region of

Rocky Mountain Geology, v. 44, no. 1, p. 71–83, 6 figs., 1 table, May, 2009 71

the Rocky Mountain–Colorado Plateau, are under-going differential incision, with Quaternary incision rates varying from 50–500 m/Ma. In the Grand Canyon region, differential incision has been posited to reflect neo-tectonic slip on normal faults (Pederson et al., 2002; Karlstrom et al., 2007), perhaps related in part to asthenosphere-driven regional epeirogenic uplift (Karlstrom et al., 2008). Alternative models to explain Quaternary drainage reorganization and youthful topography involve climatic and geomor-phic forcings involving glacial-interglacial f luctua-tions acting on previously elevated high topography (Molnar and England, 1990). In both scenarios, accompanying the observed differential incision is a system of migrating transient knickpoints (Zaprowski et al., 2001; Sandoval, 2007) that are responding to stream capture (Lohman, 1981) and/or local to regional tectonic influences (Kirkham et al., 2002, Karlstrom et al., 2005; McMillan et al., 2006). Thus,

improved data on incision rates for many reaches of the rivers are needed to help interpret the interplay among geomorphic, climatic, and tectonic influences on the entire river system.

Fluvial geomorphology in western Colorado is largely characterized by climatically controlled aggra-dational and incisional cycles superimposed on an overall regional signal of denudation/incision. Cycles of aggradation and incision are generally linked to glacial and interglacial oscillations (Sinnock et al., 1981; Dethier, 2001; Sharp et al., 2003, and references therein) and result in a series of inset strath terraces. Thus, older terraces are consistently higher in eleva-tion than younger terraces (Bull, 1991). The numer-ical ages of these terraces are becoming better con-strained through application of various Quaternary dating techniques, which allow workers to use these terraces as a tool for studying landscape evolution (Dethier, 2001; Sharp et al., 2003; Wolkowinsky and

108 W

108 W

39 N 39 N

UT

AZ NM

CO

WYID

West ElkMountains

Grand Mesa

Petrie Mesa

Black Canyon of the Gunnison

GrandJunction

Gunnison River

San Juan Mountains

BostwickkeerC -oelaP

Uncompahgre Plateau

Uncompahgre River

Colorado River

Area of Figure 2

DeltaCactusPark

Lake Fork

Unaweep Canyon

LCB ash locations

N

50 kilometers

Figure 1. Hillshade digital elevation model of study area. Kelso Gulch and Sawmill Mesa are located northwest of Delta, Colorado, on the flank of the Uncompahgre Plateau. Kelso Gulch flows into the Gunnison River downstream from confluence with the Uncompahgre River. Important source areas for gravels include the San Juan Mountains, West Elk Mountains, Grand Mesa and Black Canyon, and upstream tributaries on the Gunnison.

72 Rocky Mountain Geology, v. 44, no. 1, p. 71–83, 6 figs., 1 table, May, 2009

A. L. DARLING AND OTHERS

Granger, 2004; Sandoval, 2007). In addition to cli-matic effects, this pattern of terrace distribution is variably attributed to base-level fall due to integra-tion of lower and upper Colorado River basins within a previously elevated plateau region (Pederson et al., 2002; Frankel and Pazzaglia, 2006) and/or tec-tonic bedrock uplift of the Rockies (Steven, 2002; McMillan et al., 2006). Isostatic response to denu-dation adds additional rock uplift and is an ongo-ing feedback from regional denudation (Pederson et al., 2002). The magnitude of the isostatic feedback is small (perhaps several hundred meters), however, compared to the larger denudational signal (thou-sands of meters; Leonard, 2002; Heller et al., 2003).

Along the Gunnison River’s canyon northwest of Delta, Colorado (Fig. 1), new observations have been made regarding complex fluvial processes. This paper reports a significant series of river terraces that is interbedded with deposits of the Lava Creek B ash, known to be 639 +/-2 ka (Lanphere et al., 2002; cited elsewhere in this paper as 0.64 Ma). The goal of this study is to analyze field relationships, provide an accu-rate Quaternary incision rate for the lower Gunnison River, and correlate Lava Creek B-age terraces in the region. This paper complements other regional stud-ies of incision (Fig. 1) that have been done using ash deposits on Petrie Mesa and other mesas (Baker et al., 2002; Rider et al., 2006), in Bostwick Park (Kelley et al., 2007), and using both ash and burial-cos-mogenic dating near Black Canyon of the Gunnison (Sandoval, 2007) and in Cactus Park (Aslan et al., 2008b).

REGIONAL SETTING

The study area, near Sawmill Mesa (Fig. 2), is on a reach of the Gunnison River upstream from Grand Junction and northwest of Delta in western Colorado. Tributaries of the Gunnison River f low south from the Sawatch Range through Oligocene volcanic rocks of the West Elk Mountains and north from the San Juan Mountains. The Gunnison River passes through the Precambrian-cored Black Canyon of the Gunnison, and then flows across Cretaceous Mancos Shale near Delta, Colorado, where it joins with the Uncompahgre River. The combined rivers f low into a gorge cut into Mesozoic rocks in the study area. Bounding the river system on the north is Grand Mesa, a basalt-capped erosional remnant

of 10 Ma basaltic lava f lows. The drainage system of Figure 1 is dominated by north-f lowing rivers, like the Uncompahgre River, which flow from the Uncompahgre Plateau and San Juan Mountains and intersect a west-flowing trunk system (the Gunnison), which then joins the Colorado system near Grand Junction (Sinnock, 1981; Betton et al., 2005).

METHODS

Bedrock–strath terrace contacts (strathlines) were mapped at 1:12,000 scale based on: (1) sharp con-trasts between bedrock and overlying gravel; and (2) changes in vegetation and soil color between bed-rock and terrace gravels. Soils developed on bedrock are thin, yellow, alteration zones. Hand-tool exca-vation was used to precisely determine strath loca-tions. Strathlines were mapped using a GEO-XT GPS unit for precise elevation and spatial control. These data were differentially corrected to maintain accuracy to 1–2 m. Strath data points were projected onto a down-valley profile line. The terraces line up in map view, are correlated based on geometry, and interpreted to represent abandoned f loodplains. Two down-valley distance lines were drawn by fol-lowing the center-line of these linearly grouped ter-races (Fig. 2). Elevation data points from straths were projected onto these lines, plotted on a longitudinal profile, and compared with the modern profile of the Gunnison and Uncompahgre Rivers. Profiles for the modern rivers were taken by following the middle of the river as shown on USGS topographic maps. Note that river profiles constructed in this manner may result in profiles of lower gradient than profiles derived from terrace straths. Because the latter mea-sures paleo-floodplains, meandering paleo-channel locations cannot be observed easily. If only the down-valley distances were used for both modern rivers and terraces, then over-steepened reaches would appear in alluvial stretches of the modern river profile because the channel may meander and/or anastomose. Thus, a modern river profile is not directly comparable to an inferred profile from terraces in the absence of careful estimation of down-valley distances associ-ated with the river.

Time correlation of the terrace remnants is based on geometry of these projected straths and presence of Lava Creek B ash. Pebble-count data were collected at three locations to understand the effect of litho-


QUATERNARY INCISION RATES AND DRAINAGE EVOLUTION, WESTERN COLORADO

Qt6G

Qt6G

Qt5G

QtUlg

Qt8U

Sawmill

Mesa

Twenty

-Five

Mile

Mesa

Cottonwood C

reek

Qt3G

Kelso Gulch

Gunnison River

Qt6U

Qt3G

Qt1G

Do

wn

-val

ley

line

Down-valley line

pebble count site

pebble count sites

Subject of F

igure

3

Qt4GQt1G

Qt2G

Qt3G~

~ 0.64 Ma confluence

>0.64 Ma confluence



logic sources on bed-load composition in the drain-age basin. These counts were done on the Qt6G and Qt6U terraces, and locations are marked on Figure 2. Each count exceeded 100 clasts. Results of the pebble counts were not used directly to identify ter-races. Rather, they were used to find a few clast types that distinguish the paleo-Uncompaghre from paleo-Gunnison River source areas.

Incision rates calculated in this paper involved measuring the height difference between straths and the modern river surface at approximately mean dis-charge, but they also consider depth to bedrock below the river (Karlstrom et al., 2007). Thus, our reported incision rate (e.g., 165 m/Ma) represents a median bedrock-incision rate. For example, if we assumed a mean depth to bedrock of 10 m, the effect would be an increase in incision rates by 10 percent for our localities. Where Lava Creek B ash overlies gravel, ages represent minimal ages of straths. However this error is interpreted to be small (< a few percent) because, locally, ash-bearing units rest above gravels with no observed intervening unconformity. Most of the region’s data points for tributary incision involve uncertainties of projection in how side streams and pediments graded to the main stem. This emphasizes the importance of locations, such as the one reported in this study, in which strath terraces are directly above the modern river.

QUATERNARY UNCOMPAHGRE AND GUNNISON RIVER TERRACES

Geologic Map

Results of the field observations, including gravel roundness, size, and provenance show eight dis-tinct terrace levels of the ancestral Gunnison and Uncompahgre Rivers. Terrace remnants consist of

river deposits of gravel and sand/silt mantling bed-rock (Fig. 2). On Sawmill Mesa and Twenty-Five Mile Mesa, the river gravels are buried by locally derived, fine-grained sediments (map unit Qtfg) and/or angular to sub-rounded, sandstone gravels (map unit Qtlg, Fig. 3) that are interpreted to rep-resent local deposits post-dating abandonment of the terraces and upstream movement of the conflu-ence. River-gravel deposits are typically 2–8 m thick, and straths range in height from 5–120 m above the modern Gunnison River. Main-stem river gravels are rounded to sub-rounded cobbles with a sandy matrix and lesser amounts of boulder- and pebble-sized clasts. The base of the main-stem river gravels consists of an erosional strath cut into underlying bedrock. Remnants of terraces Qt6U and Qt7U are main-stem terraces locally trending north–south, roughly perpendicular to the modern Gunnison River. On and near Sawmill Mesa, Uncompahgre and adja-cent Gunnison River terraces (map units Qt6U and Qt6G; Fig. 2) are at similar elevations.

Remnants of younger terraces (map units Qt1G–Qt5G) of the ancestral Gunnison River are concen-trated along the present-day course of the Gunnison River and are sub-parallel to the modern channel (Fig. 2). Small gravel deposits inset near the bottom of Kelso Gulch (map unit QtUlg) consist of near-angular to subrounded sandstones and inter-mixed, sub-rounded volcaniclastics.

Ash Occurrences, Tephrochronology, and Terrace Ages

The Lava Creek B ash bed (0.64 Ma) originated from one of several Yellowstone Plateau plinian erup-tions that produced extensive ashfall over much of the west-central United States (Izett and Wilcox, 1982). In western Colorado, the Lava Creek B ash is associ-ated mainly with tributary channels and alluvial fans. These sites of rapid aggradation and burial provided areas of preservation for the ash (Dethier, 2001). Occurrences of Lava Creek B ash are rare in main-stem gravels due to peculiarities of preservation. In the study area, tephra deposits occur in two sites. The Lava Creek B ash was identified at one of the sites by petrographic and geochemical analyses.

One occurrence of ash is near the top of Twenty-Five Mile Mesa (elevation 1,608 m), in the lower half of a fine-grained, sandy-silt to silty-clay unit (Fig. 2; map unit Qtfg) that clearly overlies Uncompahgre

Figure 2, facing page. Geologic map showing terraces of paleo-Uncompahgre and paleo-Gunnison Rivers. Estimated conflu-ence locations are diagramed by circles. Incision-rate elevation datapoints denoted by stars. Relative age of terraces increase with increasing number (Qt1–Qt8). QtUlg = Quaternary ter-race of Uncompahgre River and local gravel combined; Qtlg = Quaternary terrace of local gravel; Qtfg = Quaternary ter-race of fine-grained alluvium, contains Lava Creek B Ash; Qt6G = Quaternary terrace of Gunnison River alluvium; and Qt6U = Quaternary terrace of Uncompahgre River alluvium. Grid coordinates are Universal Transverse Mercator (UTM), NAD83 Datum.



River gravels (Fig. 2; map unit Qt6U, Fig. 3). This fine-grained unit is up to eight meters thick, armored by local gravels, and pinches out to the north (Fig. 2). No unconformity was observed between the gravel and the overlying, ash-bearing, fine-grained sedi-ments, suggesting that the ash age is a good proxy for age for the gravels, assuming this ash is also LCB. No ash is directly associated with ancient Gunnison River gravels (map unit Qt6G), but the elevations of Qt6U and Qt6G in locations directly above the modern channel (stars in Fig. 2) correlate well, sug-gesting that Qt6G is also 0.64 Ma.

The second, and geochemically analyzed, occur-rence of Lava Creek B ash is in Kelso Gulch, along sloping hillsides slightly above the valley floor (Fig. 2). The tephra layer intermittently follows the contour of the hillslope at an elevation of 1,591 m. It is variably cemented with calcite and up to 5 cm thick. At this locality, geochemical confirmation of the Lava Creek B ash by co-author Wan (Table 1) comes from sample K06CO3, collected from an indurated, ca. 5-cm-thick ash bed exposed on a hillside (Fig. 2). This ash bed is thinly mantled by slope-wash. Other scattered exposures of ash are mingled with debris from rodent burrows. The ash occurs close to, and appears to be at a slightly higher elevation than, Uncompahgre River gravels (map unit Qt7U; Fig. 2). The stratigraphic

positions of the ash horizon and the Qt7U gravel are not well exposed. The strath profile of Qt7U corre-lates well with elevation of the ash based on the geom-etry presented in Figure 4. Since there is no direct gravel-to-ash association in Kelso Gulch, we infer that the ash blanketed this area and was preserved on hillslopes sometime after abandonment. Thus gravels in Kelso Gulch could be much older than LCB ash.

Processing, petrographic analysis, and geochemi-cal fingerprinting of tephra sample K06CO3 and its identification as the Lava Creek B ash was performed at the USGS Tephrochronology Laboratory and the Electron Microprobe Laboratory in Menlo Park, California. The chemical results and closest matches (similarity coefficient; Sarna-Wojcicki et al., 1984) are shown in Table 1. Based on the wet-sieved and sized (-100+200) fraction, approximately 95 percent of the volcanic glass shards of sample K06CO3 are angular to subangular, lightly to heavily coated (with carbonate, organics, clay, and FeO), with mostly solid, platy, or bubble-wall junctions. The shard mor-phology is similar to Lava Creek B samples collected at Bostwick Park (Fig. 1) by Karlstrom, except that the Kelso Gulch sample contains a greater number of vesiculated, ribbed, and webby pumiceous shards. Also, the mineral assemblage from Bostwick Park is less varied, containing fewer heavy-mineral grains.

X X X X X X X X

Ele

vatio

n(m

)

Zone with Lava Creek B ash (ca. 640 ka)

1614

1613

1608

1605

1603

Sandstone-rich local gravel (map unit Qtlg; see Fig. 2)

Tan, fine-grained, silty sandto clayey silt (map unit Qtfg; see Fig. 2)

gravel (Qt6U; see Fig. 2)

Cretaceous Dakota Sandstone

Uncompahgre River

Figure 3. Measured section showing late Quaternary stratigraphy of Twenty-Five Mile Mesa, location marked on Figure 2 (see near lower right-hand corner of map). Qtlg = local gravel; Qtfg = fine-grained, ash-bearing unit; and Qt6U = Quaternary terrace Uncompahgre River gravel.



1450

1500

1550

1600

1650

1700

60 65 70 75 80 85 90 95

Elev

atio

n (m

)

Gunnison

Uncompahgre

Qt5G

Qt6G

Qt6U

Petrie Mesa Projections

Kelso Area Ash Locations

Qt7U

Rider et al., 2006

Baker et al., 2002

A

B

Distance from Confluence with Colorado River (km)

Figure 4. Long profiles based on terrace strath measurements made with GEOXT GPS unit. Lines in support of two models (‘A’ and ‘B’, see discussion in text) are shown for alternate hypotheses of the gradient of the post-0.64-Ma Gunnison profile.



Gravel Provenance and Source Correlation

Uncompahgre River terrace gravels consist largely (ca. 70 percent) of intermediate-composition volcaniclastics derived from the San Juan volcanic field (Lipman, 2007; Fig. 5). Key index lithologies in Uncompahgre River deposits are red, arkosic sand-stone of the Permian Cutler Formation, white to gray, locally cross-bedded quartzite derived from the Paleo-Proterozoic Uncompahgre Group, and sparse, coarse-grained, reddish, alkali-feldspar granite. Tan, mature sandstone and basaltic andesite cobbles also are pres-ent (ca. 1 percent).

Gunnison River terraces hold a substantial per-centage (ca. 50 percent) of intermediate volcanic grav-els that were derived from the West Elk Mountains (Fig. 5). The volcanic rocks of the West Elk Mountains are similar in age (35–25 Ma) and composition (domi-nantly andesite and rhyolite), and they are difficult to distinguish from the San Juan volcanics. Source of the basalt was from the North Fork of the Gunnison River, which is f lanked by basaltic, gravel-rich fan complexes and terrace remnants derived from Grand Mesa (Yeend, 1969; Cole and Sexton, 1981; Betton et al., 2005; Rider et al., 2006). The most diagnos-tic compositional difference between gravels of the Uncompahgre and Gunnison River terraces is the higher abundance of metamorphic schist and gneiss (Fig. 5) in those of the Gunnison. Felsic cobbles in Figure 5 include rhyolite, diorite, granitoid, and occa-sional vein quartz. Felsic rocks such as rhyolite, granite, and diorite occur in the Gunnison River terraces, while rare vein quartz, granite, and rhyolite cobbles/pebbles occur in the Uncompahgre. The present Gunnison River erodes Precambrian crystalline rocks in the Black Canyon of the Gunnison and the Sawatch Range, and thus it transports a significant percentage of metamor-

phic clasts (ca. 12 percent; including gneissic and schis-tose) and granitoid (ca. 18 percent) cobbles (Fig. 5). In contrast, gneissic and schistose clasts were not observed in Uncompahgre River gravels, and reddish alkali-feld-spar granite clasts are rare.

Time Correlation of Terraces

Using the terrace-gravel composition to distin-guish deposits from the two paleo-rivers, Figure 4 shows proposed correlations and reconstructed long profiles of the ancestral Uncompahgre and Gunnison Rivers. Information from selected terrace remnants, as well as gradients of the modern Gunnison and Uncompahgre Rivers, were also brought to bear. The selection rule for the terraces omitted from the pro-file is based on their overall relevance to longer-term incision rates. Long-profile data points for terrace remnants younger than Qt5G were omitted, because they are younger than the widespread Lava Creek B datum that is the focus of this paper. Long profiles of paleo-Uncompahgre and paleo-Gunnison Rivers merge down-valley (Fig. 4). In map view, terrace remnants (map units Qt6U and Qt6G; Fig. 2) of the paleo-Gunnison and paleo-Uncompahgre Rivers are parallel and immediately adjacent, which indicates that the circled areas in Figure 2 represent the paleo-confluence of these two rivers.

DISCUSSION OF QUATERNARY INCISION RATES

Sawmill Mesa deposits, correlated to the Twenty-Five Mile Mesa deposits, provide an excellent oppor-tunity to evaluate Quaternary incision rates of prom-inent rivers in western Colorado as well as evolu-tion of the areas of confluence of the Gunnison and

Similarity Mean Total Sample ID Location SiO2 Al2O3 Fe2O3 MgO MnO CaO TiO2 Na2O K2O Coefficient (raw oxides) K06CO3 Kelso 76.52 12.38 1.68 0.03 0.04 0.54 0.13 3.60 5.06 1.00 94.62 ORNA-1 std. 76.62 12.30 1.70 0.02 0.03 0.53 0.13 3.63 5.03 0.99 95.48 SL-MM06-91 S. Luis 76.96 12.29 1.73 0.02 0.04 0.54 0.13 3.29 5.00 0.98 94.83 K06CO2 Bostwick 76.63 12.36 1.56 0.03 0.04 0.52 0.12 3.64 5.12 0.97 94.53

Table 1. Chemical comparison of Lava Creek B ash samples.

Notes: Numerical values from electron-microprobe analysis are shown in weight-percent oxide, normalized to a 100 percent fluid-free basis. Raw oxide totals are included to show approximate degree of hydration of volcanic glass shards. For comparison as a standard (‘std.’), we show chemical composition of ORNA-1, a ‘type’ Lava Creek sample (from USGS Tephrochronology Project database of ca. 5700 analyzed volcanic glass samples). Values for two, chemically correlated Lava Creek B ash samples (K06CO2, from Bostwick Park, Colorado and SL-MM06-91 [POP2] from San Luis, Colorado; collected by Mike Machette, USGS) are provided for comparison.



Uncompahgre Rivers. The Qt6U gravels overlain locally by the Lava Creek B ash have been traced northward to the Gunnison River, where they cor-relate with gravels of the paleo-Gunnison River (map unit Qt6G; stars on Fig. 2). The northernmost strath elevation for the Qt6U gravels is 1580 m, ca. 103 m directly above the Gunnison River (Fig. 2), yielding a maximal incision rate of 161 m/Ma. If bedrock depth is 10 m below the river (e.g., as it is on sites in the San Juan River of southeastern Utah; Wolkowinsky and Granger, 2004), bedrock-incision rate would increase to 177 m/Ma. Using the time-correlated Qt6G gravels on the north side of the Gunnison River at an elevation of 1571 m (96 m above the Gunnison River), we estimated an incision rate of 150 m/Ma, which would increase to 165 m/Ma within a situa-tion involving a 10 m mean depth to bedrock. For simplicity, we invoked a preferred bedrock-incision rate for this range (150–177 m/Ma) of ca. 165 m/Ma.

An estimated rate of 165 m/Ma is similar to regional incision rates for this part of the Rocky Mountain region (Dethier, 2001). However, this estimate is far less than rates proposed by Baker et al. (2002) or Rider et

al. (2006), who calculated incision rates of ca. 280 and 240 m/Ma for a reach that is only 10–15 km upstream. Their estimated rates were based on the presence of Lava Creek B ash underlying basaltic gravels on Petrie Mesa, an alluvial-fan complex of an ancient tributary of the Gunnison River that originated on the south flank of Grand Mesa (Fig. 1). The difference between their pro-jected points (Fig. 4) reflects uncertainties involved in projecting short segments of a tributary or fan tread pro-file toward its trunk stream.

The incision-rate data point from the present study (ca. 165 m/Ma) does not resolve such uncer-tainties, but it does introduce at least two alternative hypotheses for modeling the gradient of the post-0.64-Ma Gunnison profile (Fig. 4). Model ‘A’—if a steady 165 m/Ma incision rate also applies to the data from Petrie Mesa, earlier projections may have been too high and a more accurate projection would meet the profile shown as ‘A’ on Figure 4, found by using today’s gradient and the position of the 0.64- Ma paleo-conf luence. However, this seems to be inconsistent with the Qt6G data (triangles of Fig. 4). Alternatively, model ‘B’ (Fig. 4) suggests that the

0

10

20

30

40

50

60

70

80

Qt6U Qt6G

% c

last

s

% Felsics

% Intermediate

% Mafic

% Sedimentary

% Metamorphic

Figure 5. Histograms of generalized gravel compositions of ancestral river terraces (Qt6U and Qt6G).



observed Qt6G gradient continued upstream, which coincides with the Rider et al. (2006) Petrie Mesa projection. If ‘B’ were correct, average incision rates since 0.64 Ma may have been higher upstream, in keeping with passage of a transient knickpoint (dis-cussed below). This raises cautions and highlights the compound uncertainties associated with projection-related estimates, depth to bedrock, using modern gradients as a proxy for past gradients, and applying average incision rates within transient systems.

In upstream settings (e.g., the reach of the Gunnison River within Black Canyon), rates of inci-sion as high as 500–640 m/Ma for the same time interval (i.e., 0.64 Ma) variously have been proposed by Hansen (1965), Sandoval (2007), Kelley et al. (2007), and Aslan et al. (2008a). These estimated rates are based on mapping the paleo-Bostwick River tributary to its confluence with the Gunnison River at Red Canyon, using the Lava Creek B ash as the age control for these gravels. These data indicate that about half the depth of the Black Canyon has been incised since 0.64 Ma. These localized average inci-sion rates (ca. 500 m/Ma since 0.64 Ma) are dramat-ically higher than our estimate from Sawmill Mesa (ca. 165 m/Ma since 0.64 Ma). The estimated inci-sion rate from Sawmill Mesa is consistent with the idea that the estimates from Black Canyon are within the knickpoint of the modern Gunnison River. Rates calculated from Lava Creek B localities above the modern knickpoint, near the confluence with Lake Fork (Fig. 1), give an incision rate of 95 m/Ma over 0.64 Ma (Aslan et al., 2008a).

To explain this apparently greatly differential incision, we infer that the modern knickpoint is a transient that has migrated upstream and is now located (and perhaps hung up) in basement rocks of the Black Canyon of the Gunnison (Sandoval, 2007). If bedrock were the main control for observed dif-ferential incision rates, most observed rates would be expected to be the same or higher in softer rocks (e.g., Mesozoic sedimentary rocks at Sawmill Mesa) than in harder bedrock (e.g., Precambrian granite and schist at Black Canyon). Fault-controlled influ-ences on differential rates of river incision have been reported elsewhere (Kirkham et al., 2002; Pederson et al., 2002; Karlstrom et al., 2007), but given the absence of known Quaternary fault slip on the north-ern side of the Gunnison uplift (USGS/WGS, 2006) we find this to be a less probable explanation for the

data on differential incision. Using the transient-knickpoint hypothesis, we infer that the knickpoint migrated past the Sawmill Mesa reach prior to 0.64 Ma, as discussed in the next section.

SUMMARY OF ALLUVIAL HISTORY AND REGIONAL CORRELATIONS

We interpret the Sawmill Mesa and Kelso Gulch terraces as channel courses for distinct and recog-nizable paleo-Uncompahgre and paleo-Gunnison Rivers, thus recording a history of upstream migra-tion of the paleo-conf luence of these two rivers. Thus, in general, we support models for northeast (down-dip) migration of the Uncompahgre River on the northeastern flank of the Uncompahgre Plateau (Sinnock, 1981; Betton et al., 2005). Before 0.64 Ma, the ancestral Uncompahgre River deposited the north–south-trending terraces (map unit Qt7U; Fig. 2) in Kelso Gulch. The Gunnison River must have been entrenched in the same canyon at this time, because terraces are inset below the canyon walls. Based on these observations, the pre-0.64-Ma conflu-ence with the Gunnison River was near the mouth of Kelso Gulch (Fig. 6A). Subsequently, a drainage-reor-ganization event shifted the ancestral Uncompahgre River ca. 2.5 km east, where it joined the ancestral Gunnison River in the vicinity of the northernmost terraces of Sawmill Mesa (Figs. 2 and 6B).

The event that triggered channel relocation probably was stream piracy. We infer that the fall in local base level (produced through incision by the paleo-Gunnison River) induced a small tributary to cut headward at the location of map unit Qt6U. It cut into the floodplain of the paleo-Uncompahgre River (upstream of preserved map unit Qt7U) and thereby provided a more direct route to base level for the paleo-Uncompahgre River. This process occurred twice in the map area, thus relocating the Uncompahgre River still further upstream to the region east of the study area. Following this shift of the ancestral Uncompahgre River to the east, Lava Creek B ash blanketed the abandoned floodplains and was reworked into low-lying areas. Fine-grained sedi-ment and ash buried gravel of the paleo-Uncompah-gre River (map unit Qt6U; Fig. 2). Ash deposits near Qt7U terraces mantled hillslopes and became par-tially buried by slope-wash and colluvium. In this scenario, it is important to understand what types



of processes could have generated a transient knickpoint prior to 0.64 Ma. Our model is that both the stream capture and the pre-0.64 Ma knickpoint may ref lect transient incision as a response to downstream drainage reorganiza-tion, caused by the abandonment of Unaweep Canyon (Fig. 1) about 1 ma (Aslan et al., 2008b).

Since 0.64 Ma, the Uncompahgre River has shifted, perhaps several times, a total of 9 km east of Sawmill Mesa to its present-day confluence with the Gunnison River (Fig. 6c). Eastward migration probably is due to additional stream-cap-ture events. Furthermore, east-ward migration of the ancestral Uncompahgre River may help to explain stream abandonment from Bostwick Park, which also occurred prior to 0.64 Ma (Kelley et al., 2007). Because these events coincide temporally, stream piracy at Bostwick Park and abandon-ment of the Sawmill Mesa conflu-ence location may have been inde-pendent responses to the migrat-ing knickpoint that developed after Unaweep abandonment.

CONCLUSIONS

Four important reg iona l implications can be identif ied from our study. (1) Sawmill Mesa is a significant calibration point in the region, and it provides the most accurate incision-rate data point of ca. 165 m/Ma since 0.64 Ma. (2) This data point also is now available for comparison to other, less-well-constrained inci-sion points on nearby 0.64-Ma surfaces (e.g., Petrie Mesa). (3) Regional differential incision since 0.64-Ma has occurred along the Gunnison River system, thus

Uncom

pahgre River

Gunnison River

~ 0..64 Ma paleo-Uncompahgre River

~ 0.64 Ma paleo-Gunnison River

>0.64 Ma paleo-Gunnison River

>0.64 Ma paleo-Uncompahgre River

(c)

(b)

(a)

Delta

Delta

Delta

Figure 6. Summary of alluvial history of paleo-Gunnison and paleo-Uncompahgre Rivers. A, positions of rivers prior to 0.64 Ma; B, positions of rivers at approximately 0.64 Ma; and c, existing positions of rivers.



emphasizing that this region is important for under-standing the nature and causes of differential incision within a major river system. (4) Earlier compilations using Lava Creek B ash (Dethier, 2001) concluded that highest incision rates corresponded to upstream reaches below glaciated areas in the Rockies and that incision rates decreased downstream. Our data indi-cates variable incision rates downstream because of transient, knickpoint perturbations in this river.

ACKNOWLEDGMENTS

We thank the National Science Foundation Research Experience for Undergraduates program for support of this research (EAR-0453264). CREST (Colorado Rockies Experiment and Seismic Transect), funded by the Continental Dynamics Program of the NSF (EAR-0607808 to Karlstrom), also supported the research. We thank the staff at the U.S. Geological Survey Tephrochronology Lab for identifying the Lava Creek B ash. This research benefited from discus-sions with Magdalena Sandoval, whose master’s thesis overlapped with our study. We greatly appreciate the extremely helpful reviews by Allen Stork, Margaret McMillan, Wes Hildreth, and Robert L. Christiansen.

REFERENCES CITED

Aslan, A., Karlstrom, K., Hood, W., and 11 others, 2008a, River incision histories of the Black Canyon of the Gunnison and Unaweep Canyon: Interplay between late Cenozoic tectonism, climate change, and drainage inte-gration in the western Rocky Mountains, in Raynolds, R. G., ed., Roaming the Rocky Mountains and environs: Geological Field Trips: Geological Society of America Field Guide 10, p. 175–202.

Aslan, A., Hood, W., Karlstrom, K., and 6 others, 2008b, Abandonment of Unaweep Canyon ~1 Ma and the effects of transient knickpoint migration, western Colorado: Geological Society of America Annual Meeting and Exposition Abstracts with Programs, v. 40, no. 6, p. 220.

Baker, F., Rundell, J., Hasebi, K., and 2 others, 2002, Geomorphic evolution of Grand Mesa, western Colorado: Geological Society of America Abstracts with Programs, v. 34, no. 6, p. 472.

Betton, C., Aslan, A., and Cole, R., 2005, Late Cenozoic erosional history and major drainage changes of the Colorado–Gunnison River systems, western Colorado: Geological Society of America Abstracts with Programs, v. 37, no. 6, p. 35.

Bull, W. B., 1991, Geomorphic responses to climate change: New York, Oxford University Press, 326 p.

Cole, R. D., and Sexton, J. L., 1981, Pleistocene surficial deposits of the Grand Mesa area, Colorado: New Mexico Geological Society Guidebook no. 32, p. 121–126.

Dethier, P., 2001, Pleistocene incision rates in the west-ern United States calibrated using Lava Creek B tephra: Geology, v. 29, p. 783–786.

Frankel, K. L., and Pazzaglia, F. J., 2006, Mountain fronts, base-level fall, and landscape evolution: Insights from the southern Rocky Mountains: Geological Society of America Special Paper 398, p. 419–434.

Hansen, W. R., 1965, The Black Canyon of the Gunnison; Today and yesterday: U.S. Geological Survey Bulletin 1191, p. 1–76.

Heller, P., Dueker, K., and McMillan, M. E., 2003, Post-Paleozoic alluvial gravel transport as evidence of conti-nental tilting in the U.S. Cordillera: Geological Society of America Bulletin, v. 115, p. 1122–1132.

Hunt, C. B., 1969, Geologic history of the Colorado River (in the Colorado River region and John Wesley Powell): U.S. Geological Survey Professional Paper 669, p. 59–130.

Izett, G. A., and Wilcox, R. E., 1982, Map showing locali-ties and inferred distributions of the Huckleberry Ridge, Mesa Falls, and Lava Creek ash beds (Pearlette family ash beds) of Pleistocene age in the western United States and southern Canada: U.S. Geological Survey Miscellaneous Investigations Series I-1325: scale 1:4,000,000, 1 sheet.

Karlstrom, K. E., Crow, R. S., Peters, L., and 5 others, 2007, 40Ar/39Ar and f ield studies of Quaternary basalts in Grand Canyon and model for carving Grand Canyon: Quantifying the interaction of river incision and normal faulting across the western edge of the Colorado Plateau: Geological Society of America Bulletin, v. 119, p. 1283–1312.

Karlstrom, K. E., Crow, R., Crossey, L. J., and 2 others, 2008, Model for tectonically driven incision of the younger than 6 Ma Grand Canyon: Geology, v. 36, p. 835–838.

Karlstrom, K. E., Whitmeyer, S. J., Dueker, K., 4 others, and the CD-ROM Working Group, 2005, Synthesis of results from the CD-ROM experiment: 4-D image of the litho-sphere beneath the Rocky Mountains and implications for understanding the evolution of continental lithosphere, in Karlstrom, K. E., and Keller, G. R., eds., The Rocky Mountain region — An evolving lithosphere: Tectonics, geochemistry, and geophysics: American Geophysical Union Geophysical Monograph 154, p. 421–441.

Kelley, S. E., Hudson, A. M., Kaproth, B. M., and 2 others, 2007, Long profile analysis of the Pleistocene Bostwick River, with implications for the incision of the Black Canyon of the Gunnison: Geological Society of America Abstracts with Programs, v. 39, no. 6, p. 306.



Kirkham, R. M., Scott, R. B., and Judkins, T. W., 2002, Late Cenozoic evaporite tectonism and volcanism in west-central Colorado: Geological Society of America Special Paper 366, 234 p.

Lanphere, M. A., Champion, D. E., Christiansen, R. L., and 2 others, 2002, Revised ages for tuffs of the Yellowstone Plateau volcanic field: Assignment of the Huckleberry Ridge Tuff to a new geomagnetic polarity event: Geological Society America Bulletin, v. 114, p. 559–568.

Leonard, E. M., 2002, Geomorphic and tectonic forcing of the late Cenozoic warping of the Colorado Piedmont: Geology, v. 20, p. 595–598.

Lipman, P. W., 2007, Incremental assembly and prolonged con-solidation of Cordilleran magma chambers: Evidence from the southern Rocky Mountain volcanic field: Geosphere, v. 3, p. 42–70.

Lohman, S. W., 1981, Ancient drainage changes in and south of Unaweep Canyon, southwestern Colorado: New Mexico Geological Society Guidebook no. 32, p. 137–143.

Molnar, P., and England, P., 1990, Late Cenozoic uplift of mountain ranges and global climate change: Chicken or egg: Nature, v. 346, p. 29–34.

McMillan, M. E, Heller, P. L, and Wing, S. L., 2006, History and causes of post-Laramide relief in the Rocky Mountain orogenic plateau: Geological Society of America Bulletin, v. 118, p. 393–405.

Pederson, J., Karlstrom, K., Sharp, W., and McIntosh, W., 2002, Differential incision of the Grand Canyon related to Quaternary faulting: Constraints from U-series and Ar/Ar dating: Geology, v. 30, p. 739–742.

Rider, K., Darling, A., Gloyd, J., and Cole, R., 2006, Relative ages and origins of Grand Mesa pediments, Colorado: Geological Society of America Abstracts with Programs, v. 38, no. 6, p. 34.

Sandoval, M. M., 2007, Quaternary incision history of the Black Canyon of the Gunnison, Colorado [Master’s Thesis]: Albuquerque, University of New Mexico, 96 p.

Sarna-Wojcicki, A. M., Bowman, H. R., Meyer, and 7 others, 1984, Chemical analyses, correlations, and ages of upper Pliocene and Pleistocene ash layers of east-central and southern California: U.S. Geological Survey Special Paper 1293, p 1–40.

Sharp, W. D., Ludwig, K. R., Chadwick, O. A., and 2 others, 2003, Dating f luvial terraces by 230Th/U on pedogenic carbonate, Wind River Basin, Wyoming: Quaternary Research, v. 59, p. 139–150.

Sinnock, S., 1981, Plei stocene dra inage changes in Uncompahgre Plateau–Grand Valley region of west-ern Colorado, including formation and abandonment of Unaweep Canyon: A hypothesis: New Mexico Geological Society Guidebook no. 32, p. 127–136.

Sinnock, S., Epis, R. C., and Callender, J. F., 1981, Glacial moraines, terraces and pediments of Grand Valley, west-ern Colorado: New Mexico Geological Society Guidebook no. 32, p. 113–120.

Steven, T. A., 2002, Late Cenozoic tectonic and geomorphic framework surrounding the evaporite dissolution area in west-central Colorado, in Kirkham, R. M., Scott, R. B., and Judkins, T. W., eds., Late Cenozoic evaporite tec-tonism and volcanism in west-central Colorado: Geological Society of America Special Paper 366, p. 15–30.

U.S. Geological Survey and Colorado Geological Survey, 2006, Quaternary fault and fold database for the United States (accessed November 10, 2008, from USGS website: http//earthquakes.usgs.gov/regional/qfaults/).

Wolkowinsky, A. J., and Granger, D. E., 2004, Early Pleistocene incision of the San Juan River, Utah, dated with 26Al and 10Be: Geology, v. 32, p. 749–752.

Yeend, W. E., 1969, Quaternary geology of the Grand and Battlement Mesas area, Colorado: U.S. Geological Survey Professional Paper 617, 50 p.

Zaprowski, B. J., Evenson, E. B., Pazzaglia, F. J., and Epstein, J. B., 2001, Knickzone propagation in the Black Hills and northern High Plains: A different perspective on the late Cenozoic exhumation of the Laramide Rocky Mountains: Geology, v. 29, p. 547–550.

Manuscript Submitted May 20, 2008

Revised Manuscript Submitted January 30, 2009

Manuscript Accepted April 1, 2009



Documents

Quaternary incision rates and drainage evolution of …eps et al_RockyMtnGeo... · Quaternary incision rates and drainage evolution of the Uncompahgre and Gunnison Rivers, western