Bedrock Canyon Research Articles

40Ar/39Ar and field 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

40Ar/39Ar dates on basalts of Grand Canyon provide one of the best records in the world of the interplay among volcanism, differential canyon incision, and neotectonic faulting. Earlier 40K/40Ar dates indicated that Grand Canyon had been carved to essentially its present depth before 1.2 Ma. But new 40Ar/39Ar data cut this time frame approximately in half; new ages are all <723 ka, with age probability peaks at 606, 534, 348, 192, and 102 ka. Strategic sampling of basalts provides a semicontinuous record for deciphering late Quaternary incision and fault-slip rates and indicates that basalts flowed into and preserved a record of a progressively deepening bedrock canyon. The Eastern Grand Canyon block (east of Toroweap fault) has bedrock incision rates of 150–175 m/Ma over approximately the last 500 ka; western Grand Canyon block (west of Hurricane fault) has bedrock incision rates of 50–75 m/Ma over approximately the last 720 ka. Fault displacement rates are 97–106 m/Ma on the Toroweap fault (last 500–600 ka) and 70–100 m/Ma on the Hurricane fault (last 200–300 ka). As the river crosses each fault, the apparent incision rate is lowest in the immediate hanging wall, and this rate, plus the displacement rate, is sub-equal to the incision rate in the footwall. At the reach scale, variation in apparent incision rates delineates ∼100 m/Ma of cumulative relative vertical lowering of the western Grand Canyon block relative to the eastern block and 70–100 m of slip accommodated by formation of a hanging-wall anticline. Data from the Lake Mead region indicate that our refined fault-dampened incision model has operated over the last 6 Ma. Bedrock incision rate has been 20–30 m/Ma in the lower Colorado River block in the last 5.5 Ma, and displacement on the Wheeler fault has resulted in both lowering of the Lower Colorado River block and formation of a hanging-wall anticline of the 6-Ma Hualapai Limestone. In modeling long-term incision history, extrapolation of Quaternary fault displacement and incision rates linearly back 6 Ma only accounts for approximately two-thirds of eastern and approximately one-third of western Grand Canyon incision. This “incision discrepancy” for carving Grand Canyon is best explained by higher rates during early (5- to 6-Ma) incision in eastern Grand Canyon and the existence of Miocene paleocanyons in western Grand Canyon. Differential incision data provide evidence for relative vertical displacement across Neogene faults of the Colorado Plateau-Basin and Range transition, a key data set for evaluating uplift and incision models. Our data indicate that the Lower Colorado River block has lowered 25–50 m/Ma (150–300 m) relative to the western Grand Canyon block and 125–150 m/Ma (750–900 m) relative to the eastern Grand Canyon block in 6 Ma. The best model explaining the constrained reconstruction of the 5- to 6-Ma Colorado River paleoprofile, and other geologic data, is that most of the 750–900 m of relative vertical block motion that accompanied canyon incision was due to Neogene surface uplift of the Colorado Plateau.

Sedimentary links and the spatial organization of Atlantic salmon ( Salmo salar) spawning habitat in a Canadian Shield river

The segmenting of gravel-bed rivers flowing through mountain valleys into a number of discrete ‘sedimentary links’, each characterized by downstream fining of alluvium, is a relatively recent concept which offers promise to model the large-scale spatial organisation of many types of aquatic habitat (reproductive, feeding, refuge, etc), strongly dependent on dominant bed sediment calibre. Although, so far, the ecological application of the concept has mainly focused on benthic invertebrates, here we illustrate its application to fish (Atlantic salmon; Salmo salar). Moreover, the link concept has also been primarily applied to alpine river environments where link formation is triggered by point sources (mainly tributaries) supplying coarser sediment. However, somewhat lower relief, mountain valley landscapes of North Eastern Canada are often structured into sedimentary links triggered by non-point, ‘supply zones’ of coarse sediments, originating in bedrock canyon reaches or valley bottom deposits of glacial drift. Here, we propose an adaptation and extension of the original, sedimentary link concept to such landscapes and test its utility along one such system, the Ste Marguerite River (SMR), a salmon river draining the Canadian Shield in the Saguenay region of Québec. We first discuss a simple field and office based method of link delineation. Then we discuss potential sources of minor, sublink scale grain size variability and their effects on how sedimentary links are defined. Lastly, we demonstrate the usefulness of the link structure to model the distribution of Atlantic salmon spawning habitat (a habitat that depends critically on bed texture). Our results indicate that a revised sedimentary link typology is needed to describe longitudinal grain size patterns where non-point, valley-segment scale sources of coarse sediment are important and that consideration of the research purpose and scale is important in defining meaningful link units. We also show that salmon spawning zones can be directly predicted from the link structure: along the SMR, spawning activity is apportioned within each of the discrete links, in those sub-zones where surface sediment size and sand content are optimal for reproduction.

Ground-water surface-water interactions and long-term change in riverine riparian vegetation in the southwestern United States

Riverine riparian vegetation has changed throughout the southwestern United States, prompting concern about losses of habitat and biodiversity. Woody riparian vegetation grows in a variety of geomorphic settings ranging from bedrock-lined channels to perennial streams crossing deep alluvium and is dependent on interaction between ground-water and surface-water resources. Historically, few reaches in Arizona, southern Utah, or eastern California below 1530 m elevation had closed gallery forests of cottonwood and willow; instead, many alluvial reaches that now support riparian gallery forests once had marshy grasslands and most bedrock canyons were essentially barren. Repeat photography using more than 3000 historical images of rivers indicates that riparian vegetation has increased over much of the region. These increases appear to be related to several factors, notably the reduction in beaver populations by trappers in the 19th century, downcutting of arroyos that drained alluvial aquifers between 1880 and 1910, the frequent recurrence of winter floods during discrete periods of the 20th century, an increased growing season, and stable ground-water levels. Reductions in riparian vegetation result from agricultural clearing, excessive ground-water use, complete flow diversion, and impoundment of reservoirs. Elimination of riparian vegetation occurs either where high ground-water use lowers the water table below the rooting depth of riparian species, where base flow is completely diverted, or both. We illustrate regional changes using case histories of the San Pedro and Santa Cruz Rivers, which are adjacent watersheds in southern Arizona with long histories of water development and different trajectories of change in riparian vegetation.

Epeirogenic Controls on Canadian River Incision and Landscape Evolution, Great Plains of Northeastern New Mexico

The Canadian River has carved a deep bedrock canyon into the Great Plains of northeastern New Mexico in response to a complex interaction of post‐Laramide, epeirogenic rock‐uplift processes along the Jemez lineament and downstream baselevel fall caused by late Tertiary and Quaternary evaporite dissolution. Broad, slow‐bulging along the Jemez lineament, a lithospheric‐scale structure with prolific late Cenozoic volcanism along its trend, helps drive incision in the canyon. Canadian River canyon fluvial terraces, mapped and correlated according to soils, elevation, and rock‐type criteria, reveal longitudinal profiles that appear broadly warped across the Jemez lineament and converge to the modern channel profile both upstream and downstream of the canyon. Similarly, the modern channel long profile is distinctly and broadly convex through its Great Plains reach, a unique feature among major rivers draining the eastern flank of the southern Rocky Mountain front. Radiocarbon‐dated charcoal from terrace alluvium and an 40Ar/39Ar‐dated basalt flow within the canyon constrain long‐term average rates of bedrock incision to ∼0.06–0.07 mm/yr. River incision along the Jemez lineament has produced youthful canyons and escarpments and eroded Tertiary rocks relative to other regions along the southern Rocky Mountain front. Whereas rock‐uplift processes prevail in the canyon reach, baselevel changes driven by a well‐documented process of salt‐dissolution subsidence have controlled the evolution of reaches downstream of the canyon. Removal of bedrock along the northern flank of the Jemez lineament bulge fosters opposing dips in the valley flanks, resulting in the Canadian escarpment (northern flank) that stands up to 200 m higher than the Caprock (southern flank). These findings suggest that post‐Laramide, thermally driven epeirogeny is in part responsible for the high‐standing topography evident on the western Great Plains of northeastern New Mexico.

Sedimentology of high-stage flood deposits of the Tagus River, Central Spain

This paper details the sedimentology of high-stage flood deposits, with the definition of sedimentary environments and their characteristic sequences, along two bedrock reaches of the Tagus River (Central Spain). High-stage flood deposits accumulated in bedrock canyons include slackwater flood deposits (SWD) and other types of deposits located at flow separation zones and associated with slow-moving flow (<1 m s −1). These flood deposits are common indirect indicators of flood stages used in palaeoflood studies for estimating the discharges associated with Quaternary floods. Depositional environments of flood deposits include (1) channel widening, (2) canyon expansion, (3) bedrock obstacles, and (4) backflooded areas along tributary streams. These flood deposits can be found associated with other non-fluvial environments, namely aeolian reworked and slope washflow facies. Channel widening, due to flood stage variations, comprises internal and external zones of the channel margins, and their characteristic sequences contain similar facies to those of alluvial floodplains. Canyon expansion environments favour vertical accretion of slackwater units and the development of flood deposit benches, which contain four sequences related to bench elevation and distance from the channel's main thread of flow. At the lee side of bedrock obstacles, characteristic sedimentary sequences are dominated by reverse flow structures (e.g. climbing ripples migrating upstream) due to eddies with a high sand concentration. Flood deposits located within tributary mouths contain typical sequences of reworked floodplain deposits. Backflooding of tributaries during flood stages produces deposition from suspension of sand, silt and clay within three sequences characterised by non-structure or parallel lamination and intense bioturbation. A better understanding of the flood deposit sequences may contribute to the characterisation of flood magnitudes and flood hydraulics and can also be applied to some ancient depositional environments.

A paleohydrologic reinterpretation of the Homeb Silts, Kuiseb River, central Namib Desert (Namibia) and paleoclimatic implications

Slack water deposits are usually fine-grained (fine sand and coarse silt) flood sediments deposited in areas of the floodplain that are sheltered from high-velocity flood flows. Slack water deposits have been recognized by many investigators in a wide variety of physiographic and climatic settings. Late Quaternary slack water deposits have not been described from the extremely arid Namib Desert of southwestern Africa. Fine-grained flood sediments in the Kuiseb Valley near Homeb, accumulated between 23,000 and 19,000 14C-years BP, have been interpreted by previous workers as (i) sediments deposited behind dune dams, (ii) river endpoint accumulations, and (iii) flood deposits of an aggrading river controlled either by a base level change in the lower reaches or a change in the hydrological regime in the catchment area. Furthermore, the so-called Homeb Silts could document more arid climatic conditions in the Kuiseb catchment area during the last glacial maximum (LGM). Our research shows that the Homeb Silts are slack water deposits and that in many valleys of the Namib Desert, slack water sediments are observed most commonly at the junctions of major rivers and minor tributaries in bedrock canyon settings. Here we present evidence that these slack water deposits accumulated as a result of more humid conditions and/or more intense precipitation events in the upper reaches of the river. They do not document climatic changes in the desert itself as postulated by many researchers.

Holocene paleolakes of Lake Canyon, Colorado Plateau: Paleoclimate and landscape response from sedimentology and allostratigraphy

Middle and late Holocene paleoclimate can be inferred from lake-level fluctuations recorded in the sediment of two, fresh, groundwater‐controlled paleolakes in Lake Canyon, southeast Utah. The record from Lake Canyon contributes proxy climate data for the region as well as insight into climate controls on sedimentation. The paleolakes were formed immediately upcanyon of side-stream confluences where disproportionately large volumes of alluvium were focused into the bedrock canyon. Excellent exposure allows a depositional facies model to be developed and allostratigraphic units to be defined. A facies-based proxy for the past shoreline position is applied to the allostratigraphy and radiocarbon-based chronostratigraphy to produce a record of lake level through time that is corroborated by diatom proxy data. The earliest known lake formed during an episode of aggradation that began just before 5180 14 C yr B.P., and an arroyo-cutting event drained the lake before 3050 14 Cy r B.P. The Lake Canyon record for the past 3 k.y. is composed of wet-dry cycles that have a periodicity of ~520 yr, matching cycles resolved by the regional tree-ring record and the timing of the Medieval Warm Period and Little Ice Age. Lake level rose after 420 14 C yr B.P. and aggradation continued until an advancing arroyo headcut drained the upper lake in A.D. 1915. Fluvial and lacustrine aggradation in Lake Canyon accompanied shifts to wetter climate when sediment stored in upper reaches of the drainage basin was delivered to the canyon, whereas arroyo cutting in the canyon is more a function of sediment supply and internal complex-response mechanisms than dry climate conditions. Correlation with other records is hindered by differences in resolution and the influence of local controls on alluvial stratigraphy. The Lake Canyon record confirms that episodes of high flood frequency in Holocene time have been wetter overall, but not all wet periods have been marked by high flood frequency.

Response of extreme floods in the southwestern United States to climatic variations in the late Holocene

A regional synthesis of paleoflood chronologies on rivers in Arizona and southern Utah reveals that the largest floods over the last 5000 years cluster into distinct time periods that are related to regional and global climatic fluctuations. The flood chronologies were constructed using fine-grained slackwater deposits that accumulate in protected areas along the margins of bedrock canyons and selectively preserve evidence of the largest events. High-magnitude floods were frequent on rivers throughout the region from 5000 to 3600 14C yrs BP (dendrocalibrated age = 3800-2200 BC) and increased again after 2200 BP (400 BC), with particularly prominent peaks in magnitude and frequency around 1100-900 BP (AD 900–1100) and after 500 yrs BP (AD 1400). In contrast, the periods 3600-2200 BP (2200-400 BC) and 800-600 yrs BP (1200–1400 AD) are marked by sharp decreases in the occurrence of large floods on these rivers. In the modern record, storms that generate large floods (≥ 10-year) in the region fall into three categories: (1) winter North Pacific frontal storms; (2) late-summer and fall storms that draw in moisture from recurved Pacific tropical cyclones; and (3) summer storms, mainly convective thunderstorms. Winter storms and tropical cyclones are associated with the most severe floods on the rivers in this study, and are the most probable causes of the paleofloods over the last 5000 years. Floods from both winter storms and tropical cyclones occur when deep mid-latitude troughs steer storm systems into the region. Composite anomaly maps of daily 700-mbar heights indicate that these floods are associated with a low-pressure anomaly off the California coast and a high-pressure anomaly over the Aleutians or Gulf of Alaska. A strong connection exists between the negative phase of the Southern Oscillation Index (often associated with El Nin˜o conditions) and the large floods associated with winter storms and tropical cyclones. The paleoflood records confirm the existence of centennial-scale variations in the conditions conducive to the occurrence of extreme floods and flood-generating storms in this region. The episodes with an increased frequency of high-magnitude floods coincide with periods of cool, wet climate in the western U.S., whereas warm intervals, such as the Medieval Warm Period, are times of dramatic decreases in the number of large floods. A positive relationship between the paleofloods and long-term variations in the frequency of El Nin˜o events is evident over the last 1000 years. This relationship continues over at least the last 3000 years with warm coastal sea-surface temperatures indicative of El Nin˜o-like conditions.

Holocene slip rate and earthquake recurrence of the northern Calaveras Fault at Leyden Creek, northern California

The northern Calaveras fault traverses a heavily populated area in the eastern San Francisco Bay region and has not had a large earthquake in more than 130 years. To obtain data on the number, timing, and recurrence of large paleoearthquakes, we conducted paleoseismologic investigations at Leyden Creek, which crosses the fault in the rugged southern East Bay Hills. The site is characterized by a prominent west facing scarp and five fluvial terraces on the western (upstream) side of the fault. On the eastern (downstream) side of the fault, the creek flows through a narrow bedrock canyon that constricts the modern valley and has constrained the location of a late Pleistocene paleovalley. The margin of a buried bedrock valley west of the fault trends nearly perpendicular to the fault and is offset 54 (+18, −14) m in a right‐lateral sense from the narrow bedrock canyon. Based on radiocarbon ages for alluvial sediments predating and postdating this paleovalley margin, we estimate an age of 11.5 (+3, −1) ka for the valley margin and a Holocene slip rate of 5 ± 2 mm/yr for the fault at Leyden Creek. Slickensides exposed in multiple trenches across the fault show that the most recent movement was predominantly lateral with a minor component of down‐to‐the‐west slip. Multiple displaced scarp‐derived colluvial deposits are interpreted as results of five or six surface ruptures within the past 2500 years. Twenty‐one radiocarbon samples from scarp‐derived colluvium and interfingered alluvial deposits suggest an average interval between surface rupture earthquakes of 250 to 850 years.

A geomorphologic and hydrologic evaluation of an extraordinary flood discharge estimate: Bronco Creek, Arizona

The Bronco Creek, Arizona flood of August 19, 1971, was an extraordinary and rare flash flood. The published peak discharge estimate of 2080 m3 s−1 makes it virtually the world's largest known rainfall‐generated flood to come from a 50‐km2 basin. Previous workers have suggested that the peak discharge was overestimated based on its unlikely hydrological and hydraulic implications. In this study we reevaluate the Bronco Creek flood discharge by modeling peak discharges using relict high‐water marks in bedrock canyon reaches near the mouths of the three major subbasins of the watershed. The new estimated peak discharge of 750 to 850 m3 s−1 is based on summing the subbasin estimates and correcting for omitted drainage area. Qualitative and quantitative analyses of site‐specific and regional hydrological, meteorological, and geomorphological information support a lower peak discharge. We assessed preflood and postflood channel characteristics at the site of the original estimate using historical aerial photographs. Our analysis indicates that much of the discrepancy between the estimates may be explained by significant morphologic changes in the channel caused by the flood. The original estimate was made in a wide, high‐gradient alluvial channel where the flood changed the channel pattern from braided with vegetated bars to straight, wide, and devoid of vegetation. It is also likely that net vertical scour occurred in the channel. Although our new estimate is much lower than the original estimate, it is consistent with the dominant trend in regional flood magnitude‐drainage area relationships. By integrating hydraulic modeling of peak discharges in stable channel reaches with assessment of ancillary hydrological, meteorological, and geomorphological information, this study presents a viable approach to assessing the accuracy of extreme flood estimates in general. It also highlights potential uncertainties in estimating extreme flood discharges in steep, alluvial channels.

Flood hydrology and geomorphic effectiveness in the central Appalachians

AbstractThis paper compares hydrologic records and geomorphic effects of several historic floods in the central Appalachian region of the eastern United States. The most recent of these, occurring in November 1985, was the largest ever recorded in West Virginia, with peak discharges exceeding the estimated 500‐year discharge at eight of eleven stations in the South Branch Potomac River and Cheat River basins. Geomorphic effects on valley floors included some of the most severe and widespread floodplain erosion ever documented and exceeded anything seen in previous floods, even though comparable or greater rainfall and unit discharge have been observed several times in the region over the past 50 years. Comparison of discharge‐drainage area plots suggests that the intensity and spatial scale of the November 1985 flood were optimal for erosion of valley floors along the three forks of the South Branch Potomac River. However, when a larger geographic area is considered, rainfall totals and discharge‐drainage area relationships are insufficient predictors of geomorphic effectiveness for valley floors at drainage areas of 250 to 2500 km2.Unit stream power was calculated for the largest recorded flood discharge at 46 stations in the central Appalachians. Maximum values of unit stream power are developed in bedrock canyons, where the boundaries are resistant to erosion and the flow cross‐section cannot adjust its width to accommodate extreme discharges. The largest value was 2570 W m−2; record discharge at most stations was associated with unit stream power values less than 300 W m−2, but more stations exceeded this value in the November 1985 flood than in the other floods that were analysed. Unit stream power at indirect discharge measurement sites near areas experiencing severe erosion in this and other central Appalachian floods generally exceeded 300 W m−2; reach‐average values of 200‐500 W m−2 were calculated for valleys where erosion damage was most widespread. Despite these general trends, unit stream power is not a reliable predictor of geomorphic change for individual sites. Improved understanding of flood impacts will require more detailed investigation of interactions between local site characteristics and patterns of flood flow over the valley floor.

Origin, Structure, and Evolution of a Reattachment Bar, Colorado River, Grand Canyon, Arizona

ABSTRACT In a channel expansion, flow can separate from the bank, creating a zone of relatively weak recirculating current. Bars that accumulate in this weak flow near the point where flow reattaches to the bank are called reattachment bars. As a reattachment bar evolves, the recirculation zone may fill with sediment and restrict flow from the main channel. The increasingly restricted flow over the bar causes ripples to replace dunes and causes the sediment size to fine; the resulting vertical sequence resembles that of point bars. Seasonal and daily flow fluctuations in the Grand Canyon complicate this idealized sequence. Changes in discharge alter the geometry of recirculation zones, flow within the recirculation zones, the location of depositional and erosional sites, the kind of bedform an migration direction of bedforms on the bar, and the transported sediment size. Dunes and ripples within a recirculation zone migrate in a rotary pattern in response to the recirculating flow. Ripples near the reattachment point often resemble oscillation ripples in morphology and dynamics. The reversing flow that creates these ripples is caused by fluctuations in location of the reattachment point. These fluctuations cause flow near the reattachment point to reverse in an upstream-downstream direction, thereby producing symmetrical, reversing ripples with crests that trend normal to the bank. Low rates of ripple migration in the reversing flow, accompanied by rapid deposition, cause these ripples to climb at a high angle. At increasing distances from the reattachment point, the reversing flow is less balanced, and the ripples climb at lower angles as they migrat upstream and downstream. Although these observations were made in a bedrock canyon, the same processes operate in alluvial and tidal channels and are important in adjusting the shape of channels on point bars and concave benches and behind bedforms that become emergent at low stage. Reattachment bars can be recognized by the rotary flow patterns and by symmetrical, reversing, vertically climbing ripples.

Sediment storage and terrace formation in Coyote Gulch basin, south-central Utah

Late Holocene terraces in the Coyote Gulch basin, south-central Utah, were formed by aggradation caused by the rapid influx of alluvium produced from landslides on the Straight Cliffs. The main channel of Coyote Gulch was the primary conduit for this sediment and in places aggraded 20 m, burying the preexisting bedrock canyon. This aggradation began about 2300 B.P. and lasted for about 1300 yr. Aggradation in the main stem created a higher base level and reduced the valley gradient in a major tributary, Dry Fork Coyote Gulch, which then partially filled with fine-grained sediment. The major period of aggradation in the Dry Fork was not continuous but was interrupted by incision, presumably when a higher gradient sand-transporting stream was established across this fill. Within the study area more than 25 × 10 6 m 3 of sediment was stored during this aggradational period. Beginning about 900 B.P. this alluvium was incised, creating two terraces in upper Coyote Gulch and as many as three terraces in Dry Fork. Increased sediment production from the erosion of the Dry Fork alluvium caused further deposition downstream in Coyote Gulch, which was eventually incised to form a single terrace. Therefore, rapid sediment loading produced different fluvial responses dependent on the location of a channel segment in the drainage network. Stratigraphic studies can provide insight into the magnitude and rate of sediment storage in stream valleys as well as the timing and processes by which sediment is ultimately transported out of a basin.