Climate change and increasing societal demands for water pose challenges for the management of dam-regulated rivers. Management decisions impact the environment of these rivers, creating the need to balance societal needs with environmental conservation. Here we present a modeling framework that optimizes resource benefits within imposed water use goals for the Colorado River in Grand Canyon, where sandbars are a valued natural feature. The current sand-management paradigm utilizes controlled dam-release floods to build and maintain sandbars without exhausting the limited sand supplied by tributaries downstream from Glen Canyon Dam, which blocks all sand supplied from upriver. High monthly releases outside of controlled floods erode sandbars and cause net sand export from Grand Canyon, reducing the sand available to build sandbars. Releases are high in some months owing to the need to adjust flows to meet annual delivery targets, which can be updated throughout the year. Here, we present alternative strategies for operations that avoid high releases, while meeting water storage and delivery goals. We test these strategies using a simplified reservoir model which accounts for forecast uncertainty. We show how these strategies affect sand mass balance and sandbar size using previously developed models. Strategies optimal for sustainable sandbar building maintained sufficient reservoir elevations for implementing controlled floods, avoided high monthly releases by relaxing annual release constraints, and implemented controlled floods in fall immediately following tributary sand inputs. Coordinated modeling of reservoir operations and environmental resources is valuable for managers seeking to balance societal and environmental needs in regulated rivers worldwide.

In addition to supplying water for agriculture, cities, and industry, the Colorado River traverses the Colorado Plateau, including several of the most unique and valued National Parks and Recreation Areas in the United States. Although the water needs of these landscapes were not considered at the time water allocations were first negotiated, these needs were recognized in subsequent legislation and policy. Management goals address a range of aquatic and riparian resources, including fine sediment (sand, silt, and clay) which, in Grand Canyon, is important for ecological, cultural, and recreational resources. Over similar to 30 years, stakeholders, resource managers, and scientists collectively developed operational strategies for sediment management to meet goals outlined by an adaptive management program. However, prolonged drought, or "aridification," resulting in declining runoff and the lowest reservoir storage elevations in decades has challenged those strategies. The paradigm for sustainable sediment management relies on (1) sand accumulation on the bed of the Colorado River during periods of sediment-rich tributary floods from summer/fall thunderstorms, and (2) dam-released controlled (artificial) floods, referred to as High-Flow Experiments (HFEs), to redistribute the accumulated sand to rebuild eroded bar and floodplain deposits. The management protocol, which specifies narrowly defined sand accumulation periods and HFE implementation windows, is based on implementing HFEs in late fall during the period of greatest sediment enrichment, before higher winter releases for hydropower erode the accumulated sand from the riverbed. Low dam releases associated with drought, however, have changed the pattern of sand accumulation and low reservoir elevations have prevented HFE implementation in the defined window. An alternative strategy for HFE planning and implementation was tested opportunistically in April 2023 following lower-than-normal winter dam releases. We present findings from this HFE indicating that sand enrichment and sandbar building equaled or exceeded that of HFEs conducted under the established management protocol. These findings show that management goals for sediment under conditions of prolonged drought may be achievable but will likely require substantial changes in dam management strategies.

River flow patterns are primary drivers of lotic ecosystems, and hundreds of metrics have been developed to quantify flow attributes. Although existing metrics have been a powerful tool in designing environmental flows, they are often developed with specific resources in mind and are rarely directly comparable with each other (i.e., units are often different). Here, we focus on natural flows as the resource of interest and develop z-score metrics that measure the naturalness of regulated flows, incorporating natural means and interannual variation. These "eZ metrics" summarize whole year, subdaily, and functional flow patterns as standard deviations from natural such that their values are directly comparable. We illustrate their utility with a case study from the Colorado River downstream of Glen Canyon Dam in Arizona, USA. We calculated metrics for 1964-2022, spanning >5 decades of changing water policy, hydropower generation, and flow experimentation. We evaluate four options for estimating natural baseline flows. Across metrics, we found that subdaily stage variation deviated the most from baseline. Flows to satisfy regional water policy and power demands altered metrics more than designer flows (which target specific resource outcomes), and years with low water releases were closest to natural. Most of the designer flows have not made flow patterns more natural, due to incorrect seasonal timing, small magnitude, or short duration. By explicitly considering interannual variability and quantifying how regulated flows differ from natural using standard deviations, these metrics can inform management when the goal is to restore a natural flow regime.

Zebra Mussels (Dreissena polymorpha) were introduced into the US in the 1980s and have since spread throughout the country with concomitant impacts on aquatic ecosystems. We placed 20 m transects in a variety of sediment types and collected data on sediment size, mussel coverage (the percent of the bottom covered by mussels), and the presence of macrophytes within 0.25 m(2) quadrats in White Bear Lake, MN. To assess the degree of Zebra Mussel aggregation, we calculated Morisita's index of dispersion (I-delta) at four levels of spatial scale - 0.25 m(2), 1 m(2), 4 m(2,) and 40 m(2). Understanding the pattern of aggregation in Zebra Mussels is crucial in detecting early invasions, monitoring populations, and attempting to eradicate them. There was greater Zebra Mussel coverage within quadrats with coarser substrates (gravel/cobble and mixed sediment - including rock and boulders) compared to those with finer substrates (silt and sand). Macrophyte coverage did not significantly influence Zebra Mussel coverage on the lake bed sediments. Spatial scale had a large impact on the measure of mussel aggregation. Zebra Mussels in sand and silt had the highest degree of aggregation and high levels of macrophytes resulted in greater degrees of aggregation on the lake bed sediments. Zebra mussel coverage measured at larger spatial scales generally resulted in lower aggregation values, across both sediment types and levels of macrophyte coverage. The appropriate selection of spatial scale for sampling or observation is critical when upscaling local density to lake populations.

First posted March 8, 2022 For additional information, contact: Southwest Biological Science CenterU.S. Geological Survey2255 N. Gemini DriveFlagstaff, AZ 86001 Releases of water from Flaming Gorge Dam together with climate-related variations in runoff determine the streamflow regime of the Green River, which affects the physical characteristics of the channel and riparian ecosystem of the Green River corridor in Canyonlands National Park. The dam has decreased peak streamflows and raised base streamflows, resulting in vegetation encroachment and channel narrowing and simplification, which could be detrimental to endangered fish habitats over time. Operations of Flaming Gorge Dam are in part determined by flow recommendations provided by the Upper Colorado River Basin Endangered Fish Recovery Program that are designed to benefit native fish and disadvantage nonnative fish. These recommendations alone may not be sufficient to prevent channel narrowing and simplification. Increases in base flows may contribute to channel narrowing and simplification by increasing the water available to riparian vegetation and reducing the water volume available for increasing peak-flow magnitude or duration This report describes how proposed revisions to these flow recommendations would affect the physical characteristics of the Green River corridor in Canyonlands National Park, with a focus on riparian vegetation and channel width.Hydrologic conditions for the Green River downstream from Flaming Gorge Dam are classified by the U.S. Department of the Interior Bureau of Reclamation as dry, moderately dry, average, moderately wet, or wet. The flow recommendations for peak-flow magnitude and duration in wet years are consistent with geomorphic objectives and historical post-dam flows. In moderately wet years, although the recommended peaks may be sufficient to prevent narrowing over the short term, these peaks are lower than historical post-dam peak flows for moderately wet years and could therefore allow reduction in the occasional large peaks necessary to maintain sediment mobility and channel complexity. For average and drier years, the recommendations allow, but do not require, peak-flow magnitude and durations that are likely to achieve geomorphic objectives.

First posted September 16, 2022 For additional information, contact: Southwest Biological Science CenterU.S. Geological Survey2255 N. Gemini DriveFlagstaff, AZ 86001 Sandbars are an important resource in the Colorado River corridor in Marble and Grand Canyons, Arizona, downstream from Glen Canyon Dam. Sandbars provide aquatic and riparian habitat and are used as campsites by river runners and hikers. The study area is the Colorado River between Glen Canyon Dam and Diamond Creek, which is about 388 kilometers (241 miles) downstream from the dam. Closure of Glen Canyon Dam in 1963 and subsequent flow regulation reduced the sediment supply, limited the magnitude and frequency of floods, and increased the magnitude of baseflows. The result has been widespread erosion of sandbars and expansion of native and non-native vegetation on previously bare sand deposits in this debris-fan dominated canyon river. This study reports on the on-going long-term measurement program of Northern Arizona University, initiated in 1990 with the Bureau of Reclamation, and now also with the U.S. Geological Survey’s Grand Canyon Monitoring and Research Center. We report on all sandbar measurements made between 1990 and 2020 to demonstrate the multi-decadal response of the sandbar monitoring sites resulting from flow regulation by Glen Canyon Dam. Because only one study site is located in Glen Canyon, the 25 kilometer (15.5 miles) reach just below Glen Canyon Dam, analyses of sandbar response are only made for the next two canyon segments in the down-river direction, Marble Canyon (388 kilometers [99 miles]) and Grand Canyon (265 kilometers [165 miles]), respectively, where the majority of study sites are located.We show that a majority of monitoring sites increased in volume during a period of frequent controlled floods intended to rebuild sandbars. In the period from 2004 to 2020, which included seven controlled floods, a median discharge of 350 cubic meters per second (m3/s), and greater than average tributary sand inputs in more than half of the years, net deposition occurred at 86 percent of long-term monitoring sites. This period was preceded by a period of net erosion (1990–2003) when there was one controlled flood greater than the nominal powerplant capacity of 940 m3/s. During this period the median discharge from Glen Canyon Dam was 376 m3/s and greater than average sand inputs occurred in only 36 percent of those years. At the end of the monitoring period in 2020, 61 percent of the study sites measured since 1990 underwent a net increase in sand volume. For the entire 31-year period, these trends were statistically significant for all six sandbar types studied, indicating that increased frequency of controlled flooding maintained sandbar volume at the majority of sites monitored. These floods, also referred to as high-flow experiments (HFEs), are part of a decision-making protocol approved in 2012 for coordinating dam releases timed to occur following large sand inputs to the Colorado River by a major tributary.These findings are based on digital elevation models (DEMs) derived from approximately (~)1,800 repeat surveys of sandbar and channel bed topography made annually, or more frequently, at the 45 long-term monitoring sites, of which 31 have been monitored since 1990 and 14 were added between 1990 and 2008. This large collection of monitoring sites comprises just 7 to 9 percent of all sandbars in Marble and Grand Canyons, respectively. Nevertheless, when compared with measurements of a larger sample, these sites provide consistent characterization of average sandbar response, despite the local variability in channel and debris fan geometry. We use sand volume and normalized sand volume for tracking geomorphic changes of sandbars, because these metrics are sensitive to both changes in sandbar area and sandbar elevation. Based on checkpoint comparisons and repeat measurements, DEM elevation uncertainty was determined to be ±0.05 meter (m) and this uncertainty was used in a spatially uniform estimate of volume uncertainty. We find that the magnitudes of the topographic changes were substantially greater than the measurement uncertainty.Sandbars of similar type throughout both Marble and Grand Canyons have responded similarly during the period of the HFE protocol, despite variations in sand supply and longitudinal extent of those inputs. It should be noted that tributary-supplied sand to Glen Canyon is negligible, much of the riverbed is now armored with cobbles, and the channel bed degradation is irreversible in the current flow and sediment supply regime. Because all these HFEs have been conducted during periods of sediment enrichment, other factors such as vegetation and geomorphic setting are likely the primary causes of variation among the monitoring sites. A larger percentage of the sandbar population, predominantly located in narrow reaches where stage changes are greater, is composed of sandbar types that remain dynamic and consistently aggrade during HFEs. In contrast, wide reaches of the river corridor where stage change is not as great are characterized by sandbars that have been stabilized by vegetation and progressive aggradation during floods. In the former case, a majority of sandbars are likely to remain dynamic, requiring continued use of HFEs to achieve desired management goals. In the latter case, HFEs can do no better than replace the sediment eroded during normal dam operation between high-flow events, as they become less effective because of a diminishing amount of accommodation space available for deposition. Long-term sandbar trajectory and the continued effectiveness of HFEs are related to the differential vegetation establishment at each bar type. Future sandbar monitoring may need to consider the effects of riparian vegetation removal.

First posted October 7, 2022 For additional information, contact: Southwest Biological Science CenterU.S. Geological Survey2255 N. Gemini DriveFlagstaff, AZ 86001 Bathymetric and topographic data were collected from May 2013 to February 2016 along the 15.84-mile reach of the Colorado River spanning from Glen Canyon Dam to Lees Ferry in Glen Canyon National Recreation Area, Arizona. Channel bathymetry was mapped using multibeam and singlebeam echo sounders; subaerial topography was mapped using a combination of ground-based total stations and aerial photogrammetry. These data were combined to produce a digital elevation model (DEM), spatially variable estimates of DEM uncertainty, and bed-substrate distribution maps. This project is part of a larger effort to monitor the status and trends of sand storage along the Colorado River in Glen Canyon National Recreation Area and Grand Canyon National Park. This report documents the study methodologies (survey methods and post-processing procedures, DEM production and uncertainty assessment, and bed-substrate classification) and presents the resulting datasets.

This paper introduces a method for determining spatially‐distributed, 2‐D bedload rates using repeat, high‐resolution surveys of the bed topography. As opposed to existing methods, bedform parameters and bedload rates are computed from bed elevation profiles interpolated along the local bedform velocities. The bedform velocity fields are computed applying Large‐Scale Particle Image Velocimetry, initially developed for surface velocity measurements, to pairs of successive Digital Elevation Models (DEMs). The bathymetry data are interpolated along the direction of each bedform velocity and the mean height of the closest bedform is computed. The dune shape factor is also evaluated along each bedform direction of travel. The local bedload fluxes can be computed by multiplying the bedform velocity by its mean height averaged over the successive two DEMs, and they can be time‐averaged over a series of DEM pairs. This method is applied to a high‐resolution acoustical survey of an approximately 300 m long by 40 m wide reach of the Colorado River in Grand Canyon upstream from Diamond Creek, USA. The repeat period was about 6–10 min and bed elevation was interpolated every 0.25 m. The obtained results provide insight to the spatial and temporal variability of bedload rates, bedform parameters and bedload fluxes through cross‐sections. The method can be applied to other repeated acoustical surveys of river reaches provided that the space and time resolutions are high enough to capture the local movement of bedforms.

Streamflow regulation compounded by regional drought has resulted in up to 22% reduction in channel width, changes in channel planform, expansion of riparian vegetation, and alterations to floodplain habitat on the Colorado River in Meander Canyon, Utah. Although some changes in channel width occurred between the 1940s and 1980s, coinciding with major phases of upstream water development, larger decreases in channel width occurred between 1993 and 2006 during periods of exceptionally low annual floods. These findings illustrate that low runoff associated with regional drought and climate change may cause changes in river channel form that accelerate and compound the effects of upstream water development. Declining peak flows have also resulted in disconnection between the wetted channel and floodplains, where inundated back‐levee depressions provide habitat used by two species of threatened and endangered native fish. Despite this disconnection, some back‐levee depressions on the floodplain continue to be inundated by ~1.5‐year recurrence floods via connections created by tributary mouths, floodplain outflow channels, and levee breaches excavated by resident beaver. These changes are shown by analysis of aerial images, high‐resolution bathymetric and topographic measurements, and 2‐dimensional streamflow modeling.

Completion of Glen Canyon Dam in 1963 resulted in complete elimination of sediment delivery from the upstream Colorado River basin to Grand Canyon and nearly complete control of spring snowmelt floods responsible for creating channel and bar morphology. Management of the river ecosystem in Grand Canyon National Park now relies on dam-release floods to redistribute tributary-derived sediment accumulated on the channel bed to higher-elevation sandbars. Here, we used multivariate mixing analysis of sediment elemental compositions to evaluate the extent to which flood deposits derive from tributary-supplied sand compared to reworked, relict predam sediment. The concentrations of seven major and trace elements (Fe, Ca, K, Ti, Rb, Sr, and Zr) were measured in very fine-, fine-, and medium-grained sand from flood deposits using X-ray fluorescence and interpreted using a Bayesian mixing model to characterize the proportion of sand originating from the Paria River, the only major tributary within the study reach. Flood deposits from the 2013 and 2014 controlled floods contained 69 % +/- 16 % and 84 % +/- 20 % Paria River-derived material, respectively, with substantial variation among sites. Based on a sand mass balance, we calculated that under decreasing storage conditions since 1963, similar to 77 %-83 % of the annual Paria River sand flux needs to be retained within the mass of active sand stored in Marble Canyon each year to reach the observed concentration of Paria River sand at sample locations. This finding suggests that the use of controlled floods may continue to be effective for sandbar maintenance, provided sand inputs from the Paria River do not decline.

Turbulence‐resolving simulations elucidate key elements of fluid dynamics and sediment transport in fluvial environments. This research presents a feasible strategy for applying state‐of‐the‐art computational fluid mechanics to the study of sediment transport and morphodynamic processes in lateral separation zones, which are common features in canyon rivers where massive lateral flow separation causes large‐scale turbulence that controls sediment erosion and deposition. An eddy‐resolving model was developed and tested at the field‐scale, coupling a viscous flow and sediment transport solver using Detached Eddy Simulation techniques. A morphodynamic model was applied to the viscous flow/sediment solver to calculate erosion and deposition. A simulation of turbulence was performed at the grid resolution for a straight channel to determine the relative contributions of modeled and resolved diffusivity. The time‐dependent, energetically important, correlative, non‐stationary signals of the simulated quantities were captured at the lateral separation zone. Strong periodic signals featured by high amplitude were found at the separation zone, while low frequency pulsations were observed at the reattachment zone of the lateral separation zone. Interactions between the eddies and the loose bed boundaries resulted in erosion of sediment at the main channel followed by deposition at the primary eddy and eddy bars.

Atmospheric & Space Technology Research Associates Lidar Technologies (ASTRALiTe) has developed a polarization-based bathymetric lidar, called the edge™. The polarization approach differs from conventional lidar bathymeters that utilize waveform digitization and complex algorithms to separate laser reflections from the water surface and bottom. While discriminating and identifying these reflections in shallow depths can prove problematic for these conventional systems, a polarization-based lidar is designed to separately detect the reflection from the surface, which is polarized, and the bottom reflection which is depolarized. The U.S. Geological Survey is interested in defining the range of conditions under which the edge™ lidar can provide useful, accurate information on river depths. Potential riverine applications include collecting bathymetric data for non-contact streamflow measurement and mapping shallow water areas that pose a hazard to acoustic instruments deployed from boats. A critical factor influencing the success of a bathymetric lidar survey is water clarity. Suspended sediment and/or organic matter in the water column can scatter laser pulses and impede detection of the river bed. To gain a better understanding of how the edge™ lidar performs in various field settings, we selected several rivers with a range of water clarities. At each field site in situ measurements of the inherent optical properties of the water column (e.g., the absorption and attenuation coefficients at 532 nm, the wavelength of the edge™ laser) were collected as well as measurements of turbidity and chlorophyll and colored dissolved organic matter concentrations. The Colorado River near Lees Ferry, AZ, was selected because the turbidity is relatively low, 0.5 NTU, the water surface is generally smooth, and the substrate is predominantly sand. These conditions, along with the wide range of depths at this site, allowed the maximum depth penetration of the edge™ to be quantified. Preliminary results indicate that depths up to 5 m can be measured under such ideal conditions. The slightly more turbid (1.9 NTU) Blue River near Kremmling, CO, has a sand and gravel bed and had a smooth water surface at the time of survey. The edge™ was able to detect the bed up to the maximum depth of transects surveyed at this site, 1.2 m. The Colorado River near Parshall, CO, was also surveyed with the edge™ during a relatively high flow period in late spring. At the time of the survey, riffles were present at the water surface, the turbidity was higher than the Blue River (6.2 NTU), and the bed was composed of cobbles. Two transects surveyed with the edge™ at this site had a maximum depth of approximately 0.9 m. Despite a similar range in depths, one of the transects displayed a lower point density along the deepest portion of the transect. The reduced river bed coverage at this transect could be attributed to the elevated turbidity levels at this field site. A final field evaluation is planned for the Niobrara River near Norden, NE, to quantify the maximum turbidity under which the lidar can provide reliable bathymetry. The turbidity of the Niobrara River often exceeds 10 NTU, making this river a critical end member for evaluating the performance of the edge™ across a wide range of water clarities.

Controlled floods released from dams have become a common restoration strategy in river systems worldwide. Here we present a morphodynamic model of sandbar volume change for a subset of sandbars of the Colorado River in Grand Canyon National Park, where controlled floods are part of a management strategy focused on sandbar maintenance. We simulate sandbars as a triangular wedge, where deposition and erosion are modeled using physically based approaches that are driven by nearly continuous observations of flow and suspended sand concentration. We optimize an eddy exchange coefficient and erosion rate parameter by comparing model predictions to measured bar volumes. The model captures most of the variability in observed volume changes, and demonstrates the importance of flood frequency and sand concentration on average bar size. The model is easily implemented and adaptable, providing a means for predicting the future behavior of sandbars under a variety of streamflow and sediment supply scenarios.

This paper summarizes field trials to evaluate the performance of a prototype compact topo-bathymetric lidar sensor for surveying rivers. The sensor uses a novel polarization technique to distinguish between laser returns from the water surface and streambed and its size and weight permit deployment from a small unmanned aerial system (sUAS) or a boat. Field testing was designed to identify the range of operational conditions under which the sensor can provide accurate information on river depths. For accuracy assessment, conventional, field-based depth measurements were collected by wading and sonar. Additionally, optical properties of the rivers were measured in situ. Wading and lidar bathymetry comparisons in relatively shallow channels yielded observed versus predicted (OP) regression R-2 values ranging from 0.60 to 0.97. A comparison between sonar and lidar bathymetry in a deeper river resulted in an OP R-2 of 0.72. Absorption and attenuation coefficients at the 532 nm wavelength of the lidar were recorded in the field and the highest values of these inherent optical properties were at sites with the highest turbidity and highest concentrations of colored dissolved organic matter, chlorophyll, and suspended sediment. At these sites, which included both sand and gravel/cobble beds, the point density of riverbed returns was not uniform, with areas of sparse coverage occurring primarily in deeper water. However, submerged objects and slopes could be resolved in the lidar point clouds.

Obstacle marks are instream bedforms, typically composed of an upstream frontal scour hole and a downstream sediment accumulation in the vicinity of an obstacle. Local scouring at infrastructure (e.g. bridge piers) is a well-studied phenomenon in hydraulic engineering, while less attention is given to the time-dependent evolution of frontal scour holes at instream boulders and their geometric relations (depth to width, and length ratio). Furthermore, a comparison between laboratory studies and field observations is rare. Therefore, the morphodynamic importance of such scour features to fluvial sediment transport and morphological change is largely unknown. In this study, obstacle marks at boulder-like obstructions were physically modelled in 30 unscaled process-focused flume experiments (runtime per experiment >= 5760 min) at a range of flows (subcritical, clear-water conditions, emergent and submerged water levels) and boundary conditions designed to represent the field setting (i.e. obstacle tilting, and limited thickness of the alluvial layer). Additionally, geometries of scour holes at 90 in-situ boulders (diameter >= 1 m) located in a 50-km segment of the Colorado River in Marble Canyon (AZ) were measured from a 1 m-resolution digital elevation model. Flume experiments reveal similar evolution of local scouring, irrespective of hydraulic conditions, controlled by the scour incision, whereas the thickness of the alluvial layer and obstacle tilting into the evolving frontal scour hole limit incision. Three temporal evolution phases-(1) rapid incision, (2) decreasing incision, and (3) scour widening-are identified based on statistical analysis of spatiotemporal bed elevation time series. A quantitative model is presented that mechanistically predicts enlargement in local scour length and width based on (1) scour depth, (2) the inclination of scour slopes, and (3) the planform area of the frontal scour hole bottom. The comparison of field observations and laboratory results demonstrates scale invariance of geometry, which implies similitude of processes and form rather than equifinality.

A common impact on riparian ecosystem function following river regulation is the expansion and encroachment of riparian plant species in the active river channels and floodplain, which reduces flow of water and suspended sediment between the river, riparian area and upland ecosystems. We characterised riparian plant species occurrence and quantified encroachment within the dam‐regulated Colorado River in Grand Canyon, Arizona, USA. We mapped 10 riparian species with high‐resolution multispectral imagery and examined effects of river hydrology and geomorphology on the spatial distribution of plant species and open sand. Analysis spanned an image time series from 2002 to 2009 to 2013, a period when plant species and sand were spatially dynamic and operations of Glen Canyon Dam included daily hydro‐peaking and small episodic controlled flood releases. Plant species occurrence and encroachment rates varied with hydrology, geomorphology and local species pool. Encroachment was greatest on surfaces frequently inundated by hydro‐peaking. Seep willow ( Baccharis spp.), tamarisk ( Tamarix spp.) and arrowweed ( Pluchea sericea ) were the primary encroaching woody species. Common reed ( Phragmites australis ) and horsetail ( Equisetum xferrissii ) were the primary encroaching herbaceous species. Encroachment composition from 2002 to 2009 was similar to the entire riparian landscape, whereas encroachment from 2009 to 2013 primarily consisted of seep willow and early colonising herbaceous species. Emergence of seep willow and arrowweed after burial by sand deposited by controlled floods indicated that those species were resilient to this form of disturbance. Describing patterns of species encroachment is an important step towards designing flow regimes that favour riparian species and ecosystem functions valued by stakeholders.

Bedrock‐canyon rivers tend to be supply limited because they are efficient transporters of sediment and not because the upstream supply of sediment is small. A byproduct of this supply limitation is that the finer alluvium stored in these rivers has shorter residence times and smaller volumes than in alluvial rivers. To improve our understanding of disequilibrium sediment transport and its effect on sand storage in bedrock‐canyon rivers, we undertook a 20‐year study, synthesized herein, of the Colorado River in Grand Canyon. Despite the large loads for which it was renowned, this river exhibited evidence of natural sand‐supply limitation and became the perfect natural laboratory for studying sand transport in a bedrock canyon after upstream dam construction exacerbated this supply limitation. During our study, we made and analyzed an unprecedented ∼2.5 million measurements of the suspended and bed sediment. Results indicate that sand storage in this bedrock‐canyon river is self‐limiting owing to the physical controls of flow and grain size causing negative feedbacks that likely also operate in other bedrock‐canyon rivers. Following episodic tributary floods that supply finer sand, sand migrates quickly downstream in the form of a wave in which large systematic changes in bed‐sand grain size occur. These grain‐size changes cause discharge‐independent systematic changes in suspended‐sand concentration in excess of a factor of 20. Although the tributary supply of sand increases the amount of sand storage, it also greatly increases the downstream sand transport by causing bed‐sand fining, thus limiting the residence time and volume of sand storage.