River ecosystems are threatened by interactions among river regulation, non-native species, and climate change. Water use has exceeded supply in USA's Colorado River basin draining its two largest storage reservoirs (Lake Powell and Lake Mead). In 2022, Lake Powell began releasing water from its lower epilimnion into the Grand Canyon segment of the Colorado River, which (1) increased rates of fish passage from the reservoir through the dam and (2) created river temperatures suitable for establishment of non-native fishes. Subsequently, smallmouth bass (Micropterus dolomieu) reproduced there for the first time. To assist managers concerned about this invasion, we developed models that (1) predicted propagule pressure at different reservoir elevations and (2) linked reservoir storage/operations, water temperatures, and population dynamics to forecast smallmouth bass population growth potential. Maintaining Lake Powell elevations above 1094 m (3590 ft) would likely minimize propagule pressure from the reservoir and create downstream conditions that minimize smallmouth bass population growth. Dam and reservoir management will likely be less effective for managing smallmouth bass if smallmouth bass become abundant in far downstream reaches.

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.

AbstractPersistent overuse of water supplies from the Colorado River during recent decades has substantially depleted large storage reservoirs and triggered mandatory cutbacks in water use. The river holds critical importance to more than 40 million people and more than two million hectares of cropland. Therefore, a full accounting of where the river’s water goes en route to its delta is necessary. Detailed knowledge of how and where the river’s water is used can aid design of strategies and plans for bringing water use into balance with available supplies. Here we apply authoritative primary data sources and modeled crop and riparian/wetland evapotranspiration estimates to compile a water budget based on average consumptive water use during 2000–2019. Overall water consumption includes both direct human uses in the municipal, commercial, industrial, and agricultural sectors, as well as indirect water losses to reservoir evaporation and water consumed through riparian/wetland evapotranspiration. Irrigated agriculture is responsible for 74% of direct human uses and 52% of overall water consumption. Water consumed for agriculture amounts to three times all other direct uses combined. Cattle feed crops including alfalfa and other grass hays account for 46% of all direct water consumption.

Sediment budgets are widely used to measure reach-scale sediment accumulation and evacuation. Such measurements, however, cannot determine when the disturbance is major and the measured sediment mass imbalance is reflective of a river adjusting to a new equilibrium state, as opposed to situations when the disturbance is minor, and the mass imbalance is reflective of a river adjusting within its existing behavioral regime. Sediment sorting among channels and floodplains can have a large effect on how a river responds to a disturbance. Fine sediment may accumulate in the floodplains while coarser sediment erodes from the channel bed. We demonstrate that if a sediment budget does not account for the different behavior and destination of grain sizes, the budget cannot reveal important channel adjustments. In this study, we evaluated how a sand bed river responded to increases in sediment supply by partitioning a sediment budget among silt/clay and five sand fractions. On average, 12 metric tons/meter (downstream)/year of sand was evacuated from the system, but sorting caused channel margins to behave differently from vegetated islands, revealing how a river can slightly narrow while in deficit. Floodplain shaving and bed coarsening evacuated sediment while channel geometry barely changed, consistent with a river adjusting to a minor disturbance within its behavioral regime. This study is an important reminder that sediment mass imbalance does not always lead to channel change. Mechanisms such as floodplain shaving and bed textural change help rivers absorb minor disturbances and resist channel change. Rivers are experiencing increasing human interference, and global warming will significantly impact the amount of water and sediment rivers must convey, placing an even greater need to predict how rivers respond to disturbance. One important problem is predicting the susceptibility of a river to change. While measuring the amount of sediment that accumulates or evacuates from a reach provides clues to how rivers respond to disturbances, such measurements cannot determine whether reach-average morphology, such as width, depth, slope, and planform, will significantly change or minor adjustments may occur. In this study, we evaluated how a sand bed river responded to an increase in sediment supply that occurred similar to 60 years ago using continuous sediment transport measurements and aerial images to measure channel adjustment. We found that, on average, 12 metric tons/meter (downstream)/year of sand was evacuated from the system, but sediment sorting caused channel margins to behave differently from vegetated islands such that the overall channel characteristics barely changed. This study demonstrates that sediment accumulation or evacuation does not always lead to channel change because other processes related to sand sorting can accommodate disturbances, making sand bed rivers more resistant to change. A sediment budget must be segregated by grain sizes found in the channel and floodplain to evaluate morphologic changes in sand bed rivers Channel adjustment in sand bed rivers is strongly influenced by sand sorting in channel-floodplain components Mass imbalance doesn't necessitate channel change: floodplain shaving and bed texture absorb disturbances with minor channel adjustment

During much of the 21st century, natural runoff in the Colorado River basin has declined, while consumption has remained relatively constant, leading to historically low reservoir storage. Between January 2000 and April 2023, the amount of water stored in Lake Mead and Lake Powell, the two largest reservoirs in the United States, declined by 33.5 million acre feet (41.3 billion cubic meters). As of April 2023, total basin-wide storage was sufficient to support the 21st century average rate of basin-wide consumption for only 15 months. Runoff in spring 2023 is predicted to be large, providing a short-term reprieve. However, it will take four to five additional unusually wet years in succession to refill Lake Powell and Lake Mead if basin-wide water use remains unchanged. Increasing evapotranspiration and dry soils associated with global climate change makes such a scenario unlikely. To stabilize reservoir storage, basin-wide use needs to equal modern runoff. To recover reservoir storage, basin-wide use needs to decline even more. Based on 21st century average runoff, a 13%-20% decline in basin-wide use would allow for stabilization and some reservoir storage recovery. Future policy debate about reservoir operations will inevitably concern whether most, or all, reservoir storage should be in Lake Mead or in Lake Powell. The choice of one or the other will result in significantly different environmental and recreational outcomes for Glen Canyon and the Grand Canyon.

Water development has threatened the ecological integrity of riverine ecosystems. Increasing water demand, persistent drought, and climate change exacerbate the effects of habitat degradation and loss in altered systems such as the Colorado River basin. Today, biologists are challenged to identify management actions that benefit native fishes while not hindering water development or management. Herein, we discuss the importance of the natural flow regime for functioning riverine ecosystems and provide examples from four tributaries to the Green River, a major headwater branch of the Colorado River. These tributaries represent a gradient of impacts ranging from water abstraction to the point of complete seasonal desiccation to a relatively natural flow regime, and consequently have maintained different levels of instream habitat complexity and native fish persistence. Despite decades of management, endangered species lack self-sustaining populations and other native species have been extirpated from over half their ranges, which begs the question: can water development and fish conservation be balanced under current water laws and climate change-driven declines in runoff? Given the continued decline in freshwater biodiversity and abundance occurring across the globe, we contend that immediate designation of rivers with natural flow regimes as freshwater conservation areas will enhance native species recovery.

Dams and reservoirs significantly affect rivers in the northern third of the world, and hundreds of new dams are proposed elsewhere. Dams fragment the channel network and block the natural flux of sediment and organic matter. Reservoir operations typically seek to transform the natural flow regime to provide flood control, a relatively constant water supply, or to allow production of hydroelectricity, thereby reducing the natural diversity of flow regimes. Most large reservoirs thermally stratify in summer and release relatively cool water in summer. The changes in stream flow and sediment supply cause downstream changes in sediment mass balance and may lead to bed incision and disconnection of the post-dam flow regime from adjacent floodplains. Change in thermal regime allows proliferation of non-native species often introduced into reservoirs and tailwaters. Flood control typically leads to an expansion of riparian vegetation.

Our ability to sustainably manage the Colorado River is clearly in doubt. The Bureau of Reclamation’s 2012 Water Supply and Demand Study demonstrated the precarious balance that currently exists between water supply and the amount consumptively used by society. A future with either declining water supplies or additional consumptive uses will undoubtedly upset this balance. This balance is threatened, because:

Much has been learned about Colorado River hydrology since the severe sustained drought study in 1995. We summarize our updated understanding of plausible future drought conditions by considering historical flows, tree-ring reconstructions, and climate change. We focus on natural streamflow at Lees Ferry, the primary metric used to quantify the runoff in the Colorado River Basin. We identify drought periods using historical records and tree-ring reconstructed streamflow at Lees Ferry, which we then use to characterize potential future droughts. Resampling from past drought periods generates plausible future conditions to consider during planning. We produced three drought scenarios, each comprising 100 streamflow sequences to be used as input to systems operation and management models. We used analysis of the duration-severity and cumulative deficit relative to the mean natural flow to evaluate droughts and drought simulations and show that the current millennium drought that started in 2000 has an average flow far less than the historical record. However, the flows reconstructed from tree rings or future flows projected from climate models indicate that even more severe droughts are possible. When used as input to the Colorado River Simulation System the drought scenarios developed indicate considerable periods when Lake Powell falls below its hydropower penstocks, indicating a need to rethink management and operation of these reservoirs during these critical conditions.

A continuation of the current 23-year-long drought will require difficult decisions to prevent further decline

Drought has impacted the Colorado River basin for the past 20 years and is predicted to continue. In response, decisions about how much water should be stored in large reservoirs and how much water can be consumptively used will be necessary. These decisions have the potential to limit riverine ecosystem management options through the effect water-supply decisions have on reservoir elevations. We used projected hydrology and river temperatures to compare the outcome of combinations of water storage scenarios and consumptive use limits on metrics associated with ecosystem management of the Colorado River in Grand Canyon. Ecosystem management metrics included the ability to implement designer flows, temperature suitability for fishes, and fragmentation. We compared current water management operations to prioritizing storage in either Lake Mead or Lake Powell combined with three levels of consumptive use. Projected reservoir levels limited environmental flow delivery and increased fragmentation regardless of where water was stored if consumptive use was not limited. Warmer river temperatures associated with low reservoir levels are likely, creating suitable conditions for non-native species of concern, such as smallmouth bass. Water storage decisions provided variability and management flexibility, but water storage was less important when less water was available, highlighting the importance of keeping water in the system to provide flexibility for achieving ecosystem goals.

Managing the world’s freshwater supply to meet societal and environmental needs in a changing climate is one of the biggest challenges for the 21st century. Dams provide water security; however, the allocation of dwindling water supply among reservoirs could exacerbate or ameliorate the effects of climate change on aquatic communities. Here, we show that the relative sensitivity of river thermal regimes to direct impacts of climate change and societal decisions concerning water storage vary substantially throughout a river basin. In the absence of interspecific interactions, future Colorado River temperatures would appear to benefit both endemic and nonnative fish species. However, endemic species are already declining or extirpated in locations where their ranges overlap with warmwater nonnatives and changes in water storage may lead to warming in some of the coolest portions of the river basin, facilitating further nonnative expansion. Integrating environmental considerations into ongoing water storage negotiations may lead to better resource outcomes than mitigating nonnative species impacts after the fact.

Channels change in response to natural or anthropogenic fluctuations in streamflow and/or sediment supply and measurements of channel change are critical to many river management applications. Whereas repeated field surveys are costly and time‐consuming, remote sensing can be used to detect channel change at multiple temporal and spatial scales. Repeat images have been widely used to measure long‐term channel change, but these measurements are only significant if the magnitude of change exceeds the uncertainty. Existing methods for characterizing uncertainty have two important limitations. First, while the use of a spatially variable image co‐registration error avoids the assumption that errors are spatially uniform, this type of error, as originally formulated, can only be applied to linear channel adjustments, which provide less information on channel change than polygons of erosion and deposition. Second, previous methods use a level‐of‐detection (LoD) threshold to remove non‐significant measurements, which is problematic because real changes that occurred but were smaller than the LoD threshold would be removed. In this study, we present a new method of quantifying uncertainty associated with channel change based on probabilistic, spatially varying estimates of co‐registration error and digitization uncertainty that obviates a LoD threshold. The spatially distributed probabilistic (SDP) method can be applied to both linear channel adjustments and polygons of erosion and deposition, making this the first uncertainty method generalizable to all metrics of channel change. Using a case study from the Yampa River, Colorado, we show that the SDP method reduced the magnitude of uncertainty and enabled us to detect smaller channel changes as significant. Additionally, the distributional information provided by the SDP method allowed us to report the magnitude of channel change with an appropriate level of confidence in cases where a simple LoD approach yielded an indeterminate result. © 2020 John Wiley & Sons, Ltd.

The lower Green River episodically narrowed between the mid-1930s and present day through deposition of new floodplains within a wider channel that had been established and/or maintained during the early twentieth century pluvial period. Comparison of air photos spanning a 74-yr period (1940-2014) and covering a 61 km study area shows that the channel narrowed by 12% from 138 +/- 3.4 m to 122 +/- 2.1 m. Stratigraphic and sedimentologic analysis and tree ring dating of a floodplain trench corroborates the air photo analysis and suggests that the initial phase of floodplain formation began by the mid-1930s, approximately the same time that the flow regime decreased in total annual and peak annual flow. Tamarisk, a nonnative shrub, began to establish in the 1930s as well. Narrowing from the 1940s to the mid-1980s was insignificant, because floodplain formation was approximately matched by bank erosion. Air photo analysis demonstrates that the most significant episode of narrowing was underway by the late 1980s, and analysis of the trench shows that floodplain formation had begun in the mid-1980s during a multi-year period of low peak annual flow. Air photo analysis shows that mean channel width decreased by similar to 7% between 1993 and 2009. A new phase of narrowing may have begun in 2003, based on evidence in the trench. Comparison of field surveys made in 1998 and 2015 in an 8.5 km reach near Fort Bottom suggests that narrowing continues and demonstrates that new floodplain formation has been a very small proportion of the total annual fine sediment flux of the Green River. Vertical accretion of new floodplains near Fort Bottom averaged 2.4 m between 1998 and 2015 but only accounted for similar to 1.5% of the estimated fine sediment flux during that period. Flood control by Flaming Gorge Dam after 1962 significantly influenced flow regime, reducing the magnitude of the annual snowmelt flood and increasing the magnitude of base flows. Though narrowing was initiated by changes in flow regime, native and nonnative riparian vegetation promoted floodplain formation and channel narrowing especially through establishment on channel bars and incipient floodplains during years of small annual floods.

Predictions of river channel adjustment to changes in streamflow regime based on relations between mean channel characteristics and mean flood magnitude can be useful to evaluate average channel response. However, because these relations assume equilibrium sediment transport, their applicability to cases where streamflow and sediment transport are decoupled may be limited. These general relations also lack the specificity that is required to connect specific characteristics of the streamflow and sediment regime with the dynamics of channel morphological change that create channel complexity, which is often of ecological interest. We integrate historical records of channel change, observations of scour and fill during a snowmelt flood, measurements of sediment transport, and predictions from a two-dimensional streamflow model to describe how annual peak flow magnitude and peak-flow duration interact with the upstream sediment supply to control channel form for a 15-km study reach on the regulated Green River in Canyonlands National Park, Utah. Two major decadal-scale episodes of channel narrowing have occurred within the study area. For each of these episodes, the reduction in average channel width was consistent with the change predicted by hydraulic geometry relations as a function of average flood magnitude. However, channel narrowing occurred during periods of exceptionally low annual floods. The most recent episode of channel narrowing occurred between 1988 and 2009, during low-flow cycles when the 5-yr mean peak flow was less than 60% of the long-term (1959–2016) mean peak flow. These findings, together with findings from previous studies, demonstrate that decreases in peak-flow magnitude caused by streamflow regulation, climate change, or a combination of those factors have driven episodes of channel narrowing on the Green River. Observations of streamflow, sediment-transport, and morphologic change coupled with predictions from a two-dimensional streamflow model indicate that peak flow magnitudes of at least 75% of the long-term mean peak flow are required to transport bed-material sand in suspension in all regions of the multi-thread channel and that the ~2-month duration of the snowmelt flood played an important role in creating conditions necessary to maintain channel conveyance. These results indicate that detailed characterizations of channel response such as these are needed to predict how river channels will respond to changes in streamflow regime that affect annual peak flow magnitude and duration.

Many canyon‐bound rivers have been dammed and downstream flow and water temperatures modified. In some regions, climate change is expected to cause lower storage in reservoirs and warmer release temperatures, which may further alter downstream flow and thermal regimes. To anticipate potential future changes, we first need to understand the dominant heat transfer mechanisms in canyon‐bound river systems. Toward this end, we adapt a dynamic process‐based river routing and temperature model to account for complex shading and radiation characteristics found in canyon‐bound rivers. We apply the model to a 362 km segment of the Colorado River in Grand Canyon National Park, USA to simulate temperature over an 18‐year period. Extensive temperature and flow data sets from within the canyon were used to assess model performance. At the most downstream gaging location, root mean square errors of hourly flow routing and temperature predictions were 11.5 m3/s and 0.93°C, respectively. We found that heat fluxes controlling temperatures were highly variable over space and time, primarily due to shortwave radiation dynamics and hydropeaking flow conditions. Additionally, the large differences between air and water temperature during summer periods resulted in high sensible and latent heat fluxes. Sensitivity analyses indicate that reservoir release temperatures are most influential above the RM88 gage (141 km below Glen Canyon Dam), while a combination of discharge, shortwave radiation, and air temperature become more important farther downstream. This study illustrates the importance of understanding the spatial and temporal variability of topographic shading when predicting water temperatures in canyon‐bound rivers.