Charles M. Cannon
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Charles M. Cannon has a total of 33 co-authors. Here are the top co-authors by number of shared publications:
- David L. George
United States Geological Survey, USA
5 shared publications - James E. O'Connor
U.S. Geological Survey Portland Oregon USA
4 shared publications - Andrew Meigs
Oregon State University, Corvallis, Oregon, U.S.A.
2 shared publications Oregon State University
2 shared publications- Ray E. Wells
U.S. Geological Survey, 2130 SW 5th Avenue, Portland, Oregon, 97201, USA
2 shared publications - Richard M. Iverson
Cascades Volcano Observatory U.S. Geological Survey Vancouver WA USA
2 shared publications
Latest Publications
- Open Access
- 1 January 2025
- Published by US Geological Survey
Modeling floods, sediment entrainment, and downstream debris flows from hypothetical breaches of the blockage at Spirit Lake, Washington
This report describes a modeling investigation by the U.S. Geological Survey (USGS) of hazards in the Toutle and Cowlitz River valleys posed by hypothetical outburst floods from Spirit Lake, Washington. A massive debris avalanche resulting from the collapse of Mount St. Helens’ north flank during the May 18, 1980, eruption blocked Spirit Lake’s natural outlet into the North Fork Toutle River. Lacking a natural outlet, subsequent runoff in the Spirit Lake watershed contributed to a rising lake level, elevating the potential for debris-dam breaching or catastrophic failure. The influence of highly erodible bed sediment in the upper North Fork Toutle River on downstream flood and debris-flow dynamics and extent is assessed in this study. Simulations of clear-water (non-erosive) outburst floods were used as a baseline and compared to erosive flows that entrain large volumes of material and transition into debris flows along their flow path, revealing the influence of entrainment on hazard extent. Clear-water floods were modeled with the shallow water equations. Erosive flows were modeled with a two-phase granular fluid model that accommodates mobilization and incorporation of sediment from the bed into the overlying flow and resultant changes in flow rheology across a wide range of solid concentrations, from dilute suspensions to dense-granular debris flows. Entrainment of bed material was found to substantially increase the total flow volume (total volume of transported water and sediment is approximately 150 percent of the water volume for non-erosive flows). Erosive flows are shown to exhibit higher flow-front speeds and faster downstream arrival times than non-erosive flows, consistent with volume amplification effects near the actively mobilizing flow front. However, the larger total volume of transported material does not necessarily lead to an enhancement of total volume throughput (cumulative discharge) or inundation extent (total affected area) for all locations along the entire flow path; while entrainment leads to the displacement of a larger volume of material overall, much of this dislocated material (water and sediment) deposits upstream from the distal extent of the flows. These results are consistent with energetic considerations of initial potential energy and granular shear resistance.
- 1 January 2025
- Published by US Geological Survey
Open AccessModeling floods, sediment entrainment, and downstream debris flows from hypothetical breaches of the blockage at Spirit Lake, Washington
This report describes a modeling investigation by the U.S. Geological Survey (USGS) of hazards in the Toutle and Cowlitz River valleys posed by hypothetical outburst floods from Spirit Lake, Washington. A massive debris avalanche resulting from the collapse of Mount St. Helens’ north flank during the May 18, 1980, eruption blocked Spirit Lake’s natural outlet into the North Fork Toutle River. Lacking a natural outlet, subsequent runoff in the Spirit Lake watershed contributed to a rising lake level, elevating the potential for debris-dam breaching or catastrophic failure. The influence of highly erodible bed sediment in the upper North Fork Toutle River on downstream flood and debris-flow dynamics and extent is assessed in this study. Simulations of clear-water (non-erosive) outburst floods were used as a baseline and compared to erosive flows that entrain large volumes of material and transition into debris flows along their flow path, revealing the influence of entrainment on hazard extent. Clear-water floods were modeled with the shallow water equations. Erosive flows were modeled with a two-phase granular fluid model that accommodates mobilization and incorporation of sediment from the bed into the overlying flow and resultant changes in flow rheology across a wide range of solid concentrations, from dilute suspensions to dense-granular debris flows. Entrainment of bed material was found to substantially increase the total flow volume (total volume of transported water and sediment is approximately 150 percent of the water volume for non-erosive flows). Erosive flows are shown to exhibit higher flow-front speeds and faster downstream arrival times than non-erosive flows, consistent with volume amplification effects near the actively mobilizing flow front. However, the larger total volume of transported material does not necessarily lead to an enhancement of total volume throughput (cumulative discharge) or inundation extent (total affected area) for all locations along the entire flow path; while entrainment leads to the displacement of a larger volume of material overall, much of this dislocated material (water and sediment) deposits upstream from the distal extent of the flows. These results are consistent with energetic considerations of initial potential energy and granular shear resistance.
- 1 January 2025
- Published by US Geological Survey
Using the D-Claw software package to model lahars in the Middle Fork Nooksack River drainage and beyond, Mount Baker, Washington
Lahars, or volcanic mudflows, are the most hazardous eruption-related phenomena that will affect communities living along rivers that originate on Mount Baker. In the past 15,000 years, the largest lahars from Mount Baker have affected the Middle Fork Nooksack River drainage and beyond. Here we use the physics-based D-Claw software package to model nine lahar scenarios that are initiated as water-saturated landslides between Sherman Crater and the Roman Wall on the Mount Baker edifice and flow down the Middle Fork Nooksack River. The scenarios range in volume from 1 to 260 million cubic meters and have an initial hydraulic permeability from 10−12 to 10−10 meters squared. Model output includes data such as flow depth, velocity, runout distance, area inundated, arrival time, and sediment concentration as well as information that allows scientists to calculate other important hydrologic characteristics such as lahar discharge. These data are important to officials who have the responsibility to plan for, or take mitigation measures against, future Mount Baker lahars. To check the validity of the D-Claw results, we compare the scenarios to known geologic information. We also compare D-Claw results with empirical models that have been used in the past to determine potential inundation areas, runout distances, and arrival times. These comparisons highlight similarities and differences between empirical and physics-based models. We also present D-Claw scenario-based animations to help scientists, officials, and lay people alike to visualize how future lahars could affect communities.
- 1 January 2025
- Published by US Geological Survey
Using the D-Claw software package to model lahars in the Middle Fork Nooksack River drainage and beyond, Mount Baker, Washington
Lahars, or volcanic mudflows, are the most hazardous eruption-related phenomena that will affect communities living along rivers that originate on Mount Baker. In the past 15,000 years, the largest lahars from Mount Baker have affected the Middle Fork Nooksack River drainage and beyond. Here we use the physics-based D-Claw software package to model nine lahar scenarios that are initiated as water-saturated landslides between Sherman Crater and the Roman Wall on the Mount Baker edifice and flow down the Middle Fork Nooksack River. The scenarios range in volume from 1 to 260 million cubic meters and have an initial hydraulic permeability from 10−12 to 10−10 meters squared. Model output includes data such as flow depth, velocity, runout distance, area inundated, arrival time, and sediment concentration as well as information that allows scientists to calculate other important hydrologic characteristics such as lahar discharge. These data are important to officials who have the responsibility to plan for, or take mitigation measures against, future Mount Baker lahars. To check the validity of the D-Claw results, we compare the scenarios to known geologic information. We also compare D-Claw results with empirical models that have been used in the past to determine potential inundation areas, runout distances, and arrival times. These comparisons highlight similarities and differences between empirical and physics-based models. We also present D-Claw scenario-based animations to help scientists, officials, and lay people alike to visualize how future lahars could affect communities.
- Open Access
- 22 December 2023
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 9 (51) , eadj3357
The magmatic origin of the Columbia River Gorge, USA
Along subduction zones, high-relief topography is associated with sustained volcanism parallel to the plate margin. However, the relationship between magmatism and mountain building in arcs is poorly understood. Here, we study patterns of surface deformation and correlated fluvial knickpoints in the Columbia River Gorge to link long-term magmatism to the uplift and ensuing topographic development of the Cascade Range. An upwarped paleochannel exposed in the walls of the Gorge constrains unsteady deep magma flux, the ratio of intrusive to extrusive magmatic contributions to topography, and the impact of magmatism on Columbia River incision since 3.5 million years ago. Geophysical data indicate that deep magma influx beneath the arc axis is ongoing and not aligned with the current locations of volcanic edifices, representing a broad regional influence on arc construction.
- 22 December 2023
- Published by American Association for the Advancement of Science (AAAS) in Science Advances
- Vol. 9 (51) , eadj3357
Open AccessThe magmatic origin of the Columbia River Gorge, USA
Along subduction zones, high-relief topography is associated with sustained volcanism parallel to the plate margin. However, the relationship between magmatism and mountain building in arcs is poorly understood. Here, we study patterns of surface deformation and correlated fluvial knickpoints in the Columbia River Gorge to link long-term magmatism to the uplift and ensuing topographic development of the Cascade Range. An upwarped paleochannel exposed in the walls of the Gorge constrains unsteady deep magma flux, the ratio of intrusive to extrusive magmatic contributions to topography, and the impact of magmatism on Columbia River incision since 3.5 million years ago. Geophysical data indicate that deep magma influx beneath the arc axis is ongoing and not aligned with the current locations of volcanic edifices, representing a broad regional influence on arc construction.
- 1 January 2022
- Published by US Geological Survey
Modeling the dynamics of lahars that originate as landslides on the west side of Mount Rainier, Washington
Large lahars pose substantial threats to people and property downstream from Mount Rainier volcano in Washington State. Geologic evidence indicates that these threats exist even during the absence of volcanic activity and that the threats are highest in the densely populated Puyallup and Nisqually River valleys on the west side of the volcano. However, the precise character of these threats can be difficult to anticipate.To help predict depths and rates of possible lahar inundation in the area, this report presents the results of simulations of hypothetical future lahars that originate high on the west side of Mount Rainier and travel downstream into the Puyallup and Nisqually River valleys. Many of the results portrayed as still images in the figures of this report are also available as animated files that can be accessed at the web address provided in the figure captions. We simulated eight scenarios, including worst-case scenarios in which the simulated lahars are similar in size and mobility to the approximately 260 million cubic meter (Mm3; 340 million cubic yard) Electron Mudflow lahar that descended from Mount Rainier and inundated the Puyallup River valley about 500 years ago. The other six scenarios place the worst-case scenarios in perspective by simulating lahars that originate from the same source areas but have smaller volumes or lesser mobilities.We perform our simulations using an open-source software package that we developed called D-Claw. The numerical model composing the kernel of D-Claw solves a system of five hyperbolic partial differential equations that describe the depth-averaged dynamics of static or flowing grain-fluid mixtures interacting with three-dimensional topography. In D-Claw, the volume fraction occupied by solid grains is a dependent variable that can freely evolve, enabling simulation of landslide liquefaction and of lahar interaction with static bodies of water. The latter feature facilitates a seamless simulation of a lahar in the Nisqually River valley entering Alder Lake reservoir.In the event of an approximately 260 Mm3 high-mobility lahar originating on the west side of Mount Rainier, our results point to two areas of pronounced hazard. One area, comprising the densely populated lowlands of Orting, Washington, and environs, could be inundated by lahars originating from either the Sunset Amphitheater or Tahoma Glacier headwall areas. In the worst-case scenario we consider for the Orting lowlands, which involves a 260 Mm3 high-mobility lahar originating from a landslide in the Sunset Amphitheater, a flow front approximately 4 meters deep and traveling about 4 meters per second reaches the Orting lowlands about 1 hour after the onset of slope failure. After passing through the Orting lowlands, the simulated lahar slows down and comes to rest in the valleys surrounding Sumner and Puyallup. A second area of pronounced hazard is the stretch of the Nisqually River valley beginning in Mount Rainier National Park and extending downstream to Alder Lake reservoir and Alder Dam. This area would be substantially affected in the worst-case scenario that involves a 260 Mm3 high-mobility lahar originating from the Tahoma Glacier headwall area—the locality identified by a previous study as the sector of Mount Rainier most prone to large-scale gravitational collapse. The simulated lahar passes through the area of Ashford, Washington, within about 20 minutes of the onset of slope failure and reaches the head of Alder Lake within about 50 minutes. The lahar ultimately displaces enough reservoir water to cause overtopping of the 100 meter (330 foot) tall Alder Dam, but consequences of such dam overtopping are not addressed in this report.
- 1 January 2022
- Published by US Geological Survey
Modeling the dynamics of lahars that originate as landslides on the west side of Mount Rainier, Washington
Large lahars pose substantial threats to people and property downstream from Mount Rainier volcano in Washington State. Geologic evidence indicates that these threats exist even during the absence of volcanic activity and that the threats are highest in the densely populated Puyallup and Nisqually River valleys on the west side of the volcano. However, the precise character of these threats can be difficult to anticipate.To help predict depths and rates of possible lahar inundation in the area, this report presents the results of simulations of hypothetical future lahars that originate high on the west side of Mount Rainier and travel downstream into the Puyallup and Nisqually River valleys. Many of the results portrayed as still images in the figures of this report are also available as animated files that can be accessed at the web address provided in the figure captions. We simulated eight scenarios, including worst-case scenarios in which the simulated lahars are similar in size and mobility to the approximately 260 million cubic meter (Mm3; 340 million cubic yard) Electron Mudflow lahar that descended from Mount Rainier and inundated the Puyallup River valley about 500 years ago. The other six scenarios place the worst-case scenarios in perspective by simulating lahars that originate from the same source areas but have smaller volumes or lesser mobilities.We perform our simulations using an open-source software package that we developed called D-Claw. The numerical model composing the kernel of D-Claw solves a system of five hyperbolic partial differential equations that describe the depth-averaged dynamics of static or flowing grain-fluid mixtures interacting with three-dimensional topography. In D-Claw, the volume fraction occupied by solid grains is a dependent variable that can freely evolve, enabling simulation of landslide liquefaction and of lahar interaction with static bodies of water. The latter feature facilitates a seamless simulation of a lahar in the Nisqually River valley entering Alder Lake reservoir.In the event of an approximately 260 Mm3 high-mobility lahar originating on the west side of Mount Rainier, our results point to two areas of pronounced hazard. One area, comprising the densely populated lowlands of Orting, Washington, and environs, could be inundated by lahars originating from either the Sunset Amphitheater or Tahoma Glacier headwall areas. In the worst-case scenario we consider for the Orting lowlands, which involves a 260 Mm3 high-mobility lahar originating from a landslide in the Sunset Amphitheater, a flow front approximately 4 meters deep and traveling about 4 meters per second reaches the Orting lowlands about 1 hour after the onset of slope failure. After passing through the Orting lowlands, the simulated lahar slows down and comes to rest in the valleys surrounding Sumner and Puyallup. A second area of pronounced hazard is the stretch of the Nisqually River valley beginning in Mount Rainier National Park and extending downstream to Alder Lake reservoir and Alder Dam. This area would be substantially affected in the worst-case scenario that involves a 260 Mm3 high-mobility lahar originating from the Tahoma Glacier headwall area—the locality identified by a previous study as the sector of Mount Rainier most prone to large-scale gravitational collapse. The simulated lahar passes through the area of Ashford, Washington, within about 20 minutes of the onset of slope failure and reaches the head of Alder Lake within about 50 minutes. The lahar ultimately displaces enough reservoir water to cause overtopping of the 100 meter (330 foot) tall Alder Dam, but consequences of such dam overtopping are not addressed in this report.
- Open Access
- 10 November 2021
- Published by Geological Society of America in Geological Society of America Bulletin
- Vol. 134 (7-8) , 1834-1844
Major reorganization of the Snake River modulated by passage of the Yellowstone Hotspot
The details and mechanisms for Neogene river reorganization in the U.S. Pacific Northwest and northern Rocky Mountains have been debated for over a century with key implications for how tectonic and volcanic systems modulate topographic development. To evaluate paleo-drainage networks, we produced an expansive data set and provenance analysis of detrital zircon U-Pb ages from Miocene to Pleistocene fluvial strata along proposed proto-Snake and Columbia River pathways. Statistical comparisons of Miocene-Pliocene detrital zircon spectra do not support previously hypothesized drainage routes of the Snake River. We use detrital zircon unmixing models to test prior Snake River routes against a newly hypothesized route, in which the Snake River circumnavigated the northern Rocky Mountains and entered the Columbia Basin from the northeast prior to incision of Hells Canyon. Our proposed ancestral Snake River route best matches detrital zircon age spectra throughout the region. Furthermore, this northerly Snake River route satisfies and provides context for shifts in the sedimentology and fish faunal assemblages of the western Snake River Plain and Columbia Basin through Miocene–Pliocene time. We posit that eastward migration of the Yellowstone Hotspot and its effect on thermally induced buoyancy and topographic uplift, coupled with volcanic densification of the eastern Snake River Plain lithosphere, are the primary mechanisms for drainage reorganization and that the eastern and western Snake River Plain were isolated from one another until the early Pliocene. Following this basin integration, the substantial increase in drainage area to the western Snake River Plain likely overtopped a bedrock threshold that previously contained Lake Idaho, which led to incision of Hells Canyon and establishment of the modern Snake and Columbia River drainage network.
- 10 November 2021
- Published by Geological Society of America in Geological Society of America Bulletin
- Vol. 134 (7-8) , 1834-1844
Open AccessMajor reorganization of the Snake River modulated by passage of the Yellowstone Hotspot
The details and mechanisms for Neogene river reorganization in the U.S. Pacific Northwest and northern Rocky Mountains have been debated for over a century with key implications for how tectonic and volcanic systems modulate topographic development. To evaluate paleo-drainage networks, we produced an expansive data set and provenance analysis of detrital zircon U-Pb ages from Miocene to Pleistocene fluvial strata along proposed proto-Snake and Columbia River pathways. Statistical comparisons of Miocene-Pliocene detrital zircon spectra do not support previously hypothesized drainage routes of the Snake River. We use detrital zircon unmixing models to test prior Snake River routes against a newly hypothesized route, in which the Snake River circumnavigated the northern Rocky Mountains and entered the Columbia Basin from the northeast prior to incision of Hells Canyon. Our proposed ancestral Snake River route best matches detrital zircon age spectra throughout the region. Furthermore, this northerly Snake River route satisfies and provides context for shifts in the sedimentology and fish faunal assemblages of the western Snake River Plain and Columbia Basin through Miocene–Pliocene time. We posit that eastward migration of the Yellowstone Hotspot and its effect on thermally induced buoyancy and topographic uplift, coupled with volcanic densification of the eastern Snake River Plain lithosphere, are the primary mechanisms for drainage reorganization and that the eastern and western Snake River Plain were isolated from one another until the early Pliocene. Following this basin integration, the substantial increase in drainage area to the western Snake River Plain likely overtopped a bedrock threshold that previously contained Lake Idaho, which led to incision of Hells Canyon and establishment of the modern Snake and Columbia River drainage network.
- 22 July 2017
- Published by American Geophysical Union (AGU) in Geophysical Research Letters
- Vol. 44 (14) , 7276-7284
New methodology for computing tsunami generation by subaerial landslides: Application to the 2015 Tyndall Glacier landslide, Alaska
Landslide‐generated tsunamis pose significant hazards and involve complex, multiphase physics that are challenging to model. We present a new methodology in which our depth‐averaged two‐phase model D‐Claw is used to seamlessly simulate all stages of landslide dynamics as well as tsunami generation, propagation, and inundation. Because the model describes the evolution of solid and fluid volume fractions, it treats both landslides and tsunamis as special cases of a more general class of phenomena. Therefore, the landslide and tsunami can be efficiently simulated as a single‐layer continuum with evolving solid‐grain concentrations, and with wave generation via direct longitudinal momentum transfer—a dominant physical mechanism that has not been previously addressed in this manner. To test our methodology, we used D‐Claw to model a large subaerial landslide and resulting tsunami that occurred on 17 October 2015, in Taan Fjord near the terminus of Tyndall Glacier, Alaska. Modeled shoreline inundation patterns compare well with those observed in satellite imagery.
- 22 July 2017
- Published by American Geophysical Union (AGU) in Geophysical Research Letters
- Vol. 44 (14) , 7276-7284
New methodology for computing tsunami generation by subaerial landslides: Application to the 2015 Tyndall Glacier landslide, Alaska
Landslide‐generated tsunamis pose significant hazards and involve complex, multiphase physics that are challenging to model. We present a new methodology in which our depth‐averaged two‐phase model D‐Claw is used to seamlessly simulate all stages of landslide dynamics as well as tsunami generation, propagation, and inundation. Because the model describes the evolution of solid and fluid volume fractions, it treats both landslides and tsunamis as special cases of a more general class of phenomena. Therefore, the landslide and tsunami can be efficiently simulated as a single‐layer continuum with evolving solid‐grain concentrations, and with wave generation via direct longitudinal momentum transfer—a dominant physical mechanism that has not been previously addressed in this manner. To test our methodology, we used D‐Claw to model a large subaerial landslide and resulting tsunami that occurred on 17 October 2015, in Taan Fjord near the terminus of Tyndall Glacier, Alaska. Modeled shoreline inundation patterns compare well with those observed in satellite imagery.
- 1 January 2011
- Published by US Geological Survey
Channel change and bed-material transport in the Umpqua River basin, Oregon
The Umpqua River drains 12,103 square kilometers of western Oregon; with headwaters in the Cascade Range, the river flows through portions of the Klamath Mountains and Oregon Coast Range before entering the Pacific Ocean. Above the head of tide, the Umpqua River, along with its major tributaries, the North and South Umpqua Rivers, flows on a mixed bedrock and alluvium bed, alternating between bedrock rapids and intermittent, shallow gravel bars composed of gravel to cobble-sized clasts. These bars have been a source of commercial aggregate since the mid-twentieth century. Below the head of tide, the Umpqua River contains large bars composed of mud and sand. Motivated by ongoing permitting and aquatic habitat concerns related to in-stream gravel mining on the fluvial reaches, this study evaluated spatial and temporal trends in channel change and bed-material transport for 350 kilometers of river channel along the Umpqua, North Umpqua, and South Umpqua Rivers. The assessment produced (1) detailed mapping of the active channel, using aerial photographs and repeat surveys, and (2) a quantitative estimation of bed‑material flux that drew upon detailed measurements of particle size and lithology, equations of transport capacity, and a sediment yield analysis. Bed-material transport capacity estimates at 45 sites throughout the South Umpqua and main stem Umpqua Rivers for the period 1951–2008 result in wide-ranging transport capacity estimates, reflecting the difficulty of applying equations of bed-material transport to a supply-limited river. Median transport capacity values calculated from surface‑based equations of bedload transport for each of the study reaches provide indications of maximum possible transport rates and range from 8,000 to 27,000 metric tons per year (tons/yr) for the South Umpqua River and 20,000 to 82,000 metric tons/yr for the main stem Umpqua River upstream of the head of tide; the North Umpqua River probably contributes little bed material. A plausible range of average annual transport rates for the South and main stem Umpqua Rivers, based on bedload transport capacity estimates for bars with reasonable values for reference shear stress, is between 500 and 20,000 metric tons/yr. An empirical bed-material yield analysis predicts 20,000–50,000 metric tons/yr on the South Umpqua River and main stem Umpqua River through the Oregon Coast Range, decreasing to approximately 30,000 metric tons/yr at the head of tide. Surveys of individual mining sites in the South Umpqua River indicate minimum local bed-material flux rates that are typically less than 10,000 metric tons/yr but range up to 30,600 metric tons/yr in high-flow years. On the basis of all of these analyses, actual bedload flux in most years is probably less than 25,000 metric tons/yr in the South Umpqua and main stem Umpqua Rivers, with the North Umpqua River probably contributing negligible amounts. For comparison, the estimated annual volume of commercial gravel extraction from the South Umpqua River between 2001 and 2004 ranged from 610 to 36,570 metric tons, indicating that historical in-stream gravel extraction may have been a substantial fraction of the overall bedload flux.
- 1 January 2011
- Published by US Geological Survey
Channel change and bed-material transport in the Umpqua River basin, Oregon
The Umpqua River drains 12,103 square kilometers of western Oregon; with headwaters in the Cascade Range, the river flows through portions of the Klamath Mountains and Oregon Coast Range before entering the Pacific Ocean. Above the head of tide, the Umpqua River, along with its major tributaries, the North and South Umpqua Rivers, flows on a mixed bedrock and alluvium bed, alternating between bedrock rapids and intermittent, shallow gravel bars composed of gravel to cobble-sized clasts. These bars have been a source of commercial aggregate since the mid-twentieth century. Below the head of tide, the Umpqua River contains large bars composed of mud and sand. Motivated by ongoing permitting and aquatic habitat concerns related to in-stream gravel mining on the fluvial reaches, this study evaluated spatial and temporal trends in channel change and bed-material transport for 350 kilometers of river channel along the Umpqua, North Umpqua, and South Umpqua Rivers. The assessment produced (1) detailed mapping of the active channel, using aerial photographs and repeat surveys, and (2) a quantitative estimation of bed‑material flux that drew upon detailed measurements of particle size and lithology, equations of transport capacity, and a sediment yield analysis. Bed-material transport capacity estimates at 45 sites throughout the South Umpqua and main stem Umpqua Rivers for the period 1951–2008 result in wide-ranging transport capacity estimates, reflecting the difficulty of applying equations of bed-material transport to a supply-limited river. Median transport capacity values calculated from surface‑based equations of bedload transport for each of the study reaches provide indications of maximum possible transport rates and range from 8,000 to 27,000 metric tons per year (tons/yr) for the South Umpqua River and 20,000 to 82,000 metric tons/yr for the main stem Umpqua River upstream of the head of tide; the North Umpqua River probably contributes little bed material. A plausible range of average annual transport rates for the South and main stem Umpqua Rivers, based on bedload transport capacity estimates for bars with reasonable values for reference shear stress, is between 500 and 20,000 metric tons/yr. An empirical bed-material yield analysis predicts 20,000–50,000 metric tons/yr on the South Umpqua River and main stem Umpqua River through the Oregon Coast Range, decreasing to approximately 30,000 metric tons/yr at the head of tide. Surveys of individual mining sites in the South Umpqua River indicate minimum local bed-material flux rates that are typically less than 10,000 metric tons/yr but range up to 30,600 metric tons/yr in high-flow years. On the basis of all of these analyses, actual bedload flux in most years is probably less than 25,000 metric tons/yr in the South Umpqua and main stem Umpqua Rivers, with the North Umpqua River probably contributing negligible amounts. For comparison, the estimated annual volume of commercial gravel extraction from the South Umpqua River between 2001 and 2004 ranged from 610 to 36,570 metric tons, indicating that historical in-stream gravel extraction may have been a substantial fraction of the overall bedload flux.