Declining geomorphic diversity and potential adaptive management opportunities on a highly regulated reach of the Bighorn River, Montana

ABSTRACT

Rivers downstream from dams often experience decreased flow variability and disrupted sediment transport. We investigated a highly regulated 35.5-kilometre reach of the Bighorn River, downstream from Yellowtail Dam, emplaced in 1965. This dam created a thriving trout fishery, but more recently, side channel networks and habitat diminished. We document how the Bighorn River's anabranching morphology responded to flow regulation and a near cessation of sediment supply. Geomorphic diversity drastically decreased since 1939 and the Bighorn River abandoned numerous side channels. By 1980, geomorphic loss slowed, and the river became laterally static. The 1980 river and side channel network could represent the maximum areal extent of a downscaled morphology, maintained through restoration and adaptive management. However, side channel restoration alone will not return natural alluvial processes to this river reach. The Bighorn River needs sediment to reinstate dynamic lateral movement, even within a downscaled morphology. Sediment augmentation is an option, but more complex to implement on rivers where potential downstream impacts and landowners must be considered. The decline of geomorphic diversity and potential adaptive management solutions on the Bighorn River has wide applicability to numerous dammed rivers, where similar changes to the flow and sediment transport regime are common.

KEYWORDS:

• Geomorphic change
• side channels
• dammed rivers
• regulated rivers
• Bighorn River
• adaptive management

1. Introduction

Mid-twentieth century development in the western United States (U.S.) was marked by the building of numerous large dams and reservoirs (Graf 1993, Billington et al. 2005). Reservoirs located above dams trap both water and sediment, transforming the hydrologic and sediment transport regimes of western U.S. river systems. Large dams typically operate to maximize hydropower production, allocate water for cities and farming irrigation, and mitigate flood risk to downstream infrastructure and populations. The release of water from dams is dictated by these operating criteria, often resulting in reduced magnitude and altered timing of peak flow events (e.g. Petts 1980, Mei et al. 2017), increased discharge for base flows (e.g. Merritt and Cooper 2000, Kondolf and Batalla 2005), a reduced range of the daily discharge, and reduced ratio between annual peak flows and mean daily flows (Graf 2006). In summary, the regulated hydrology below dams is typically less variable than the natural hydrology.

Large reservoirs tend to have high trap efficiencies (e.g. Williams and Wolman 1984), especially for coarse sediment trapped in the upper delta (Strand and Pemberton 1982). Sediment accumulates in the upstream reservoir through time while the downstream river becomes sediment starved, unless significant sediment is input from tributaries or sediment contribution from the river’s erosion of bars, islands, and channel banks. As a response to reduced sediment supply, rivers may degrade and scour immediately below dams or coarsen and armour their beds, as sediment-starved water selectively erodes finer material but is not capable of moving coarser grains (e.g. Williams and Wolman 1984, Chien 1985, Dietrich et al. 1989, Kondolf 1997). Changes to the flow and sediment regime affect a river's planform and geomorphic complexity. A river’s morphological response likely depends on how the balance between water discharge, sediment discharge, grain size, and slope is disrupted (Lane 1955, Brandt 2000). Graf’s (2006) study of 72 river reaches below large dams found a 50% reduction in high-flow channel area, a 79% reduction in active floodplain area, and a 37% reduction in geomorphic complexity, defined by the river’s interaction with functional surfaces (i.e. channels, bars, islands, and floodplains). Reduced peak flows can also result in vegetation encroachment, as the flows are unable to scour out vegetation (e.g. Williams and Wolman 1984, Johnson 1994, Petts and Gurnell 2005). In addition, increased summer base flows further promote vegetation growth (Shafroth et al. 2002).

Our study focuses on how changes to the flow and sediment regime, due to the emplacement of upstream dams, affected the Bighorn River’s planform and geomorphic complexity. Three large dams exist in the Bighorn River watershed: (1) Buffalo Bill Dam, emplaced in 1910; (2) Boysen Dam, emplaced in 1952; and (3) Yellowtail Dam and Afterbay (collectively referred to as ‘Yellowtail Dam’), emplaced in 1965 and 1966, respectively. Historically, the Bighorn River below Yellowtail Dam included a diverse geomorphic corridor, defined by the presence of numerous, dynamic side channels, overflow channels, unvegetated and vegetated bars and islands, and laterally dynamic behaviour.

Regulated flow releases from Yellowtail Dam, located at 0 river-kilometres (river-km), created a cold-water trout fishery along the downstream river. The 26 river-km reach immediately below the dam is the highest-use reach for fishing recreation. In the late 1990s through early 2000s, fishermen expressed concern over diminishing and degrading side channel habitat. To address these concerns, the Bureau of Reclamation (Reclamation) initiated a series of side channel investigations, predominately focused on the 35.5 river-km below Yellowtail Dam, including: (1) a literature review of geomorphic and hydraulic relationships in the Bighorn River fishery (Klumpp 1997); (2) a geomorphic side channel analysis (Godaire 2010); and, (3) a hydraulic modelling and sediment transport investigation of side channels (Hilldale 2012). Reclamation performed the initial field studies in April and August 2009 following a period of lower precipitation and reduced flows. We conducted additional geomorphic mapping and further site visits in 2012 and 2019, during a period marked by higher peak flows and higher mean-daily releases, to investigate the effects of the high flow events on geomorphic complexity, side channel connectivity, and the movement of tracer gravels. This paper synthesizes geomorphic mapping and repeat field assessments to analyse side channel connectivity, geomorphic complexity, and identify opportunities to adaptively manage side channel connections to maintain a downscaled geomorphic diversity.

2. Study area

2.1. Regional setting

The headwaters of the Bighorn River originate in mountain ranges in Wyoming. The Wind River drains the Wind River Range and flows into Boysen Reservoir, located upstream from the Wedding of the Waters near Thermopolis, Wyoming, where the Wind River changes name to the Bighorn River. The Shoshone River drains the Absaroka Mountain Range and is impounded by Buffalo Bill Dam, located upstream from its confluence with the Bighorn, which is impounded by Yellowtail Dam (Figure 1). Smaller tributary headwaters enter the Bighorn River from the Pryor Mountains to the west and Bighorn Mountains to the east. Yellowtail Dam is located at the end of Bighorn Canyon, surrounded by steeply dipping and resistant Mississippian to Jurassic age sedimentary rocks. Below the dam, the landscape is less confined, flowing onto more easily eroded Cretaceous shales for the entirety of our 35.5 km study reach (Hamilton and Paulson 1968). Pliocene to Quaternary strath terraces, with inset Holocene floodplains consisting of sandy gravels, surround the Bighorn River along our study reach (Alden 1932, Hamilton and Paulson 1968, Agard 1989).

Figure 1. (A) Drainage basin above Yellowtail Dam, showing elevations and hillshade imagery derived from a 30-m DEM. Rivers and reservoirs are shown in blue. (B–C) Relative elevation maps along the study reach from 0 river-km to 21 river-km (B) and 20 river-km to 35 river-km (C). River distance is measured downstream from the afterbay. The anabranching system of side and overflow channels are visible as relatively low elevations (blue). Note the north arrow and rotation of each subplot.

The Bighorn River likely contained a much higher proportion of sand, like other Great Plains rivers, prior to the emplacement of Yellowtail Dam. Great Plains rivers typically exhibit semi-braided to meandering planforms, wide and sandy beds, and lack confining valley walls (e.g. Matthews 1988, Costigan 2013, Costigan et al. 2014). Frequent deposition of fresh alluvium by dynamic Great Plains rivers supports pioneer species, such as cottonwood and willows (Friedman et al. 1998). However, others note that historically, large Great Plains rivers originating in the Rocky Mountains might have exhibited sparse woody vegetation, similar to historical observations of the Platte and North Platte River (e.g. Williams 1978, Currier 1982).

2.2. Altered hydrology and sediment transport

The hydrology of the Bighorn River was substantially altered by dam construction over the past century. There are no pre-dam hydrology data, but high flow variability exists within the region. Great Plains rivers often exhibit both periods of drought and large magnitude flood events (Osterkamp et al. 1982, Friedman et al. 1998). In 1910, Buffalo Bill disconnected approximately 8% of the of the contributing drainage area above our study reach. The flood of record (1059 m³/s (cms)) occurred in 1935 (Figure 2), prior to the construction of Boysen Dam in 1952. Boysen Dam disconnected an additional 39% of the upstream watershed (Figure 1). The closure of Yellowtail Dam in 1965 significantly affected the Bighorn River, as reservoir operating criteria during wet and dry periods dictate downstream flow patterns. As a result, flows exhibit decreased flood magnitude, less variability in daily discharge between September and May (Figure 2), and fewer low flow periods because minimum flow releases from the dam likely exceed natural flow during periods of drought (e.g. Boyd 2019). The pre-Boysen period (1939–1951) was characterized by a strong spring runoff signature, with average mean daily flows in late June exceeding 400 cms. Following the construction of Boysen Dam, the average peak spring runoff discharge was reduced by approximately 40% (1952–1965), with mean daily flows in June barely exceeding 200 cms. Dam releases during spring runoff continued to decline after Yellowtail Dam was closed in the mid-1960s. The dry years between 2000 and 2008 were characterized by exceptionally low flows, with an average annual peak discharge (115 cms) just 20% of the average annual peaks pre-Boysen (577 cms). The drought severity waned in 2009 and peak flows above 400 cms occurred in water years 2011, 2015, and 2018. Following our study, peak flows for 2021 and 2022 were 91 and 205 cms, respectively. In addition, high, sustained peak flows between 450 and 459 cms persisted from June 22 through July 3, 2023, which could mark the longest period of flows above 400 cms since Yellowtail Dam’s closure.

Figure 2. Peak flows (top plot) and mean daily discharge (bottom plot) on the Bighorn River near St. Xavier, Montana from USGS gage 06287000. Daily discharge data are grouped by periods of time between dam construction. The post-2000 data are grouped by wetter and drier time periods.

Reservoirs in the upstream drainage basin store more than 300 × 10⁶ m³ of sediment, based on the most recent reservoir surveys (Blanton 1991, Ferrari 1996, Hilldale 2020). Without upstream reservoirs, this sediment would be available for downstream transport. Yellowtail Dam has a trap efficiency of approximately 87% (following Brown 1944), meaning that only approximately 13% of the incoming sediment load, likely wash load, is transported downstream of the dam. Surprisingly, the river’s longitudinal profile below Yellowtail Dam has been largely stable since the dam closure (Godaire 2010), although the river does exhibit an armoured bed consisting of large, coarse cobbles. Pebble counts at 22 points on the mainstem channel between 0 and 19 river-km resulted in D₅₀ values between 24 and 101 mm and a median D₅₀ value of 63 mm (Hilldale 2012). Side channels of the Bighorn River still exhibit coarse substrate beds, but the D₅₀ values from 41 pebble counts at side channel entrances and point bars are lower, between 11 and 94 mm with a median D₅₀ value of 45 mm (Hilldale 2012).

2.3. Previous studies

Koch et al. (1977) conducted a channel morphology assessment of the Bighorn River from Yellowtail dam to its confluence with the Yellowstone River using aerial imagery between 1939 and 1974. During this period, the Bighorn River experienced significant area losses for vegetated islands, island gravel bars, and lateral gravel bars. Bank riparian area increased, likely due to lower flows and less frequent flood inundation, which allowed vegetation to invade previously unvegetated gravel bars. They demonstrated that losses in island gravel bars decreased with distance from the dam (Koch et al. 1977).

Following observations of fine sediment impacts at side channel entrances (Godaire 2010), Hilldale (2012) conducted two-dimensional hydraulic modelling to investigate sediment transport along the upper 25.5 river-km of the Bighorn River at its side channels. Hilldale (2012) determined threshold criteria for marginal to vigorous sediment transport, following Andrews (1994) and Pitlick and Van Steeter (1998). Numerical modelling results indicated that frequent intermediate to high flow (170–425 cms) releases from Yellowtail Dam could likely retard or stop further aggradation at side channel entrances. However, high flows alone were unlikely to efficiently scour and reopen disconnected side channels.

Recent Channel Migration Zone (CMZ) mapping demonstrated that mean channel migration rates on the Bighorn River below Yellowtail Dam decreased between 45% and 70% since 1979 (Boyd and Thatcher 2020). As a result, the turnover rate for riparian vegetation was reduced, coupled with an overall expansion of riparian vegetation on the Bighorn River below Yellowtail Dam. Most cottonwood forests lack diverse age classes and are skewed towards older trees. The lack of channel movement is accompanied by the infestation of Russian olive trees that form thick stands on streambanks (Boyd and Thatcher 2020).

2.4. Field observations

During the April 2009 field study, we observed thick accumulations of fine sediment at the inlets of many side channels. Previously active side channels were disconnected from the mainstem, meaning that they now only inundate during large flood events and do not maintain a seasonal flow connection to the mainstem river. Several channels were also at risk for disconnection due to aggraded and constricted inlets. The Bighorn River entered a wetter decade, starting in 2009. Higher flow releases in the following years allowed the river to erode several bars and islands and scoured fine sediment at many, but not all, side channel entrances.

In 2019, we noted narrower and possibly deeper channel entrances that lacked the thick, fine sediment accumulations we observed in 2009. Many channels scoured vegetation from the side channel beds (e.g. Figure 3), but vegetation along the banks was not removed by flows overtopping the surface. Channels inlets were typically narrower than in 2009 and surrounded by well-established vegetation. We qualitatively observed an overall lack of spawning-size gravels between 6- and 50-mm in the mainstem channel, but more frequently observed this target grain size in side channels.

Figure 3. Side channel 4-1, located on river-right between 3 and 4 river-km, in August 2009 (top) and in September 2019 (bottom) with labelled trees for comparison (A, B, C). Both photos are looking downstream from the inlet.

2.5. Side channel restoration efforts

In 2012, Cline’s Channel, near 15 river-km, was mechanically excavated to remove sediment that was deposited at the inlet and reconnect flow from the mainstem Bighorn River (Figure 4). In 2019, the channel entrance remained open and flowing, with some narrowing and deepening at the entrance. By 2021, the inlet excavation remained stable, but the aggradation of a large gravel bar 140 m downstream resulted in shallow flows and reduced discharge entering the Cline’s Channel. Therefore, the Bighorn River Alliance (BRA) conducted further excavation on Cline’s Channel and utilized the excavated material to create bars downstream (bighornriveralliance.org). Between 2019 and 2023, BRA completed the excavation of 7 additional side channel entrances within our study reach.

Figure 4. Side channel 12-1 (Cline's channel) in October 2011 before inlet excavation (top) and in February 2012 following excavation (bottom). Hillslope peak ‘A’ noted in both photos. Photos look downstream from the inlet.

3. Methods

3.1. Geomorphic analyses

We mapped six geomorphic features on aerial and satellite imagery datasets (Table 1) using ESRI ArcGIS along the 35.5 km study reach, downstream of the Yellowtail Dam. Mapped geomorphic features include:

1. Mainstem channel (Qac) is mapped at the wetted perimeter or the defined channel bank. We defined the main channel as the widest channel in areas of multiple channels. Occasionally, the mainstem channel split around small mid-channel bars.

2. Side channels (Qsc) are channels with defined bed and banks that exhibit inlet and outlet connections to the mainstem channel at discharges between 55 and 85 cms, based on field observations and hydraulic modelling results by Hilldale (2012).

3. Overflow channels (Qoc) are channels with defined bed and banks that only receive flow during larger flood events. Overflow channels may originate from another channel or originate at the transition between unvegetated alluvium and a defined channel.

4. Vegetated islands (Qb1) are bars vegetated with dense shrubs or mature trees and surrounded by defined channels (mainstem channel, side channels, or overflow channels).

5. Unvegetated bars (Qb2) are areas without vegetation or only small shrubs and grasses, indicating frequent inundation. This category includes lateral bars, point bars, and mid-channel bars.

6. Holocene alluvium (Qa) includes valley alluvium that lies outside of the margins of the active channel and dynamic geomorphic features listed above, ranging in age from historically abandoned floodplain to early Holocene. The boundary of the Holocene floodplain follows mapping from Agard (1989). Holocene alluvium is also typically vegetated.

   Table 1. Aerial and satellite imagery details.

   Year	Resolution	Daily mean Q (cms)^a
   1939	1:20,000	47–56
   1954	1:20,000	51–77
   1961	1:20,000	37–67
   1970	1:40,000	65
   1980	1:40,000	113
   1991	1:40,000	100–113
   2006	2.0 m	56–57
   2011	1.0 m	396
   2017	0.6 m	89–265

   ^a Daily discharges USGS gage 06287000.

Flow releases from Yellowtail Dam preclude valley filling or inundation beyond the extent of the Holocene alluvium, barring a major disruption to the mainstem channel’s flow path, such as an avulsion. Therefore, the area available for geomorphic change predominantly lies within the boundary of the Holocene alluvium. We use the channels (Qac, Qsc, Qoc) and unvegetated bars and islands (Qb2) to represent the system’s geomorphic diversity or ‘active geomorphic corridor,’ as these features are frequently inundated and dynamic. Conceptually, the active geomorphic corridor is similar to ‘process space’ defined by Ciotti et al. (2021) and ‘freedom space’ defined by Biron et al. (2014), which are also used to quantify the portion of the valley bottom occupied by connected floodplains and active channels.

If the system is gaining geomorphic diversity through time, the river is removing vegetation through scouring flows or reconnecting areas of Holocene alluvium. If the system is losing geomorphic diversity through time, active features such as side channels or unvegetated alluvium are becoming increasingly vegetated or fully abandoned and converted to Holocene alluvium that no longer interacts with the active channel. When a side channel no longer exhibits an inlet and outlet connection to the Bighorn River at mainstem discharges between 55 and 85 cms, its classification changes to an overflow channel. Similarly, if an overflow channel exhibits vegetation encroachment and no longer has a defined flow path, we reclassify the channel as vegetated or Holocene alluvium and no longer part of the active geomorphic corridor. We use the term ‘disconnection’ to describe the evolution of side channels to overflow channels and the complete abandonment of overflow channels.

We note that vegetated floodplains, particularly on mid-channels bars, are often part of a healthy and diverse geomorphic system, as they promote split-channel flows and diverse fauna and food webs. We focus on active channels and bare alluvium because vegetated floodplains are more difficult to identify in historical imagery. In addition, the regulated flows and the lack of sediment sources are more likely to produce static morphology and aging vegetation on river bars and floodplains, rather than a dynamic and active system of floodplain recruitment and vegetation progression. Side channels and overflow channels can exist in isolation, but most often occur within a network of anabranching channels that share channel junctions with each other and with the mainstem channel. We grouped the side channels into distinct complexes that include any channels that interact or are located within a short distance of one another. We labelled individual channels within each complex to allow comparison of the channels between 1980 and 2017 (e.g. Figure 5).

Figure 5. Example of geomorphic mapping and channel width measurements near side channel complex 9.

We measured channel widths at 160 locations, approximately every 0.25 km, on the mainstem Bighorn River for both the 1980 and 2017 datasets (Figure 5). If side channel inlets or outlets coincided with the 0.25 km interval for channel width measurement, we moved the measurement to the closest location upstream or downstream to avoid the junction and maintain the same measurement location in both datasets. We did not label individual channels or measure channel widths prior to 1980 because of the large differences in the Bighorn River’s channel centreline location.

3.2. Gravel tracers

In February 2012, we installed 398 tracer particles with passive integrated transponder (PIT) tags in side channels 8-2, 9-1, 10-1, 10-2, and 12-1 (Cline’s channel, following inlet excavation) to determine: (1) the travel distances of coarse gravel and cobbles in response to high flow events; and (2) the transport threshold flow for a range of grain sizes to compare with model results, where applicable (Hilldale 2012, see locations on Figure 6). We selected three size classes ranging from very coarse gravel to small cobbles based on pebble count data (Hilldale 2012): (1) 32–45 mm, (2) 45–64 mm, (3) 56–84 mm, and (4) 64–90 mm (Table 2). We obtained the 56–84 mm and 64–90 mm size groups at different times, leading to overlapping size groups. PIT tags equip each tracer particle with a unique identifier that can be read with an antenna and detector at a distance up to approximately 1 metre. PIT-tagged tracers are an increasingly common method for tracking fluvial sediment (e.g. Hassan and Bradley 2017, Liébault et al 2024).

Figure 6. The Bighorn River (black) and selected side channels (white), which correspond to the location of topographic surveys at channel inlets. Hillshade imagery is shown in the background.

Table 2. The number of tracer particles in each size class seeded in side channels.

Channel^a	Tracer Size Class
	32–45 mm	45–64 mm	56–84 mm	64–90 mm
8-2	34	34	15	17
9-1	34	34	14	16
10-1	34	34	15	16
12-1	35	34	15	17

^a Channel complex-channel number.

We installed the tracer particles in the side channel entrances on 29 February and 1 March 2012, prior to spring high flow releases. We pressed the tracers into loosely packed native gravel to reduce the likelihood that tracer particles are artificially mobile relative to the native material (e.g. Hassan and Ergenzinger 2003), perpendicular to the flow direction across the deepest and swiftest portion of the channel. We surveyed the end points of the seeding-line location (Hilldale 2012) and surveyed individual gravel tracer locations during subsequent recovery trips on 11–12 April 2012, and 6–11 September 2019.

3.3. Topographic surveys

We collected topographic survey points using Real Time Kinematic (RTK) Global Positioning System (GPS) equipment during wading surveys in 2009 and 2019, as well as limited survey points during the gravel tracer recovery in 2012. From the topographic data, we created longitudinal profiles of 19 side and overflow channels located within the uppermost 16 river-km below Yellowtail Dam, as this area is the most heavily utilized by trout (Figure 6). We concentrated survey points along the thalweg and snapped survey points to the 2009 channel centreline to compare profiles across years.

4. Results

4.1. Geomorphic analyses

The Bighorn River exhibits a dominant mainstem channel that alternates between single-thread reaches and multi-thread reaches. Side channel complexes are typically separated from the mainstem channel by vegetated, alluvial islands. The 1939 and historical river morphology might have been a sand-dominated, island-forming river, but the current morphology is consistent with a gravel-dominated, laterally active anabranching river (Nanson and Knighton 1996) or wandering gravel bed river (i.e. Church 1983, Burge 2005). Rivers within this category typically have high stream power values, between 30 and 100 W/m² (Nanson and Knighton 1996), and tend to exhibit dynamic lateral movement.

Geomorphic diversity in the Bighorn River, measured by the areal extent of the active geomorphic corridor, was greatest in 1939 and decreased in each subsequent photoset (Figure 7), declining by 47% between 1937 and 2017 (Table 3). The initial reduction in geomorphic diversity between 1939 and 1954 might be due to channel recovery following the 1935 peak flow event of 1059 cms, which would have inundated a large spatial area, potentially scouring new side and overflow channels. Between 1939 and 2017, the river evolved from an extensive anastomosing morphology towards a single-thread morphology with limited side channel complexes (Figure 8). Despite the Bighorn River’s changing planform, the area fluctuations on the mainstem Bighorn River are complicated to untangle, as the changing morphology of bars, islands, and channel junctions impacts mainstem channel area. Because the Bighorn River lost several mid-channel island bars, we expected a gain in the mainstem channel area. However, a 4.5% loss in the mainstem channel area occurred between 1937 and 2017, which is likely within the mapping uncertainty. In the older photosets, the Bighorn River was surrounded by extensive unvegetated alluvium, and it was more difficult to differentiate the channel bed and banks from surrounding alluvium in many of the photosets.

Figure 7. Stacked area plot of geomorphic units. Total area of the active geomorphic corridor was greatest in 1939 and reduced greatly between 1939 and 1980. The entire area of the Holocene alluvium is not shown and would extend up to approximately 23.5 km² on the y-axis of the stacked area plot for all of the mapping years.

Figure 8. Example of geomorphic mapping near river-km 8 to 15 for 1939, 1980, 2006, and 2017. The side channel complex numbers associated with each side channel network are shown in pink. Restoration efforts to manually excavate the entrance to Cline’s channel occurred in 2012 (complex 12-channel 1, flowing north at river-km 15), converting it from an overflow channel to a side channel.

Table 3. Geomorphic unit area, mapped from imagery, through time.

Year	Qac (km^2)	Qsc (km^2)	Qoc (km^2)	Qcall ^a (km^2)	Qb1 (km^2)	Qb2 (km^2)	Qactive ^b (km^2)	Qa (km^2)
1939	3.11	1.03	0.81	4.95	3.67	2.51	7.46	12.64
1954	2.97	0.86	0.80	4.63	3.96	2.20	6.83	12.93
1961	2.86	0.57	0.69	4.12	3.95	1.96	6.08	13.62
1970	2.81	0.63	0.46	3.90	3.47	1.36	5.26	14.85
1980	3.09	0.74	0.15	3.98	2.37	0.71	4.69	16.63
1991	3.00	0.74	0.13	3.87	2.55	0.57	4.44	16.65
2006	3.06	0.65	0.22	3.93	2.78	0.28	4.21	16.69
2011	3.06	0.63	0.25	3.94	3.35	0.12	4.06	16.27
2017	2.97	0.65	0.19	3.81	3.17	0.12	3.93	16.57
%Δ39-17^c	−5	−37	−77	−23	−14	−95	−47	31
%Δ39-80^d	−1	−28	−81	−20	−35	−72	−37	32
%Δ80-17^e	−4	−12	+27	−4	+34	−83	−16	0

^a Qc_all is the sum of the active channel, side channels, and overflow channels.

^b Qactive represents the active geomorphic corridor and is the sum of Qc_all and Qb2.

^c %Δ_39-17 is the percent change in area between 1939 and 2017.

^d %Δ_39-80 is the percent change in area between 1939 and 1980.

^e %Δ_80-17 is the percent change in area between 1980 and 2017.

Significant channel loss occurred between 1939 and 1980 due to the disconnection of side (Qsc) and overflow (Qoc) channels (Figure 8), with the most drastic loss occurring in overflow channel area (−82%, Table 3). However, an increase in side channel area occurred between 1961 and 1980, driven by the development of defined channel beds and banks within areas previously mapped as unvegetated alluvium. For example, we mapped large swaths of unvegetated alluvium near side channel complex 10 (9 river-km) in 1939. By 1980, vegetation encroached onto the unvegetated alluvium and flow was concentrated into a defined side channel (Figure 8). In this example, even the development of a side channel was associated with an overall loss within the active geomorphic corridor.

The loss of geomorphic diversity significantly slowed between 1980 and 2017 (Table 3, Figure 7). Side channels and unvegetated bars are the only two individual units that continued to decrease in area when comparing 1980 and 2017 mapping sets, with an area loss of 12% and 83% for side channels and unvegetated bars, respectively. Eight side channels transitioned to overflow channels between 1980 and 2017 imagery. Three new, small side channels also formed during this period, associated with the movement of gravel bars (Figure 9). The overall trend was marked by a decrease in side channel area and increase in overflow channel area between 1980 and 2006, stabilizing between 2011 and 2017 during a period of higher flow (Table 3). Satellite imagery taken during higher flows in 2011 (317 cms, Table 1) revealed the reactivated and new overflow channels. These new or reconnected channels typically exhibited very narrow widths and very small channel areas relative to other overflow channels. For the majority of these overflow channels, we observed defined channels in 2017 imagery and during the 2019 field visit.

Figure 9. Area and channel type changes for side channels (SC) and overflow channels (OC) between 1980 and 2017; channel numbering is consistent between years. Example explanations: (1) Channel 15-1 was a small- to moderately sized side channel in 1980 and in 2017 the side channel was smaller, and two overflow channels developed within its original footprint. (2) A new overflow channel, 24-2, developed after 1980 (absent from top plot) and before 2017 (present on bottom plot).

Between 2006 and 2017, a period marked by several high peak flows, one new side channel formed. The decrease in overflow channel area during this period was in part due to restoration efforts to reconnect Cline’s Channel in 2012 (0.07 km²; channel 12-1). However, the disconnection of side channels still dominated the overall trend. Comparing the geomorphic mapping from 2006 and 2017, in combination with field observations, we identified channels at risk for disconnection (Table 4). These channels transitioned from side channels to overflow channels and appear at risk for complete abandonment or have narrowing inlets and appear to receive less flow in 2017 than in 2006.

Table 4. Channels at-risk for disconnection.

Channel	2017 unit
4-1	Qsc
6-4	Qoc
9-1	Qsc
10-1	Qsc
11-3	Qoc
11-4	Qsc
11-5	Qoc
11-8	Qoc
21-1	Qoc
23-2	Qoc
25-1	Qoc

The area of inactive Holocene alluvium (Qa) exhibits the opposite trend of the active geomorphic corridor, greatly increasing in area between 1939 and 1980. After 1980, the area of Holocene alluvium largely stabilized (Table 3, Figure 7). The area of vegetated bars and islands (Qb1) initially dropped as side and overflow channels were disconnected from the Bighorn River, and these features were no longer surrounded by water. However, after 1980 this trend reversed, as vegetation began to encroach on the remaining unvegetated alluvium (Qb2) (Supplemental Data), resulting in a decrease in unvegetated bars and an increase in vegetated bar area (Table 3). The reactivation of overflow channels in 2011 surrounded portions of alluvium with channels, changing the classification of many areas from Holocene alluvium in 2006 to vegetated alluvium and bars in 2011. This is evident from the increase in vegetated alluvium between 2006 and 2011 (Table 3, Figure 7).

4.1.1. Lateral changes and channel widths

The Bighorn River’s lateral mobility greatly decreased between 1939 and 2017. Prior to 1970, channel movement was most often due to a change in the channel configuration between the mainstem, side channels, and overflow channels. Between 1970 and 1980, we mapped only minor changes in the morphology of side channel complexes and only minor shifts in mainstem channel banks at few locations. Between 1980 and 2017, the location of the mainstem channel and prominent side channel complexes remained static.

To assess the locations and magnitude of lateral migration, we identified reaches where the 1980 Bighorn River centreline did not fall within the mapped extent of the Bighorn River (Qac) or active geomorphic corridor (Qactive) for each photoset. For 1970 and after, the 1980 centreline falls completely within the mapped extent of the mainstem river (Table 5). In addition, we qualitatively observed that the 1980 centreline was a good approximation for the river centreline for photo years from 1970 through 2017. Prior to 1970, we noted several locations where the 1980 centreline fell completely outside of the mapped extent of the active geomorphic corridor. More than half of the total stream length identified as experiencing lateral migration occurred between 26.0 and 35.5 river-km downstream (Supplemental Data). After 1970, the mainstem river no longer exhibited this type of significant lateral movement. Only very minor changes in channel morphology occurred after 1980, including (1) the narrowing or disconnection of side channels from the mainstem, without migration of the channel locations; (2) splitting a side or overflow channel into multiple, smaller channels within its 1980 footprint (Figure 9); (3) the reopening or identification of small overflow channels following the 2011 peak flow; and, (4) minor changes to the bank lines (Figure 8, Supplemental Data).

Table 5. Lateral migration based on a comparison of 1980 centreline and mapped perimeter of the mainstem Bighorn River.

Photo year	Length 1980 centreline, outside Qac, (km)^a	Length 1980 centreline, outside Qactive, (km)^a
1939	8.84	3.75
1954	4.73	0.81
1961	3.2	1.18
1970–2017	0	0

^a The 1980 Bighorn River centreline was measured in river kilometres along the centreline. We only quantified the length reaches where at least 0.05 km fell outside of the mapped mainstem river or the active geomorphic corridor (including all channels and unvegetated alluvium). We did not count mid-channel islands as falling outside of the mainstem channel. Specific locations are shown in Supplemental Data.

Our comparison of repeat channel width measurements in 1980 and in 2017 produced scattered results (Figure 10). The minimum, maximum, and mean channel width measurements were similar between 1980 and 2017 (40.0, 143.3 m, and 83.4 in 1980 and 40.0, 149.7, 82.9 m in 2017). We observed three visual groupings in the data: (1) between 0 and 5.5 km downstream there is a slight increase in the mean channel width (4.0%); (2) between 5.5 and 26.0 km downstream, there is a slight decrease in the mean channel width (−3.7%); (3) and between 26.0 and 35.5 km downstream there is a small increase in mean channel width (5.7%). The moving mean of percent channel change, averaged across 4 measurements, is negative for most of the measurements, consistent with the small loss in the mainstem channel area between 1980 and 2017. However, most individual measurements fell within a 5% bound, which is likely a reasonable representation of measurement error. Between 17 and 23 river-km, the moving mean falls outside of this 5% range, but the mean is influenced by a large amount of channel change at 5 locations near narrowing or disconnecting side channel junctions. The increase in channel widths between 26.0 and 35.5 river-km downstream is more consistent, with 17 measurements increasing over 5% from 1980 to 2017 (Figure 10). The channel width increases in this reach coincide with areas of significant past lateral migration, indicating that the river might be able to widen more easily in zones of past migration.

Figure 10. Top plot: Percent change in channel width from 1980 to 2017; positive numbers indicate an increase in channel width. Bottom plot: Channel width measurements. In both plots, the lines represent a moving mean across 4 measurements.

4.1.2. Uncertainty in mapping and channel width measurements

Measurement error undoubtedly exists in both the areal measurements and the channel widths. Unfortunately, we do not know the individual errors at control points Reclamation used to georectify historical imagery from 1939 through 1991 for the Godaire (2010) study. As such, we are unable to conduct a thorough and quantitative error analysis for geomorphic area and channel width change (e.g. Mount et al. 2003). Error estimates are important to indicate whether channel change has truly occurred, or if the channel change is due to errors in georectification. Using a GIS-based approach to assess river planform change and associated error, Downward et al. (1994) concluded that lateral channel movement in excess of 5 m was required to ensure genuine channel change. Lateral movement of the Bighorn River prior to 1970 exceeds 5 m and the historic centreline of the river often falls completely outside of the modern active geomorphic corridor. Lateral changes to the river location between 1970 and 2017 are very minor and cannot be verified as true channel change. Apart from the mainstem Bighorn River, the change in geomorphic unit area from 1939 to 2017 is between a magnitude of 14 and 77 percent (Table 3). We believe these values represent genuine channel change due to their large magnitude. The smaller changes noted in the mainstem Bighorn River (5%) and in channel width measurements (typically within 5%) could be due to inherent error in geomorphic mapping, image quality, and the georectification process.

4.2. Gravel tracers

In 2012, all gravel tracers were recovered (Table 6), allowing us to observe the total movement of tracers between February and April 2012. The highest flow release during this period was 192.6 cms. Channel 8-2 exhibited the greatest mobilization of particles, where 27 particles moved downstream from the seeding line location, although only 15 particles moved more than 10 m downstream. Tracer movement was consistent with hydraulic model predictions for marginal sediment transport based on critical shear stress values between 0.031 and 0.047 for a 192.6 cms flow (Hilldale 2012). Particles in channels 9-1 and 10-1 were much less mobile in 2012 (Figure 11). In channel 9-1, 16 particles moved from their seeding location, but only 1 particle moved more than 5 m. In channel 10-1, 5 particles moved downstream and 4 of these moved more than 3 m (Figure 11). Sediment transport in channels 9-1 and 10-1 was consistent with weakly marginal transport (Hilldale 2012). We did not seed particles into channel 12-1, Cline’s Channel, until after the channel inlet excavation.

Figure 11. (Left panels) Tracer positions in April 2012 following a 200 cm flow release from Yellowtail Dam. Tracers were originally seeded in February 2012. The backdrop is 2011 imagery. (Right Panels) Tracer positions shown in 2019 on 2017 imagery. Flow is from left to right in all images. Histograms of cumulative tracer displacement in 2019 are shown below and the stacked bars indicate gravel size.

Table 6. Gravel tracer recovery counts.

Channel	Tracer Count
	Installed Feb. 2012	Recovered April 2012	Recovered Sept. 2019
8-2	101	101	83
9-1	98	98	26
10-1	99	99	48
12-1	101	101	76

We were not able to recover all the tracers in 2019 (Table 6). It is possible that tracers were buried or lost. However, the mainstem river is largely armoured and there are no coarse sediment sources upstream of the channels with tracers, which limits the amount of mobile sediment available to bury tracers. We find it more likely that tracers completely washed out of the channels in the 7 years between surveys.

In all channels, the fraction of installed tracers that we recovered is similar for all sediment size groups (Supplemental Data), indicating that size groups are equally mobile. In channels 8-2, 9-1, and 10-1, recovered tracers were predominantly near the channel inlet (Figure 11) with very few tracers distributed along the downstream channel. In channel 8-2, more than 80% of the tracers were within 40 metres of the channel inlet. In channels 9-1 and 10-1, we recovered only 26.5% and 48.5% of the installed tracers on the channel bed, all but a few located within 100 metres of the inlet. In contrast, we found most of the tracers in channel 12-1 (Cline’s Channel) deposited on a gravel bar about 150 m downstream of the installation position (Figure 11). Cline’s Channel was the only side channel where we found tracers on a point bar, as opposed to the channel bed. The other channels with mobile tracers lack geomorphic elements (such as gravel bars) where the tracers might be trapped and buried.

4.3. Topographic surveys

Of the side channel entrances we surveyed, 3 showed minor aggradation, 10 showed minor incision, and 6 were inconclusive (Table 7, Figure 12, Supplemental Data). Bed elevation changes along the longitudinal profiles were typically within 10 to 20 cm of the previous survey. As we did not observe large deposits of fine or organic sediments at the channel entrances, we conclude that the flows between the 2012 and 2019 surveys have prevented deposition or eroded fine sediment at most of the channel entrances. Despite minor incision or inconclusive channel change, channels 11-3, 11-5, and 11-8 all transitioned from side channels to overflow channels. Individual channel width measurements on the Bighorn River showed small amounts of narrowing, that could be associated with vertical bed incision or bed reorganization near these side channel inlets, requiring higher discharge to connect these channels to the mainstem. However, Godaire (2010) found no evidence of vertical incision on the mainstem river, with the exception of potential scour within 0.6 km downstream of the afterbay, and new data have not been collected to evaluate if recent higher flows resulted in mainstem incision. Regardless, the amount of longitudinal profile change at side channel entrances is small. Other factors, such as vegetation encroachment, might influence the threshold discharge required to create flow connections between the mainstem Bighorn River and side channels.

Figure 12. Example long profile results from channel 1-1 (aggradation), channel 11-5 (inconclusive), and channel 9-1 (incision). The changes in the profiles are minor, often 10 cm or less. 2012 data are limited and only collected associated with gravel tracer recovery in April 2012.

Table 7. Observed change in longitudinal profiles.

Minor aggradation	Minor incision^a	Inconclusive
Channel # 1-1, 3-1, 4-1	Channel # 5-1, 6-1, 8-1, 9-1, 9-3, 10-2, 11-3, 11-4, 12-1, 12-2	Channel # 8-2, 10-1, 11-5, 11-8, 13-1, 15-1

^a Channel 12-1, Cline’s channel, was excavated in 2012. Profiles post-date excavation.

5. Discussion and recommendations

Higher sediment load and a more variable flow regime prior to the emplacement of dams enabled dynamic lateral movement and channel avulsions along the mainstem Bighorn River and created the anabranching planform. The emplacement of Yellowtail Dam greatly decreased flow variability and halted input of coarse sediment into the 35.5 km study reach located directly below the dam, resulting in an increasingly static channel morphology through time. Thus, the anabranching morphology is a relic of past lateral movement and not indicative of current processes. Coincident with increasing stability, geomorphic diversity along our study reach decreased in each subsequent mapping year. The evolutionary trajectory is that side channels are continuing to disconnect through time, but at a much slower rate since 1980. Our observations of decreasing area in the active geomorphic corridor are consistent with the losses of floodplains and channel connectivity observed in other impacted systems (e.g. Arnaud et al. 2015; Eder et al. 2022), although we did not observe significant channel narrowing or vertical bed degradation common in other rivers below dams (e.g. Kondolf 1997, Ma et al. 2012).

Yellowtail Dam releases clear, cold water into the downstream Bighorn River and its active side channels. This enabled the creation of a blue-ribbon trout fishery, which has implications for the local economy and recreation. However, the loss of side channels and increased armouring along the channel bed through time has also negatively impacted fish habitat and recreation. Several studies highlight the importance of societal relevance in river restoration projects (e.g. Dofour and Piégay 2009, Deffner and Haase 2018). These studies indicate that to be successful, potential river management projects should prioritize local societal needs while also aiming to promote diversity in riverine habitat, riparian vegetation, and geomorphology. On the Bighorn River, such projects must also work within the operational constraints at Yellowtail Dam.

Some evolutionary changes on rivers are reversible, while others are not (e.g. Brierly and Fryirs 2022). Given the altered flow and sediment regime, the Bighorn River will not recover the spatial extent or geomorphology observed in 1939. An achievable goal could be scaling down the geomorphic corridor so that the regulated flow regime can meet key geomorphic and ecological goals (e.g. Brown 2022). The active geomorphic corridor was passively downscaled through flow regulation, rather than actively downscaled by engineering a suite of geomorphic features accessible by regulated flows. As a result of passive downscaling, the mainstem channel and many side channels exhibit an armoured bed, many floodplains inundate less frequently, and aging riparian vegetation is not being replaced by young vegetation recruited on freshly deposited sediment.

Past implementation projects to reconnect side channels on heavily regulated systems provide reference examples, including large-scale implementation projects on the Rhine and Danube Rivers (refer to case studies reviewed in Schneider 2010). On the Bighorn River, the 1980 network of side channels might represent the maximum spatial extent of a downscaled, anabranching river morphology that could be reclaimed or maintained with an adaptive management approach. Side channels (Qsc) that regularly receive flow but have experienced inlet aggradation between 1980 and 2017, or are at risk for disconnection (Table 4), would likely require minimal mechanical excavation at channel inlets. Channel reopening efforts could then target overflow channels at-risk for disconnection with high potential for habitat creation, then target fully disconnected channels. Bighorn River Alliance (BRA), a local community group, has already excavated several side channel entrances, including two channels we identified as at-risk (10-1 and 21-1). These projects have increased access to spawning habitat in side channels (Dennis Fischer, BRA, persoral. comm., October 17, 2024), where a higher proportion of spawning-size gravels is available. In addition, following the 2012 restoration at Cline’s Channel, our study indicates that bar-to-bar sediment transport was re-established and sediment was recruited from the mainstem channel.

Maintaining side channel connections will require adaptive management. Our study, coupled with hydraulic modelling by Hilldale (2012), indicates that side channel inlets will likely remain open and experience little fine sedimentation and vegetation encroachment during wetter periods with larger flow releases. Repeat topographic profiles indicate that minor incision or no change to the longitudinal profile is common during wet periods with peak flows above 400 cms. During drier periods, marked by several years of lower flow releases with moderate peak flows less than 200 cms, it is likely that vegetation will encroach and sediment will aggrade at channel inlets and might require additional mechanical excavation. Riquier et al. (2017) provide a framework for assessing fine sedimentation rates and the persistence of restored or excavated side channels, while Eschbach et al (2021) detail geodetic and geomorphic monitoring methods on restored side channels. Such monitoring and forecasting guidelines could be used to inform an adaptive management plan.

Trush et al. (2000) outlined important geomorphic, ecological, and hydrological considerations for alluvial river restoration, which they employed to actively downscale the Trinity River. An underlying core principle is that alluvial rivers have the capacity to exhibit dynamic behaviour, which our study reach lacks: (1) the Bighorn River’s planform is static, (2) the sediment budget is not in balance, and (3) floodplains are homogeneous and do not support the recruitment of new seedlings. A restoration plan focused solely on maintaining or restoring past channel morphology ignores dynamic lateral movement. The ability to erode and deposit sediment are critical aspects of natural rivers (e.g. Brierly and Fryirs 2022). Creating static channel features in what should be a dynamic environment could work against, rather than with, natural river processes (Beechie et al. 2010).

Sediment and flow variability are key to restoring and maintaining a diverse geomorphic and ecological river corridor, even as a downscaled morphology. Infrequent, large floods can reshape the river and sustain channel complexity (Trush et al. 2000). In contrast, the Bighorn River has an upper threshold for flow release at approximately 425 cms. Therefore, during wet years, the maximum release is sustained for weeks at a time, resulting in a plateau-shaped hydrograph. Under this flow regime, channel complexity in areas above this inundation threshold cannot be restored. However, in drier years when a large volume of water does not need to be passed through the dam, flow releases could mimic a more natural peak flow. Hydraulic modelling by Hilldale (2012) suggested annual intermediate flows (170–285 cms) are required to hinder vegetation encroachment and higher flow releases (285–425 cms) every 2 to 5 years would increase habitat diversity and maintain excavated side channel inlets.

As no incoming sediment enters the Bighorn River immediately below Yellowtail Dam, sand and gravel (sediment) augmentation could be considered, similar to recommendations on other river systems (e.g. Tockner et al. 1998, Arnaud et al. 2015). Sediment augmentation at multiple locations may be even more effective (Gaeuman et al. 2017). Potential outcomes associated with sediment augmentation include: (1) restoring bedload transport; (2) encouraging channel dynamics, promoting lateral migration and flow complexity; (3) promoting a reorganization of riverbed structure, creating new pools and riffles; or (4) increasing interstitial habitat to support spawning (Mörtl and De Cesare 2021). Sediment augmentation on reaches of the Trinity River in California and the Nunome River in Japan, both downstream from dams, resulted in increased geomorphic activity and channel change, new bar formation, increased bed mobility, and the removal of algae from the bed (Ock et al. 2013). The recruitment of new gravel into Cline’s Channel, following restoration to reconnect the inlet, indicates that augmented sediment is likely to interact synergistically with other restoration activities. Fresh sediment on bars should also enable pioneer species recruitment.

Our gravel tracer data demonstrated sediment mobility for particles between 32 and 90 mm, but sediment particle size for augmentation would likely be finer than the armoured bed cobbles (e.g. Sklar et al. 2009). The timing of sediment augmentation could be tricky, as sustained high-flow releases could wash out gravel. If possible, sediment augmentation prior to a more natural hydrograph release might be the most effective. Additional studies with updated bathymetric data, hydraulic modelling, and tracking additional PIT-tagged gravels could better constrain how supplemental gravel is transported through the mainstem river and side channels, enabling an adaptive approach for future sediment augmentation based upon sediment transport goals.

There are likely obstacles to an effective sediment augmentation plan on the Bighorn River. Sediment augmentation can be expensive and is not a one-time solution. In addition, dynamic rivers move, potentially affecting landowners or infrastructure, and reorganizing the channel morphology (e.g. Gaeuman 2012). Some well-known fishing spots could disappear. But, when channel loss is balanced by dynamic movement, the creation of new channels, and deposition of fresh sediment, the system would be likely to experiencean overall gain in geomorphic diversity, which, in turn, would likely support more complex food webs and habitat.

6. Conclusions

Geomorphic diversity on the 35 river-km reach of the Bighorn River, immediately below Yellowtail Dam, significantly decreased between 1939 and 2017. The initial reduction in geomorphic diversity might be associated with the channel’s recovery following the 1935 peak flow event of 1059 cms, but the continued reduction of geomorphic diversity is influenced by the emplacement of upstream dams. By 1980, the altered hydrologic and sediment regime resulted in a cessation of lateral movement immediately downstream from Yellowtail Dam. Side channels continue to disconnect from the mainstem Bighorn River through sediment deposition at side channel inlets and vegetation encroachment, but the rate of side channel disconnection and loss of geomorphic diversity have greatly decreased since 1980. The anabranching morphology present in 1980 might represent a downscaled morphology that could be restored or maintained with adaptive management techniques. Thus far, restoration techniques along this reach of the Bighorn River have focused on mechanical excavation of side channel inlets to reconnect flow and promote scour. These efforts alone will not reinstate dynamic channel movement nor address the loss of unvegetated sediment bars, which are needed to recruit new tree seedlings. The combination of side channel restoration and sediment augmentation, adaptively implemented through time, could reinstate dynamic alluvial river behaviour on the Bighorn River.

Acknowledgements

The authors would like to thank Brianna Benjamin for assistance with digital map compilation and Erin Connor for assistance with data management. Anne Marie Emery and the Bighorn River Alliance provided a wealth of background information and collaborative discussions. Dennis Fischer provided invaluable field assistance and historical knowledge of the Bighorn River. Mike Ruggles, Earl Radonski, and Ryan Colloton assisted with field work. The authors are grateful to Aaron Hurst, Katherine Skalak, and two anonymous reviewers for helping to improve the quality of this article. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

ArcGIS shapefiles from our study are available at https://rise.bor.doi.net/admin/rise/catalog-item/view/72346.

Additional information

Funding

This work was supported by the US Bureau of Reclamation Science and Technology Research Program [grant number 19306] and the US Bureau of Reclamation Montana Area Office. Funding for Karin Boyd was provided by the Bighorn River Alliance. All other authors were funded by Reclamation’s Science and Technology Program and Reclamation’s Montana Area Office. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.

Notes on contributors

Melissa A. Foster

Dr. Melissa A. Foster is a geomorphologist with the US Bureau of Reclamation. Her work focuses on the crossover between engineering and river science. Her professional interests include reservoir sedimentation, hydraulic modeling, Quaternary dating techniques, and landscape evolution. Prior to joining Reclamation, she received her Ph.D. from CU Boulder in Geological Sciences and worked as a consultant in river restoration and sedimentation.

Jeanne E. Godaire

Jeanne E. Godaire has a bachelor's degree in geology from Bucknell University and a master's degree in geosciences from the University of Arizona. Ms. Godaire has 25 years of experience in fluvial geomorphology applied to flood hazards, river restoration and sedimentation/erosion issues along rivers in the western U.S. She is currently a senior scientist with the U.S. Geological Survey, Rocky Mountain Region in Lakewood, CO, where she coordinates scientific investigations between geologic science centers and other partners, working closely with leadership, partners and science staff in the centers.

Robert C. Hilldale

Robert C. Hilldale has a bachelor's and master's degree from Washington State University in civil engineering, specializing in fluid mechanics and sediment transport. He has worked for the Bureau of Reclamation's Technical Service Center since 2001. His primary experience is applying 2-dimensional hydraulic and sediment transport numerical models to solve issues related to dam removal, fish passage, aquatic habitat, and structural impact.

D. Nathan Bradley

Dr. D. Nathan Bradley is a geomorphologist with the U.S. Bureau of Reclamation. His work focuses primarily on hydraulic and sediment transport modeling in support of river restoration projects that aim to improve fish habitat. He has a bachelor's degree in astronomy from Northwestern University and a Ph.D. in geology from the University of Colorado.

Karin Boyd

Karin Boyd received an MS in geology from the University of Wyoming in 1986. For the past 36 years, she has worked as a private-sector fluvial geomorphologist, specializing in stream assessment and restoration.