Introduction

The Upper Mississippi River Restoration (UMRR) program covers 2.7 million acres across the Upper Mississippi River System (UMRS), including the Illinois River (fig. 1). The 1986 Water Resources Development Act (Public Law 99–662) recognized the importance of the UMRS for both aspects and authorized the U.S. Army Corps of Engineers (USACE) to implement the UMRR Program in consultation with the Secretary of Interior and the adjacent States. The program has two elements: monitoring, research, and analysis (long term resource monitoring [LTRM]), and habitat protection rehabilitation and enhancement (HREP).

The LTRM and HREP components of the UMRR acknowledged a need for more context of the river’s hydrogeomorphology across aquatic and terrestrial settings, especially regarding areas that are prone to hydrogeomorphic change. The LTRM is done annually in six study reaches and trend pools and includes water quality, aquatic vegetation, fish, standardized study design and methods, and centrally stored and publicly available data in a variety of formats. The six study reaches include pools 4, 8, and 13; La Grange Pool; pool 26; and open river Jackson, Missouri. Initially the program had a component for monitoring bathymetry, but this was disbanded. Long-term monitoring is needed in the UMRS because of its complex ecologic and navigational significance. In addition, a high-level synthesis of existing studies and more recent data collection and mapping efforts was needed to better design and prioritize future research and monitoring associated with the HREP element of UMRR.

In 2018, a core team of scientists was assembled to develop a conceptual model for hydrogeomorphic change and begin to outline a hierarchical classification scheme with the combined knowledge gained from a broad panel of expert geomorphologists, engineers, and physical scientists with regional expertise in a workshop setting (appendix 1). The workshop panel covered the background and history of geomorphic classification schemes and UMRR-related geographic information system (GIS) datasets available for possible use in the classification. The approach included incorporation of research results and restoration activities that had taken place after the completion of a cumulative effects study (CES) in 2000 (WEST Consultants, Inc., 2000). A feasibility investigation of hydrogeomorphic modeling and analyses for the UMRS had also been published with relevant information (Heitmeyer, 2007). The 2018 workshop included the following specific goals and outcomes:

The idea for the hydrogeomorphic change conceptual model and classification was introduced by Schumm (1977) to describe the river continuum of geomorphic processes and channel forms in relation to predictable zones of erosion, transport, and deposition in a stream network. During the last couple of decades, there have been several classifications developed for large rivers that describe geomorphic response potential; relative stability of channel types related to type and amount of sediment load, sediment size, flow velocity, and slope; and floodplain-river interactions (Nanson and Croke, 1992; Fryirs and Brierley, 2000; Church, 2002; Thorne, 2002; Fryirs, 2003; Buffington and Montgomery, 2013; Wheaton and others, 2015; Fryirs and Brierley, 2022). Integration of hydrogeomorphology with ecology has been a long-term priority for science and management actions at international levels for more than a decade (Vaughan and others, 2009).

A system-wide conceptual framework helps river managers and decision makers with a context-based understanding of how the current geomorphology of the river is potentially changing. Characterization of geomorphic forms with processes can be assessed relative to historical changes and how human alterations or climate change may affect future trajectories of change (Grabowski and others, 2014; Belletti and others, 2017). This understanding can facilitate future rehabilitation techniques and project selection, prioritization, and expected longevity regarding areas with the potential for erosion and sedimentation. For project implementation, challenges included predicting backwater sediment deposition, altered connectivity, shoreline erosion, and floodplain forest mortality. Opportunities included development of a qualitative baseline sediment deposition and erosion map and analysis of the experimental design of restoration features compared with geomorphic scale and how they interact to potentially accelerate deposition in backwaters and affect habitat resilience. The hydrogeomorphic change framework provides critical information for river managers and decision makers to prioritize restoration actions within the HREP. Key needs for restoration are identification of areas of habitat needs and habitat resilience. Key habitat questions include what habitats are most affected by hydrogeomorphic change, where are they, how are they changed, and were restoration efforts successful?

Of special interest for the UMRR is how historical (for example, lock and dam systems) and new structures (for example, constructed islands) affect flows and sedimentation patterns. Also, potentially changing tributary inputs of sediment and flow are of interest, with such inputs considered boundary conditions for studies focused on the UMRS main stem and its valley. Causes for and trajectories of sedimentation of pools in backwater lakes was also a major interest (Rogala and others, 2020b).

This report documents the approach and results from the development of a hydrogeomorphology-based conceptual model and hierarchical classification system for the UMRS. The report contains a summary of previous studies and existing geospatial data layers that were reviewed and considered for helping to build a classification system. Also included are a draft application of the conceptual model and classification, and future needs and next steps for full development. The resulting model and classification system were developed from the combined efforts of a core team of scientists and engineers from the U.S. Geological Survey (USGS) and USACE aided by a wider panel of workshop participants that helped to cover the large spatial extent and range of environmental settings along the UMRS, as well as connect with other related ongoing and past work. An important component of this project was to begin to explore and update the UMRS GIS database and query tools.

Previous Studies and Existing Geospatial Data

As part of the core meetings and 2018 workshop, existing studies and geospatial datasets were examined and summarized with the goal of potential use in a hydrogeomorphic change mapping system. The following descriptions of previous studies and existing datasets were extracted from core team and workshop meeting minutes. The descriptions are not all-inclusive but are focused on the potential application of the datasets to the development of the conceptual model and hierarchical classification.

Cumulative Effects Study

A CES (WEST Consultants, Inc., 2000) for the UMRS addressed how the navigation projects affected the river environment. In addition, one goal of the CES was to predict future conditions for the year of 2050, including potential changes related to tributaries, levees, land use, climate change, and floods. The first volume of the CES report contained the results of a broad geomorphic assessment compiled by a board of consultants and a USACE support team consisting of engineers, hydrologists, and geomorphologists familiar with the Upper Mississippi River. The CES identified UMRS geomorphic reaches by longitudinal profile and location of dams, including 10 reaches on the Mississippi River and two on the Illinois River (fig. 2). When identifying reaches, effects of tributaries and the associated deltas were considered. The longitudinal profile had not been updated since 2000 (fig. 3; WEST Consultants, Inc., 2000). The CES described the geologic background that informed the reach designations. These geomorphic reaches were slightly refined later with more information on geomorphic controls, including levees, for floodplain inundation studies (Theiling and Burant, 2013).

The CES included sediment budgets for two-time periods, 1930s–1950s and 1950s–2000, for the main channel and pools 11 through 26. Unknowns in the sediment budget included off-channel sediment storage and changes in tributary loads over time. Bedload was measured for tributaries including the Chippewa, Black, and Wisconsin Rivers (Rose, 1992). If no other bedload data were available, tributary inputs were calculated assuming bedload was 10 percent of the total sediment load. Dredging records from USACE District Offices helped in completing the sediment budgets. Tributary and main-stem loads may have decreased over time because of storage in tributary reservoirs and improved watershed land use, including soil conservation practices.

Conclusions and recommendations from CES included the need for more research on the following topics:

• • Effects of climate change and global warming on hydrology and sediment transport

• • Sedimentation in backwater areas and possible loss of habitat diversity

• • Loss of continuous and isolated backwaters

• • Role of secondary channels in changing backwater habitats

• • Several topics specific to sediment transport

  • o Contributions of suspended-sediment loads and bedload from gaged and ungaged tributaries

  • o Contributions from bank erosion

  • o Changes in reservoir trapping efficiency

During the 2018 workshop, attendees were able to re-examine the CES’s 50-year projections for geomorphic conditions and to refine mechanisms driving the patterns in geomorphic change. The following considerations related to the CES were identified during the 2018 workshop:

• • The 2050 CES predictions need to be reviewed and the CES’s descriptions of the mechanisms driving the patterns in geomorphic change need to be refined. This could include a review of the forecasts for 1950–2000 and analysis of decadal based changes perhaps driven by climate change.

• • Physiographic regions need to be grouped by their glacial geology to determine if there are geographic differences that were not apparent for physiographic regions.

• • Available data for the 2000–20 period need to be examined to determine if overall floods and tributary sediment loads are decreasing as expected from land-use changes, with exceptions such as Lake Pepin owing to documented increasing sediment loads on the Minnesota River.

• • Determine if isostatic rebound of glacial bulge areas (terminus of Wisconsin, Illinoian, and pre-Illinoian glacial margins) are evident in longitudinal profile deviations in tributaries and the main stem and need to be considered as a driver of geomorphic change.

• • How hydraulic models and sediment data and models can be used for describing the potential for geomorphic change needs to be explored. There are two-dimensional hydraulic models available for pools 5, 8, 13, 21, and 26 for the 2-year flood event. Pool areas were assigned velocity groupings of high (greater than 0.45 meter per second), medium (0.45 to 0.15 meter per second), and low (less than 0.15 meter per second) velocity. Sediment data are available for all pools from navigation dredging projects to describe particle size and sediment chemistry. Most of the navigation channel was medium-sized sand, coarsening upstream and fining downstream. Pools 4, 5, 6, and 8 had sediment transport models (River Resources Forum–Water Level Management Task Force, 2013; Nelson, 2020; Nelson and others, 2022).

• • Causes for spatial heterogeneity in backwater sedimentation rates need to be investigated.

• • How climate cycles and potential changes in flood characteristics have interacted with land-use changes need to be considered.

• • The previous assumptions that reservoir trapping rates were constant, and that bank erosion was a small fraction of total sediment load and therefore was considered negligible need to be explored.

• • Projections for geomorphic changes, mostly along the main channel, need to be described.

The 2018 workshop participants also identified more near-term followup actions for the CES, which included updating the longitudinal profile with available water-level datasets and overlaying the more detailed profile with existing reaches to see if any followup was needed for reach designations.

Planform Change Mapping

Coincident with this study was a study focused on mapping planform changes through overlays of 1989, 2000, and 2010 land-cover/land-use maps (Rogala and others, 2020a; Rohweder, 2019). Changes in the vegetation classes in the land-cover/land-use maps were the basis for identifying new areas of land in the planform change mapping. Willow, and secondarily wet meadow, were especially important for reflecting new fluvial landforms. The dataset was determined to be important for developing the conceptual model because it described the location and mechanisms of hydrogeomorphic change in the UMRS. A particular focus of the study was depositional features common in large rivers, especially those with numerous impoundments.

Planform changes during two time periods (1989–2000 and 2000–10) were classified based on a combination of geomorphic form, origin of sediment, geomorphic process, and proximity to existing terrestrial land (Rogala and others, 2020a). Classes included tributary delta, crevasse splay/delta, impounded delta, and bar-tail limbs. The first three delta-related areas had lenticular forms arranged in fan shapes and were generally downstream from distributary channels. The bar-tail limbs were along the margins of a main or side channel and had more linear forms. Many new depositional features were not classified into one of the four main types. The amount of area of new fluvial landforms was similar between the two time periods, with about 350 hectares of depositional bars added across the UMRS in each period (fig. 4). These features illustrated areas of hydrogeomorphic change resulting from a potential combination of hydrologic drivers, hydraulics, and sediment inputs at multiple spatial scales. Important factors affecting the spatial distribution of change areas included location within the impoundments, tributary entrants along valley sides, and proximity to the main channel and levees.

Obstacles to overcome in the change mapping were georeferencing errors, water-level effects, effects of seasonality and age on new vegetation, and subjectivity in methods used to derive polygons. Areas with planform change were mapped if larger than 0.1 hectare. The adjacency of new landforms to existing terrestrial land was also tracked in this dataset.

Conceptual Model Development for Upper Mississippi River System Hydrogeomorphic Change

Conceptual models are constructed to help visualize, describe, and communicate the key components within a complex system of interest and how they interact (Jacobson and others, 2015). No studies with a specific focus on terrestrial and aquatic areas of the river system that are prone to hydrogeomorphic change from the combination of hydrology, flows, and sediment dynamics had been done for the UMRS. Geomorphology of the river is one of five major ecosystem descriptors for the UMRS and UMRR program with the other groupings of hydraulics and hydrology, biochemistry, habitat, and biota (De Jager and others, 2018). Existing conceptual models and classifications that were examined during the project included general hierarchical classification system for stream habitat (Frissell and others, 1986), river styles systems (Brierley and Fryirs, 2005), dam/interdam sequencing (Skalak and others, 2013), classifications of linearly connected geomorphic forms related to floodplain sedimentation (Lewin and others, 2017), and the lower Columbia River estuary classification system (Simenstad and others, 2011). Two other models of particular significance for identifying drivers of geomorphic processes include the model of drivers of flow and sediment in geomorphic responses and multiple feedback loops from Charlton (2008) and Newson and others (1998).

The wide range of expertise and disciplines of the 2018 workshop participants allowed for active discussion and inclusion of the major processes of hydrogeomorphic change in the UMRS that helped to construct the beginnings of a conceptual model (fig. 5). Extra attention was given to immediate needs for furthering interaction among USACE’s restoration and management activities and the USGS’s science in support of restoration. The brainstorming groups tallied important spatial scales, within-corridor drivers and controls, drivers and feedbacks of processes, river processes and rates, evolutionary pathways, hydrogeomorphic responses, outcomes of processes, and indicators of processes. Simplistic conceptual models of hydrogeomorphic change were aided in development by familiarization with other larger river conceptual models, such as the conceptual ecological model developed for pallid sturgeon (Scaphirhynchus albus) population dynamics in the Missouri River pallid sturgeon effects analysis (Jacobson and others, 2015).

This first attempt at a conceptual model for hydrogeomorphic change included recognition of the importance of sediment sources and interactions of flows in describing the dynamics of fluvial landforms (fig. 5). The ultimate focus of the conceptual model was on those geomorphic processes that produce dynamic and continually changing landforms, acknowledging that the processes reflect human and biological interactions that can change over time and understanding the processes of hydrogeomorphic change can better link science and research activities with restoration and management associated with the UMRR. The workgroup identified six scales of interest that spanned the three scales shown in figure 5 and acknowledged the importance of the dams—entire basin, subbasins (hydrologic unit code 4 level; Jones and others, 2022), the four major river reaches, the 12 geomorphic reaches, interdam sequences, and hydrogeomorphic complexes. These six nested spatial scales have elements of conceptual models and classifications of Frissell and others (1986), Brierley and Fryirs (2005), Skalak and others (2013), and Lewin and others (2017), and formed a starting point for the discussions on and the development of the UMRS hierarchical classification.

The relatively simple conceptual model that was developed illustrates drivers and boundaries at three major spatial scales: the basin, a segment of the stream network, and a section or reach of the Upper Mississippi River valley corridor (ascending order in fig. 5). The three spatial scales span the scales at work for altered hydrology, sediment budgets, and topography (Lord and others, 2009). Drivers of change at the basin scale include vegetation and land cover, topographic relief, climate, and human modifications that speed or inhibit the delivery of water and sediment to downstream areas and affect the hydrology at the segment scale. In turn, the processes may be controlled by boundary conditions of soil types, glacial landforms, and natural base-level features such as bedrock types and glacial end moraines. At the segment scale, tributary flow inputs, as well as valley or floodplain slopes and the interaction of slopes with bedrock outcrops and valley features, and dams, can be quantified. It was thought that the subbasin level would be useful for looking at climate change scenarios and tributary inputs to the valley corridor.

The most granular valley corridor scale allows for accounting for how the hydrology and sediment inputs from the basin and segment scales locally affect flow characteristics, the distribution of suspended and bedload, and channel slopes. Those interactions depend on the proximity of the main channel and artificial structures, floodplain vegetation, and channel hydraulics. Thus, resulting landforms derived from hydrogeomorphic change reflect the local hydraulics and velocities and their interaction with sediment (for example, for a project location or a river reach) as well as broader scale hydrology and sediment inputs from the surrounding landscape and tributary watersheds. The type of, and potential for, hydrogeomorphic change is mediated by local variations in vegetative roughness, proximity to tributaries and dams, valley slope and width, and hardened structures added throughout the years for navigation and rehabilitation. Hydrogeomorphic changes reflect several time scales ranging from episodic, immediate changes from flood events to more multidecadal responses to interdam base-level rises associated with the 29 lock and dam structures constructed in the 1930s.

Hydrogeomorphic processes that produce depositional and erosional changes in fluvial landforms in the UMRS are related to the interaction of flows and sediment sources within a surrounding environment of alluvial sediment, vegetation, relict glacial meltwater landforms, and anthropogenic structures (fig. 5). In large mapping efforts based on remote sensing data sources, depositional forms are easier to distinguish in repeat mapping efforts than are erosional features, especially erosion in the vertical direction. However, even for depositional forms, it is difficult to link the form back to the process and a rate and source for a change because of the complexities involved in geomorphic complex-response mechanisms (Lane and Richards, 1997; Buffington, 2012).

Large lowland river systems tend to be dominated by depositional environments within and transcending channels onto floodplain surfaces (Lewin and others, 2017) and the UMRS is no exception. The lock and dam system adds further base-level control that continues to promote sediment deposition upstream from the dams. Depositional processes of hydrogeomorphic change include aggradation, lateral channel migration, floodplain sedimentation, levee building, backwater filling, delta and fan expansion, and bar growth, including downstream extension of islands. The presence of depositional landforms and their spatial heterogeneity lend evidence for determining the processes responsible for them. Notable erosional processes include lateral channel migration, new channel formation, island loss, incision, wave shoaling, scour, widening, levee breaching, and bank erosion. The distance and elevation of a landform relative to the nearest channel gives some indication of the potential processes involved in its formation (Fryirs and Brierley, 2022). In addition to physical changes, biological activity associated with beavers may also lead to erosion and sedimentation processes including the formation of new side or accessory channels and damming of old.

Indicators of hydrogeomorphic changes identified during the 2018 workshop included both measured (repeat surveys) and predictive (modeling). Examples of indicators included hydraulic geometry (width, depth, area) of the main channel, side channels, and connector channels; longitudinal profiles of the channels and aquatic and terrestrial overlays; areas that need frequent dredging; specific gage analyses (for example, Biedenharn and others, 2017); variations in fluvial landforms; cumulative distribution of elevations (topobathy diversity); velocity/depth combinations; lateral connectivity indices of main channel to side channel and backwaters; variations in aquatic and terrestrial vegetation; channel bedforms; and bed sediment texture. Losses, gains, or changes in geomorphic types can also be indicative of change. For example, sequences of bedforms, such as riffle-pool or dune-ripple, give indications of the hydraulics and type of sediment transport (Montgomery and Buffington, 1993; Wheaton and others, 2015).

Other geomorphic concepts that are difficult to show on the simplified conceptual diagram (fig. 5) included time scales involved in the outcomes of geomorphic change, such as the interplay of historical drivers and controls, evolutionary pathways, and geomorphic feedback mechanisms. Timescales of importance identified during the 2018 workshop included yearly to seasonal, 20 years in the past (for example, 2000–20), 30–50 years in the future (within 30-year epoch climate cycles and 50-year design life for HREP), and 100 years in the future (long-range climate predictions). Hydrogeomorphic changes at the reach scale might be abrupt or gradual, depending on the duration and frequency of hydraulic energy required to erode or deposit sediment. The changes may represent exceedance of a threshold (such as bank stability and tree toppling along banks) or be gradual or episodic (such as flood-related overbank sedimentation). An example of a geomorphic feedback mechanism might be bar formation and backwater sedimentation related to a levee breach and new side channel formation. The evolutionary trajectory associated with that side channel might be eventual channel filling (from main channel sedimentation) and building of a delta in the backwater that eventually becomes a relict feature once the side channel fills with sediment. Mechanisms associated with a longer time scale would be the decadal and long-term geomorphic response to the construction of the dams in the 1930s, which changed the valley slope and backwater sedimentation and likely resulted in new connector channel formation, splays, deltas, sand filling, and transport of silts and clays downstream. Geomorphic responses to the rise in base level associated with the dam construction may continue for centuries and affect responses to decadal changes in hydrology and sediment inputs. The location of the dams may alter tributary channel migration through the valley bottoms and delta formation. Wing dams, constructed for easier navigation, cause main channel narrowing and floodplain building.

Not shown on this diagram (fig. 5) are important steps of conceptual model building that require more ecological details that could be related to habitat needs and biological functions and endpoints (Jacobson and others, 2015; De Jager and others, 2018). These include development of indicators of change and disturbance, trajectories of recovery, and restoration potential. The past and future trajectories and probabilities of the processes and their effects on ecosystems and biogeochemical processes could be developed. Science topics of immediate interest are aquatic vegetation and sedimentation, island forests and bank erosion, the role of large wood in geomorphic function and habitat, mussel distribution relative to sedimentation, and prolonged flooding effects on floodplain sedimentation and channels connecting backwaters.

Uses and products of the conceptual model linked to geomorphic change include the following:

Components of a Hydrogeomorphic Change Hierarchical Classification System

Large river systems require a systematic and organizational approach for mapping geomorphic features that will be useful for application in scientific studies and restoration activities (Thorne, 2002). An initial hierarchical classification system for hydrogeomorphic change in the UMRS was developed based the conceptual model and its associated spatial scales of interest (fig. 6). During the workshop, the core team provided overviews of the existing datasets that could be useful in developing a classification system for mapping fluvial landforms in the UMRS with a high potential to change. The classification builds off the matrix style HGM floodplain hydrogeomorphology approach (Heitmeyer, 2007; Theiling and others, 2012).

The following six hierarchical classes were included, listed from broadest to finest spatial and temporal scale: (1) physiographic province, (2) floodplain reach, (3) geomorphic segment, (4) longitudinal process zone, (5) hydrogeomorphic catenae, and (6) hydrogeomorphic unit (fig. 6). The first three classes are from existing GIS-based datasets (Fenneman and Johnson, 1946; U.S. Army Corps of Engineers, 2011; WEST Consultants, Inc., 2000, respectively) and the last three classes are proposed as new data sets. The fourth class, an initial longitudinal process zone, was constructed after the 2018 workshop and is included in this report for the first time, adapting the interdam sequences of Skalak and others (2013) for UMRS impounded sections. The details of an approach for development of the hydrogeomorphic catenae and hydrogeomorphic units are proposed for a followup study. The hydrogeomorphic unit forms the basic mapping unit of the classification with a goal of being able to add attributes related to sensitivity to hydrogeomorphic change, mainly erosion and deposition, that would span terrestrial and aquatic settings across the valley corridor. Once areas of potential geomorphic change are identified, the hierarchical classification can potentially help to identify basin-wide changes that happen synchronously, possibly owing to a climatic shift, or long-term continuous change, such as from dam construction. In addition, risk-related changes can be identified at a reach scale, such as shifts in side channel flow and associated sedimentation near bridges or critical habitats (Sear and Newson, 2003).

Physiographic Provinces

The physiographic provinces for the UMRS cover large geographic areas (Fenneman and Johnson, 1946) (fig. 7). The use of the provinces as the highest level of the classification stemmed from its inclusion in defining the geomorphic reaches of the CES (WEST Consultants, Inc., 2000) and reflect overall differences in natural landscape features and glacial/post-glacial history affecting the physical characteristics of the river basin (Knox, 2007). Patterns in land cover/land use, drainage density, basin slope, soils, glacial history, and related tributary hydrology and sediment inputs associated with the provinces may affect valley corridor-scale hydrogeomorphic processes. The UMRS overlaps with mainly the Central Lowland of the Interior Plains, with subdivisions in the Western Lake, Wisconsin Driftless, Till Plains, and Dissected Till Plains sections.

Longitudinal Process Zones (Interdam Sequences)

For geomorphic segments 1–8 in the upper and lower impounded sections of the UMRS, an interdam sequencing of longitudinal flow and sediment interactions was adopted based on the river corridor study of two dams on the Upper Missouri River (Skalak and others, 2013). The Upper Missouri River study described zones of river geomorphic conditions between dams based on their erosion and depositional tendencies and geomorphic forms, although the Missouri River model is based on a system of reservoirs that have more regulated water-level fluctuations and flows than the UMRS. The major river zones in the Upper Missouri River include dam proximal, attenuating, river-dominated transitional, reservoir dominated transitional, and reservoir, which are all organized along a continuum of sediment supply-limited to transport limited conditions. A first attempt at a longitudinal process zone diagram for the UMRS was done near the end of the study and was adapted from the Upper Missouri River using pool 8 as an example (fig. 8).

Following the Upper Missouri River model, the flow and sediment dynamics associated with the river zones can be described in terms of their possible effects on the spatial distribution of hydrogeomorphic units that have a potential to change (table 1). Proceeding downstream from a dam, the sequences move through potentially erosional or unchanging to depositional until the portion of the river becomes dominated by impounded water, where most of the pre-dam fluvial features have been submerged or perhaps eroded by wave action associated with a longer fetch across the open water sections. Most of the hydrogeomorphic change likely happens in the river-dominated transitional zone, where there is enough sediment contributed from upstream erosion or from tributaries, the amplitude of the water-level hydrograph is increasing, and the number of hydraulic connections is high to drive erosion and deposition within the channel and floodplain. Lateral floodplain slopes are steep enough to cause erosion of existing side channels with sediment deposition in the receiving backwater. The effects of sediment from tributaries depend on the proximity of their mouths to the main channel. If the tributaries connect directly to the main navigation channel, it is likely that sediment contributions will be transported downstream or dredged. If they connect to a backwater environment, then most of the sediment may be deposited and form new land (Rogala and others, 2020a). It is important to note that these longitudinal zones will likely have different lengths and not all zones may be present, depending on the lock and dam configuration and the geologic setting.

Table 1.    

River zones descriptions related to interdam sequences of river zones along geomorphic segments 1–8 in the Upper Mississippi River System, in comparison to those developed for the Upper Missouri River (Skalak and others, 2013).

Table 1.    River zones descriptions related to interdam sequences of river zones along geomorphic segments 1–8 in the Upper Mississippi River System, in comparison to those developed for the Upper Missouri River (Skalak and others, 2013).

River zone	Geomorphic indicators (Skalak and others, 2013)	Hydrogeomorphic change characteristics for (adaption for this study)
Dam proximal	Removal of islands in channel, new vegetation growth and stabilization of point bars.	Flow inputs across valley bottom controlled by location of lock and dam structures, sediment starved, lack of new depositional features.
Attenuating	Sand bar islands remain but island movement is steady in managed flows	Intact pre-dam floodplain landforms, distributary flows and side channels, sediment inputs from tributaries variable but potentially large and cause of deposition in backwaters. Proportion of sand to fines from tributaries varies.
River dominated transitional	Creation of large islands on outside of bends from sediment drop off and backwater effects	Transitional between pre-dam floodplain landforms and backwater effects from next downstream dam. Depositional bars common in highly diverse floodplain environments. Tributary deltas common in backwater and cause of backwater sedimentation. Levee breaches and side channels to backwaters common.
Reservoir/impoundment dominated transitional	Inundated scroll bars, large tree die-off from changing reservoir levels, submerged delta front	Limited distributary pre-dam landforms. Depositional bars in more open water settings. Potential for more submerged deposition from main channel and tributaries. Wave erosion more common. Subsurface erosion possible.
Reservoir/impoundment	Minor deposition	Limited to no pre-dam landforms visible. Depositional landforms from tributary deltas may be present. Submerged deposition possible from tributaries. Subsurface erosion possible.

Hydrogeomorphic Units and Hydrogeomorphic Change Catenae

Hydrogeomorphic units (HGUs) form the basic building block and mapping unit of the classification scheme and collectively represent a new combination of terrestrial and aquatic geomorphic forms that are informed by existing mapping sets of topobathy (U.S. Army Corps of Engineers’ Upper Mississippi River Restoration Program Long Term Resource Monitoring element, 2016) aquatic areas (Wilcox, 1993), planform change (Rogala and others, 2020a), and floodplain inundation (Van Appledorn and others, 2021). Geomorphic mapping of fluvial landforms in a GIS has been developed using high resolution bathymetry datasets (Wheaton and others, 2015; Bangen and others, 2017; Kramer and others, 2017). The Wheaton and others (2015) geomorphic unit technique approach has a four-tiered framework that differentiates forms within a channel based on their ties to water levels (tier 1), shape (tier 2), additional key shape attributes (tier 3), and roughness or vegetation modifiers (tier 4). For terrestrial areas the geomorphon approach (automated mapping of shapes with similar morphometrics) was developed by Jasiewicz and Stepinski (2013). Both approaches would use the existing UMRS valley-wide topobathy data. Classifications based on form alone are difficult to associate with geomorphic processes or sensitivity to geomorphic change unless the features include evidence of flow energy and sediment particle size. Other horizontal and vertical associations are also needed, such as proximity of the features to the main channel or tributary mouths, how often they are flooded, and where they are located within the interdam longitudinal process zones.

The individual HGUs can be grouped into catenae, or assemblages of units linked by related hydrogeomorphic processes, based on commonalities in flow and sediment dynamics as well as pre-impounded geomorphology. The HGUs defined by forms in fluvial environments have been successfully grouped into zones of similar geomorphic processes for other large rivers (Lewin and others, 2017; Fryirs and Brierley, 2022). Assemblages of geomorphic units have been created based on river character, behavior, condition, and recovery in the River Styles approach linked by morphology and sediment type, which can be related to available hydraulic energy associated with locations (Fryirs and Brierley, 2022).

The approach for the hydrogeomorphic change catena or process assemblage is different for this UMRR application because it attempts to connect HGUs based on their common origin and formation from the interaction of flows and sediment. The combination of flow direction and inundation relative to the main channel or tributaries, coupled with potential sediment sources, helps to define units more sensitive to erosion and depositional processes. An example catena is the crevasse channel and splay that are common in levee breaches (Lewin and others, 2017). Erosive flows originating from the main or side channel cause crevasse channel formation and subsequent deposition of a delta or fan on the floodplain or backwaters (delta or splay bars) (fig. 9). The crevasse splay deposits which form on floodplains or in backwaters are linked with accessory channels connected to the main stem from which the sediment originated. The splay HGUs are thus linked with the crevasse channel HGU by their common flow and sediment origin (the main or side channel) and by the crevasse channel serving as a conduit and potential additional source for the sediment forming the splay. Crevasse splays and the side channels that feed them are dynamic features that can change after a single flood, and new side channels can form or old channels be filled over time. The linkage of the splay to the source of the flows and sediment help to characterize how fast they might build or where they might form.

Application of Draft Conceptual Model and Hierarchical Classification System to Pool 8

A simple example for crevasse splays (fig. 9) shows how the hierarchical classification system can be applied using available data and pre-existing maps (fig. 10). Crevasse delta bars were one of four types of new land masses identified in Rogala and others (2020a) (Rohweder, 2019). These above-water features are growing and changing shape over time, indicating a hydrogeomorphic setting sensitive to change. The availability of sediment for the bar deposition is linked to proximity of an accessory or side channel. These channel features have been mapped in the UMRR aquatic areas dataset (De Jager and others, 2018; U.S. Army Corps of Engineers’ Upper Mississippi River Restoration Program Long Term Resource Monitoring element, 2018) but not specifically linked to the crevasse delta bars in a mapping system. The combination of the accessory channel linked to the bars forms the assemblage or catena. These usually originate from the main channel as a source for flows to transport sediment from the main channel or to erode additional sediment from a side channel. If the hydraulics of the flows change from either natural or anthropogenic causes, a new side channel may form, grow larger, or fill with sediment. The geomorphic process zone, where the side channels and crevasse splays are most likely to form, is in the river dominated transitional zone (mid-pool; fig. 10). In pool 8, the abundance of crevasse splays is in the river dominated transitional zone coincident with where the valley widens.

Future Needs for Classification, Mapping, and Visualization

As part of the 2018 workshop, panelists identified short- and long-term needs for future classification, mapping, and visualization of hydrogeomorphic change in the UMRS. The exact method to quantify sensitivity to hydrogeomorphic change for the newly mapped HGUs was not determined during the workshop but was proposed to be explored once the HGUs were mapped and overlaid with other geospatial data like the planform change maps.

A summary of the 2018 workshop findings was presented to the larger group of scientists involved in Focal Area 1 planning group, HREP, and LTRM studies in 2020. As part of the 2020 round of UMRR science in support of restoration proposals, an experimental GIS database and query tool was proposed and funded to explore innovative visualization tools for mapping temporal variability in dynamic geomorphic features. Immediate potential applications included floodplain forest dynamics and health, mussel habitat, and fish survival by mapping areas prone to high erosion and sedimentation rates that may cause tree die-off, bury mussel beds, and fill overwintering pools for fish. Possible long-term studies included applications of mapping hydrogeomorphic units and their sensitivity to geomorphic change in terms of sediment source and sink variability from watershed sources, water-level variations, and climate change. In addition, the hydrogeomorphic change classification system should give context to locally derived water chemistry and biological monitoring data. Remotely sensed datasets and models that cover large geographic areas continue to grow. These additions will add granularity to characterizing the drivers, boundary conditions, and temporal context of hydrogeomorphic change in the UMRS.

Below is a list that summarized the workshop findings into short-term and long-term needs to help build the process toward hydrogeomorphic change mapping.

Summary

From 2018 through 2020 the U.S. Geological Survey and U.S. Army Corps of Engineers led a series of calls and meetings, and a workshop to develop an Upper Mississippi River System hydrogeomorphic change conceptual model and hierarchical classification scheme. This report contains a summary of the process used to develop the conceptual model and a description of how the hierarchical classification was assembled. Three of the six datasets used in the hierarchical classification are new. A longitudinal process zone related to interdam sequences was proposed as a fourth level. Hydrogeomorphic catenae and the basic building block of hydrogeomorphic units were proposed to be developed during a follow-on study. The catenae are hydrogeomorphic units linked by flows and sediment sources that resulted in relatively recent (mid-1990s to mid-2010s) changes in fluvial landforms of the valley bottomlands. These landforms are most likely different types of depositional bars but also include erosional features such as side channels and crevasse splays. Short-term and long-term data needs for mapping and classification identified during the 2018 workshop were included in the report for future reference as the mapping process associated with the hierarchical classification and possible applications continue to be of interest to the for the habitat protection rehabilitation and enhancement and long term resource monitoring studies.