**3. Results**

174 Studies on Environmental and Applied Geomorphology

The HGM process of evaluating ecosystem restoration and management options relies heavily on eight types of data, most of which require geospatial digital information usable in an ArcGIS/ArcMAP format. These data include historic and current information about: 1) soils, 2) geomorphology, 3) topography/elevation, 4) hydrology/flood frequency, 5) aerial photographs and cartography maps, 6) land cover and vegetation communities, 7) presence and distribution of key plant and animal species, and 8) physical anthropogenic

The three-stages of HGM are as follows: first, the historic condition and ecological processes of an area and its surrounding landscapes are determined from a variety of historical and current information such as geological, hydrological, and botanical maps and data. Public Land Survey (PLS) maps and notes are especially useful to understand historic vegetation composition and distribution. A key element of HGM is developing a "matrix" of understanding of which plant communities historically occurred in different geomorphological, soil, topographic, and flood-frequency settings (Table 1). For example, in the Mississippi-Missouri River Confluence Area, wet bottomland prairie that was dominated by prairie cordgrass historically occurred at elevations greater than 417 feet, on relict alluvial floodplain terrace surfaces, on silt loam soils, and between the two- and fiveyear flood frequency zones (Heitmeyer and Westphall, 2007). Contemporary areas that offer these conditions, especially surface, soil, and flood frequency attributes now offer the best

Second, alterations in hydrological condition, topography, vegetation community structure and distribution, and resource availability to key fish and wildlife species are determined by comparing historic vs. current landscapes. This analyses is essentially a qualitative "best professional judgment" assessment of current condition and the types and magnitudes of changes, including assessment of which communities are most resilient and which types of

Third, options and approaches are identified to restore specific habitats and ecological conditions. The foundation of ecological history coupled with assessment of current conditions helps to determine which system processes (e.g., periodic dormant season flooding) and habitats (e.g., forest composition) can be restored or enhanced, and where this is possible, if it is at all. Obviously, some landscape changes are more permanent and less reversible (e.g., mainstem levees on the Mississippi and Illinois rivers) than others (e.g., clearing of bottomland forest). Through development of the HGM matrix conservation planners can identify: 1) which, and where, habitat types have been lost or altered the most and establish some sense of priority for restoration efforts; 2) where opportunities exist to restore habitats in appropriate geomorphic, soil, hydrological, topographic settings including both public and private lands; 3) how restoration can replace lost functions and values including system connectivity; and 4) what management types and intensity will be needed to sustain restored communities. HGM can be an iterative process that is well-coupled with adaptive ecosystem management (Christensen et al., 1996; Palmer et al., 2005) because new monitoring and research can be used to refine

edaphic conditions for restoring wet bottomland prairie communities.

**2.5 Hydro-geomorphic methodology** 

change are the most/least reversible.

HGM models and restoration plans.

features.

#### **3.1 Gemorphology**

Land Sediment Assemblage abundance plotted by rive mile illustrates the distribution of each class and the relative width of the floodplain (Figure 5, top). Geomorphic reach overlays helped identify characteristics that separated reaches in a multivariate analysis (Theiling, 2010). The Chippewa River Reach (RM 650 – 750) is separated downstream by the narrower Wisconsin River Reach (RM605-650) which runs through resistant dolomite valley walls (Knox, 2007). The floodplain widens again through erosive shale in the Maquoketa

Hydro-Geomorphic Classification

rivers (Starrett, 1972).

discharge alone can maintain navigable depths.

very important for many flora and fauna.

**3.2 Hydrology** 

and Potential Vegetation Mapping for Upper Mississippi River Bottomland Restoration 177

The Illinois River floodplain presents a diverse land sediment assemblage (Figure 5; Hajik, 2000). The Upper Illinois (> river mile 245) is deeply flooded by dams and only Sandy Terraces remain visible. The Lower Illinois River has not been subdivided into reaches here, but other authors have defined three or more reaches (Starrett, 1972; Sparks, 1992). Terraces are the most abundant floodplain feature, but natural levees are also widely distributed. Active floodplain surfaces increase at river mile 100 below the confluence with the Sanganois River, a major tributary. The Lower Illinois River is slightly narrower than the Mississippi (Figure 5) and it has a much lower gradient than most

The abundance of water mapped at low flow periods was relatively constant in the river in 1890 (Figure 6, bottom). Several large aquatic areas: Lake Pepin – River Mile 765, Lima Lake – River Mile 350, MMR Backwaters <River Mile 200, were notable features of the floodplain in 1890, but now only Lake Pepin and degraded and disconnected MMR backwaters persist. The contemporary distribution of surface water (Figure 6) reflects the impact of navigation dams completed ~1940 (Theiling and Nestler, 2010). Water surface area increases in impounded reaches upstream of RM400 and a repeating pattern of dam effects are apparent. UMRS navigation dams are only required to maintain low flow navigation, and their impoundment effect only extends partway up each navigation pool (Theiling and Nestler, 2010). Dam gates are raised out of the river during flood stage, except at Dam 19 (hydropower), about 15 percent to 50 percent of the time (USACE, 2004c, 2004a) when

The change in distribution of aquatic classes is quite striking in the floodplain upstream from the Rock Island Gorge (~River Mile 500) where impoundment effects are pronounced (Figure 7). Sandbars were lost throughout the river system coincident with increased river stages. Wooded islands were lost in the upper river reaches during the post-dam era because of wind-wave erosion of former floodplain ridges and levees exposed following impoundment (Rohweder et al., 2008). The increase in contiguous, or connected, backwaters is a very prominent change in the upstream reaches, but not very important in lower reaches. Isolated backwaters were not prominent in either period, but they are considered

Floodplain inundation differs throughout river valleys in response to many natural and anthropogenic drivers. Major tributary rivers demark most geomorphic reaches and each contributes flow and its unique sediment signature to the mainstem Mississippi and Illinois Rivers. The wider banded segments in Figure 8 (bottom) represent areas of greater floodplain inundation diversity which typically occurred at tributary fans and in steep valley reaches. Areas where all the flood stages are compressed (e.g., below river mile 125) are primarily influenced by frequent floods that would fill most of the valley. The impact of the navigation system is apparent in the amount that "Pool Stage" increases as a proportion of maximum inundated area upstream from river mile 400. The distribution of the 2-year flood is prominent along the entire river where it commonly exceeds 70 percent of the total floodplain area and 90 percent in a few locations. This is a characteristic of floodwater

distribution across a range of streams and rivers (Leopold et al., 1964).

River Reach (RM510-605) to the Rock Island Gorge (RM465-510) which presents another constrained, resistant dolomite reach (Trowbridge, 1959). Significant widening occurs just below the gorge where the Mississippi Valley intersects an ancient bedrock channel (Iowa River Reach RM420-465). Sandy terraces are abundant in the Iowa Reach and broader reaches upstream (Figure 5, bottom), but they are buried below Holocene sediments downstream of Quincy Illinois near river mile 325. Alluvial/Colluvial apron is ubiquitous, but uniquely abundant in the Des Moines River, Quincy Anabranch, and Sny Anabranch Reaches (RM240-400) where perched wetlands were once present. Paleofloodplain created from Missouri River outwash in the early Holocene is the dominant LSA class at the confluence with and south of the Missouri River (RM200; Bettis et al., 2008). Active floodplain abundance and distribution is relatively constant among reaches. The abundance of aquatic area is higher upstream from river mile 400 because of the effect of dams increasing surface water area in a series of shallow navigation pools (Theiling and Nestler, 2010).

Fig. 5. Geomorphic class distribution in acres and as proportion of total floodplain area for the Upper Mississippi River System.

The Illinois River floodplain presents a diverse land sediment assemblage (Figure 5; Hajik, 2000). The Upper Illinois (> river mile 245) is deeply flooded by dams and only Sandy Terraces remain visible. The Lower Illinois River has not been subdivided into reaches here, but other authors have defined three or more reaches (Starrett, 1972; Sparks, 1992). Terraces are the most abundant floodplain feature, but natural levees are also widely distributed. Active floodplain surfaces increase at river mile 100 below the confluence with the Sanganois River, a major tributary. The Lower Illinois River is slightly narrower than the Mississippi (Figure 5) and it has a much lower gradient than most rivers (Starrett, 1972).

#### **3.2 Hydrology**

176 Studies on Environmental and Applied Geomorphology

River Reach (RM510-605) to the Rock Island Gorge (RM465-510) which presents another constrained, resistant dolomite reach (Trowbridge, 1959). Significant widening occurs just below the gorge where the Mississippi Valley intersects an ancient bedrock channel (Iowa River Reach RM420-465). Sandy terraces are abundant in the Iowa Reach and broader reaches upstream (Figure 5, bottom), but they are buried below Holocene sediments downstream of Quincy Illinois near river mile 325. Alluvial/Colluvial apron is ubiquitous, but uniquely abundant in the Des Moines River, Quincy Anabranch, and Sny Anabranch Reaches (RM240-400) where perched wetlands were once present. Paleofloodplain created from Missouri River outwash in the early Holocene is the dominant LSA class at the confluence with and south of the Missouri River (RM200; Bettis et al., 2008). Active floodplain abundance and distribution is relatively constant among reaches. The abundance of aquatic area is higher upstream from river mile 400 because of the effect of dams increasing surface

water area in a series of shallow navigation pools (Theiling and Nestler, 2010).

Fig. 5. Geomorphic class distribution in acres and as proportion of total floodplain area for

the Upper Mississippi River System.

The abundance of water mapped at low flow periods was relatively constant in the river in 1890 (Figure 6, bottom). Several large aquatic areas: Lake Pepin – River Mile 765, Lima Lake – River Mile 350, MMR Backwaters <River Mile 200, were notable features of the floodplain in 1890, but now only Lake Pepin and degraded and disconnected MMR backwaters persist. The contemporary distribution of surface water (Figure 6) reflects the impact of navigation dams completed ~1940 (Theiling and Nestler, 2010). Water surface area increases in impounded reaches upstream of RM400 and a repeating pattern of dam effects are apparent. UMRS navigation dams are only required to maintain low flow navigation, and their impoundment effect only extends partway up each navigation pool (Theiling and Nestler, 2010). Dam gates are raised out of the river during flood stage, except at Dam 19 (hydropower), about 15 percent to 50 percent of the time (USACE, 2004c, 2004a) when discharge alone can maintain navigable depths.

The change in distribution of aquatic classes is quite striking in the floodplain upstream from the Rock Island Gorge (~River Mile 500) where impoundment effects are pronounced (Figure 7). Sandbars were lost throughout the river system coincident with increased river stages. Wooded islands were lost in the upper river reaches during the post-dam era because of wind-wave erosion of former floodplain ridges and levees exposed following impoundment (Rohweder et al., 2008). The increase in contiguous, or connected, backwaters is a very prominent change in the upstream reaches, but not very important in lower reaches. Isolated backwaters were not prominent in either period, but they are considered very important for many flora and fauna.

Floodplain inundation differs throughout river valleys in response to many natural and anthropogenic drivers. Major tributary rivers demark most geomorphic reaches and each contributes flow and its unique sediment signature to the mainstem Mississippi and Illinois Rivers. The wider banded segments in Figure 8 (bottom) represent areas of greater floodplain inundation diversity which typically occurred at tributary fans and in steep valley reaches. Areas where all the flood stages are compressed (e.g., below river mile 125) are primarily influenced by frequent floods that would fill most of the valley. The impact of the navigation system is apparent in the amount that "Pool Stage" increases as a proportion of maximum inundated area upstream from river mile 400. The distribution of the 2-year flood is prominent along the entire river where it commonly exceeds 70 percent of the total floodplain area and 90 percent in a few locations. This is a characteristic of floodwater distribution across a range of streams and rivers (Leopold et al., 1964).

Hydro-Geomorphic Classification

and Potential Vegetation Mapping for Upper Mississippi River Bottomland Restoration 179

Fig. 7. Pre-development (top) and contemporary proportional distribution of aquatic area.

The UMRS geomorphic reaches neatly superimpose on our floodplain inundation simulation (Figure 8). The Minnesota (XVI) and Chippewa River (XIV) Reaches show diverse inundation patterns, with the influence of the Chippewa River delta diminishing about mid-reach. The Wisconsin River Reach (XIII) is dominated by frequent floods, but the geomorphically diverse Maquoketa River Reach (XII) influences a diverse floodplain hydrology. The importance of the 2-year flood increases through the Iowa River (X), Des Moines River (VIII), and Quincy (VII) and Sny (VI) Anabranch Reaches until it meets the massive alluvial fan deposited by the Missouri River at Columbia Bottoms (V). Hydrology is similar to upstream reaches in the Jefferson Barracks Reach (IV) between the Missouri River and the Kaskaskia River (III) where the low elevation floodplain is greatly influenced by the 2-year flood. The Illinois River shows a relatively diverse flood stage distribution that is consistent in most of the reach (Figure 8). The influence of the higher head dams above river mile 150 is apparent, whereas the influence of dams is much less in most of the rest of the

Fig. 6. Surface water impacts from impoundment differ in the northern and southern parts of the system as represented by acres of surface water (bottom) and the map of the Lock and Dam 13 area at River Mile 522. Dam effects in the upper pools are similar to the upper portion of the 1989 image with large contiguous backwaters created by dams, whereas dam effects in downstream pools are more similar to their pre-dam form as shown in the bottom part of the 1989 image.

Fig. 6. Surface water impacts from impoundment differ in the northern and southern parts of the system as represented by acres of surface water (bottom) and the map of the Lock and Dam 13 area at River Mile 522. Dam effects in the upper pools are similar to the upper portion of the 1989 image with large contiguous backwaters created by dams, whereas dam effects in downstream pools are more similar to their pre-dam form as shown in the bottom

part of the 1989 image.

Fig. 7. Pre-development (top) and contemporary proportional distribution of aquatic area.

The UMRS geomorphic reaches neatly superimpose on our floodplain inundation simulation (Figure 8). The Minnesota (XVI) and Chippewa River (XIV) Reaches show diverse inundation patterns, with the influence of the Chippewa River delta diminishing about mid-reach. The Wisconsin River Reach (XIII) is dominated by frequent floods, but the geomorphically diverse Maquoketa River Reach (XII) influences a diverse floodplain hydrology. The importance of the 2-year flood increases through the Iowa River (X), Des Moines River (VIII), and Quincy (VII) and Sny (VI) Anabranch Reaches until it meets the massive alluvial fan deposited by the Missouri River at Columbia Bottoms (V). Hydrology is similar to upstream reaches in the Jefferson Barracks Reach (IV) between the Missouri River and the Kaskaskia River (III) where the low elevation floodplain is greatly influenced by the 2-year flood. The Illinois River shows a relatively diverse flood stage distribution that is consistent in most of the reach (Figure 8). The influence of the higher head dams above river mile 150 is apparent, whereas the influence of dams is much less in most of the rest of the

Hydro-Geomorphic Classification

**3.3 Hydrogeomorphic methodology** 

and Potential Vegetation Mapping for Upper Mississippi River Bottomland Restoration 181

have been hugely successful in preventing inundation during high frequency flood events with only a few significant disasters (Belt, 1975; Interagency Floodplain Management Review Committee, 1993; Galloway 2008). Levees and the development they protect have

Our HGM maps are relatively simple deterministic models that select various combinations of hydrology, geomorphology, and soil to map individual community distribution (Figure 9) which are integrated to produce potential vegetation estimates (Figure 9). Potential vegetation (HGM) maps (Figure 10) have been produced for several Mississippi River

Reaches (Heitmeyer, 2008a; 2010) and many individual refuges or restoration sites

Fig. 9. Hydrogeomorphic Data layers and examples of deterministic model results.

greatly altered hydro-ecological drivers and land cover in the floodplain.

river. Dam effects on the Illinois River are exhibited by much larger and permanent backwater lakes compared to isolated lake and channel networks present at low flow prior to development (Mills et al., 1966).

Fig. 8. Simulated floodplain inundation (bottom) and levee distribution by river mile.

Levees impede the flooding simulated above and prevent floodwater distribution in the floodplain south of river mile 450 (Figure 8). Most UMRS levee districts were established more than 100 years ago, and they occur as independent, quasi-political entities that have taxation and other authority for residents within their boundaries (Thompson., 2002). They have been hugely successful in preventing inundation during high frequency flood events with only a few significant disasters (Belt, 1975; Interagency Floodplain Management Review Committee, 1993; Galloway 2008). Levees and the development they protect have greatly altered hydro-ecological drivers and land cover in the floodplain.
