Des Moines River Watershed Characterization Report

Watershed Characterization Report

MINNESOTA DEPARTMENT OF NATURAL RESOURCES

DIVISION OF ECOLOGICAL AND WATER RESOURCES

2016

Contents Table of Figures ...... 3 Table of Tables ...... 6 List of Acronyms ...... 7 Executive Summary ...... 9 Introduction ...... 10 Watershed Characterization ...... 10 Geology ...... 11 High Value Resources ...... 16 Rare Natural Features ...... 17 Native Plant Communities ...... 18 Study Background ...... 21 Hydrology ...... 22 Connectivity ...... 23 Geomorphology ...... 24 Methods ...... 27 Hydrology ...... 27 Discharge Analysis ...... 27 Precipitation ...... 27 Double Mass Curve ...... 27 Web-based Hydrograph Analysis Tool ...... 27 Ground Water Usage ...... 28 Connectivity ...... 28 Longitudinal Connectivity ...... 28 Lateral Connectivity ...... 28 Geomorphology ...... 29 Field Methods ...... 29 Office Methods ...... 30 Results ...... 32 Hydrology ...... 32 Discharge Analysis ...... 32 Precipitation ...... 37 Double Mass Curve ...... 38

Web-based Hydrograph Analysis Tool ...... 40 Ground Water Usage ...... 41 Anticipated Trend in Appropriation ...... 42 Connectivity ...... 45 Longitudinal Connectivity ...... 45 Lateral Connectivity ...... 51 Geomorphology ...... 51 Des Moines River Headwaters Watershed ...... 52 Upper Des Moines River Watershed ...... 72 East Fork Des Moines River Watershed ...... 75 Restoration and Protection Strategies ...... 89 Correlated Planning Efforts ...... 91 Minnesota Prairie Conservation Plan ...... 91 Minnesota State Wildlife Action Plan ...... 93

Table of Figures Figure 1. Light Detection and Ranging (LiDAR) imagery depicting the north-to-south slope of the watershed’s landscape with the major and minor tributaries of the watersheds...... 10 Figure 2. Geomorphic association of the three Des Moines River watersheds within southern Minnesota (from Minnesota Geological Survey). The Des Moines Lobe association comprises much of the landscape and very little exposed bedrock exists within the watersheds...... 11 Figure 3. Sedimentary association of the Des Moines River watershed within Minnesota (SSURGO). .... 12 Figure 4. Soil Survey Geographic database (SSURGO) map of soil organic matter within the Des Moines River watersheds of Minnesota...... 13 Figure 5. Marschner pre-settlement vegetation map of the Des Moines River watersheds within Minnesota...... 13 Figure 6. Soil Survey Geographic database (SSURGO) depiction of soil drainage capacity within the Des Moines River watersheds of Minnesota...... 14 Figure 7. Index represents the proportion of the watershed that has been drained and converted out of wetland coverage (WHAF)...... 14 Figure 8. 2014 land use data for the Des Moines River watersheds (CropScape)...... 15 Figure 9. The relative proportion of total watershed acres harvested annually has risen over time...... 16 Figure 10. The total number of acres harvested for each crop type farmed within the Des Moines River watersheds. Over time, an increase in corn and soybeans has occurred while small grains have decreased...... 16 Figure 11. Rare features, native plant communities, and protected lands in the Des Moines River watersheds...... 19 Figure 12. Five components of a healthy watershed. All components are interrelated; a disruption of any one of the components can have an effect on the rest of the components...... 21 Figure 13. Three spatial dimensions of connectivity: longitudinal (red), lateral (green), and riparian (orange). Dams and bridges are two of several structures that disrupt each of these three dimensions of connectivity. Rapidan Dam, Blue Earth River image courtesy of Google Earth, 2014...... 23 Figure 14. Measurements used to classify a representative stream reach. Once measurements of entrenchment, bankfull width to depth ratio (w/d ratio), sinuosity, and slope at a riffle cross section have been established, a conclusion on stream type can be made. Additional measurements are taken to determine whether the stream is stable within its current state of if it is in a successional state to adapt to its current climate, hydrology, and land use (Rosgen 1997) ...... 25 Figure 15. Location of discharge data collection site on the Des Moines River...... 32 Figure 16. Hydrograph depicting discharge over the period of record for the Des Moines River watersheds at the Jackson gage with the linear discharge and 5 year average trends represented...... 33 Figure 17. Flow duration curve for the Des Moines River gage...... 34 Figure 18. Flow duration curve normalized by watershed area...... 35 Figure 19. Annual precipitation and the total number of days each year that flows were at or above flood level flows (i.e. Q10 – 1180 CFS)...... 36 Figure 20. Hydrograph depicting discharge over the period of record at the Jackson, MN gage with 1.5 and 100 year return intervals identified...... 36 Figure 21. Annual precipitation and the total number of days each year that flows were at or below low flows (i.e. Q90 – 6 CFS)...... 37 Figure 22. Annual precipitation trend analysis for the Des Moines River watershed...... 37

Figure 23. Annual deviations from precipitation averages within the Des Moines River watershed...... 38 Figure 24. Double mass curve for Des Moines River near Jackson...... 38 Figure 25. Double mass curve plotted using the calculated runoff from the WHAT model...... 39 Figure 26. Annual monthly discharge for period of record split at the date identified by the change in relationship within the double mass curve and then plotted together...... 39 Figure 27. Calculated baseflow and runoff volumes...... 40 Figure 28. Calculated baseflow and runoff volumes...... 40 Figure 29. Appropriation by resource type in the Des Moines River watersheds...... 41 Figure 30. Authorized groundwater use in the Des Moines River watersheds...... 41 Figure 31. Total authorized water use in the Des Moines River watersheds...... 43 Figure 32. Volume of water appropriated by use type for the years 2010-2015...... 44 Figure 33. Spatial location of twenty six structures and potential fish passage barriers within the Des Moines River watersheds...... 46 Figure 34. Outline depicting Okabena Creek’s contributing watershed as well as the Whiskey Creek watershed that contributes to the Okabena Creek watershed when the structure within the city of Worthington is opened...... 46 Figure 35. Lakes, rivers, streams, and restorable wetlands of the Des Moines River watersheds...... 47 Figure 36. Relative proportion of altered stream miles within the minor sub-watersheds of the Des Moines River watersheds...... 48 Figure 37. Location of bridges and culverts as identified by the MNDOT shapefile, as well as road/stream intersections, throughout the Des Moines River watersheds...... 48 Figure 38. Longitudinal connectivity of riparian corridors within the sub-watersheds of the Des Moines River watersheds as assessed by the Watershed Health Assessment Framework (WHAF)...... 49 Figure 39. Perennial vegetative cover within the sub-watersheds of the Des Moines River watersheds as assessed by the watershed health assessment framework (WHAF)...... 50 Figure 40. Spatial location and channel classification of geomorphology survey sites within the Des Moines River watersheds...... 51 Figure 41. Aerial photo showing the channel straightening that occurred at the upper Jack Creek survey location...... 61 Figure 42. Restorable depressional wetlands within the Jack Creek survey site's watershed (RWI)...... 62 Figure 43. The shallow root system of the riparian corridor leaves streambanks vulnerable to erosion and bank sloughing is prevalent...... 64 Figure 44. Study bank cross section resurveys from 2013, 2014, and 2015 show the lateral bank erosion that occurred over the course of the study...... 65 Figure 45. Channel succession scenarios documented by Dave Rosgen (Rosgen 2014)...... 67 Figure 46. Watershed boundaries of the 137 square mile Okabena Creek watershed with the inclusion of the 14.43 square mile Whiskey Creek watershed that occasionally contributes to Okabena Creek when the gates of a structure are lowered...... 70 Figure 47. Analysis of aerial photography indicates that the overall pattern of the stream channel has remained relatively unchanged since 1938...... 74 Figure 48. A meander bend abutting the valley wall shows evidence of bank stratification with a sand and gravel lens separating layers of clay and top soil...... 74 Figure 49. Stream crossings that are not properly sized for the stream channel do not have adequate sediment transport capacity and therefore sediment deposits before, after, and within the crossings.

Three crossings near the study location show signs of deposition where as a clear span bridge in the vicinity allows for proper stream function...... 77 Figure 50. Bank sloughing was prevalent within the study reach. Some instances bank sloughs filled all or parts of adjacent pools...... 80 Figure 51. The abandoned bridge and road grade create flood flow confinement upstream of the study location...... 80 Figure 52. The abandoned road grade and bridge creates flood flow confinement. The bridge constricts flow and disconnects the river's floodplain causing flooding upstream of the bridge (A) but not downstream (B)...... 81 Figure 53. Light Detection and Ranging (LiDAR) imagery identifying an abandoned road grade (A) that confines flood flows (i.e. FCC), a berm that confines the rivers floodplain width (B), and another road grade (C) that creates FCC...... 84 Figure 54. County Ditch 53 riffle cross section depicting the floodplain that the channel built within the original trapezoidal channel...... 86 Figure 55. County Ditch 53 developed a floodplain bench within the original trapezoidal channel creating a small sinuous stream with habitat features (A). After the initial survey, the channel was re- excavated to its original design which removed the floodplain bench and all habitat features (B)...... 87 Figure 56. Longitudinal profile of County Ditch 53 in 2014 and 2015. The excavation of the floodplain bench and thalweg removed all habitat features from the stream reach...... 87 Figure 57. County Ditch 53 pool cross section before and after excavation...... 88 Figure 58. Minnesota Prairie Conservation Plan as it pertains to the Des Moines River watersheds...... 92 Figure 59. Des Moines River Headwaters watershed Wildlife Action Plan priorities...... 94 Figure 60. Lower Des Moines River watershed Wildlife Action Plan priorities...... 95 Figure 61. East Fork Des Moines River watershed Wildlife Action Plan priorities...... 96

Table of Tables Table 1. Native plant communities in the Des Moines watersheds...... 20 Table 2. Dimension, pattern, and profile, measurements used within the Rosgen methodology for channel classification...... 25 Table 3. Valley type descriptions with stable and unstable stream types exhibited (from Rosgen 2014)...... 26 Table 4. Listing of structures within the Des Moines River watersheds including the structures name, type, potential as barrier, number of IWM sites impacted, county, and UTMs...... 47 Table 5. Number and density of bridges and culverts as identified by the MNDOT shapefile throughout the Des Moines River watersheds broken down by study reach drainage area...... 49 Table 6. Vegetation type, bank height, root depth, root density, weighted root density, and BEHI rating for each study site study bank...... 50 Table 7. Mean streambank erosion rates (tons/yr/ft) for each survey site's pool study bank...... 53

List of Acronyms AUID – Assessment Unit Identification

BANCS – Bank Assessment for Non-point source Consequences of Sediment

BEHI – Bank Erosion Hazard Index bgy – Billion Gallons a Year

BHR – Bank Height Ratio

DEMs- Digital Elevation Models

DMC – Double Mass Curve

FFC – Flood Flow Confinement

GNIS – Geographic Names Information System

GPS – Global Positioning System

HLWD – Heron Lake Watershed District

HUC – Hydrologic Unit Code

IWM – Intensive Watershed Monitoring

JD – Judicial Ditch

LiDAR – Light Detection and Ranging mgy – Million Gallons a Year

MNDNR – Minnesota Department of Natural Resources

MNDOT – Minnesota Department of Transportation

MPCA – Minnesota Pollution Control Agency

NBS – Near-Bank Sheer Stress

NID – National Inventory of Dams

QBAA – Quaternary Buried Artesian Aquifers

QWTA - Quaternary Water Table Aquifers

SGCN – Species of Greatest Conservation Need

SID – Stressor Identification

SWCD – Soil and Water Conservation District

SWMA – State Wildlife Management Area

USDA – United States Department of Agriculture

USGS – United States Geologic Survey w/d – Width-to-Depth Ratio

WAN – Wildlife Action Network

WHAF – Watershed Health Assessment Framework

WMA – Wildlife Management Area

WRAPS – Watershed Restoration and Protection Strategies

Executive Summary The Des Moines River watershed within the state of Minnesota is comprised of all or part of three Hydrologic Unit Code (HUC) 8 watersheds. Together, the watersheds drain approximately 1,537 mi2 of agricultural land in a northwest to southeast direction. The following report analyzes the hydrology, connectivity, and geomorphology components of the Des Moines River watersheds as well as general watershed characteristics. Historical gage data, stream crossing data, and applied fluvial geomorphology assessments were analyzed to characterize conditions of the watershed and find relationships to help understand water quality and biological impairments throughout the watershed.

According to the Minnesota Pollution Control Agency (MPCA), the degree of channel alteration varies among the three watersheds. Roughly 53% of the Des Moines Headwaters watershed channels have been altered (i.e. channelized or impounded) while 75.2% of the Upper Des Moines, and 79.9% of the East Fork Des Moines River watershed channels have been altered. Channelized systems have limited floodplain connectivity and are often incorrectly sized (e.g., cross sectional area to drainage area, width/depth ratio), not allowing the channel to effectively transport the sediment of its watershed. Rivers and streams in the Des Moines River watershed that are still able to access their floodplains and dissipate energy during high flows are showing greater signs of stream stability and habitat quality. Altered hydrology is currently a significant driver of geomorphic change, water quality deterioration, and habitat loss. Precipitation and discharge data indicated the amount of water delivered to streams per inch of precipitation has increased over time. This increased volume of water delivery can further destabilize the rivers and streams of the watershed and is a contributing factor to the geomorphic response within the watershed.

Longitudinal connectivity assessments within the watersheds indicated an overall road crossing density of 0.96/mi2. Significant numbers of road crossings were identified within various portions of the watershed (e.g. Interstate 90 corridor), while several abandoned bridge crossings dotted other more remote locations. Road crossings (i.e. bridges and culverts) can have drastic impacts on rivers and streams, especially when improperly sized. Improperly sized road crossings create flood flow confinement (FFC), which in turn can cause channel widening, alter sediment transport capacity, and sediment deposition (Zytkovicz and Murtada 2013). Furthermore, thirty dams or lake outlet structurers were located within the watershed with eleven of the structures potentially acting as fish passage barriers during certain flows.

Four of the survey site channels were classified as deeply incised, while two were moderately incised, four were slightly incised, and one was not incised. The degree of incision is a measure of the degree that the channel has down cut (i.e. low bank height/bankfull height). As a channel down cuts its increases is probability of becoming entrenched (i.e. floodprone width/bankfull width). Entrenched channels cannot access their floodplains and increases the shear stress within the channel. Increases in shear stress reduces channel stability and leads to excessive bank erosion and channel widening. Eight of the survey site channels were classified as slightly entrenched, three channels were moderately entrenched, and one channel was fully entrenched (i.e. cannot access floodplain at 2 times bankfull flow elevation).

Survey assessment results indicated systemic issues within the watershed. Restoration and protection strategies within the Des Moines River watersheds should focus of system wide issues.

Introduction Watershed Characterization The Des Moines River Watershed resides almost entirely within the state of Iowa, however, a small portion (i.e. the headwaters) is positioned in southern Minnesota. One Hydrologic Unit Code (HUC) 8 watershed, and parts of two other HUC 8 watersheds are above the Minnesota and Iowa border. The Des Moines River Headwaters watershed [07100001] is entirely within Minnesota and together with the Upper Des Moines River watershed [07100002] are locally known as the West Fork Des Moines River watershed. The third watershed is the East Fork Des Moines River watershed [07100003] and flows across the Minnesota and Iowa border before joining with the West Fork Des Moines River watershed.

The three HUC 8 watersheds of the Des Moines River watershed are separated from the Watonwan River watershed, and thus the Minnesota River basin, by the Algona Moraine. The Algona Moraine constitutes the western most boundary of the Watonwan River watershed (i.e. eastern most boundary of the Des Moines watersheds) while the Bemis moraine marks the furthest extent of the Des Moines glacial lobe and the western extent of the Des Moines watersheds (Lusardi 1997) Together the three watersheds cover portions of seven different southern Minnesota counties (i.e., Cottonwood, Jackson, Lyon, Martin, Murray, Nobles, and Pipestone). The West Fork Des Moines River watershed is drained by the main stem Des Moines River as well as 9 primary tributaries [Figure 1 (i.e. Beaver, Division, Elk, Jack, Lime, Okabena, and Scheldorf Creeks and Heron Lake watershed)]. The East Fork Des Moines River watershed is drained only by the East Fork Des Moines River and one primary tributary (i.e. Fourmile Creek).

Figure 1. Light Detection and Ranging (LiDAR) imagery depicting the north-to-south slope of the watershed’s landscape with the major and minor tributaries of the watersheds.

Geology Bedrock Geology Bedrock outcroppings can be a relatively common landscape feature within certain watersheds of southern Minnesota. Outcroppings, however; are not well represented within the three HUC 8 Des Moines River watersheds. The primary geomorphic association within the watersheds is the Des Moines Lobe association, while others such as lake and pond sediments and fluvial associations are also well represented (Figure 2).

Figure 2. Geomorphic association of the three Des Moines River watersheds within southern Minnesota (from Minnesota Geological Survey). The Des Moines Lobe association comprises much of the landscape and very little exposed bedrock exists within the watersheds.

Surficial Geology The surficial geology of the Des Moines River watersheds was greatly influenced by the advance of the Des Moines Lobe during the Wisconsin glacial period (i.e. roughly 14,000 years B.P.; Lusardi 1997). As the Laurentian ice sheet receded across the state of Minnesota, glacial meltwater rivers and lakes became prevalent on the landscape. Cutting through the supraglacial drift complexes and till plains that dominate the surficial geology of the Des Moines River watersheds are numerous glacial outwashes left by the meltwater rivers (Figure 3).

Figure 3. Sedimentary association of the Des Moines River watershed within Minnesota (SSURGO).

Soils Till within the Des Moines lobe is gray to brown as it contains Cretaceous shale transported from North Dakota and Canada (Lusardi 1997). The Soil Survey Geologic database (SSURGO) identifies that much of the watersheds’ soils are high in organic matter (Figure 4). As the glacial retreat of the last ice age proceeded, a prairie ecotype established itself (Figure 5). The soils left behind from the glacier range from poorly drained to well drained with a prevalence of lesser drained soils (Figure 6). These soils in accord with the tall grass prairie led to a landscape abundant in lakes, wetlands, and small seasonally saturated meadows [i.e. type 1 and 2 wetlands (BWSR 1999)] and thus a building of organic matter over the thousands of years.

Similar to many of the highly organic soils found in southern Minnesota, those of the Des Moines River watersheds are agriculturally productive. Due to the poor soil drainage classifications and wetlands throughout the watersheds, expansive drainage complexes (i.e. private and public drainage ditches, as well as county and private tile systems) were implemented to lower the water table. The draining of the landscape has thus reduced the number of seasonally saturated meadows, wetlands, lakes (Figure 7), and impacted the overall organic matter within the soils of the watershed.

Figure 4. Soil Survey Geographic database (SSURGO) map of soil organic matter within the Des Moines River watersheds of Minnesota.

Figure 5. Marschner pre-settlement vegetation map of the Des Moines River watersheds within Minnesota.

Figure 6. Soil Survey Geographic database (SSURGO) depiction of soil drainage capacity within the Des Moines River watersheds of Minnesota.

Figure 7. Index represents the proportion of the watershed that has been drained and converted out of wetland coverage (WHAF).

Land Use The Des Moines River watersheds were historically dominated by tallgrass prairie with numerous wet prairie islands and complexes (Figure 5; Marschner pre-settlement map). After European settlement of the landscape, efforts to drain the land for agricultural production began. Today, vast drainage systems exist within the watersheds and have drained many of the lakes, wetlands, and seasonally saturated soils. Currently land use is dominated by agriculture as 85% of the landscape is used for agricultural production (Figure 8; CropScape). Corn and soybean production accounts for 77.2% of the land cover, small grain production accounts for 0.1%, and other agriculture accounts for 7.7%. Water covers roughly 8.1% of the land while development covers 5.6%, and perennial vegetation accounts for 1.4%. Since settlement, the overall proportion of watershed acres farmed has increased (Figure 9; NASS 2016). Furthermore, a significant decrease in total small grain acres harvested has occurred through the years as a significant increase in soybean acres has occurred (Figure 10; NASS 2016).

Figure 8. 2014 land use data for the Des Moines River watersheds (CropScape).

Relative Proportion of Total Watershed Acres Harvested Annually 100 90 80 70 60 50 40 30 20

% of % Total Watershed Acres 10 0 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020 Year

Figure 9. The relative proportion of total watershed acres harvested annually has risen over time.

Acres of Crops Harvested Annually 1,400,000 1,200,000 1,000,000 800,000 600,000 400,000 Acres Acres Harvested 200,000 0 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020 Year

Soybeans Corn Hay/Alfalfa Barley Oats Wheat Flax

Figure 10. The total number of acres harvested for each crop type farmed within the Des Moines River watersheds. Over time, an increase in corn and soybeans has occurred while small grains have decreased.

High Value Resources The Des Moines watershed is in the prairie parkland ecological province and two ecological subsections: the Coteau Moraines and the Minnesota River Prairie. The Coteau Moraines are a transition zone between areas of shallow wind-blown silt (i.e. loess) to deeper deposits of loess (MNDNR 2016). This translates to a rolling landscape with dry to moist prairie soils (MNDNR 2016). The majority of the Coteau Moraines subsection drains northeast into the Minnesota River or southeast to the Des Moines River (MNDNR 2016). The Minnesota River consists of a gently rolling ground moraine about 60 miles wide (MNDNR 2016a). The Minnesota River occupies a broad valley that splits the subsection in half.

The valley was created by Glacial River Warren, which drained Glacial Lake Agassiz (MNDNR 2016a). The Des Moines River watershed retains a variety of rare and unique features that primarily occur along the Des Moines and Minnesota Rivers. The health of these natural resources can be impacted by watershed activities, land-use changes, and hydrologic changes. The following features occur within the Des Moines River watershed:

 Twenty-four mapped native plant communities  Nine designated calcareous fens  One designated trout stream (i.e. Scheldorf Creek; DNR Hydro ID: 102146)  Forty-three rare plant and animal species that are listed as threatened, endangered, or special concern. The list includes state and federally listed species. The following hydrologic features within the Des Moines River watershed are also important from a rare, natural, and high value resources standpoint:

 Lime Creek (DNR Hydro ID: 102438) and Jack Creek (DNR Hydro ID: 102223)  Heron Lake (DNR Hydro ID: 61425), Lake Shetek (DNR Hydro ID: 66328), North Badger Lake (DNR Hydro ID:54502), South Badger Lake (DNR Hydro ID: 55122), Big Slough Lake (DNR Hydro ID: 58913), as well as other important smaller lakes The Des Moines River watershed was historically dominated by tallgrass prairie, with many islands of wet prairie in the far southeast portion of the watershed (MNDNR 2016). Forests on the majority of the area were restricted to ravines along streams. In the southeast, forests of silver maple, elm, cottonwood, and willow grew on floodplains in the Minnesota River Prairie subsection (MNDNR 2016a). Only a small percentage of these native systems remain. These types of natural system features have a social and/or an economic value for society. This value may be direct or indirect. Valuation and documentation of ecosystem changes helps determine whether policy changes that alter these systems result in a net cost or benefit to society. Rare Natural Features Rare features (Figure 11) contribute to the health of the habitat and environment that surrounds us. Some even contribute directly to local economies in the form of recreation, including hunting, fishing, canoeing, kayaking, wildlife viewing, and camping. The DNR has a statutory responsibility to conserve rare features (Minnesota Stat. 84.0895). Rare features can include species of unique plants and animals as well as native plant communities (i.e. habitats). Rare features are often key indicators of the health of our environment. When they decline, it is usually an indication that a natural process or element is not functioning well.

There are 21 endangered and threatened species known to exist in the Des Moines River watershed (MNDNR 2016c). There are an additional 26 species in suspected decline and are listed as special concern or watch list species (MNDNR 2016c, Table 1). These species are often tied to native plant communities that may also be in decline. In addition to the number of rare species, there is also a high density of these species occurring along the main channel of the Des Moines River and around Heron Lake in Jackson County.

Three of the listed species are associated with calcareous fen habitats such as the 9 designated calcareous fens in the Des Moines watershed. Calcareous fens are rare and distinctive wetlands characterized by a substrate of non-acidic peatland dependent on a constant supply of cold, oxygen- poor groundwater rich in calcium and magnesium bicarbonates (MNDNR 2008). This calcium-rich environment supports a plant community dominated by calcium-loving species. Because these types of wetlands are one of the rarest natural communities in the United States, they are a specially protected resource in Minnesota (Minnesota Wetlands Conservation Act, Minnesota Statutes, section 103G.222 - .2373 and Minnesota Rules, chapter 8420). Native Plant Communities A native plant community (NPC) is a group of native plants that interact with each other and with their environment in ways not greatly altered by modern human activity or by introduced organisms (MNDNR 2016b). These groups of native plant species form recognizable units, such as prairies, oak forests, or marshes. The Des Moines River watershed is dominated by wet to dry prairie, marshes, and mesic- hardwood forest communities. Over one-third of the plant communities in the Des Moines are considered imperiled or critically imperiled. Zero of the twenty-four types are considered abundant or secure.

Native Plant Communities in Minnesota have been assigned conservation status ranks (S-ranks) that reflect their risk of elimination from the state (MNDNR 2009). There are five ranks that are determined using methodology developed by the conservation organization NatureServe and its member natural heritage programs in North America. Ranks in Minnesota are based on information compiled by DNR ecologists.

Connections between wildlife species, native plant communities, lakes and wetland features are many and often complex. In order to conserve these features, a tiered approach should be used—preserving native communities, restoration and enhancement to create larger habitat networks, and incorporating best management practices such as soil health into the agricultural landscape. All three tiers can be implemented at the same time and focusing on these three levels of restoration and protection strategies maximizes our conservation benefits. Remaining clusters of rare or sensitive natural features are indicative of habitat quality while their scarcity elsewhere in the watershed signal the need for restoration or adaptive management. Maintaining and restoring biological diversity, abundance, and resiliency is a component of integrated watershed health. The more diverse an area is, the better chance it has at long-term health and self-sustainability. Over the years, there will be variations in invasive species pressure, soil conditions, and climate such as extreme drought or extreme moisture. Having a diversity of communities and species ensures that more of these will become established and adapted to these extremes and can therefore meet the ebb and flow of change.

The Des Moines River watershed is unique in southwest Minnesota because it has areas with dense concentrations of high value ecological features. This is a hot spot in terms of conservation potential. Figure 11 shows large opportunities to create a connected corridor of native and restored plant communities along the main channel of the Des Moines River; Heron Lake, Lake Shetek, North Badger, South Badger, and Big Slough Lakes; and along Lime and Jack creeks. These communities, which include priority fish and wildlife habitat areas, wetland/upland complexes, and natural areas not only provide

18 quality habitat, but sequester carbon, provide a home for rare species, contribute to clean water, and offer many recreational opportunities.

Figure 11. Rare features, native plant communities, and protected lands in the Des Moines River watersheds.

Table 1. Native plant communities in the Des Moines watersheds.

Native Plant Conservation Community Native Plant Community Name Status Rank Code S1 = critically MRp83 Prairie Mixed Cattail Marsh imperiled MRp83a Cattail - Sedge Marsh (Prairie) S1 MRp93a Bulrush Marsh (Prairie) S1 MRp93b Spikerush - Bur Reed Marsh (Prairie) S1 OPp93b Calcareous Fen (Southwestern) S2 = imperiled UPs13d Dry Hill Prairie (Southern) S2 UPs23a Mesic Prairie (Southern) S2 WMs92 Southern Basin Wet Meadow/Carr S2 WPs54b Wet Prairie (Southern) S2 S3 = vulnerable FFs68 Southern Floodplain Forest to extirpation MHs38 Southern Mesic Oak-Basswood Forest S3 MHs38b Basswood - Bur Oak - (Green Ash) Forest S3 MHs49a Elm - Basswood - Black Ash - (Hackberry) Forest S3 WMp73 Prairie Wet Meadow/Carr S3 WMp73a Prairie Meadow/Carr S3 WMs83a Seepage Meadow/Carr S3 CMX=Complex MHS_CX Southwestern Rich Mesic Hardwood Forest Complex Community MMS_CX Meadow - Marsh - Fen -Swamp Complex CMX PWL_CX Prairie Wetland Complex CMX SNR=State Not FFs59 Southern Terrace Forest Ranked MHs39 Southern Mesic Maple-Basswood Forest SNR MHs49 Southern Wet-Mesic Hardwood Forest SNR UPs13 Southern Dry Prairie SNR WMs83 Southern Seepage Meadow/Carr SNR WPs54 Southern Wet Prairie SNR

Study Background The Minnesota Pollution Control Agency (MPCA) initiated the Intensive Watershed Monitoring (IWM) process for the Des Moines River watershed in 2014 to assess the overall health of the watershed and identify areas for restoration and protection efforts. The MPCA and Minnesota Department of Natural Resources (MNDNR) use a five component “healthy watershed” framework to understand and describe how complex ecological systems are functioning. The five components of a healthy watershed consist of hydrology, geomorphology, connectivity, water quality, and biology (Figure 12). Within this five component framework, the MPCA is charged with assessing the biology and water quality components. The MNDNR thus analyzes the current and historical hydrology trends of the watershed, assesses the fluvial geomorphology and stability of rivers and streams within the system, and investigates connectivity (i.e., longitudinal, lateral, and riparian). All of the components are interrelated, and the disruption of any of them can result in undesirable results deeming the stream impaired for one or more condition.

Once all of the components have been assessed, the MPCA completes a stressor identification (SID) document to show what stressors are causing listed impairments within each Assessment Unit ID (AUID). The SID document is one component to help develop Watershed Restoration and Protection Strategies (WRAPS); a report summarizing the water quality monitoring and assessment, pollutant and stressor source identification, civic engagement and public participation, and restoration and protection strategy development. Ultimately, WRAPS can be used to inform local plans and guide conservation work within the watershed.

Figure 12. Five components of a healthy watershed. All components are interrelated; a disruption of any one of the components can have an effect on the rest of the components.

Hydrology Hydrologic conditions (e.g., precipitation, runoff, storage, and annual water yield) and the disturbance of natural pathways (e.g., tiling, ditching, land use changes, and loss of water storage) has become the driver of many impairments in other Minnesota watersheds (MPCA 2012). These disturbances coupled with an increase in precipitation (i.e., total, frequency, and magnitude) have resulted in issues with: increased bank erosion, excess sediment, habitat degradation, and disturbance of natural flow regime.

Hydrologic modification is the alteration or addition of water pathways and associated changes in volume by human activity. Those modifications can dramatically alter discharge due to changes in volume, timing, connectivity, or flow rates, particularly if the area was not a flow pathway in the past. The types of hydrologic modifications are vast, including the draining and filling of wetlands and lakes, ditching or draining formerly hydrologically disconnected basins, adding impervious surfaces across the basin, increasing drainage for increased transport of water (i.e., in urban and agricultural areas), straightening or constricting a natural flow path or river, and changing the timing and rate of delivery within the hydrologic system. Any increase in stream power (e.g., due to change in peak flows or increased frequency of bankfull flows) will generate an increase in water yield (Lane 1995). Reduced surface storage, increased conveyance, increased effective drainage area along with altered crop rotations supporting soybeans over perennial grasses and small grains have all altered the dynamics of, and generally increased the annual water discharge from, these watersheds while also dramatically altering the return interval for various flow stages (Schottler 2014).

In extensively drained landscapes, such as the agricultural Midwest of the United States, the connection of isolated basins has inflated total surface water discharge and increased the density of linear drainage networks (Ter Haar & Herricks 1989, Haitjema 1995, Magner et al. 2004). Many streams in the region are in disequilibrium due to past and current land-use change with corresponding hydrologic responses, as well as direct channel modifications (Lenhart 2007).

These modifications have not occurred at a constant rate, but in episodes or events, such as construction of the public drainage system from 1912-1920 (Lenhart 2007, 2008) and continue today through repair, upgrade, and increased amount of impervious surfaces and subsurface drainage. Construction of subsurface tile and surface ditch drainage systems in the early 1990s increased contributing drainage areas, resulting in greater amounts of water delivered to rivers (Leach and Magner 1992, Kuehner 2004, Lenhart 2008). The effects of these suites of changes are cumulative, interrelated, and tend to compound across different spatial and temporal scales (Spaling & Smit 1995, Aadland et al. 2005, Blann et al. 2009). The contribution of subsurface drainage to aquatic ecosystem affects may be difficult to isolate relative to other agricultural impacts (Blann et al. 2009). Cumulatively, these changes in hydrology, geomorphology, nutrient cycling, and sediment dynamics have had profound implications for aquatic ecosystems and biodiversity (Blann et al. 2009)

The hydrologic analysis found in this report focuses on surface-water components of the hydrologic cycle, rainfall-runoff relationships, open-channel flow, flood hydrology, and statistical and probabilistic methods in hydrology. Furthermore, some analysis of groundwater and watershed appropriations are summarized.

Connectivity Connectivity is defined as the maintenance of lateral, longitudinal, and vertical pathways for biological, hydrological, and physical processes within a river system (Annear 2004). Connectivity is thus the water mediated transfer of energy, materials, and organisms across the hydrological landscape (Pringle 2003). The transport of these integral components within a river travel in four dimensions: longitudinal, upstream and downstream; lateral, channel to floodplain; vertical, hyporheic to groundwater zones; and temporal, continuity of transport over time (Annear 2004; Figure 13). Due to the objectives of this study, vertical connectivity was not directly assessed.

Longitudinal connectivity of flowing surface waters is of the utmost importance to fish species. Many fish species life histories employ seasonal migrations for reproduction or overwintering. Physical barriers such as dams, waterfalls, perched culverts, and other instream structures disrupt longitudinal connectivity and often impedes seasonal fish migrations. Disrupted migration not only holds the capacity to alter reproduction of fish, it also impacts mussel species that utilize fish movement to disperse their offspring. Structures, such as dams, have been shown to reduce species richness of systems, while also increasing abundance of tolerant or undesirable species (Winston et al. 1991, Santucci et al. 2005, Slawski et al. 2008).

Figure 13. Three spatial dimensions of connectivity: longitudinal (red), lateral (green), and riparian (orange). Dams and bridges are two of several structures that disrupt each of these three dimensions of connectivity. Rapidan Dam, Blue Earth River image courtesy of Google Earth, 2014.

Longitudinal connectivity of a system’s immediate riparian corridor is an integral component within a healthy watershed that promotes the free movement of aquatic terrestrial species both up and down stream within a hydrologic system. Continuous corridors of high quality riparian vegetation work to sustain stream stability and play an important role in energy input and light penetration to surface waters. Lateral riparian connectivity provides habitat for terrestrial species as well as spawning and refuge habitat for fish during periods of flooding and represents one of the most significant attributes for maintaining stream stability. Improperly sized bridges and culverts hinder the role of longitudinal riparian connectivity as they reduce localized floodplain access, disrupt streambank vegetation, create impassable flow velocities, stand as tunnel lengths that restrict species movement, reduce or eliminate natural migration pathways for terrestrial and aquatic organisms, and cause streambed and floodplain aggradation upstream and streambed degradation or aggradation downstream.

Lateral connectivity represents the connection between a river and its floodplain. The dynamic relationship amongst terrestrial and aquatic components of a river’s floodplain ecosystem comprises a spatially complex and interconnected environment (Ickes et al. 2005). The degree to which lateral connectivity exists is both a time-dependent phenomenon (Tockner et al. 1999) and dependent upon the physical structure of the channel. Stable river systems are hydrologically dynamic systems where their floodplain inundation relates to prevailing hydrologic conditions throughout the seasons. Riverine species have evolved life history characteristics that exploit flood pulses for migration and reproduction based on those seasonally predictable hydrologic conditions that allow systems to access their floodplains (Welcomme 1979, McKeown 1984, Scheimer 2000). When a river system degrades to a point where it can no longer access its floodplain, the system’s capacity to dissipate hydrologic energy is lost. Without adequate dissipation of hydrologic energy through floodplain access, the hydrologic shear stress on streambanks and streambed increase. Increasing shear stress within the channel causes the channel to widen or streambed to scour, depending upon which feature has the lowest shear strength based upon the materials the feature is comprised of. Channel widening or channel incision driven by this hydrologic shear reduces channel stability and decreases effective bedload transport capacity. Reduced bedload transport capacity causes aggradation of the streambed, leads to the loss of important aquatic habitat, increases water temperature, lowers oxygen levels, and leaves smaller and less diverse bed materials that in turn reduce biotic integrity of the aquatic and terrestrial communities in the system. Geomorphology Fluvial geomorphology pertains to the way land has formed and continues to form by flowing water (Leopold et al. 1964). The principle methods utilized in this study to describe the geomorphology of the watershed follow the Rosgen (1994) classification system (Figure 14). Within the Rosgen classification system, the dimension, pattern, and profile of a stream are measured through the use of dimensionless ratios in order to classify the stream regardless of its size (Table 2). Boundary conditions [i.e. valley type (Table 3)] of stream type are also documented as they can strongly influence channel evolution. Subsequently, measurements such as bank height ratio, erosion rate, and sediment competence are used to assess whether the channel is in a stable or transitional state (i.e. evolving to or from a disturbed channel type).

Equilibrium and river stability are terms that are often used interchangeably within the context of geomorphology. River stability is defined as a river’s ability, in the present climate, to transport the

24 flows and sediment of its watershed, over time, in such a manner that the channel maintains its dimension, pattern, and profile without either aggrading or degrading (Rosgen 1996, 2001a, 2006). When components of a healthy watershed are disturbed, successional changes in rivers occur in order to adjust to the new conditions. Typically, these adjustments impact habitat and water quality due to an imbalance of sediment and bedload supply and transport that results in biotic and turbidity impairments.

Figure 14. Measurements used to classify a representative stream reach. Once measurements of entrenchment, bankfull width to depth ratio (w/d ratio), sinuosity, and slope at a riffle cross section have been established, a conclusion on stream type can be made. Additional measurements are taken to determine whether the stream is stable within its current state of if it is in a successional state to adapt to its current climate, hydrology, and land use (Rosgen 1997)

Table 2. Dimension, pattern, and profile, measurements used within the Rosgen methodology for channel classification.

Table 3. Valley type descriptions with stable and unstable stream types exhibited (from Rosgen 2014).

Confined (C) - Valley Width/Bankfull Width = <7 Boundary Material Description Slope Abbreviation Stable Stream Types Unstable Stream Types Alluvial (AL) Fluvial Deposition, Narrow Floodplain <5% C-AL-FD B, C, E A, D, F, G Alluvial (AL) Inner-Gorge, Entrenched, Meandering <2% C-AL-IG Bc, C, F D, Gc Bedrock (BR) Bedrock-Controlled Landscape Varies C-BR-BC Aa+, A, B, F, G n/a Eolian (EO) Loess Hills >2% C-EO-LH Aa+, A, B, Cb, Eb D, Fb, G Eolian (EO) Sand Hills >2% C-EO-SH Aa+, A, B Cb, Eb D, Fb, G Eolian (EO) Fluvial-Dissected Landscape >2% C-EO-FD Aa+, A, B, Fb, G n/a Colluvial (CO) Fluvial-Dissected Landscape >2% C-CO-FD Aa+, A, B, Fb, G n/a Colluvial (CO) V-Shaped, Steep, and Narrow >6% C-CO-VS Aa+, A, Ba n/a Colluvial (CO) U-Shaped, Moderately Steep <6% C-CO-US A, Ba, B Fb, G Glacial (GL) Glacial Trough, U-Shaped Valley <5% C-GL-GT B, C, D F, G Glacial (GL) Till Plain with Glacial Terraces <5% C-GL-TP B, C, E D, F, G Lacustrine (LA) Abandoned Beaches, Over-Steepened >4% C-LA-AB Aa+, A, B D, Fb, G Marine (MA) Abandoned Beaches, Fossil Beds >2% C-MA-AB Aa+, A, B, Cb, Eb D, Fb, G

Unconfined (U) - Valley Width/Bankfull Width = >7 Boundary Material Description Slope Abbreviation Stable Stream Types Unstable Stream Types Alluvial (AL) Fluvial Deposition, Terraces, Floodplain <3% U-AL-FD Bc, C, E A, D, F, Gc Alluvial (AL) Active Alluvail Fan >2% U-AL-AF D A, Fb, G Alluvial (AL) Inactive Alluvial Fan >2% U-AL-IF Ba, B A, D, Fb, G Alluvial (AL) River Deltas, Gentle Slope <2% U-AL-RD C, DA, E D, F, Gc, Bedrock (BR) Bedrock-Controlled Landscape <2% U-BR-BC C, D n/a Eolian (EO) Sand Dunes, Gentle Slopes <2% U-EO-SD Bc, C, D F, Gc Glacial (GL) Glacial Outwash Plain <4% U-GL-GO Bc, C, D F, Gc Glacial (GL) Till Plain, Moraine Materials <4% U-GL-TP B, C, E D, F, G Periglacial (PE) Cryoplanated Surfaces, Extremely Cold Climates <4% U-PE-CS Bc, C, E F, Gc Lacustrine (LA) Lacustine Deposition - Broad, Gentle, Valley <2% U-LA-LD C, DA, E D, F, Gc Note: a and b = high slope range for that particular stream type; c = low slope for that stream type

Methods Hydrology In order to understand and evaluate the hydrologic processes within a watershed, several types of analysis are used to examine the relationships between flow (discharge) and precipitation. Ground water levels and usage over time is also reviewed. The analysis methods can evaluate and measure changes within a system by reviewing statistical variations and trends over time. Discharge Analysis Flow/discharge data sets are collected by the United States Geologic Survey (USGS) and MPCA/DNR stream gage network for the various watersheds. Site specific streamflow data are calculated using continuous stream stage measurements and periodic streamflow measurements. These data are plotted and charted to allow for statistical analysis and are used to create hydrographs, flow duration curves and other visual representations of the period of record.

Watershed discharge data can be used to review daily, monthly, seasonal, annual and long term trends within a watershed and examine changes in the discharge characteristics such as periods of low or zero flow, flood frequency, base flow volume, and seasonal variability. Discharge data from the Des Moines River were reviewed from the USGS station number 5476000 at Jackson, Minnesota. Precipitation Precipitation data are based on long-term data collection and the gridded data set created based on watershed boundaries. All precipitation data are acquired through the “High Density Radius Retrieval” website maintained by the Minnesota State Climatology Office. Precipitation data are used to examine long term trends within a watershed, and the relationship and response of discharge, runoff, and baseflow conditions relative to recorded precipitation totals. Precipitation data is used on both a monthly and annual timescale. Double Mass Curve A Double Mass Curve is an analysis based on a cumulative comparison of one independent variable with a cumulative dependent variable. This is useful in hydrologic data as it allows the examination of the relationship between two variables. This technique was used to compare precipitation and stream discharge relationships (annual and seasonal) and well elevation fluctuations relative to precipitation. When plotted, a straight line indicates consistency in the relationship, a break in the slope would mean a change in the relationship.

When used with long term discharge data sets, the curve can demonstrate when the change in the relationship began to occur. All double mass curves presented are runoff (Discharge/Watershed area) and monthly precipitation in inches. All discharge values are converted to inches by dividing total volume by the watershed area (the annual discharge converted to acre–ft. and then to inches of runoff over the watershed). Other methods, such as removing baseflow values, allows additional analysis. Additional information on double mass curve development and interpretation can be found on the following website: http://pubs.usgs.gov/wsp/1541b/report.pdf Web-based Hydrograph Analysis Tool The Web-based Hydrograph Analysis Tool (WHAT) was developed by Purdue University and designed to separate baseflow and direct runoff using digital filtering algorithms from user specified flow data. Data

27 can be automatically uploaded from the USGS database or manually entered by the user. The analysis can be run over the entire period of record or for dates specified by the user. Subsets of the data can be used to look for a change in the relationship as indicated by the double mass curve or precipitation records. This tool is beneficial to examine the baseflow discharge relationship over time, and can be used to look at long term and seasonal variations.

The supplied dataset is analyzed using a recursive digital filter, based on a groundwater system with “Perennial Streams with Porous Aquifers”. The tool and additional information can be found on the following website: https://engineering.purdue.edu/~what/ Ground Water Usage Permitted groundwater usage was reviewed to examine changes in type of usage and volume over time. Data was collected through the Minnesota Permitting and Reporting System (MPARS) and used to review total volume appropriated, volume appropriated by county, aquifer type, and well level fluctuations relative to precipitation. Observation well data was also used to examine groundwater trends, and potential ground/surface water interaction when data was available. Connectivity Longitudinal Connectivity Longitudinal connectivity within the Des Moines River watershed was assessed through the use of desktop reconnaissance tools. ArcMap, Geographic Names Information System (GNIS), Watershed Health Assessment Framework (WHAF), National Inventory of Dams (NID), MNDNR’s dam safety records, and Minnesota Department of Transportation’s (MNDOT) bridges and culverts inventory were utilized to assess longitudinal connectivity. Because culvert inventories do not document whether individual structures are perched, dams were the only barriers with adequate data to analyze continuity within the surface waters of the Des Moines Watershed.

Similarly, the same tools were used to assess longitudinal connectivity of riparian corridors throughout the watershed. Bridges and culverts were divided within the watershed, and amongst subwatersheds, to assess abundance and density (i.e., bridges and culverts/mi2).

Furthermore, riparian corridor and habitat quality was qualitatively assessed at each geomorphology survey site. Vegetation type, root depth, root density, and weighted root density [i.e., (root depth/study bank height) * root density] were all measured and documented. Information gathered was later used in site-specific assessment of stream stability and potential of sediment supply through bank erosion. Lateral Connectivity Flood-prone area (i.e. active floodplain) is defined as the area adjacent to the stream channel that is under water in flow events that are 2x maximum bankfull depth at the riffle cross section (Rosgen 1996, 2006). Bankfull, in the context of this report, refers to the normal high water flow; typically relating to ~1.5 year return interval flow (i.e. channel forming flow). Field surveys are required to calibrate bankfull elevations at a riffle within a study reach, as well as determine flood-prone elevation. Due to the need for field surveys to acquire bankfull and flood-prone elevations, only geomorphology sites were subjected to lateral connectivity analysis. Light Detection and Ranging (LiDAR) imagery digital elevation models (DEMs) were used in conjunction with surveyed elevations to determine flood-prone width at sites with particularly wide flood-prone areas.

Geomorphology Field Methods Site selection worked to fulfill several objectives set forth through multi-level coordination that included the MNDNR, MPCA, Heron Lake Watershed District (HLWD), and local soil and water conservation districts (SWCDs). Sites were selected based on their capacity to identify specific stressors and investigate impairments, represent various channel and valley types found within the watershed’s geology, depict stable and unstable stream reaches, and represent channels of varying slopes (i.e. sites stratified across watershed).

At each survey site elevation data was collected to characterize the dimension, pattern, and profile of the reach. At sites with minimal or no canopy cover, a Trimble R6 Global Positioning System (GPS) grade surveying receiver was used to calculate elevations based on distance and angle from multiple satellites; data that is then corrected through a signal from a local base station. At sites where canopy cover hampered the ability of the R6 to connect with multiple satellites, a Trimble S3 total station was used to measure elevation.

In order to characterize the dimension, pattern, and profile of each reach, methods consistent with those taught by Dave Rosgen were employed. A longitudinal profile of at least 20x bankfull width in length was surveyed. Longitudinal profiles consist of thalweg (i.e., deepest portion of the channel), water surface, bankfull, and low bank height (i.e., actual “floodplain if located above bankfull) elevations throughout the reach. Those elevations incorporate water slope, bankfull slope, channel bed features (e.g., glide, riffle, run, pool), and rate of incision (i.e., low bank height/bankfull height; if greater than 1). These data are necessary to help classify the stream through use of the Rosgen (1994) stream classification system.

Surveying at each site also included a riffle cross section that allowed for the analysis of width-to-depth ratio (i.e., bankfull width/average bankfull depth), entrenchment ratio (i.e., flood-prone width/bankfull width), flood-prone width, bankfull cross-sectional area, and to calibrate bankfull elevations for the reach. Cross sections were started on the left bank (facing downstream) and elevations were surveyed to include all changes in slope across the channel. Starting and ending points for cross sections were positioned so that flood-prone width and entrenchment ratio could be calculated from the data.

Generally, a second cross section was surveyed and monumented within the study reach. Methods used were similar to the methods used for surveying riffle cross sections; however, benchmarks (i.e. rebar posts) were placed at the start and end of the cross sections to allow for annual resurveys and assessment of change over time. Furthermore, a toe pin was placed at the base of a study bank where three pins were driven horizontally into the bank face. The toe pin at each study bank serves as the starting point for the bank survey, whereas the benchmark on top of the bank serves as the ending point. The placement of these benchmarks allows for annual bank erosion assessment, where exposed portions of the horizontal banks pins are used to help validate measured versus modeled bank erosion.

The Bank Assessment for Non-point source Consequences of Sediment (BANCS) model is a combination of the Bank Erosion Hazard Index (BEHI) and Near-Bank Shear-stress (NBS) utilized to predict annual sediment inputs from a specific bank. Bank Erosion Hazard Index and NBS methods developed by Dave Rosgen (2001b) were employed to estimate bank erosion at the study bank, as well as other models: Colorado, Yellowstone, and North Carolina models. Modeled estimates varied with actual erosion rates

29 after one year of study, therefore, the Colorado model was used for all sites in order to standardize the data. Since the model was developed in Colorado, measured bank erosion at each site will help develop a regional model for southern Minnesota rivers.

Within each study reach, 100 active stream bed particles were measured (i.e. pebble counts) throughout the reach (for classification; Wolman 1954, Rosgen 2012) and 100 particles were measured through the riffle cross section (for hydraulic analysis, Rosgen 2012). The D50 particle (i.e. 50% of particles are smaller than the D50 particle) in the representative pebble count helps to classify the reach. For example, a C4 stream is a C channel type with a reach D50 particle representing gravel substrate. The D84 particle in the riffle cross section is used to calculate roughness coefficients and bankfull discharge estimation.

At the completion of surveying at each study reach, a modified Pfankuch stability rating was conducted. The Pfankuch stability rating is a qualitative assessment predicting stability of the representative channel based on upper, middle, and lower banks, and channel characteristics (Pfankuch 1975, Rosgen 1996, Rosgen 2001c). After each metric is scored a final score is calculated and an adjective rating is given (i.e., poor, fair, or good) based on the stream type for the respective study reach. Office Methods Office methods for the analysis of survey elevation data utilized the RIVERMorph Professional, version 5.2 software; developed by Stantec. RIVERMorph was used to develop cross sections, longitudinal profiles, dimensional and dimensionless ratios, and other graphs that were subsequently used to help classify the channel type of respective study reaches. Other measures of pattern, profile, and dimension such as radius of curvature, stream meander length, belt width, sinuosity, and linear wavelength were measured within RIVERMorph’s geographic information system (GIS) tool.

ArcMap was another GIS tool used to investigate study reach characteristics. LiDAR data was used in correlation with ArcMap to create valley cross sections at study reaches in order to confirm valley type. Valley type defines boundary conditions of stream channels and helps to understand lateral confinement. LiDAR was also used to assess local slope, stream power, terrain analysis, and locate historical depressional areas. ArcMap was also used in correlation with aerial photography to further assess lateral confinement and stability. Historical aerial photos from 1991 were used to develop stream line shape files that were later overlaid upon current aerial photographs (e.g., 2013 and 2015). This allowed for the assessment of lateral stability as well as changes in radius of curvature, stream meander length, belt width, sinuosity, linear wavelength, and dimensional changes such as channel widening over the last 20 years.

Bankfull elevations (i.e. ~1.5 year discharge return interval) are integral measurements taken during field surveys. Using RIVERMorph, predicted bankfull discharge was calculated through the assessment of water slopes and roughness coefficients. The USGS StreamStats tool was then used to validate bankfull calls made during field surveys (Lorenz et al. 2009). StreamStats provided drainage area and the predicted flows, with confidence estimates, for catchments upstream of study reaches. Validation of bankfull field survey calls were attained when bankfull discharge predictions aligned with ~1.5 year discharge return intervals calculated by StreamStats.

Regional curves (Appendix I) were another tool used to help validate field survey bankfull elevation calls. Regional curves graphically depict the mathematical relationships that exist amongst drainage area and

30 bankfull dimensions of width, depth, and cross-sectional area (Smith and Turrini-Smith 1999). Although, regional curves are currently under development for south-central Minnesota, the information was still helpful to validate relationships between cross sectional area and drainage area. Regional curves correlate a variety of variables, however, the most commonly used set of variables are cross sectional area and drainage area. Though other factors such as slope and channel type can affect how close a site is compared to predicted cross sectional area, often the regional curve provides a strong estimate of what the cross sectional area of a riffle cross section should be based on drainage area. Development of the curves is regionally based as many factors can affect the dimension of the channel (e.g., precipitation, runoff potential, local geology, local storage capacity, etc.).

Results Hydrology Continuous stream data collection at Jackson began in 1909 (Figure 15), but was only periodically or seasonally collected until 1936. This long-term data set (i.e. >30 years) allows for in-depth analysis of changes over time. Long-term data allow for better analysis within a watershed and can help show trends or pinpoint when relationships began to change. Additional data including daily, monthly, annual, and peak flow statistics have been computed and compiled by the USGS for the site.

Figure 15. Location of discharge data collection site on the Des Moines River.

Discharge Analysis Streamflow, or discharge, is the volume of water (i.e. volume per unit time) in an open channel. The amount of water and how quickly it moves through a watershed is related to the overall health and stability of the watershed. Alterations in basin yield, peak flows, low flows, and total annual discharge have significant implications for stream channel geometry and form, and by extension, stream habitats (Blann et al. 2009). Overall, the Des Moines River has seen increasing flow volumes over its period of record (Figure 16).

The exact cause of the increased average flow is unknown. Evidence has been presented for multiple causes, ranging from increased drainage on the landscape to climatic variation. Streamflow has been found to be sensitive to precipitation in the Midwest of the United States, including most of Southern Minnesota (Sankarasubramanian et al. 2001) have shown an increase in the intensity of rain events.

Figure 16. Hydrograph depicting discharge over the period of record for the Des Moines River watersheds at the Jackson gage with the linear discharge and 5 year average trends represented.

Numerous factors, such as stream channelization, increased watershed connectivity wetland drainage, land use changes, and reduction in upland water storage are generally believed to have affected the flow and water velocity in area water ways and contribute to channel instability. A change in contributing drainage area will result in a concordant change in discharge (Magner et al. 2004). At larger scales and event magnitudes, the effects of subsurface drainage on peak flows tend to be dominated by other variables, including the pattern, magnitude, and timing of precipitation, the design and layout of surface and subsurface drainage networks, and the capacity and conveyance of the surface drainage network (Moore & Larson, 1980; Robinson & Rycroft, 1999).

Streamflows in Minnesota reflect observed changes in precipitation with increases in mean annual precipitation, a larger number of intense rainfall events, more days with precipitation and earlier and more frequent snowmelt events (Novotny & Stefen, 2007).

Discharge data is also used to create a duration curve. Duration curves are used to examine the discharges and determine when a specific flow volume was exceeded or equaled in a given period, such as how often the flow volume exceeds high (i.e. 10th percentile) and low (i.e. 90th percentile) flow conditions for the watershed (Figure 17). You can also use the data to calculate relative frequency if the data set is large enough.

Figure 17. Flow duration curve for the Des Moines River gage.

A flow duration curve (Figure 18) is used to calculate the ability or frequency of the magnitude of flow conditions found in the watered. Information about the amount of time that are likely to equal or exceed a specified value of interest is extremely useful in engineering and analysis of the watershed and river characteristics.

The shape of the curve characterizes the ability of the basin to sustain its various flow conditions. For example, the flood to high flow conditions (i.e. >Q10) indicates the relative length and volume of the flooding regime. A steep curve (i.e. high flows for short periods) would be expected for precipitation derived floods on small or highly altered watersheds. Flooding caused by snowmelt runoff will have a flatter slope near the upper limit.

The Des Moines River curve has a relatively flat slope even at high flows, indicating longer prolonged events like snowmelt or other watershed storage may be moderating flood and high flow conditions. The relatively flat slope throughout the curve suggests the presence of surface or groundwater interaction. This may also indicate moderate water storage/ground water interaction in the watershed. Some zero flow conditions are noted in the record, so additional examination of the total number of low/zero flow (i.e. Q10) is below.

Figure 18. Flow duration curve normalized by watershed area.

Using the duration data, trends can be analyzed for various flow conditions. Plotting the flow duration normalized by watershed area indicates that the watershed delivers more runoff per square mile during the large events than the upper but is very similar to the upper during baseflow events.

The number of days at or above flood flow conditions (i.e. Q10) have increased in the Des Moines River watershed. Historically, the combined effect of agricultural surface and subsurface drainage, channelization, and land use change has been to increase streamflow peaks and shorten catchment response times (Robinson & Rycroft, 1999; Wiskow & van der Ploeg, 2003). Sixty percent of streams or rivers in the United States have experienced major changes in high or low flow (i.e. >75% change) or in the timing of these flows (i.e. >60 day shift) in the 1970s through 1990s compared to a 1930–reference period (Heinz Center 2002). Similar trends are found when plotting the recurrence interval.

Statistical techniques, through a process called frequency analysis, are used to estimate the probability of the occurrence of a given event. The recurrence interval is based on the probability that the given event will be equaled or exceeded in any given year. The 1.5 year event is often referred to as the channel forming flow, and is typically when a river is at bankfull, or the water level, or stage, at which a stream, river or lake is at the top of its banks and any further rise would result in water moving into the floodplain. The frequency of which the Des Moines River reaches the 1.5 year event return interval has increased over time (Figure 20).

The effect of subsurface drainage is generally to increase baseflows (i.e., that portion of streamflow that is derived from seepage or shallow groundwater, as opposed to surface runoff), regardless of whether peak flows are increased or decreased (Moore & Larson, 1980; Robinson, 1990; Schilling & Libra, 2003). The number of days at or below Q90 in the Des Moines River have decreased over time, reflecting an increase in baseflow conditions (Figure 21). Examination of the flow duration curves plotted seasonally also demonstrates that zero flow conditions have decreased over time.

Figure 19. Annual precipitation and the total number of days each year that flows were at or above flood level flows (i.e. Q10 – 1180 CFS).

Figure 20. Hydrograph depicting discharge over the period of record at the Jackson, MN gage with 1.5 and 100 year return intervals identified.

Figure 21. Annual precipitation and the total number of days each year that flows were at or below low flows (i.e. Q90 – 6 CFS).

Precipitation Data collected within the Des Moines River watershed indicate that the area had dry to drought conditions from 1925 until 1940 (Figures 22 & 23). Since then the yearly precipitation totals have been widely variable, with slowly trending upwards until approximately the 2000s. Even with the variability of the annual total values, the seven year average is largely within the 25th-75th percentile values, indicating a fairly stable precipitation in the region.

Figure 22. Annual precipitation trend analysis for the Des Moines River watershed.

Figure 23. Annual deviations from precipitation averages within the Des Moines River watershed.

Double Mass Curve Double mass curves (DMC) were developed for the Des Moines River data (Figure 24). Precipitation and discharge data are used to develop the DMC to examine the relationship between precipitation and discharge over time. Precipitation data was collected from the state climatology office using the gridded precipitation data set developed for the watershed.

Figure 24. Double mass curve for Des Moines River near Jackson.

The curve is based on continuous discharge data starting in 1936. Data from 1933 through 1935 were excluded due to only open water season (i.e. April through November) data being collected. The data show a change in the relationship around 1980. A double mass curve using the calculated runoff data from the WHAT model was also plotted, and shows a similar break in slope and slightly higher R2 values with a break in slope around 1983 (Figure 25). The annual monthly average discharges were also plotted based on the change in the relationship seen in the double mass curve.

Figure 25. Double mass curve plotted using the calculated runoff from the WHAT model.

Figure 26. Annual monthly discharge for period of record split at the date identified by the change in relationship within the double mass curve and then plotted together.

Web-based Hydrograph Analysis Tool The discharge data sets were then analyzed for changes for runoff and baseflow conditions by uploading the data into the Web-based Hydrograph Analysis Tool. This tool is beneficial to examine the baseflow discharge relationship over time, and can be used to look at long-term and seasonal variations. Additional analysis shows that the amount of baseflow discharge is increasing faster than runoff discharge.

Figure 27. Calculated baseflow and runoff volumes.

Figure 28. Calculated baseflow and runoff volumes.

Ground Water Usage Lastly, groundwater usage for the watershed was reviewed by compiling all reported permitted usage. All permit data was collected through the Minnesota Permitting and Reporting System (MPARS). The largest appropriation/usage category in the Des Moines River Watershed is groundwater (Figures 29 and 30).

Figure 29. Appropriation by resource type in the Des Moines River watersheds.

Figure 30. Authorized groundwater use in the Des Moines River watersheds.

In the Des Moines Valley, there are multiple aquifers that are utilized, but the use leans heavily to the Quaternary Buried Artesian Aquifers (QBAA) and the Shallow Quaternary Water Table Aquifers (QWTA). To a lesser extent, the Cretaceous Bedrock Aquifer (KRET) and the Sioux Quartzite (PMSX) are utilized, but due to available quantity in the Quartzite and quality in the Cretaceous, they are not dependable sources of water. Many bedrock aquifer systems such as the Ordovician Limestones and Cambrian Sandstones are not found in most corners of the Des Moines River watershed.

Most appropriators are located along the Des Moines River specifically between Talcott Lake and Windom. The two spots where a majority of the water allocated is in the Lake Augusta Area (2015 use: 344mg, all from the QWTA) and in the City of Windom (2015 use: 444.5mg, mostly from the QBAA). Currently, DNR authorizes a total of 3.3 billion gallons of water per year (bgy) in the Des Moines River watershed. Since the early 1990’s, total appropriation has been increasing from an average of 1.1bgy to between 1.7 and 1.8bgy in the watershed, with larger than average volumes during dryer periods. Most of this increase in appropriation was on the groundwater side. Surface water appropriations have remained relatively stable during the period of available records.

There have been two significant water use changes in the watershed over the past 20 years. In 1993, the volume of water used for irrigation reached its lowest since 1988 (i.e. when MPARS records begin) before climbing to uses over 200mgy. In 2003, the first bioenergy plant came online. Water use for bioenergy production began a steady increase to over 350mgy in 2010 before sliding off to around 275mgy in 2015. These two uses account for nearly a quarter of the water used in the Des Moines River watershed.

The primary water use is still for public water supply. This is combined in Rural Water provided to domestic users in the country, and municipal water supply systems serving communities such as Windom and Jackson. Water used in these supplies may also supply industry and livestock facilities. Use in these two categories show different trends. Increasing use by the rural water systems show an increased customer base, where community water supplies have seen a decrease of nearly 300mgy since seeing a peak use in 2003. Anticipated Trend in Appropriation Appropriations will likely continue to grow (Figure 31). One reason for this is the efforts by DNR to ensure all users that meet permitting requirements have the necessary permits and are reporting water use. This includes the livestock permitting effort currently underway. It is anticipated that this use could be in the 100mgy range in each county, depending on rural water connections.

Rural Water use will also likely continue to grow. Depending on the location, water is not easily obtained, and in some locations, it could be more economical to connect to a rural water system instead of drilling a well due to quality and quantity of water that may be available. This includes uses for domestic and livestock.

There is potential for increased irrigation, but this is dependent on the agricultural economy. Areas where this is likely to expand include the Des Moines River valley where sands and gravels make up the surficial material.

The one area that is likely to see a continued decrease in use is in the municipal water supply category (Figure 32). As residents continue to leave rural areas for larger communities, it decreases the demand

Figure 31. Total authorized water use in the Des Moines River watersheds. for water. The other factor is water use efficiency. The increase in water saving fixtures over the past decade has seen a decline in demand for water utilities.

There are six active DNR Observation Wells located within the Des Moines River Watershed. The wells are either completed in the QWTA or the QBAA systems. Three of the four wells completed in the QBAA have noted declines in the water surface elevations. Declines in each of these wells appear to have begun in the mid 1990’s, after the record floods of 1993. The lone exception is well #51005 near Slayton. This well is completed in the upper reaches of the watershed and appears to have distinctive wet and dry periods with a 6ft variation in elevation with a long term increasing trend. This well is located on the Engebretson WMA with no known appropriation in close proximity to the well.

Most of the QBAA observation wells are completed along the Des Moines River between Talcott Lake and just south of Windom. This is an area where we see the most permited appropriaiton for all uses within the watershed.

Figure 32. Volume of water appropriated by use type for the years 2010-2015.

Connectivity Longitudinal Connectivity An extensive search through GNIS, WHAF, NID, and MNDNR’s dam safety records indicate that 30 structures exist within the 3 Des Moines River watersheds (Figure 33). Twenty-three of the existing structures are potential barriers to fish passage. Sixteen of the structures identified within the watershed are lake outlet structures (Table 4) where 15 of the structures likely have minimal effect on 23 MPCA fish community assessment sites. One lake outlet structure (i.e. Desilation Project 73-2) exists as a barrier and affects one upstream MPCA fish assessment site. An additional lake structure (i.e. Shetek Lake inlet) was identified that together with the Dog Creek dam and Current Lake outlet exclude fish from any lake refuge and ultimately may affect fish communities at 5 MPCA assessment sites. Five structures identified within the watersheds are low head dams where 4 of the structures exist as barriers to fish passage but do not affect any assessment sites. Five MPCA assessment sites were located upstream of the fourth dam, however, the stop logs have been removed from the structure and allows passage. Two of the structures located within the Des Moines River watershed were stream diversion structures. Though no fish assessment sites were affected by these structures, one structure (i.e., Okabena Creek Diversion) allows flow from a subwatershed of the Missouri River basin to enter the Okabena Creek watershed (Figure 34). One culvert was identified as a barrier (i.e. devil’s run) and affects one upstream MPCA fish assessment site.

Prior to settlement the landscape held abundant numbers of lakes, wetlands, and wetland complexes. After European settlement, lakes, wetlands, and depressional areas within the Des Moines River watersheds were altered (i.e. outlet structures) or drained [i.e. public and private drainage systems (Figure 35)]. Extensive drainage (Figure 36) and outlet structures have had drastic impact on the longitudinal connectivity, natural drainage network, and quality of aquatic resources within the Des Moines River watersheds.

The MNDOT bridge and culvert dataset in ArcMap indicated that there are 275 bridges (0.18/mi2) and 266 culverts (0.17/mi2) within the watershed (Figure 37; Table 5). An intersection, however, of stream lines and road lines within ArcMap indicated that there were 1,593 (1.04/mi2) road and stream intersections which likely have some form of crossing within the Des Moines River watersheds. Bridges and culverts can have drastic impacts on rivers and streams, especially when improperly sized. Improperly sized bridges and culverts can create flood flow confinement (FFC), which can cause channel widening, alter sediment transport capacity, and sediment deposition (Zytkovicz and Murtada 2013).

The road and stream line intersection dataset was also used to indicate longitudinal connectivity of the riparian corridor. The shapefile created allows for spatial assessment of road crossings and thus breaks in longitudinal connectivity. Longitudinal connectivity of riparian corridors was also assessed locally during field surveys. One of the study sites (i.e. County Ditch 53) had less than adequate riparian corridor connectivity and minimal riparian vegetation width. An additional site (i.e. Upper Beaver Creek) also had minimal riparian vegetation width on at least one side of the stream. All other study sites had intact riparian vegetative corridors and adequate vegetative widths. Longitudinal connectivity of riparian corridors and perennial cover within subwatersheds of the Des Moines River watersheds were assessed using the watershed health assessment framework tool [WHAF (Figure 38 and 39)] and state and federal easement datasets. Finally vegetation type, root depth, root density, and weighted root density was also assessed at each study bank location (Table 6).

Figure 33. Spatial location of twenty six structures and potential fish passage barriers within the Des Moines River watersheds.

Figure 34. Outline depicting Okabena Creek’s contributing watershed as well as the Whiskey Creek watershed that contributes to the Okabena Creek watershed when the structure within the city of Worthington is opened.

Table 4. Listing of structures within the Des Moines River watersheds including the structures name, type, potential as barrier, number of IWM sites impacted, county, and UTMs.

# IWM Sites Name Structure Type Barrier Potential Impacted County UTMX UTMY Talcot Lake F Area Lake Outlet Barrier 0 Cottonwood 302485 4859723 Talcot Lake WMA Lake Outlet Barrier 29 Cottonwood 304272 4861975 Talcot Lake WMA - Frerichs Lake Outlet Barrier 0 Cottonwood 303333 4863204 Warren Lake Lake Outlet Barrier 1 Cottonwood 331113 4861696 Gruhlke Lake Outlet Barrier 0 Jackson 336900 4835107 Heron Lake Outlet Lake Outlet Not a Barrier 0 Jackson 316073 4851903 Heron Lake Fish Barrier Fish Barrier Dam Barrier 0 Jackson 317088 4859098 Tubola Detention Lake Outlet Barrier 0 Jackson 336971 4839409 Jackson City Tributary Structure Dam Barrier 0 Jackson 339158 4831139 Dog Creek Dam Potential, however stop logs removed 4 Lyon 273389 4900781 Yankton Lake Lake Outlet Barrier unless wooden screw gate deteriorated 0 Lyon 271877 4901298 Desilation Project 73-2 Lake Outlet Barrier 1 Nobles 287919 4853562 East Graham Lake Lake Outlet Potential barrier under low flows 2 Nobles 300683 4853525 Kremer-Leiner-Goedtke Pond Lake Outlet Barrier 0 Nobles 285697 4851650 Current Lake Lake Outlet Potential barrier under most flows 0 Murray 265155 4894219 Fulda Lake Lake Outlet Barrier 0 Murray 291497 4859069 Hjermstad Slough WMA Lake Outlet Barrier 0 Murray 262718 4895101 Long Lake Lake Outlet Not a barrier 4 Murray 278247 4896986 Maria Lake Lake Outlet Potential barrier under most flows 1 Murray 277116 4893942 Sarah Lake Lake Outlet Barrier 1 Murray 278773 4891682 Shetek Lake Inlet Lake Inlet Barrier 5 Murray 280663 4895122 Shetek Lake Outlet Lake Outlet Barrier 5 Murray 285281 4884278 County Ditch 4 Culvert Barrier 1 Murray - - Judicial Ditch 34 Pump Station Barrier 0 Martin 353001 4833890 Jack Creek Diversion Stream Diversion Barrier 0 Jackson 312030 4850057 Okabena Creek Diverson Stream Diversion Barrier except when gate is open 0 Nobles 289212 4834253 NA Dam Barrier 0 Murray 295173 4880132 NA Dam Barrier 0 Murray 293768 4881201 Shetek Lake Diverson Lake Diversion Barrier 0 Murray 284093 4883646 Fox Lake Diverson Stream Diversion Barrier, sealed shut 0 Martin 357692 4835700

Figure 35. Lakes, rivers, streams, and restorable wetlands of the Des Moines River watersheds.

Figure 36. Relative proportion of altered stream miles within the minor sub-watersheds of the Des Moines River watersheds.

Figure 37. Location of bridges and culverts as identified by the MNDOT shapefile, as well as road/stream intersections, throughout the Des Moines River watersheds.

Table 5. Number and density of bridges and culverts as identified by the MNDOT shapefile throughout the Des Moines River watersheds broken down by study reach drainage area.

Drainage Number Density of Number of Density of Total # of Density of Study Stream Reach Area (mi2) of Bridges Bridges (#/mi2) Culverts Culverts (#/mi2) Crossings Crossings Upper Beaver Creek 21.90 0.00 0.00 40.00 1.83 40.00 1.83 Lower Beaver Creek 177.00 20.00 0.11 158.00 0.89 178.00 1.00 Lime Creek WMA 85.47 15.00 0.18 17.00 0.20 98.00 1.15 Upper Jack Creek 31.02 1.00 0.03 47.00 1.51 48.00 1.55 Jack Creek HWY 59 46.36 4.00 0.09 61.00 1.32 68.00 1.47 Jack Creek Graham Lakes 138.34 18.00 0.13 177.00 1.28 195.00 1.41 Okabena Creek 134.29 41.00 0.31 159.00 1.84 200.00 1.49 Judicial Ditch 11 16.40 2.00 0.12 15.00 0.91 17.00 1.04 E.F. Des Moines - I-90 Channel 21.10 1.00 0.05 8.00 0.38 31.00 1.47 E.F. Des Moines - Sherburn 74.20 17.00 0.23 58.00 0.78 75.00 1.01 E.F. Des Moines - Tuttle Lake 133.00 31.00 0.23 99.00 0.74 130.00 0.98 County Ditch 53 13.10 0.00 0.00 12.00 0.92 12.00 0.92 Des Moines River - Headwaters 1248.00 217.00 0.17 223.00 0.17 1330.00 1.07 Upper Des Moines River 87.00 19.00 0.22 8.00 0.09 93.00 1.07 East Fork Des Moines River 202.00 39.00 0.19 35.00 0.17 170.00 0.84

Figure 38. Longitudinal connectivity of riparian corridors within the sub-watersheds of the Des Moines River watersheds as assessed by the Watershed Health Assessment Framework (WHAF).

Figure 39. Perennial vegetative cover within the sub-watersheds of the Des Moines River watersheds as assessed by the watershed health assessment framework (WHAF).

Table 6. Vegetation type, bank height, root depth, root density, weighted root density, and BEHI rating for each study site study bank.

Bank Root Root Weighted Root Site Dominant Vegetation Type Height Depth Density Density BEHI Rating Upper Beaver Creek Brome/Reed Canary Grasses NA NA NA NA NA Lower Beaver Creek Brome/Reed Canary Grasses 10.5 1.5 15 2.14 Very High Lime Creek WMA Brome/Reed Canary Grasses 6 2 20 6.67 High Upper Jack Creek Brome/Reed Canary Grasses 7 2 10 2.86 Very High Jack Creek -Graham Lakes Brome Grass 10.5 1.5 5 0.71 High Jack Creek - Highway 59 Brome Grass 8.5 1 10 1.18 Very High Okabena Creek Brome Grass 9 1 5 0.56 High Judicial Ditch 11 Native Grasses/Forbs and Reed Canary 8 7 40 35 Moderate East Fork Des Moines River - I-90 Channel Reed Canary Grass 4 1.5 20 7.5 Low East Fork Des Moines River - Sherburn Native/Reed Canary Grasses/Willows 8 6 30 22.5 Moderate East Fork Des Moines River - Tuttle Lake Reed Canary Grass 5 4 70 56 Low County Ditch 53 Brome Grass/Mixed Hardwoods 3 1 20 6.67 High

Lateral Connectivity Seven of the twelve fluvial geomorphology study reaches (i.e., E.F. Des Moines River - Tuttle Lake, E.F. Des Moines River – Sherburn, E.F. Des Moines River – Tuttle Lake, Jack Creek – HWY59, Lime Creek, JD11, and Okabena Creek) have sufficient lateral connectivity to access their floodplains, recharge oxbows, and provide refuge to biota during high flow events. An additional site (i.e. Upper Beaver Creek) was also categorized as slightly entrenched but has less floodplain prone width as the initial seven reaches. Three assessed reaches (i.e., Lower Beaver Creek, Jack Creek – Below Proposed Dam, and CD53) were moderately entrenched and had limited access to their floodplains. One reach (i.e. Jack Creek Graham Lakes) was completely entrenched and did not have access to its floodplain.

Flood-prone elevations are two times maximum bankfull depth at a riffle cross section and are typically comprised from the approximate 1.5 year return interval flows. The Jack Creek - Graham Lakes site was classified as a G5c channel. G channels such as this are deeply incised channels that are extremely sensitive to disturbance, have a very high streambank erosion potential and sediment supply, and have a very poor recovery potential (Rosgen 1996). Geomorphology Twelve reaches were surveyed during the 2013, 2014, or 2015 field seasons. The following map shows the location and channel classification of each site (Figure 40). This section provides an in-depth look at the characterization and stabilization of survey sites, starting in the northern portion of the Des Moines River Headwaters watershed and progressing south and downstream through the Upper Des Moines River watershed. A north to south progression through sites follows for the East Fork Des Moines River Watershed.

Figure 40. Spatial location and channel classification of geomorphology survey sites within the Des Moines River watersheds.

Des Moines River Headwaters Watershed Upper Beaver Creek Stream Information Stream Name Beaver Creek Drainage Area 22 mi2 AUID NA Stream Type E5 County Murray Valley Type U-AL-FD Section, Township, Range S10, T107N, R34W Water Slope 0.00108 ft/ft Entrenchment Ratio 5.61 -Slightly Entrenched Sinuosity 1.51 Width/Depth Ratio 8.11 Mean Bank Erosion Rate 0.0870 tons/yr/ft Bank Height Ratio 1.71 - Deeply Incised Pfankuch Stability Rating 123 - Poor

The upper Beaver Creek study site is located approximately 7 miles north of the town of Lake Wilson. The 22 square mile watershed contributing to the study site comprises the headwaters of Beaver Creek. Land use within the headwaters consists of 66.62% cultivated land, 25.22% perennial cover, 4.23% water, and 3.9% development (NLCD 2011). Similar to other subwatersheds of Beaver Creek, more than half of the stream miles within the catchment have been straightened or channelized. The MPCA has not previously assessed this stream reach’s AUID for impairments.

The channel at the study reach location was classified as an E5 stream type. The E5 stream types are channels with slight to moderately steep gradients, very low width/depth ratios, where the dominant channel materials are composed of finer materials [e.g. sand (Rosgen 1996)]. Channel slopes are typically less than 2% while the bed and banks are inherently stable (Rosgen 1996). Riparian vegetative influence is very important in E5 channels (Rosgen 1994; Appendix II) as root mats help the relatively deep channels to maintain high resistance to plan form adjustment (Rosgen 1996). The E5 stream type is an efficient channel type that maintains high sediment transport capacity and are very stable unless streambanks are disturbed and/or significant changes in sediment or streamflow occur (Rosgen 1996). Having a high sensitivity to disturbance also makes the streambank erosion potential high (Rosgen 1994). The E5 stream type, however, does have a good recovery potential if disturbance does occur (Rosgen 1994).

The bank height ratio [BHR (i.e. lowest bank height/max bankfull depth)] at the riffle cross section was measured to be 1.71. A BHR of 1.71 indicates that the channel is deeply incised, however, the entrenchment ratio (i.e. flood-prone width/bankfull width) of 5.61 indicates that the channel is only slightly entrenched. The channel at the riffle cross section can readily access its floodplain (i.e. 2x bankfull flows) and would need to incise another one foot (i.e. roughly 33% of its bankfull max depth at the riffle) before the channel would become completely entrenched (i.e. flood-prone width <1.4X bankfull width). The pattern of the channel downstream of the township road has remained relatively unchanged since 1938 and is likely due to the higher entrenchment ratio in accordance with a small contributing watershed and low slope.

Sediment supply from streambank erosion was estimated using the BANCS (i.e. BEHI matched with NBS) model. The mean streambank erosion estimates were 0.087 tons (i.e. 174 pounds) of sediment per linear foot of streambank annually (Table 7) when using the Colorado erosion rate curve (Rosgen 2001). A monumented cross section was not established at the upper Beaver Creek survey site as it was only assessed during the 2015 survey season. The riparian corridor at the survey location was comprised of primarily Reed Canary and Brome grasses. With an adequate buffer provided, the grass species were covering and rooted down through a majority of the stream’s banks. The current vegetation, root depth, and weighted root densities are a strong component to existing bank stability within upper Beaver Creek.

Table 7. Mean streambank erosion rates (tons/yr/ft) for each survey site's pool study bank.

Mean Bank Erosion Rates Site tons/yr/ft Lower Beaver Creek 0.2592 Upper Beaver Creek 0.087 County Ditch 53 0.0186 E.F. Des Moines River - Sherburn 0.1348 E.F. Des Moines River - Tuttle Lake 0.0067 E.F. Des Moines River - I-90 0.0242 Upper Jack Creek 0.1078 Jack Creek - Graham Lakes 0.1137 Jack Creek - HWY59 0.9236 Lime Creek WMA 0.236 Lower Des Moines - JD11 0.1637 Okabena Creek 0.1534

Restoration and Protection Strategies The upper Beaver Creek survey site is located within a stretch of stream that is unchannelized. Much of the upper Beaver Creek watershed, however, has previously been altered by direct excavation or straightening. The natural pattern and profile of the study location should be preserved to hold some natural characteristics within the upper Beaver Creek watershed. Furthermore, other previously channelized stretches of stream within the watershed should be restored in order to provide natural function and thus habitat within the stream.

One side of the upper Beaver Creek study reach has been protected by native grass and forb plantings (i.e. easement or contract) that will help to maintain the structural stability of the streambanks. The opposing side of the channel has been narrowly buffered in Brome and Reed Canary grasses, however, a wider buffer would provide protection further into the future as channel adjustments occur. The existing vegetation should also be maintained in order to maintain channel width. If the channel is allowed to widen or incise further, the entrenchment ratio and ability to access the streams floodplain will change.

The upper Beaver Creek watershed provides an important starting point for any progress to restoration of the Beaver Creek watershed. Often issues are systemic and are best addressed on a system wide, top-down basis. By starting restoration efforts in the headwaters of a watershed you are minimizing any impacts to other downstream efforts from upstream alterations. Furthermore, restoring the headwaters of a watershed will benefit downstream stream reaches as their headwaters will evolve around the natural processes of a river environment. Specific restoration efforts should include, but not be limited to, perennial vegetative easements, wetland restorations, and channel remeandering.

Lower Beaver Creek Stream Information Stream Name Beaver Creek Drainage Area 177 mi2 AUID 07100001-503 Stream Type C5c- County Murray Valley Type U-AL-FD Section, Township, Range S18, T107N R40W Water Slope 0.00152 ft/ft Entrenchment Ratio 1.72 - Moderately Entrenched Sinuosity 1.28 Width/Depth Ratio 25.85 Mean Bank Erosion Rate 0.2592 tons/yr/ft Bank Height Ratio 1.92 - Deeply Incised Pfankuch Stability Rating 130 - Poor

The lower Beaver Creek survey site is located approximately 1.5 miles west of Currie, south of state highway 30 (i.e. 14DM014). This portion of stream was previously determined to be impaired for turbidity and Fecal Coliform. Land use within the 177 square mile watershed that contributes to the survey location is comprised of 83.86% cultivated land, 8.62% perennial cover, 5.08% development, and 2.39% water (NLCD 2011). Much of the headwaters to the lower Beaver Creek site have been intensively channelized. All of the sub-watersheds within the study site’s catchment have had at least 40% of their stream miles channelized, straightened, or altered in a fashion other than impoundment.

The channel at the study location was classified as a C5c- stream type. This stream type is typically described as a slightly entrenched, meandering, sand-dominated, riffle/pool channel with a well- developed floodplain (Rosgen 1996). Within the Rosgen stream classification system, C5 channels have a slight gradient of less than 2%. Streams such as the one found at this study location are further classified as C5c- streams as they have a gradient of less than 0.1%. C5 stream types have a very high sensitivity to disturbance, sediment supply, and streambank erosion potential; while vegetative influence is very important and recovery potential is fair (Rosgen 1994; Appendix II). The rate of lateral erosion within C5 streams is controlled by the presence and quality of riparian vegetation (Rosgen 1996). C5 channels are very susceptible to shifts in vertical and lateral stability due to direct channel disturbances and alterations in the flow and sediment regimes of the contributing catchment (Rosgen 1996).

Typically C5 channels have riffle/pool sequences found to be 5 to 7 times the bankfull channel width (Rosgen 1996). However, the lower Beaver Creek survey site had riffle/pool sequences roughly 1 to 2

55 times the bankfull channel width. These results may be indicating instability within the various channel features (i.e., riffles and pools), but also coincide with the fact that the channel at the riffle cross section is over widened. Over widening of the channel is likely due to alterations of the catchment’s flow regime in accordance with shallow riparian vegetative root depth. Changes within the flow regime are likely attributed to multiple drivers such as changing precipitation patterns, land use, and drainage practices (e.g. channelization of headwaters). The riparian vegetative community was comprised primarily of Brome and Reed Canary grasses (Table 5). Root depth at the BEHI study bank location was only 1.5 feet in depth with a root density of only 15%. These vegetative root attributes leave banks more susceptible to near bank sheer stress and bank failure.

The bank height ratio [BHR (i.e. lowest bank height/max bankfull depth)] at the riffle cross section was measured to be 1.92. A BHR of 1.92 indicates that the channel is deeply incised and is less than four tenths of a foot vertically from becoming completely entrenched. Entrenchment ratio is measured as flood-prone width/bankfull width. When rivers incise (i.e. down cut) to the point where flood-prone elevations (i.e. 2x bankfull elevation) cannot access the river’s floodplain, the channel is classified as fully entrenched. When flows above bankfull are contained within the banks of a river, extra energy is contained within the channel and subjects banks to excess stress. The over widening of the lower Beaver Creek study location is likely a result of excess bank stress caused by entrenchment. The further the channel becomes entrenched, the more bank failure and over widening will likely be seen.

Sediment supply from streambank erosion was estimated using the BANCS (i.e. BEHI matched with NBS) model. The average streambank erosion estimates were 0.2592 tons (i.e. 518.4 pounds) of sediment per linear foot of streambank annually when using the Colorado erosion rate curve (Rosgen 2001). A monumented study bank cross section was not setup at this study reach as it was only assessed during the last field season. Excess sediment was abundant within the channel. This sediment was deposited on point bars, side bar formations, and mid channel bars as the over widened channel does not have the same capacity to transport sediment of its watershed as a stable channel would.

Restoration and Protection Strategies The stream channel at the lower Beaver Creek study location was classified as a C5 stream type. C5 stream types are very susceptible to changes in lateral and vertical stability due to direct channel disturbance and changes to flow and sediment regimes of their watershed (Rosgen 1996). Changes in flow regime and sediment regimes, land use, and drainage practices in combination with low riparian vegetative quality have been strong components to lateral and vertical instability at the study location. Further channel instability will be caused if entrenchment ratios continue to decline to complete entrenchment. Once the channel has become completely entrenched, very little can be done until the channel has over widened to the point a channel and floodplain can develop within the old channel.

Deep rooted native vegetation management should be a focus within the riparian corridor of the lower Beaver Creek study location. Though the channel is well buffered, poor root depth and density leave the streambanks highly susceptible to flows and sheer stress. Seeding of native deep rooted perennial grasses (e.g. Big Bluestem) and/or dogwood or willow stakes would greatly benefit streambank vegetation quality.

Lime Creek WMA Stream Information Stream Name Lime Creek Drainage Area 85 mi2 AUID 0710000-1-535 Stream Type E5 County Murray Valley Type U-AL-FD Section, Township, Range S18, T105N, R39W Water Slope 0.0004 ft/ft Entrenchment Ratio 75.3 - Slightly Entrenched Sinuosity 2.72 Width/Depth Ratio 9.95 Mean Bank Erosion Rate 0.2360 tons/ft/yr Bank Height Ratio 1.43 - Moderately Incised Pfankuch Stability Rating 118 - Poor

The Lime Creek WMA survey site is located roughly 0.5 miles north of the town of Lime Creek. The AUID that the Lime Creek WMA lies within was previously found to be impaired for turbidity and Fecal Coliform. Land use within the Lime Creek watershed consists of 84.47% cultivated land, 3.96% perennial cover, 6.46% developed, and 5.14% water (NLCD 2011). The relative proportion of channel alteration within the Lime Creek watershed is less than that of the Beaver Creek watershed. On average, 30% of the channel miles within the Lime Creek subwatersheds have previously been altered.

Similar to the Upper Beaver Creek survey site, the channel at the Lime Creek WMA survey site was also classified as an E5 channel. The Lime Creek survey location, however, was only moderately incised and slightly entrenched. The entrenchment ratio at the riffle cross section was 75.3 indicating that Lime Creek has a wide floodplain when compared to the channel’s bankfull width. Bank height ratio at the riffle cross section was 1.43, where an additional foot of incision would entrench the channel completely. Some previous incision may have been attributed to the channel straightening that removed roughly 1,000 feet of channel at the immediate downstream road crossing. Channel cut offs, both natural and man-made, steepen the slope of the channel. The over steepened portion of channel, as well as the channel upstream of that point, begin to incise (i.e. down cut) as a stable slope works to be reestablished by the watersheds hydraulics.

Sediment supply from streambank erosion was estimated using the BANCS (i.e. BEHI matched with NBS) model. The average streambank erosion rate estimates were 0.236 tons (i.e. pounds) of sediment per linear foot of streambank annually when using the Colorado erosion rate curve (Rosgen 2001). In order to validate streambank erosion estimates, a monumented cross section was established at an eroding bank. Bank erosion estimates at the monumented study bank was 0.1097 tons/ft/yr during both the

2013 and 2014 assessments. Erosion rates at the study bank were measured to be 1.083 feet per year on average from 2013 to 2014, and 0.145 feet per year on average from 2014 to 2015.

The monumented study bank cross section was located across a pool feature that was relatively shallow when compared to other pool features within the stream reach. Cross sections of the study bank appear to show minimal water depth and inadequate habitat during the 2013 and 2014 survey dates. The longitudinal profile of the survey reach, however, indicates that some deeper pools exist (i.e. ~2.5 feet in depth during base flow) within the channel that would provide refuge and habitat for aquatic species during base flow.

Restoration and Protection Strategies The channel at the Lime Creek WMA survey location was classified as an E5 stream type. The resilience of E channels is highly dependent upon the stream’s riparian vegetation. Due to the reliance on high quality vegetative riparian corridors, the natural vegetation at the site and within upstream buffers should be protected. Furthermore, upstream riparian areas that are over grazed should have rotational grazing plans implemented.

Future incision should be prevented in order to ensure the channel does not become fully entrenched. If the channel becomes entrenched, the stream will be susceptible to channel widening and increased rates of erosion and sediment contribution. Water storage on the landscape would help to stabilize the current hydrology of the watershed and thus prevent incision. Two large wetland complexes could be restored to provide the water storage needed to offset the rate of hydrologic change within the watershed. The Big Slough State Wildlife Management Area (SWMA) could provide additional water storage if the eastern most drained basin was restored. Another large drained wetland two miles east of Slayton could also hold a significant volume of water. Increased wetland habitat within the watershed would also help support the Northern Pike spawning that currently occurs in the Lime Creek watershed.

Floodplain connectivity throughout the Lime Creek watershed is confined in several locations. To restore natural function of the floodplain and reduce floodplain confinement several abandon road and railroad grades could be removed. A railroad grade exists just upstream from the Lime Creek WMA while several abandoned road grades and bridges exist spatially upstream to the approximate location of the Avoca waste water treatment plant. These longitudinal floodplain barriers could easily be removed at a low cost in order to restore floodplain connectivity.

Jack Creek – Upper Jack Creek Stream Information Stream Name Jack Creek Drainage Area 31 mi2 AUID 07100001-549 Stream Type E4 County Nobles Valley Type C-AL-FD Section, Township, Range S29, T104N, R40W Water Slope 0.002 ft/ft Entrenchment Ratio 2.12 - Moderately Entrenched Sinuosity 1.53 Width/Depth Ratio 8.78 Mean Bank Erosion Rate 0.1078 tons/yr/ft Bank Height Ratio 1.82 - Deeply Incised Pfankuch Stability Rating 112 - Poor

The upper Jack Creek survey site is located approximately 6.25 miles south west of Fulda, MN. This portion of stream was previously assessed for aquatic life impairments but insufficient data was acquired to make an impairment determination. The upper Jack Creek AUID was not assessed for other impairments so other listings do not exist. Land use within the 31 square mile watershed that contributes to the study reach location is comprised of 89.44% cultivated land, 3.68% perennial cover, 5.23% development, and 1.5% water (NLCD 2011). Similar to upper Beaver Creek, much of the headwater subwatersheds to the upper Jack Creek site have been intensively channelized (i.e. as great as 90%).

The attributes measured at the study site classified the stream channel as an E4 stream type. Characteristics of E4 channels are identical to E5 channels (e.g. upper Beaver Creek channel) with the exception that E4 channels have predominantly gravel sized substrates in contrast to the finer substrates of an E5 channel (Rosgen 1996). The BHR at the riffle cross section was measured to be 1.82. A BHR of 1.82 classifies the upper Jack Creek channel as deeply incised, however, the entrenchment ratio of 2.12 classifies the channel as only moderately entrenched.

The entrenchment ratio at this site is sensitive to minimal amounts of incision. Currently the flood- prone width is slightly two times the width of the bankfull channel. However, if the channel were to incise just 0.6 feet, the entrenchment ratio would become 1.39 (i.e. entrenched). One channel alteration that likely resulted in previous channel incision was the straightening of the downstream

59 portion of this stream reach for the road crossing (Figure 41). This forced the stream to make up the change in elevation over a shorter length of stream likely causing down cutting of the streambed.

Vegetation within the riparian corridor at the upper Jack Creek survey site was primarily Reed Canary and Brome grass. Banks that had a moderate to gentle angle of repose appeared to be well vegetated and relatively stable due to the vegetative cover and root density. Banks with steeper angles, and outside bends, were much more unstable. These banks showed evidence of mass failure where steep angled bank faces were primarily non-vegetated. The riparian corridor in this location had adequate width; however, incision, entrenchment ratio, and vegetative species have culminated in tall vulnerable outside bend banks.

Sediment supply from streambank erosion was estimated using the BANCS (i.e. BEHI matched with NBS) model. The mean streambank erosion estimates were 0.1078 tons (i.e. 215.6 pounds) of sediment per linear foot of streambank annually when using the Colorado erosion rate curve (Rosgen 2001). A monumented cross section was established to validate the BEHI erosion estimates. The model estimated that 0.1281 tons (i.e. 256.2 pounds) of sediment were eroding from the study bank per foot each year. Measured erosion rates indicated that the study bank (i.e. from the toe of the bank to the top) eroded 0.7979 feet from 2013 to 2014, and 0.5274 feet from 2014 to 2015.

Another area of high sediment contribution is the previously straightened portion of the channel. Upstream of the road, the channel is working to re-meander itself and is creating tall cut banks on the outside bends. Downstream of the culvert, the road authority left a very tight radius of curvature where the straightened channel intersected the old channel. This angle, in accordance with increased velocities due to flood flow confinement (FFC) caused by the culverts, has increased erosion and meander bend migration rates downstream of the road. The increase is erosion rate was high enough that public funding was used in order to place best management practices (BMPs) at the outside bends to slow the rate of land loss and instream sediment input.

Initial survey efforts were conducted in the fall of 2013 when water levels were below base flow. The intermittent channel had shallow pools of less than 1.5 feet in depth. Return visits to resurvey the study bank were also conducted in the fall but flows were at least base flow or above. Deeper pool depths were witnessed during these return visits, however, biota may struggle to find adequate refuge habitat during base flow or lower flows. Shallow pool depths in part are due to the w/d ratio of the channel. Rosgen (1996) identifies that E4 channels have an average w/d ratio of 5.86 and that w/d ratios between 8 and 10 represent less than 25% of the channels included in the data set used in the publication. Therefore, pools become shallower as a similar volume of water does not fill the channel to the same depth as when the w/d ratio is lower. Furthermore, the less constricted flow in an over widened channel has less capacity to scour sediment in order to create pools of greater depth.

Restoration and Protection Strategies Restoration and protection strategies should be implemented within the Upper Jack Creek watershed in order to increase ecosystem health and prevent the channel from becoming completely entrenched. Currently, the channel at the study location is only 0.5 feet of further incision away from becoming entrenched. Entrenchment of the channel would contain all flows up to the flood-prone elevation within the channel, thus exerting increased stress on streambanks. Entrenchment of the Jack Creek channel would lead to over widening of the channel much like the Jack Creek at Graham Lake survey site.

In order to prevent further incision, hydrologic conditions need to be stabilized and additional drainage efforts decreased. Because E channels are susceptible to changes in streamflow, these changes would need to occur in order to protect the channel from entrenchment (Rosgen 1996). Increasing water storage within the watershed would be a strong restoration and protection strategy. Seven separate large wetland features, as well as numerous small wetlands, exist within the watershed and could be targeted for restoration (Figure 42). The size of the 7 larger wetlands holds the potential to stabilize the hydrologic regime within the watershed.

Similar to many of the survey locations within the Des Moines River watersheds, riparian vegetation management or restoration should be employed. At the Upper Jack Creek survey location, more native deep rooted grass species should be managed for in the riparian corridor. Increased root depth and density will help to prevent or slow the rate of excessive bank erosion and downstream meander migration. Additionally, buffer widths should be increased where narrow, and rotational grazing plans implemented on pastured riparian lands. Riparian vegetative management is important as vegetation is an integral component to stability within E channels (Rosgen 1996).

Figure 41. Aerial photo showing the channel straightening that occurred at the upper Jack Creek survey location.

Figure 42. Restorable depressional wetlands within the Jack Creek survey site's watershed (RWI).

Jack Creek – Highway 59 Stream Information Stream Name Jack Creek Drainage Area 46 mi2 AUID 07100001-549 Stream Type E5 County Nobles Valley Type U-AL-FD Section, Township, Range S25, T104N, R40W Water Slope 0.00035 ft/ft Entrenchment Ratio 39.57 - Slightly Entrenched Sinuosity 1.97 Width/Depth Ratio 7.63 Mean Bank Erosion Rate 0.9236 tons/ft/yr Bank Height Ratio 1.2 - Slightly Incised Pfankuch Stability Rating 135 - Poor

The Jack Creek Highway 59 survey site is located approximately 6.25 miles south of the town of Fulda. The AUID that the survey site is located within has previously been assessed for aquatic life impairments, however, insufficient data was attained to make an impairment determination. The AUID was not assessed for other impairments, therefore, no listings occur for this survey location. The land use within the 46 square mile watershed that contributes to the survey location is comprised of 89.52% cultivated land, 3.42% perennial cover, 5.19% developed land, and 1.76% water (NLCD 2011).

Data collected at the Jack Creek Highway 59 survey site classified the channel as an E5 stream type. The BHR was 1.2, indicating that the channel at the riffle cross section is only slightly incised. The entrenchment ratio was 39.57, indicating that the channel is also only slightly entrenched. As indicated by the entrenchment ratio, the stream at the survey site location can readily access its floodplain. Furthermore, an entrenchment ratio of 39.57 indicates that the floodplain is wide in relation to the bankfull channel allowing for energy to dissipate as flood flows access the floodplain.

The riparian vegetative corridor was primarily Brome grass with a shallow root system (i.e. one foot). Bank heights at the survey site were roughly 8 to 8.5 feet in height and vulnerable to erosion due to the shallow root system and low weighted root density. Even with the slight incision and entrenchment ratios, bank erosion and bank sloughing was prevalent due to these riparian characteristics (Figure 43).

Sediment supply from streambanks was estimated using the BANCS model. The mean streambank erosion estimates were 0.9236 tons (i.e. 1,847.2 pounds) of sediment per linear foot of streambank annually when using the Colorado erosion rate curve (Rosgen 2001). A monumented cross section was put in place at the Jack Creek Highway 59 survey site and resurveyed each year of the study. Bank

Erosion Hazard Index assessments were conducted on the pool study bank within the monumented cross section. The BEHI estimates were 0.1555 tons (i.e. 311 pounds) a year from 2013 to 2014 and 0.2353 tons (i.e. 470.6 pounds) a year from 2014 to 2015. Resurveys of the cross section indicated that the study bank had eroded on average 2.2766 feet laterally from 2013 to 2014 and 0.2911 feet from 2014 to 2015 (Figure 44). Furthermore, the thalweg at the monumented cross section migrated laterally and incised or scoured roughly a half foot.

During the initial survey effort in 2013, the channel at the Jack Creek Highway 59 location was intermittent due to dry conditions. During the following study years, baseflows were also experienced during resurveys. During these conditions, as depicted by the longitudinal profile data, pool depths are very shallow (i.e. ~0.75 to 1.5 feet in depth). For a drainage area of 46 square miles, the pools within this reach of Jack Creek are very shallow and of poor habitat quality. Similar to the Upper Jack Creek survey location, the w/d ratio is within the upper range for an E5 channel.

Restoration and Protection Strategies The Jack Creek at highway 59 survey location is only ~6.5 river miles downstream from the upper Jack Creek survey location. Therefore, many of the restoration strategies outlined within the upper Jack Creek subsection can be applied for this channel. The only large difference is that the highway 59 location has much better floodplain accessibility and therefore priority restoration efforts may shift more towards the riparian corridor rather than the water retention because incision is less of a threat to this channel.

Figure 43. The shallow root system of the riparian corridor leaves streambanks vulnerable to erosion and bank sloughing is prevalent.

Figure 44. Study bank cross section resurveys from 2013, 2014, and 2015 show the lateral bank erosion that occurred over the course of the study.

Jack Creek – Graham Lakes Stream Information Stream Name Jack Creek Drainage Area 138 mi2 AUID 07100001-514 Stream Type G5c County Nobles Valley Type U-AL-FD Section, Township, Range S36, T104N R39W Water Slope 0.00018 ft/ft Entrenchment Ratio 1.37 - Entrenched Sinuosity 1.81 Width/Depth Ratio 10.33 Mean Bank Erosion Rate 0.1137 tons/ft/yr Bank Height Ratio 2.12 - Deeply Incised Pfankuch Stability Rating 123 - Fair

The Jack Creek Graham Lakes survey location is located approximately 3.75 miles south east of Dundee, MN. The AUID that the study site is located in has not previously been assessed to determine whether impairments exist. Land use for the watershed contributing specifically to this study location was not available, however, land use within the Jack Creek watershed up to MN highway 60 is 88.08% cultivated land, 3.12% perennial cover, 5.23% development, and 3.51% water (NLCD 2011).

Pattern, profile, and dimension measurements collected at the survey site identified the channel as a G5c stream type. G5 stream types are entrenched, moderately steep, step/pool channels that have incised deeply into sandy materials (Rosgen 1996). Width to depth ratios and sinuosities within G5 channels are relatively low and typically transport large volumes of sediment due to the inherent nature of the surrounding soils [i.e. easily detached and entrained (Rosgen 1996)]. Active, extensive, and consistent channel erosion is typically observed as G5 stream types are generally in a degradation mode due to continuous channel adjustments (Rosgen 1996). A ‘c’ is added to the label of G5 channels when slopes are measured to be less than 2% to indicate a low gradient system. Overall, G5 channels are extremely sensitive to disturbance, have a very high sediment supply and streambank erosion potential, a high dependence upon vegetative influence, and a very poor potential for recovery (Rosgen 1994).

Riffle cross section measurements identified that the BHR was 2.12 and entrenchment ratio was 1.37. These ratios indicate that the channel at the study location is both entrenched and deeply incised. Flood-prone elevations are roughly 1 foot lower than the top of the low bank height and means that flood flows (i.e. 2 times bankfull max depth at the riffle) cannot access the rivers floodplain. The confinement of flood flows within the channel is a leading factor as to why the channel is in such poor

66 shape. The Pfankuch for the survey location had a score and adjective rating of 123 and ‘fair.’ However, the scores are to be applied to the stream’s potential stream type and not its current stream type if not one in the same. Therefore, this channel is most likely supposed to be either an E5 or C5 stream type which would change the Pfankuch rating to ‘poor.’

As noted by Rosgen (1996), G5 channels are in a degradation stage. Furthermore, Rosgen has outlined different channel succession scenarios that he has observed over time (Rosgen 2014; Figure 45). Depending upon what the original Jack Creek channel type was, one of the two channel succession scenarios outlined in Figure 45 is more than likely occurring at the Jack Creek Graham Lakes survey location. As time progresses, the confinement of flood flows within the channel will lead to further streambank erosion and channel widening. The current channel has many of the same attributes of an F channel already, and could have been classified as an F already had the width to depth ratio been 12 or greater (i.e. currently 10.33). As widening continues, a floodplain will develop within the evolved channel and begin to revert back to a C channel. Depending upon future hydrologic and boundary conditions, the channel may continue to progress from a C to an E, or may remain a C channel.

The average w/d depth ratio for a G5 channel was 7.18 in Rosgen’s (1996) data set. The w/d ratio at the survey location thus indicates that the channel is over widened for a majority of G5 channels and thus leads to other poor channel qualities. A high w/d ratio is one attribute that has left pool depths shallow throughout the reach. Poor pool depth and habitat quality may be adequate for smaller tolerant species, however, it most likely does not provide enough refuge for larger individuals or more sensitive species during low flow periods. The channel throughout all of the features (i.e. glide, riffle, run, and pool) were over widened and of poor quality. Inundated with fine sediments, the channel had several side channel bars and other sediment deposits.

Sediment supply from streambanks was assessed using the BANCS model. The mean streambank erosion estimates were 0.1137 tons (i.e. 227.4 pounds) of sediment per foot of streambank each year when using the Colorado model (Rosgen 2011). A monumented pool study bank was established in order to validate streambank erosion estimates. The monumented study bank was estimated to erode 0.25 feet a year or 0.1264 tons (i.e. 252.8 pounds) a foot each year according to the BEHI model. Resurveys of the cross section indicated that the study bank eroded 0.3427 feet laterally between 2013 and 2014, and 0.2195 between 2014 and 2015.

Figure 45. Channel succession scenarios documented by Dave Rosgen (Rosgen 2014).

Restoration and Protection Strategies Restoration and protection strategies for the Jack Creek Graham Lakes channel needs to focus on a watershed-wide, headwaters-down systemic approach. The state of the channel at the Graham Lakes site is degraded and to a point where the channel is no longer connected to its floodplain. Because the channel cannot access its floodplain and thus dissipate energy within the channel, excess shear stress is applied to the streambanks. This shear stress results in excess bank erosion and sediment supply, as well as increased rates of channel widening. Due to these detrimental attributes, any instream structures placed within the channel are highly susceptible to failure.

A systemic, watershed-wide approach is the only way to effectively progress the evolution of the channel to a more stable state. Reversing the trend of the hydrologic regime within the watershed could cause the channel to begin to narrow up through deposition. This deposition of sediment on the margins of the channel would create a small floodplain within the current channel. However, this process would likely not occur at a large enough scale to create a stable stream.

It is much more likely that the channel will continue to widen and progress evolutionarily. Therefore, restoration and protection within the watershed would be working to build a more healthy watershed and stable flow regime that would create a more stable channel once the channel has progressed to its next non-G or F channel type (i.e. most likely a C or E channel). These restoration efforts include many options such as channel restoration, wetland restoration, increasing perennial cover, widening of buffers, and in-field and near channel BMPs (e.g., WASCOBs, rock tile inlets, grassed waterways, saturated buffers, drain tile management, etc.).

Okabena Creek Stream Information Stream Name Okabena Creek Drainage Area 134 mi2 AUID 07100001-602 Stream Type E5 County Jackson Valley Type U-AL-FD Section, Township, Range S7, T103N, R37W Water Slope 0.00007 ft/ft Entrenchment Ratio 23.23 - Slightly Entrenched Sinuosity 1.96 Width/Depth Ratio 10.5 Mean Bank Erosion Rate 0.1534 tons/ft/yr Bank Height Ratio 1.31 - Moderately Incised Pfankuch Stability Rating 97 - Poor

The Okabena Creek survey site is located on the northern edge of the town of Okabena just west of county road 9 (i.e. 1 mile downstream of 14DM010). Previous assessments of the AUID containing the Okabena Creek survey site have identified Turbidity and Fecal Coliform impairments for the reach. Land use within the 134 square mile contributing watershed is comprised of 88.12% cultivated land, 2.22% perennial cover, 8.18% developed land, and 1.17% water (NLCD 2011).

Measurements of the channel’s pattern, profile, and dimension classified the stream as an E5 stream type. The BHR was 1.31, depicting that the channel is moderately incised. Though the channel was moderately incised, the entrenchment ratio was only slightly entrenched at 23.23 due to an expansive, flat floodplain. A large entrenchment ratio and very low slope has helped to maintain a relatively stable reach through this portion of Okabena Creek. Downstream meander migration can be seen at most outside bends since 1938, however, most banks have only moved roughly 40 to 50 feet laterally over the last 77 years. The channel downstream of 370th avenue is very low slope and is influenced by the water elevation of South Heron Lake. Upstream of 370th avenue, outside of the lake influence, the dimensions of the channel decrease and the slope increases. Even though the w/d ratio is high when compared to other E5 channels, the influence from South Heron Lake maintained good pool depth throughout the survey reach.

A pool cross section was monumented and a study bank established within the Okabena Creek survey site. The riparian corridor within the study reach consisted primarily of Brome grass where root depth at the study bank was estimated to be roughly 1 foot. The study bank was 9 feet tall with a 5% root density and a high BEHI rating. The study bank, however, had a low NBS rating based off its near bank

69 max depth (i.e. 6.51 feet) in relation to the mean depth (i.e. 4.37 feet). Sediment supply from the banks within the study reach was estimated using the BANCS model. The mean streambank erosion rate within the reach was 0.1534 tons per foot annually. The study bank erosion rate was estimated to be 0.1083 tons per year when using the Colorado erosion rate curve (Rosgen 2001). Measured average erosion rates were 0.4727 feet between 2013 and 2014, and 0.2615 between 2014 and 2015. Measured erosion rates were lower between 2014 and 2015 due to the installation of 5 J-hook structures.

A diversion structure exists on Whiskey Ditch in the City of Worthington that diverts water into the headwaters of Okabena Creek (i.e. County Ditch 12). Typically the water within Whiskey Ditch flows into Okabena Lake, both of which lie within the Little Sioux watershed of the Missouri River Basin. The diversion structure has gates so the flow into Okabena Creek is variable depending upon the status of the gates. According to staff operating the structure, most of the time the gates are closed because CD12 is prone to flash flooding. When the gates are open, staff believe that a majority of the flow still travels to Okabena Lake rather than the creek. Okabena Creek typically has a 137 square mile contributing watershed. Whiskey Ditch’s watershed upstream of the diversion is approximately 14.43 square miles according to USGS stream stats. In order to prevent invasive carp from entering the Des Moines River Headwaters watershed from the Missouri River basin, MNDNR fisheries blocked a portion of the Whiskey Creek watershed off from the remainder (i.e. isolated 6.5 square miles). Therefore, the contributing watershed to Okabena Creek at times can be as large as 144.93 square miles when the diversion is open (Figure 46).

Figure 46. Watershed boundaries of the 137 square mile Okabena Creek watershed with the inclusion of the 14.43 square mile Whiskey Creek watershed that occasionally contributes to Okabena Creek when the gates of a structure are lowered.

Restoration and Protection Strategies Given the size of the watershed contributing to the Okabena Creek survey location, restoration and protection strategies should also be systemic watershed-wide efforts. As with many of the watersheds in southern Minnesota, many of the originally existing wetlands have been drained. The draining of the landscape has altered the hydrology of watersheds and led to evolutionary changes in rivers and streams. Many drained wetlands exist within the Okabena Creek watershed. Restoration of wetlands throughout the watershed would work to slow the rates of evolutionary change within area rivers. Many large drained wetlands exist within the eastern portion of the watershed and could provide sizable water retention areas as well as wildlife habitat. Other small farmed wetlands that fail to produce a crop during marginally wet years would also be easy targets for restoration.

Review of aerial photography shows that much of Okabena Creek and its small tributaries lack adequate buffer widths. Implementation of buffers will help to reduce sediment inputs to the stream in areas where overland runoff enters the river or stream. Many of these areas could be identified through review of LiDAR. Furthermore, many abandoned meander scrolls exist and are currently farmed through. Through review of aerial photography, many of these areas drown out crops and are areas where sediment, nutrients, and the river all intersect during precipitation and high water events. Nutrient and sediment inputs could be reduced by planting these meander scrolls in perennial cover such as the conservation reserve program (CRP).

Numerous miles of river and stream channel within the watershed have been straightened and channelized over time. Restoration of the original channels, when opportunities arise, should be pursued. By restoring the original channel, a more natural hydrologic regime and river function is restored. Furthermore, channel restoration replaces a channel devoid of habitat with a channel that has glides, riffles, runs, and pools; habitat features that aquatic life relies upon. Similarly, channelized reaches that have naturally built a small meandering channel and bench should be protected. These streams that have naturally formed a two-stage ditch should not be cleaned out. The small channel that has evolved within the bottom of these ditches holds many of the same habitat features that other natural channels have.

Finally, nutrient management on lands within the watershed should be strongly considered. Eutrophication of surface waters occur when excess nutrients enter the water. Excess nutrients spark the rapid growth of excess algae and periphyton. When excess algae dies, dissolved oxygen within the water is consumed as the organic material deteriorates. The depletion of oxygen can occur quickly and at a magnitude large enough to kill off biota living within the water. Stagnant or non-flowing waters are more susceptible to oxygen depletions, therefore, nutrient management should be considered for the health of Heron and South Heron lakes as well as the rivers and streams. Numerous nutrient management options exist for both agricultural and urban lands. Management and planning options can be found at the University of Minnesota Extension, Natural Resource Conservation Service, Soil and Water Conservation Districts, or Minnesota Pollution Control Agency stormwater management resources.

Upper Des Moines River Watershed Judicial Ditch 11 Stream Information Stream Name Judicial Ditch 11 Drainage Area 16 mi2 AUID 07100002-501 Stream Type E4 County Jackson Valley Type C-AL-FD Section, Township, Range S26, T101N, R34W Water Slope 0.00273 ft/ft Entrenchment Ratio 3.21 - Slightly Entrenched Sinuosity 1.9 Width/Depth Ratio 7.68 Mean Bank Erosion Rate 0.1637 tons/ft/yr Bank Height Ratio 1.2 - Slightly Incised Pfankuch Stability Rating 68 - Good

The Judicial Ditch (JD) 11 survey site is located approximately 1.75 miles south east of the small town of Petersburg in Jackson County. Previous assessments within this AUID found that the biological community met standards for all assessed parameters. Land use within the watershed that contributes to the survey site is consisted of 89.73% cultivated lands, 5.28% perennial cover, 4.17% development, and 0.79% water (NCLD 2011). Judicial Ditch 11 is a small, moderately confined, tributary to the West Fork Des Moines River in the Upper Des Moines River watershed.

The MPCA altered water course layer identifies that 81% of the stream miles within the watershed are considered to be altered. Stream miles within the lower 1.25 square miles of the watershed remain unaltered and encompassed by large tracts of native grasses and forbs that date back to the 1938 aerial photo. The remaining portion of the watershed, however, is channelized and travels through a series of depressional areas and wetlands. Altogether, the historical ditch maps show that 21 wetlands and 2 lakes (i.e. Goose Lake and Wolf Lake) have been drained in the upper watershed of JD11.

Pattern, profile, and dimension measurements collected at the JD 11 study site classified the stream channel as an E4 stream type. The BHR was 1.2 and the entrenchment ratio was 3.21 meaning the stream is both slightly incised and slightly entrenched. Analysis of historical aerial photography indicated that the general path of the stream channel has remained relatively unchanged over time, however, the stream has gained sinuosity through the development of small tortuous meanders within the overall path of the channel (Figure 47).

The BANCS model was used to estimate sediment supply from the study reach within JD 11. The mean streambank erosion estimates were 0.1637 tons (i.e. 360.89 pounds) per foot annually. Mean streambank erosion estimates for all banks within the study reach is heavily influenced by the meander bends that encounter the valley walls. Most banks positioned away from the valley walls and historic terraces had low to moderate BEHIs and very low to low NBS adjective ratings. Meander bends encountering the valley walls were typically taller (i.e. 8-9 feet tall versus 4-5 feet tall) and had moderate to extreme BEHI ratings. One particular bank abutting the valley wall was stratified and showed the presence of an unconsolidated sand and gravel lens [Figure 48 (i.e. highly erodible)]. A monumented cross section and pool study bank were never established at this location as it was originally only to be surveyed once in 2014.

The longitudinal profile taken at JD11 identified deep pools and steep riffles. Pool depth varied from 1.3 to 2.4 feet in depth during low flow periods. High quality E channels are distinguished by exhibiting narrow and deep channels (Rosgen 1996), the pool depth for this 16 mi2 drainage was greater than the depth found in other degraded E channels of much larger drainage area (e.g. Jack Creek at Highway 59 study location) within the Des Moines River Headwaters watershed. The deep pool depths provide for high quality habitat and ample refuge from fish eating birds, temporal temperature fluctuation, and low flow periods.

Restoration and Protection Strategies Protection within the JD11 watershed should focus on protecting any existing natural riparian vegetation as well as any perennial cover throughout the watershed and along stream corridors. Priority within those areas should focus on establishing permanent easements for any of the land that has never been plowed. The existence of historically unturned land is rather rare in southern Minnesota and efforts should be made to preserve that land.

E4 channels are inherently stable channel types unless there are significant changes in sediment or flow (Rosgen 1996). Therefore, increases in upstream drainage should carefully consider flow and volume increases. Although this site exhibited some stability, it was still slightly incised and entrenched meaning increased flows could have dramatic effects on the stream. Effects from drainage practices could result in more significant impacts than expected due to the historical drainage of nearly all water storage within the watershed and an increasing precipitation trend. Increases in flow will work to further incise and possibly widen the channel that in turn will deteriorate the quality of habitat in the stream.

Figure 47. Analysis of aerial photography indicates that the overall pattern of the stream channel has remained relatively unchanged since 1938.

Figure 48. A meander bend abutting the valley wall shows evidence of bank stratification with a sand and gravel lens separating layers of clay and top soil.

East Fork Des Moines River Watershed East Fork Des Moines River – I-90 Channel Stream Information Stream Name East Fork Des Moines River Drainage Area 21 mi2 AUID 07100003-502 Stream Type E5 County Martin Valley Type C-GL-TP Section, Township, Range S7, T102N, R33W Water Slope 0.00057 ft/ft Entrenchment Ratio 18.38 - Slightly Entrenched Sinuosity 2.26 Width/Depth Ratio 8.53 Mean Bank Erosion Rate 0.0242 tons/yr/ft Bank Height Ratio 1.24 - Slightly Incised Pfankuch Stability Rating 105 - Poor

The Interstate 90 (I-90) survey site located on the East Fork of the Des Moines River (i.e. HUC 07100003) is located roughly 1.75 miles northeast of the town of Alpha (i.e. just upstream of 14DM093). Previous assessments completed within this specific AUID identified impairments for dissolved oxygen and turbidity. The drainage area contributing to this survey location is comprised of 89.65% cultivated land, 1.83% perennial cover, 7.63% development, and 0.81% water (NLCD 2011).

Pattern, profile, and dimension measurements taken at the survey location classified this portion of the East Fork Des Moines River as an E5 channel. The BHR (i.e. 18.38) and entrenchment ratio (i.e. 1.24) indicated that the channel was both slightly incised and slightly entrenched even with the very wide flood-prone width. Flood-prone width at the riffle cross section was roughly 300 feet, while the upstream portions of the survey reach were more than double that. Excess flows easily access the floodplain at the survey location as the floodplain elevation is only 0.5 feet above bankfull elevation.

A high entrenchment ratio and intact riparian corridor are two attributes that have strongly influenced the pattern of the channel in this location. Much of the unaltered stream length near the I-90 corridor has remained basically unchanged since 1956 and relatively similar to 1938 (i.e. 1938 aerial shows the river flowing through several wetlands). The construction of I-90 in the mid 1970’s, however, lead to the straightening and shortening of the stream channel in several locations. The density of road crossings is very high within this stream reach with 6 road crossings in roughly 2.7 miles of channel.

Culvert sizing is having an influence on sediment transport throughout this portion of the river. In the 14,200 feet of stream containing the 6 crossings, 4 different stream crossing structures were employed.

Both a double box and triple box culvert were employed, as well as a double culvert and clear span bridge. The clear span bridge appears to be the only structure that is not accumulating sediment within the crossing. Sediment deposition within parts of the structures indicates that one culvert was likely the correct size to properly transport sediment within the stream (Figure 49). The second and third culverts are plugged with sediment and only see flows during bankfull and higher stages. What is not certain, however, is whether these extra culverts are adequately sized to be floodplain culverts. These crossings may require more floodplain connectivity to naturally transport water and sediment onto and throughout the floodplain.

The BANCS model (i.e. BEHI matched with NBS) was used to estimate streambank sediment contribution through this stretch of stream. The mean streambank erosion estimates was 0.0242 tons (i.e. 48.4 pounds) of sediment per linear foot annually when applying the Colorado erosion rate curve (Rosgen 2001). A study bank was established within a monumented cross section. The model estimated that the study bank would erode 0.014 tons (i.e. 28 pounds) of sediment per linear foot of streambank annually. Measured erosion rates taken from the monumented study bank identified that on average 0.5117 feet of erosion occurred between 2014 and 2015.

Restoration and Protection Strategies The channel at the East Fork Des Moines River I-90 survey location was classified as an E5 channel. E channels are highly dependent upon their riparian vegetative corridor to maintain resilience. Therefore, protection of the riparian corridor of this channel is vitally important. In many locations upstream from the study location, efforts to increase the riparian corridor width will help to protect the channel from destabilizing.

Multiple restoration efforts could be employed within the watershed of the I-90 channel study site. First and foremost, many miles of upstream river channel could be restored to its natural pattern. Restoring straightened or channelized streams provides for multitudes more habitat for aquatic and semi aquatic biota. Furthermore, restoring natural pattern profile and dimension would restore the natural hydrologic function and sediment transport capacity of the stream.

Floodplain connectivity within the watershed is also an issue that should be addressed. Adequate lateral floodplain connectivity exists at the study site location. However, throughout the watershed longitudinal connectivity is often disrupted by road crossings. Numerous road crossings and road grades exist within the watershed and several crossings (i.e. 6 within 2.7 miles of channel) exist near the study location. Road crossings stop the down valley flow of flood waters and create FFC. Furthermore, many of the road crossings are not properly sized (Figure 49). When the replacement of culverts occurs, culverts should be sized for the bankfull channel so that proper sediment transport occurs. Floodplain relief culverts should then be placed to provide for down valley floodplain flow.

Figure 49. Stream crossings that are not properly sized for the stream channel do not have adequate sediment transport capacity and therefore sediment deposits before, after, and within the crossings. Three crossings near the study location show signs of deposition where as a clear span bridge in the vicinity allows for proper stream function.

East Fork Des Moines River – Sherburn Stream Information Stream Name East Fork Des Moines River Drainage Area 74 mi2 AUID 07100003-501 Stream Type E5 County Martin Valley Type U-AL-FD Section, Township, Range S7 & S8, T101N, R32W Water Slope 0.00064 ft/ft Entrenchment Ratio 22.01 - Slightly Entrenched Sinuosity 1.27 Width/Depth Ratio 8.87 Mean Bank Erosion Rate 0.1348 tons/yr/ft Bank Height Ratio 1.16 - Slightly Incised Pfankuch Stability Rating 113 - Poor

The East Fork Des Moines River Sherburn survey site is located approximately 5.5 miles south of the town of Sherburn. Previous assessments conducted identified that the AUID that the survey site is positioned in is impaired for dissolved Oxygen and turbidity. Land use within the contributing catchment of the Sherburn site is comprised of 89.8% cultivated land, 6.4% development, 2.58% perennial cover, 1.12% wetlands (NLCD 2011). Survey data collected within the study reach classified this portion of the East Fork of the Des Moines River as an E5 channel (i.e. the same as the I-90 study site channel).

The measured BHR at this survey location was 1.16 and the entrenchment ratio was 22.01. Both of the measurements indicate that the channel is slightly incised and slightly entrenched. Heightened flows can easily access the river’s floodplain at this location as the low bank height is only approximately 0.8 feet above the bankfull elevation. Longitudinal connectivity of the floodplain at this location is affected, however, by an abandoned bridge crossing. At this crossing, the elevation of the road blocks downstream flows on the floodplain and causes flood flow confinement. The flood flow is thus bottle necked down and forced to travel through the much narrower bridge. Often, these confined flows cause excess erosion just downstream of the crossing and widen the channel.

The riparian corridor at the study location was a mix of native grasses and forbs, willow trees, and reed canary grass. Several banks were assessed for sediment contribution using the BANCS model. Streambanks assessed ranged in bank angle from 60 to 85 degrees with a root depth of 3 to 6 feet deep. Several streambanks were raw even though root densities were relatively high (i.e. 30 to 60%) when compared to other survey sites within the Des Moines River watershed. Most of the banks were rated

78 moderate or high in both BEHI and NBS adjective ratings. A monumented cross section was established at a pool in order to validate bank erosion estimates. The BANCS model estimated that the monumented study bank would erode 0.473 feet a year (i.e. 0.182 tons/ft/year). Measurements taken at the cross section, however, showed that the study bank only eroded 0.2015 feet between 2014 and 2015.

Initial survey efforts at the study location were conducted during base flows. Pool depths averaged 1.25 feet in depth with deeper pools falling short of 2 feet in depth. These pool depths are shallower than those found further upstream at the East Fork Des Moines River I-90 channel study location. Both channels were classified as E5 stream types and have width to depth ratios close to 8.75. Drainage area for the Sherburn site was 74.2 square miles compared to 21.1 square miles at the I-90 location. Pool depths throughout the longitudinal profile surveyed at this location indicate that adequate, deep, refuge pool habitat for fish may be limited and sparsely located throughout this reach of the East Fork Des Moines River. Many pools viewed within the reach were subject to filling due to bank sloughing [Figure 49 (e.g. monumented pool study bank)].

Restoration and Protection Strategies Restoration strategies for the East Fork Des Moines River at the Sherburn study location could begin with the removal of the upstream abandoned bridge and road crossing. The crossing creates flood flow confinement and backs water up within the upstream section (Figure 51 and 52). The retention of floodwaters and constriction of flood flows thus alters the natural function of the floodplain. Once the road grade is removed, options to re-meander the straightened section of river could be explored.

Protection of the riparian corridor should be prioritized. The existing native buffer should be managed for deep rooted vegetation as many bank sloughs were witnessed within the channel. Managing the native grasses and introducing more willows along the channel will help to protect the streambanks from mass wasting. Preventative measures should also be taken if the channel gets any closer to meandering into the upstream gravel pit. Vegetation management may be enough to accomplish that objective.

Maintaining the BHR and stabilizing any changes in hydrology will benefit the channel. Any changes to hydrology could potentially lead to down cutting and incision that would in turn change the BHR and entrenchment ratios. One way to prevent hydrologic change is to provide for upland water storage through wetland restoration. Restoring straightened or altered channels in the headwaters (i.e. many are 40-60% altered) will help to restore natural hydrologic function to the system as well.

Figure 50. Bank sloughing was prevalent within the study reach. Some instances bank sloughs filled all or parts of adjacent pools.

Figure 51. The abandoned bridge and road grade create flood flow confinement upstream of the study location.

Figure 52. The abandoned road grade and bridge creates flood flow confinement. The bridge constricts flow and disconnects the river's floodplain causing flooding upstream of the bridge (A) but not downstream (B).

East Fork Des Moines River – Tuttle Lake Stream Information Stream Name East Fork Des Moines River Drainage Area 133 mi2 AUID 07100003-501 Stream Type C5c- County Martin Valley Type U-LA-LD Section, Township, Range S13, T101N, R32W Water Slope 0.00003 ft/ft Entrenchment Ratio 4.58 - Slightly Entrenched Sinuosity 1.27 Width/Depth Ratio 17.27 Mean Bank Erosion Rate 0.0067 tons/yr/ft Bank Height Ratio 1 - Stable Pfankuch Stability Rating 81 - Good

Upstream of Little Tuttle Lake is another portion of the East Fork Des Moines River that was surveyed in 2014. The study site is located roughly 1.5 miles north of the town of Ceylon and 1.5 river miles upstream of Little Tuttle Lake (i.e. 14DM002). Previous assessments conducted on the AUID the study site is located within identified Dissolved Oxygen and Turbidity impairments. Land use within the 133 square mile contributing watershed is comprised of 87.36% cultivated land, 6.57% development, 2.79% perennial cover, and 3.23% water (NLCD 2011).

Pattern, profile, and dimension measurements surveyed within the study site identified the channel as a C5c- stream type. This stream type is the same as the Lower Beaver Creek study location and is of very low slope (i.e. 0.00003 ft/ft). The entrenchment ratio indicated that the channel is slightly entrenched (i.e. ratio 4.58) and the bank height ratio was 1, indicating that the bankfull elevation and low bank height are the same elevation. A bank height ratio of 1 indicates that the channel in this location has not incised over time and is likely due to the channel’s very low slope and wide floodplain.

The floodplain width at the study location has been reduced over time due to historical drainage efforts. Adjacent to the East Fork Des Moines River channel, a ditch was dug in what appears to be an effort to straighten this portion of river. Beside the ditch is the ditch spoil that in turn creates a berm that crests at an elevation above the flood-prone elevation of the river. This berm, and the valley on the west side of the channel, constricts floodplain width through this portion of the East Fork Des Moines River. Longitudinal disruption of the floodplain also exists both upstream and downstream of the survey location. Upstream, an abandoned township bridge and road grade constricts floodplain flow.

Downstream of the study location, another road grade constricts floodplain flows and creates FFC (Figure 53).

The riparian corridor at the study location was intact and continuous throughout the stream reach. The primary vegetation type was Reed Canary grass, however, some hardwoods were located throughout the floodplain. Sediment contribution from banks were estimated to be very low (i.e. 0.000 to 0.073 feet a year) when the BANCS assessment was conducted. Bank erosion hazard index adjective ratings ranged from very low to low, and NBS adjective ratings ranged from very low to moderate. The mean streambank erosion rate was 0.0067 tons per foot annually, which was the second lowest rate of all sites assessed throughout the three Des Moines River watersheds. These low erosion rates can be attributed to the combined effects of an intact riparian vegetative corridor, very low channel slope, very low BHR, and lateral floodplain connectivity.

The study reach is heavily influenced by Little Tuttle Lake which is roughly 1.5 river miles downstream. Little Tuttle Lake’s surface elevation was less than a foot lower than the water surface at the study site location as measured on LiDAR. This low gradient system thus may be impacted during periods of time when Little Tuttle Lake elevations are high. Rivers and streams that are backed up by downstream lakes or impoundments often have issues with sediment transport. During high lake levels velocities in the river or stream are reduced and sediment falls out of suspension thus causing excess deposition within the reach. The surveyed reach did show some evidence of this backwater effect, however, not to the point at which pool habitats were filled. Pool depth was more than adequate for refuge for aquatic life during low flow periods.

Restoration and Protection Strategies In order to maintain the morphological integrity of the East Fork Des Moines River at the Little Tuttle Lake survey location, the riparian vegetative corridor should be protected. The existing riparian community and BHR are attributes that lead to low streambank sediment contributions. Another attribute that helps to maintain the channel at the survey location is the rivers floodplain. Although the floodplain is relatively wide and easily accessed by the river, some connectivity attributes could be improved.

Lateral connectivity adjacent to the surveyed channel is disconnected due to an abandoned ditch. Though, the ditch elevation is below the flood-prone elevation, the excavated materials were used to create an earthen berm beside the ditch that is higher than the flood-prone elevation. By simply filling the abandoned channel in with the berm material the floodplain width would be increased and thus the entrenchment ratio as well.

Longitudinal connectivity could be improved both upstream and downstream of the survey location. Upstream of the survey location, an abandoned bridge and road grade exists. The bridge creates a point of constriction for flood flows and creates FFC. By removing the bridge and road grade, longitudinal connectivity could be restored. Similarly, downstream from the survey location the road grade again creates a barrier across the floodplain. Because this specific road is still in use, floodplain culverts could be placed across the floodplain to help pass water.

Upstream from the survey location is County Ditch 1 and 11. Within the drainage area of these two ditch systems are six drained lake basins. Restoration of some, or all, of these basins would help to stabilize the changing hydrology. Ensuring a stable hydrologic pattern will help to maintain the channels

83 dimensions and floodplain connectivity. Prioritization could focus on restoring those basins that have largely been converted to grasslands first.

Figure 53. Light Detection and Ranging (LiDAR) imagery identifying an abandoned road grade (A) that confines flood flows (i.e. FCC), a berm that confines the rivers floodplain width (B), and another road grade (C) that creates FCC.

County Ditch 53 Stream Information Stream Name County Ditch 53 Drainage Area 13 mi2 AUID 07100003-506 Stream Type G5c County Martin Valley Type C-GL-TP Section, Township, Range S36, T101N, R33W Water Slope 0.00035 ft/ft Entrenchment Ratio 2.09 - Moderately Entrenched Sinuosity 1 Width/Depth Ratio 9.93 Mean Bank Erosion Rate 0.0186 tons/yr/ft Bank Height Ratio 2.47 - Deeply Incised Pfankuch Stability Rating NA

The County Ditch 53 survey site is located approximately 3.75 miles south east of the town of Dunnell and 0.5 miles north of the Iowa border (i.e. 14DM097). The AUID that the survey site is located within had not had any other previous assessments conducted in order to determine whether impairments exist. Land use within the CD 53 watershed is 91.98% cultivated land, 2.05% perennial cover, 5.53% developed land, and 0% water (NLCD 2011). Analysis of the MPCA’s altered watercourse layer identified that 76% of stream miles within CD 53’s watershed have been channelized.

Data collected at the County Ditch 53 survey site classified the channel as a G5c channel type. The G5c stream type is typically referred to as a sand bed gully stream (Rosgen 1996), however, this particular stream exists as a G5c due to historical channelization (i.e. between 1938 and 1954) where the dimensions of the end result fit within the criteria of a G channel. Because of this, the G5c channel at this location does not hold some of the same characteristics as a naturally evolved G5 channel (i.e. active, extensive, and consistent channel erosion).

Analysis of aerial photography identified that the straightening of the original stream in this location occurred sometime between 1938 and 1954. Over time, CD 53 began to build a bankfull bench within the bottom of the trapezoidal channel. Often, historical channelization and straightening efforts resulted in over widened drainage ditches that struggle to transport the sediment of their watersheds. Because of sediment transport issues, excess sediment is deposited within the bottom of the channel. Furthermore, the channel sometimes (i.e. as in this case) is wide enough for a floodplain to begin to develop within the trapezoidal channel (Figure 54 and 55). During these instances, a bankfull bench forms and the stream begins to meander back and forth within the floodplain. The stream that evolves

85 is a natural form of the two stage ditch concept. The two stage concept is built around the idea of transporting elevated flows for drainage but maintaining a small sinuous stream during low flows in order to transport sediment.

The small sinuous channel that had developed in the bottom of CD 53 had features (i.e., glides, riffles, pools, and runs) of a natural river (Figures 55 and 56). These natural features can be seen in the longitudinal profile that was surveyed at the study location (Figure 56). After the initial survey of CD 53 in 2014, the channel was excavated to the originally designed trapezoidal shape. The reconfiguration and excavation back to the original engineer’s design removed the floodplain bench, substrate, and subsequently all of the natural stream features that had developed within the stream (Figures 56 and 57). Removal of these features removed all of the various habitat for aquatic and semi aquatic biota that had developed within and beside the small sinuous channel.

Restoration and Protection Strategies Due to the recent excavation of the naturally formed two stage ditch, little physical resource value exists to protect. Restoration strategies for this reach would include nutrient management and channel restoration. Channel restoration would provide a more natural hydrologic function to the stream as well as habitat for aquatic biota. Currently very little habitat exists within the ditched portions of this watershed.

Figure 54. County Ditch 53 riffle cross section depicting the floodplain that the channel built within the original trapezoidal channel.

Figure 55. County Ditch 53 developed a floodplain bench within the original trapezoidal channel creating a small sinuous stream with habitat features (A). After the initial survey, the channel was re-excavated to its original design which removed the floodplain bench and all habitat features (B).

Figure 56. Longitudinal profile of County Ditch 53 in 2014 and 2015. The excavation of the floodplain bench and thalweg removed all habitat features from the stream reach.

Figure 57. County Ditch 53 pool cross section before and after excavation.

Restoration and Protection Strategies A system-wide approach should be utilized to restore watershed health and system stability within the Des Moines River watersheds. Restoration efforts should focus on the sources (e.g., altered hydrology or land use practices) of water quality, watershed health, and stream stability degradation as opposed to the effects (e.g. streambank erosion). The following strategies are recommended, but are not limited to:

 Increase water storage throughout the watersheds and protect the existing water features (e.g. East Fork Des Moines watershed lakes).  Restore drained lake beds as well as shallow lakes where temporary drawdowns are feasible.  Target marginal land that frequently floods (e.g. drained wetlands) to restore to hold water on the landscape and thus meter out runoff and flows.  Target water storage projects in areas that provide additional floodplain/lateral connectivity (e.g. grade control to restore oxbow recharge).  Target water storage projects in areas that provide water quality (e.g. nutrient removal) and ecological benefits (e.g. waterfowl habitat).  Land use practices that increase organic matter in the soil will benefit future land uses and store water as every 1% increase in organic matter can hold roughly 1 inch of precipitation (U of M Extension).  Establish, maintain, and/or protect deep rooted native perennial vegetation (e.g., Big Bluestem, willows) in the riparian corridor. Many E channels exist within the Des Moines watersheds and are highly dependent upon vegetative riparian corridors (Rosgen 1996).  Establish adequate buffer widths and vegetation type for the size of river system and bank height ratio to allow for the development of bank stability.  Avoid hard armoring banks (e.g., rip rap or gabion baskets) unless infrastructure is in danger. Bank stabilization projects that employ hard armoring only deflect energy.  Re-slope and vegetate susceptible banks that are prone to sloughing and/or mass failure as an alternative to armoring.  Where channel restoration is applicable, utilize natural channel design techniques to restore the stream to its stable pattern, profile, and dimension.  Restore marginal cropland back into native prairie (e.g. Conservation Reserve Program) to increase water storage and allow for ground water infiltration.  Establishing additional native plants (e.g. native forbs) can provide additional ecological benefits (e.g. pollinators).  Road crossing projects should implement proper culvert and bridge sizing for the river or stream to allow for water and sediment movement throughout the watershed.  Improperly sized culverts and bridges can affect the river or stream channel downstream and lead to excess sediment supply and habitat degradation.  Floodplain culverts should be placed at bankfull elevations across the floodplain in order to restore longitudinal connectivity of the floodplain and reduce flood flow confinement. See Zytkovicz and Murtada (2013) for further guidance. Proper bridge sizing and floodplain culverts will help to restore travel corridors for riparian animals in many instances so that they do no need to cross busy highways; a situation dangerous to humans and animals.

 Abandoned road and railroad bridges should be removed in order to reduce channel constriction. Furthermore, the associated road and railroad grades should be leveled in order to restore floodplain connectivity.  Implement grassed waterways, conservation tillage, and cover crops to slow water down, reduce excess nutrient and sediment runoff, increase soil organic matter, and allow for greater infiltration.  Implement other Ag BMPs, as appropriate for the site, to reduce nutrients, sediment, and surface runoff into surface waters or open tile intakes.  Livestock should be excluded from rivers and streams by fencing where applicable. Supplying an additional water source will prevent livestock from trampling banks and supplying E.coli and other bacteria and pathogens to the stream (e.g., Cryptosporidium, Campylobacter, Giardia, or Fecal Coliform).  Pursuit of re-establishing natural river and stream channels, where historically channelized, should be prioritized in order to restore the natural physical and ecological function of the system.  Protection opportunities may seem sparser than areas to restore; however, options and opportunities do exist. Lands providing multiple ecosystem services, or environmental benefits, should have highest priorities for protection. Critical habitat areas, wetland/upland complexes, and natural areas not only provide quality habitat, but sequester carbon, provide a home for rare species, produce clean water, and offer many recreational opportunities.

Correlated Planning Efforts Minnesota Prairie Conservation Plan Prairie habitats once covered one third of Minnesota but presently less than 2% remain. Native prairie, other grasslands, and wetlands provide habitat for many species and are key components of functional landscapes. The Minnesota Prairie Conservation Plan is a habitat plan for native prairie grassland, and wetlands in the Prairie Region of western Minnesota with the goal to protect, restore, and enhance remaining native prairie, other grassland, and wetland habitat. In strategic locations, the Prairie Plan has identified key prairie core areas (i.e. high concentration of native prairie), corridors, and habitat complexes to create a connected landscape for wildlife and provide opportunities for sustainable grass- based agriculture such as grazing and haying.

There are six main aspects of the work:

 Implementation by multi-disciplinary Local Technical Teams in prairie Focus Areas  Secure permanent protection of high quality prairie landscapes, including native prairies, wetlands, and other habitats  Retain restored and natural grassland in these landscapes  Enhance the quality and function of prairie habitat using prescribed fire, conservation grazing, haying, invasive species control and woody plant removal  Secure the resources needed to monitor progress, assess results and implement adaptive strategies that increase success and efficiency  Integrate the efforts of the Prairie Plan Local Technical Teams to increase success and efficiency The Des Moines River watershed includes two local technical teams (see map): Prairie Coteau Team to the west and Red Rock Team to the east. These established and active Prairie Plan Local Technical Teams are available to assist and provide support to the Des Moines River watershed and its landowners to achieve wildlife value and water quality goals through targeted placement of perennial vegetation or other agricultural best management practices.

Figure 58. Minnesota Prairie Conservation Plan as it pertains to the Des Moines River watersheds.

Minnesota State Wildlife Action Plan Minnesota’s Wildlife Action Plan (2015-2025) was recently updated by the MNDNR and focuses on conservation and protection for rare, declining, or vulnerable to decline nongame wildlife species, including certain birds, mammals, reptiles, amphibians, fish, mussels and other invertebrates. The plan focuses on prioritizing efforts within connected habitat networks to assist species movement and adaption as a result of climate change. It also provides a framework to advocate for the preservation of biological diversity through the acquisition, preservation, and management of important wildlife habitats. The Wildlife Action Network (WAN) within the plan comprised of terrestrial and aquatic habitat cores and corridors to support biological diversity and ecosystem resilience with a focus on Species of Greatest Conservation Need (SGCN). The mapped WAN illustrates high, medium-high, medium, low- medium, and low scores based on SGCN population viability, SGCN richness, spatially prioritized Sites of Biodiversity Significance, Lakes of Biological Significance, and Stream Indices of Biological Integrity. Focusing conservation efforts within the mapped WAN, especially the high to medium priority zones (i.e. red, yellow and orange polygons in the following maps), will result projects and practices with multiple environmental benefits (i.e. protecting and restoring perennial vegetation for habitat enhancement and for clean water). Figures 55 through 57 indicate the WAN boundaries and scores in each of the three Des Moines HUC8 watersheds. The Des Moines River Headwaters (Figure 55) also includes the proposed conservation focus area for DNR non-game staff to place extra emphasis on. Additional information on the Minnesota Wildlife Action Plan can be found on the following webpage: http://www.dnr.state.mn.us/eco/

Figure 59. Des Moines River Headwaters watershed Wildlife Action Plan priorities.

Figure 60. Lower Des Moines River watershed Wildlife Action Plan priorities.

Figure 61. East Fork Des Moines River watershed Wildlife Action Plan priorities.

Appendix I

Appendix I. Southern Minnesota regional curve and Des Moines River watershed curve by stream type; developed through MNDNR geomorphology surveys. Bankfull cross sectional area was taken at a representative riffle cross section at each site. These are draft curves and subject to change.

Appendix II

Appendix II. Management implications for individual stream types (Rosgen 1994).

Streambank Stream Sensitivity to Recovery Sediment Erosion Vegetation Influence d Type Disturbance a Potential b Supply c Potential A1 Very Low Excellent Very Low Very Low Negligible A2 Very Low Excellent Very Low Very Low Negligible A3 Very High Very Poor Very High Very High Negligible A4 Extreme Very Poor Very High Very High Negligible A5 Extreme Very Poor Very High Very High Negligible A6 High Poor High High Negligible B1 Very Low Excellent Very Low Very Low Negligible B2 Very Low Excellent Very Low Very Low Negligible B3 Low Excellent Low Low Moderate B4 Moderate Excellent Moderate Low Moderate B5 Moderate Excellent Moderate Moderate Moderate B6 Moderate Excellent Moderate Low Moderate C1 Low Very Good Very Low Low Moderate C2 Low Very Good Low Low Moderate C3 Moderate Good Moderate Moderate Very High C4 Very High Good High Very High Very High C5 Very High Fair Very High Very High Very High C6 Very High Good High High Very High D3 Very High Poor Very High Very High Moderate D4 Very High Poor Very High Very High Moderate D5 Very High Poor Very High Very High Moderate D6 High Poor High High Moderate DA4 Moderate Good Very Low Low Very High DA5 Moderate Good Low Low Very High DA6 Moderate Good Very Low Very Low Very High E3 High Good Low Moderate Very High E4 Very High Good Moderate High Very High E5 Very High Good Moderate High Very High E6 Very High Good Low Moderate Very High F1 Low Fair Low Moderate Low F2 Low Fair Moderate Moderate Low F3 Moderate Poor Very High Very High Moderate F4 Extreme Poor Very High Very High Moderate F5 Very High Poor Very High Very High Moderate F6 Very High Fair High Very High Moderate G1 Low Good Low Low Low G2 Moderate Fair Moderate Moderate Low G3 Very High Poor Very High Very High High G4 Extreme Very Poor Very High Very High High G5 Extreme Very Poor Very High Very High High G6 Very High Poor High High High

100

Appendix III

101

Appendix III. MNDNR Stream Habitat Program Resources Sheets.

102

103

104

105

Aadland, L. P., T. M. Koel, W. G. Franzin, K. W. Stewart, and P. Nelson. 2005. Changes in fish assemblage structure of the Red River of the north. American Fisheries Society Symposium 45:293-321.

Annear, T. I., I. Chisholm, H. Beecher, A. Locke, P. Aarrestad, C. Coomer, C. Estes, J. Hunt, R. Jacobson, G. Jobsis, J. Kauffman, J. Marshall, K. Mayes, G. Smith, R. Wentworth, and C. Stalnaker. 2004. Instream flows for riverine resource stewardship. Instream Flow Council. Cheyenne, WY.

Blann, K. L., J. L. Anderson, G. R. Sands, and B. Vondracek. 2009. Effects of agricultural drainage on aquatic ecosystems: a review. Critical Reviews in Environmental Science and Technology 39(1):909-1001.

BWSR (Board of Water and Soil Resources). 1999. Wetlands in Minnesota. BWSR, St. Paul, MN.

CropScape. 2015. United States Department of Agriculture: National Agricultural Statistics Service.

Heinz Center. 2002. The state of the nation’s ecosystems 2002: measuring the lands, waters, and living resources of the United States. H. John Heinz III Center for Science, Economics, and the Environment. Cambridge University Press, Cambridge, United Kingdom.

Haitjema, H. M. 1995. On the idealized time distribution in idealized groundwatersheds. Journal of Hydrology 172:127-146.

Ickes, B. S., J. Vallazza, J. Kalas, and B. Knights. 2005. River floodplain connectivity and lateral fish passage: a literature review. U.S. Geological Survey, Upper Midwest Environmental Sciences Center, La Crosse, WI June 2005. 25pp.

Kuehner, K. J. 2004. An historical perspective on hydrologic changes in Seven Mile Creek watershed. ASAE paper no 701P0904. Proceedings of the 2004 Self Sustaining Solutions for Streams, Wetlands, Watersheds Conference, St. Paul, MN.

Lane, E. W. 1955. The importance of fluvial morphology in hydraulic engineering. Proceedings of the American Society of Civil Engineers 81:811-817.

Leach, J., and J. A. Magner. 1992. Wetland drainage impacts within the Minnesota River Basin. Currents 3(2):3-10.

Lenhart, C. F., K. Brooks, D. Henely, and J. A. Magner. 2007. Contributions of organic matter and suspended sediment to turbidity: muddying TMDLs in the Blue Earth River Basin. Hydrological Science and Technology 23(1-4):127-136.

Lenhart, C. F. 2008. The influence of watershed hydrology and stream geomorphology on turbidity, sediment, and nutrients in tributaries of the Blue Earth River, Minnesota, USA. PhD Thesis, University of Minnesota, Twin Cities, St. Paul, MN.

Leopold, L. B., G. M. Woman, and J. P. Miller 1964. Fluvial processes in geomorphology. San Francisco, CA. W. H Freeman.

Lorenz, D. L., C. A. Sanocki, and M. J. Kocian. 2009. Techniques for estimating the magnitude and

106

frequency of peak flows on small streams in Minnesota based on data through water year 2005. U. S. Geological Survey Scientific Investigations Report: 2009-5250. 54 pp.

Lusardi, B. A. 1997. Minnesota at a glance: quaternary glacial geology. Minnesota Geological Survey. University of Minnesota, St. Paul, MN. 4 pp.

Magner, J. A., G. A. Payne, and L. J. Steffen. 2004 Drainage effects on stream nitrate-n and hydrology in south-central Minnesota (USA). Environmental Monitoring and Assessment 91:183-198.

McKeown, B. A. 1984. Fish migration. Croom Helm, London. 225 pp.

Moore, I. D., C. L. Larson, and D. C. Slack. 1980. Predicting infiltration and micro-relief surface storage for cultivated soils. Water Resources Research Center, University of Minnesota, Graduate School, St. Paul, MN.

MPCA (Minnesota Pollution Control Agency). 2012. Pomme de Terre River watershed biotic stressor identification. MPCA, Report wq-iw7-36n, St. Paul, MN.

MNDNR (Minnesota Department of Natural Resources). 2005. Field guide to the native plant communities of Minnesota: the prairie parkland and tallgrass aspen parklands provinces. Ecological Land Classification Program, Minnesota Biological Survey, and Natural Heritage and Nongame Research Program, MNDNR, St. Paul, MN.

MNDNR (Minnesota Department of Natural Resources). 2008. Calcareous fen fact sheet. MNDNR, St. Paul, MN.

MNDNR (Minnesota Department of Natural Resources). 2009. Conservation status ranks for native plant community types and subtypes. MNDNR, St. Paul, MN.

MNDNR (Minnesota Department of Natural Resources). 2015. State wildlife action plan. MNDNR, St. Paul, MN.

MNDNR (Minnesota Department of Natural Resources). 2016. Coteau moraines subsection. MNDNR, St. Paul, MN.

MNDNR (Minnesota Department of Natural Resources). 2016a. Minnesota River prairie subsection. MNDNR, St. Paul, MN.

MNDNR (Minnesota Department of Natural Resources). 2016b. Minnesota native plant communities. MNDNR, St. Paul, MN.

MNDNR (Minnesota Department of Natural Resources). 2016c. Natural heritage information system database. MNDNR, St. Paul, MN. Heritage review conducted January 2016, MNDNR, Region IV, New Ulm, MN.

MNDNR (Minnesota Department of Natural Resources). 2016d. Minnesota prairie conservation plan. MNDNR, St. Paul, MN.

107

NASS (National Agricultural Statistics Service). 2016. United States Department of Agriculture. www.nass.usda.gov/.

NLCD (National Land Cover Database) 2011. Homer, C. G., J. A. Dewitz, L. Yang, S. Jin, P. Danielson, G. Xian, J. Coulston, N. D. Herold, J. D. Wickham, and K. Megown. 2015. Completion of the 2011 national land cover database for the conterminous United States – representing a decade of land cover change information. Photogrammetric Engineering and Remote Sensing. 81(5):345- 354.

Novotny, E. V., and H. G. Stefan. 2007. Stream flow in Minnesota: indicators of climate change. Journal of Hydrology 334.3:319-333.

Robinson, J. M., and K. G. Hubbard. 1990. Soil water assessment model for several crops in the high plains. Agronomy Journal 82:1141-1148.

Robinson, M., and D. W. Rycroft. 1999. The impact of drainage on streamflow. in Skaggs, R. W., and J. V. Schilfgaardge, editors. Agricultural Drainage. Agronomy Monographs 38:767-800.

Rosgen, D. L. 1994. A classification of natural rivers. Elsevier, Catena. 22:169-199

Rosgen, D. L. 1996. Applied river morphology. Pagosa Springs, CO. Wildland Hydrology Books.

Rosgen, D. L. 1997. A geomorphical approach to restoration of incised rivers. Proceedings of the Conference of Management of Landscapes Disturbed by Channel Incision. 11 pp.

Rosgen, D. L. 2001a. A hierarchical river stability/watershed-based sediment assessment methodology. in Proceedings of the Seventh Federal Interagency Sedimentation Conference. Vol. 1. pp. II-97-II- 106. Reno, NV. Subcommittee on Sedimentation.

Rosgen, D. L. 2001c. A stream channel stability assessment methodology. in Proceedings of the Seventh Federal Interagency Sedimentation Conference. Vol. 1 pp. II (9) – 11 (15). Reno, NV: Subcommittee on Sedimentation.

Rosgen, D. L. 2012. Applied river morphology: training manual. Wildland Hydrology, Fort Collins, CO. 268 pp.

Sankarasubramanian, A., R. M. Vogel, and J. F. Limbrunner. 2001. Climate elasticity of streamflow in the United States. Water Resources Research 37(6):1771-1781.

Santucci Jr., V. J., S. R. Gephard, and S. M. Pescitelli. 2005. Effects of multiple low-head dams on fish, macroinvertebrates, habitat, and water quality in the Fox River, Illinois. North American Journal of Fisheries Management 25:975-992.

Schottler, S. P., J Ulrich, P. Belmont, R. Moore, J. Wesley Lauer, D. R. Engstrom, and J. E. Almendinger. Twentieth century agricultural drainage creates more erosive rivers. Hydrological Processes 28(4):1951-1961.

Slawski, T. M., F. M. Veraldi, S. M. Pescitelli, and M. J. Pauers. 2008. Effects of tributary spatial position,

108

urbanization, and multiple low-head dams on warmwater fish community structure in a Midwestern stream. North American Journal of Fisheries Management 28:1020-1035.

Smith, D. P., and L. Turrini-Smith. 1999. Western Tennessee fluvial geomorphic regional curves. United States Environmental Protection Agency, Region IV, Water Management Division. Atlanta, GA.

Spaling, H., and B. Smit. 1995. Conceptual model of cumulative environmental effects of agricultural land drainage. Agriculture, Ecosystems and Environment 53:299-308.

Ter Harr, M. J. and E. E. Herricks. 1989. Management and development of aquatic habitat in agricultural drainage systems. University of Illinois, Water Resource Center, Champaign, IL. 145 pp.

Tockner, K., D. Pennetzdorfer, N. Reiner, F. Scheimer, and J. V. Ward. 1999. Hydrological connectivity and the exchange of organic matter and nutrients in a dynamic river floodplain system (Danube, Austria). Freshwater Biology 41(3):521-535.

U of M Extension (University of Minnesota Extension). 2016. Organic matter management. University of Minnesota, St. Paul, MN.

USDA (United States Department of Agriculture). 2011. Cropland data layer. United States Department of Agriculture. National Agriculture Statistics Service.

WHAF (Watershed Health Assessment Framework). 2014. Watershed health assessment framework – interactive web based information system. Minnesota Department of Natural Resources.

Welcomme, R. L. 1979. Fishery management in large rivers. Food and Agriculture Organization of the United Nations, Rome, Italy. Food and Agriculture Organization Fisheries Technical Paper 194. 60 pp.

Winston, M. R., C. M. Taylor, and J. Pigg. 1991. Upstream extirpation of four minnow species due to damming of a prairie stream. Transactions of the American Fisheries Society 120:98-105.

Wiskow, E., and R. R. van der Ploeg. 2003. Calculation of drain spacings for optimal rainstorm flood control. Journal of Hydrology 272.2:163-174.

Wolman, M. G. 1954. A method of sampling coarse riverbed material. Transactions of the American geophysical Union 35:951-956.

Zytkovicz, K., and S. Murtada. 2013. Reducing localized impacts to river systems through proper geomorphic sizing of on-channel and floodplain openings at road/river intersections. Minnesota Department of Natural Resources. 56 pp.

109