U.S. Army Corps of Engineers Detroit District. Boardman River SIAM Modeling Base-case Scenario

Transcription

1 U.S. Army Corps of Engineers Detroit District Boardman River SIAM Modeling Base-case Scenario May 2009

2

3 Boardman River SIAM Modeling Base-case Scenario Report Prepared for U.S. Army Corps of Engineers Detroit District Prepared by NTH Consultants, Ltd. and Baird W.F. Baird & Associates Ltd. For further information please contact Jim Selegean at (313) Cover photograph courtesy of U.S. Army Corps of Engineers Detroit District: Brown Bridge Dam on the Boardman River near Mayfield, MI, 2008

4

5 TABLE OF CONTENTS Figures... iii Tables... v Executive Summary... vi 1.0 Introduction Methods Literature Review Field Reconnaissance Boardman River Sediment Data Bank Erosion Hydraulic and Sediment Impact Modeling Hydraulic Modeling Sediment Impact Assessment Model (SIAM) Background Analytical Methods Sediment Data and Analyses Sediment Loading from Major Tributaries Sediment Loading from Bank Erosion Sediment Trapping Efficiency of Dams SIAM Implementation Model Checking Results Discussion SIAM Results for Sediment Reaches SIAM Results with Bank Erosion Conclusions and Recommendations Cited Literature Relevant Literature Appendix A: Hydraulic Model i

6 ACRONYMS FEMA GIS HEC-RAS MDEQ MiGDL SIAM USACE USDA USGS Federal Emergency Management Agency Geographic Information Systems Hydraulic Engineering Centers River Analysis System Michigan Department of Environmental Quality Michigan Geographic Data Library Sediment Impact Assessment Model United States Army Corps of Engineers United States Department of Agriculture United States Geological Survey ii

7 FIGURES Figure 1: Boardman River Watershed and Project Dams... 2 Figure 2: Major sediment reaches for sediment transport modeling of the Boardman River. Small reaches were also established to represent the deltas for Brown Bridge and Boardman Ponds, B and F, respectively Figure 3: Representative particle size distributions for each sediment reach... 5 Figure 4: Example of eroding bank downstream of Beitner Road. Total eroding bank approximately 50 meters long and 4 meters high Figure 5: Regression of 2 to 500-year discharge values on watershed area for the Boardman River hydraulic model. Watershed area explains more than 95% of variability observed in discharge estimates Figure 6: Sediment loading at Ranch Rudolph based on data collected by the USGS Figure 7: Flow-frequency analysis for USGS gauge upstream of Brown Bridge Pond Figure 8: Precipitation and Water Yield for the Boardman River Figure 9: Sediment loading data from USGS data collection at gauging station upstream of Brown Bridge Pond at Ranch Rudolph (Regression equations for sediment loading rates based on data collected from the USGS at gauge number (USGS, 2008)) Figure 10: Average annual sediment yield based on observed daily flow at Ranch Rudolph gauge Figure 11: Observed bed load grain size data (Adapted from USGS, 2008) Figure 12: Observed total sediment grain size data (Adapted from USGS, 2008) Figure 13: Soil map of bank erosion example. Soil type is GrA (Gladwin-Richter gravelly sandy loams, 0 to 2 percent slopes). (Image Source: USDA Web Soil Survey, websoilsurvey.nrcs.usda.gov) Figure 14: Bank material estimation based on percent clay, silt, and sand values available from the web soil survey (NRCS Soil Survey Staff, USDA, 2009). Values of soil sizes between particle size classification thresholds were estimated using trend lines between known size classes Figure 15: 1916 imagery of Keystone (Boardman) Pond over current (2006) aerial photo Figure 16: Boardman Pond storage curves. Current volume data are based on available bottom surveys and topography data. Volumes for the pond in 1916 were calculated by estimating surface areas from the 1916 drawing by E.P. Waterman Co. Surveyors. 22 Figure 17: Estimated magnitudes of deposition/erosion for each reach as per SIAM modeling. Series name indicates maximum washload diameter used for all reaches with the exception of reaches B, F, G, and C Figure 18: Results of scenarios examining impact of bank erosion on sediment volume in Boardman Pond. Bank erosion was estimated based on the dimensions of a large eroding bank upstream of the impoundment Figure 19: Longitudinal profile of the Boardman River HEC-RAS model showing the 2-year event iii

8 Figure 20: Upstream cross-section of Ranch Rudolph to Brown Bridge delta sediment reach Figure 21: Upstream cross-section of Brown Bridge delta to Brown Bridge Dam sediment reach Figure 22: Upstream cross-section of Brown Bridge Dam to East Creek sediment reach Figure 23: Upstream cross-section of East Creek to Jaxon Creek sediment reach Figure 24: Upstream cross-section of Jaxon Creek to Keystone (Boardman) Pond sediment reach Figure 25: Upstream cross-section of Keystone (Boardman) Pond to Boardman Dam sediment reach Figure 26: Upstream cross-section of Boardman Dam to Sabin Dam sediment reach Figure 27: Upstream cross-section of Sabin Dam to Boardman Lake sediment reach iv

9 TABLES Table 1: Watershed Areas of Major Tributaries Downstream of Brown Bridge Dam... 1 Table 2: Watershed Areas and River Miles of Boardman River Dams... 1 Table 3: Retention time and theoretical minimum particle size for sediment retention by Boardman River Dams under three flow scenarios. Low Mean and High Mean represent average bounds from USGS flow data (scaled from Brown Bridge) and Peak is the largest event of record Table 4: Boardman River sediment reach descriptions Table 5: Estimates of volume loss for Boardman and Brown Bridge Ponds Table 6: SIAM modeling results of deposition/incision by sediment reach Table 7: Results examining impact of bank erosion on sediment volumes in Boardman Pond; differences in values were only equivalent to differences in depth of hundredths of an inch per year Table 8: Estimated annual sediment loading rates in tons per year based on total failure time of 2 years and 5 years v

10

11 EXECUTIVE SUMMARY Four dams currently impound water on the main stem of the Boardman River: Brown Bridge Dam, Boardman Dam, Sabin Dam, and Union Street Dam. Power generation at Brown Bridge, Boardman and Sabin dams has ceased due to a lack of economic viability. Studies have commenced to examine the fate of these structures with many alternatives proposed, including complete removal. In the first phase of assessing potential response of the river to each of these alternatives, a numerical model representing the existing river and dams was developed and applied to simulate the sediment budget characteristics along the river. The model development process included field observation, field data collection, and review and analyses of historic and current data. Extensive field data was collected by the USACE during 2007 and 2008 to determine the sediment characteristics of the Boardman River throughout the study area from the USGS gauge at Ranch Rudolph (# , approximately 4 miles upstream of Brown Bridge Pond) to Boardman Lake (Figure 1). Additional field reconnaissance was conducted in October 2008 to observe fluvial processes, verify sediment sources, and assess river stability along the study reach. A hydraulic model of the Boardman River, developed by the USACE, was adjusted to model an annualized flow time series using observed flow data from the USGS stream gauge located at the upstream end of the model domain. This updated model was used in conjunction with sedimentdischarge relationship data collected by the USGS and bed material data collected by the USACE to develop a SIAM model for the entire study area. Based on available data and field observations, the model was subdivided into distinct sediment reaches that had similar hydraulic and sediment characteristics. The results from the SIAM modeling showed minor sediment supply limitations of less than 2 inches per year along most reaches of the Boardman River; because many of these were known to be stable, supply limitations of this small magnitude are considered within the range of stable conditions. While supply limitations just below Brown Bridge dam and in a steep section of the Boardman River were relatively larger, these sections also appeared to be stable. vi

12 Supply exceeded transport capacity for Brown Bridge and Boardman Ponds which were predicted to have annual sediment excesses ranging from 0.3 to 1.2 in/yr and 1.5 to 2.9 in/yr, respectively. These results agree with observed sedimentation values from each reservoir that showed 0.5 and 1.6 in/yr of deposition in Brown Bridge Pond and Boardman Pond, respectively. While bank erosion was only rarely observed, it was included as a component in the SIAM sediment budget analysis. Observed bank erosion data were not available consequently, a couple of hypothetical scenarios were developed using estimated bank volumes from one large eroding bank. Sediment from erosion of this bank over two time periods showed very little impact (0.3% increase) in the total sediment volume of Boardman Pond. Although the results show that the river is relatively stable with the exception of some deposition in Brown Bridge and Boardman Ponds, and a significant sediment shortage below Brown Bridge Dam, alterations to the structures may change the sediment dynamics given the amount of sediment that has accumulated since the early 1900s. The deltas in these ponds would be likely exposed to increased shear and mobilize if the dams were removed. Determining the fate of these mobilized sediments was beyond the study scope of this report. Given that SIAM is a sediment budget model, rather than a transport model, these results indicate only relative tendencies of surplus or deficit of material in each sediment reach, and therefore should only be used as a screening tool to determine which potential future dam modification/removal scenarios would be the best alternatives until a more detailed sediment transport model can be developed. This model can be used to make comparisons of future Boardman River scenarios to identify likely changes in the sediment budgets along the river and be used to select the best alternatives for more detailed modeling analysis. Sediment reaches of concern, especially those that would be impacted by modifications to the dams, should be subject to greater scrutiny in more detailed sediment transport modeling and morphodynamic analysis to determine potential changes on the fluvial regime. vii

13 1.0 INTRODUCTION The Boardman River is located in the northwestern portion of Michigan s Lower Peninsula and originates in central Kalkaska County. The river flows southwest into Grand Traverse County then turns north and ultimately discharges to West Grand Traverse Bay, Lake Michigan in Traverse City, Michigan. The watershed area is approximately 291 square miles and includes 179 lineal stream miles, a couple of major tributaries, 12 natural lakes, and four dams (Table 1, Table 2 and Figure 1). The study area is a ~24 mile section of the Boardman River s main stem from the USGS Gauge located upstream of Brown Bridge Pond to Boardman Lake in Traverse City, MI (Figure 2). The river a designated State of Michigan Natural River and includes 36 lineal miles of Blue Ribbon Trout Stream (Michigan Department of Natural Resources). From upstream to downstream, the four dams that impound the water on the main stem of the Boardman River are: Brown Bridge Dam, Boardman Dam, Sabin Dam, and Union Street Dam (Table 2). Power generation at Brown Bridge, Boardman, and Sabin was ceased by Traverse City Light and Power (TCLP) due to a lack of economic viability. The dams had previously been leased to TCLP for hydropower generation by Traverse City (owner of Brown Bridge and Union Street dams) and Grand Traverse County (owner of Boardman and Sabin dams). Table 1: Watershed Areas of Major Tributaries Downstream of Brown Bridge Dam Watershed Area Dam (sq. mi) Jaxon Creek 11.7 Swainston Creek (Downstream of Mayfield Dam) 2.3 East Creek 31.6 Table 2: Watershed Areas and River Miles of Boardman River Dams Dam River Mile (from Lake Michigan) Watershed Area (sq. mi) Union Street (Boardman Lake) Sabin Boardman (Keystone) Brown Bridge

14 Boardman River SIAM Modeling Base Case Scenario Report Figure 1: Boardman River Watershed and Project Dams 2

15 Ranch Rudolph Figure 2: Major sediment reaches for sediment transport modeling of the Boardman River. Small reaches were also established to represent the deltas for Brown Bridge and Boardman Ponds, B and F, respectively. 3

16 2.0 METHODS 2.1 Literature Review An extensive literature review was conducted to gather information pertinent to the dams and sedimentation issues on the Boardman River. Sources of information included geospatial data, scientific literature and publications, agency reports, local media sources, historic imagery, and information from past and current studies conducted on the Boardman River. Cited and relevant literature are listed in sections 6 and 7. A detailed literature review was conducted on the existing Boardman River Feasibility Study reports drafted by Environmental Consulting and Technology, Inc. (ECT) in late 2007 and early These reports focused on the existing conditions of existing habitats, wetlands, wildlife, sediments, structures, economics, and fisheries in the context of the possible removal and/or modification of Brown Bridge, Boardman, and Sabin dams. 2.2 Field Reconnaissance Field reconnaissance was conducted along the Boardman River project study area during the summer and autumn of Fieldwork consisted of flow and sediment transport modeling, river channel sediment sampling and analysis, and the mapping of significant bank erosion sites. Flow and sediment sampling were conducted by the USGS; the analyses of these data are described in Section Channel sediment sampling and bank erosion site mapping were conducted by the USACE and are described below Boardman River Sediment Data Reservoir and channel sediment data were sampled at numerous locations along the Boardman River from USGS Gauge # at Ranch Rudolph to downstream of Sabin Dam in 2008 (Figure 2). Particle size distributions were developed for all of the sites and compared to identify reaches of the river having similar channel bed material sediments. These results were used to develop the sediment reaches for sediment transport modeling described in Section 2.3 (shown in Figure 2). The representative particle size distributions for each reach are shown in Figure 3. 4

17 A B_Delta B C D E F G H % Finer Grain Size (mm) Figure 3: Representative particle size distributions for each sediment reach Bank Erosion Bank erosion was not frequent along the Boardman River; riverbanks were largely stable along the entire project length. A few localized sites of active bank erosion and over-steepening were identified where the river meandered against terraces. A typical site was located downstream of Beitner Road and upstream of Keystone Pond (Figure 4). As no historic data on bank erosion rates or stability were available, the potential magnitude of bank erosion was estimated for this site based upon the size of the eroding bank (length, height, and depth) and soil characteristics. This is described in Section

18 Figure 4: Example of eroding bank downstream of Beitner Road. Total eroding bank approximately 50 meters long and 4 meters high. 2.3 Hydraulic and Sediment Impact Modeling In order to estimate the total volume of sediment going through the Boardman River System under the existing conditions, a combination of methods was employed including hydraulic modeling and impoundment volume analysis described below. All pertinent model files are included with the project DVD, while only the most pertinent tables are included in Appendix A Hydraulic Modeling The USACE developed a complete hydraulic model for the Boardman River using a hydraulic model called HEC-RAS (USACE, 2002, ver ). This model was developed using a combination of existing hydraulic and flood models for the river and additional topography and structure data (USACE, 2008). HEC-RAS is a one-dimensional hydraulic model developed by the USACE for natural and constructed channels. The model can compute both steady and unsteady flow, sediment transport capacity, and water quality. Several functions are built into the model to calculate sediment transport. 6

19 The Boardman River HEC-RAS model was developed under the USACE Section 506 Great Lakes Fishery and Ecosystem Restoration program. The hydraulic model includes approximately 25 miles of the river from its mouth at Grand Traverse Bay to almost 6,000 feet upstream of Brown Bridge Road (upstream of Brown Bridge dam). This model was developed using the existing conditions, including Boardman Pond being at full-pool. Data for model development came from a variety of sources: FEMA Flood Insurance Study within Traverse City, survey data collected by Gordie Fraser and Gosling Czubak, GIS contour data from Grand Traverse County, USACE bathymetry data of the impoundments, design drawings of bridges and structures, and the MDEQ Boardman River hydraulic model. Discharge data for the hydraulic model were gathered from a variety of sources. The model included estimated annual flood events with recurrence intervals of 2, 5, 10, 25, 50, 100, 200, and 500-years that were obtained from the MDEQ Hydrologic Studies unit for a number of locations along the Boardman River. These events were estimated using regional flood and precipitation relationships to predict flow as a function of a number of watershed and landscape parameters including, but not limited to: watershed area, watershed slope, contributing area, percentage watershed as lakes, percentage watershed as swamps, and soils and geologic properties (Farrand and Bell, 1984; Holtschlag and Croskey, 1984). The discharge values increased with watershed area and recurrence interval as shown in Figure 5. Results from simple linear regression indicated watershed area explained more than 95% of the observed variability in peak discharge. Due to the limited flow record for observed data at Ranch Rudolph, only the 2 and 5-year recurrence interval design flows were included in the model. The reasoning for exclusion of larger events is explained in Section

20 Discharge (cfs) Watershed Area (sq. miles) Figure 5: Regression of 2 to 500-year discharge values on watershed area for the Boardman River hydraulic model. Watershed area explains more than 95% of variability observed in discharge estimates. The model includes four dam structures (Union Street dam, Sabin dam, Boardman dam, and Brown Bridge dam), bridges, and culverts. Data for the Brown Bridge dam was obtained from the 2005 Supporting Technical Information Document prepared by Gannett Fleming, Inc., and the Union Street dam structural information was taken from construction design drawings. Data for both Sabin and Boardman dams were obtained from the MDEQ hydraulic model Sediment Impact Assessment Model (SIAM) Background SIAM is a sediment budget tool, now available through the Version 4.0 Beta HEC-RAS interface, which estimates sediment budgets along user-specified sections of rivers by comparing annualized sediment reach transport capacities to supplies and indicates reaches of overall sediment surplus or deficit. The intended use of SIAM is as a screening level tool to compute relative responses in sediment budgets to various alternative river conditions (which should then be modeled in more detail). The algorithms in SIAM evaluate sediment impact caused by local changes on the system 8

21 from a sediment continuity perspective. The results map potential imbalances and instabilities in a channel network and provide the first step in designing or refining remediation. SIAM performs sediment transport capacity computations to determine potential imbalances and instabilities in a channel network. SIAM does not predict intermediate or final morphological patterns and does not update cross-sections, but rather indicates trends of locations in the system for potential sediment surpluses or deficits. This is because SIAM is not a sediment routing model. SIAM is only intended as a screening tool for sediment budget assessment. Results from SIAM indicate general trends of surplus or deficit, not volumes of eroded or deposited material. The intended use of SIAM is to assess the impacts of alternative river scenarios on relative sediment budgets to select the best alternatives for detailed modeling analysis (USACE, 2006) Analytical Methods The boundary conditions required for SIAM include water surface elevations and flow duration data. The USGS stream gauging station at Ranch Rudolph (gauge no , approximately 4 miles upstream of Brown Bridge Pond) served as the upstream boundary for this project (Figure 2). The HEC-RAS model developed by the USACE (2008) only included design events, but as Figure 6 shows, a significant portion of the sediments being transported in the Boardman River occur with flows less than 200 cfs. 9

22 Daily Sediment Load Cumulative Sediment Load Cumulative Sediment Load w/ Flows <200 cfs Daily Sediment Load (tons) Cumulative Sediment Load (tons) 0 12/1/1996 4/30/1997 9/27/1997 2/24/1998 7/24/ /21/1998 5/20/ /17/1999 3/15/2000 8/12/2000 1/9/2001 6/8/ /5/2001 4/4/2002 9/1/2002 1/29/2003 6/28/ /25/2003 4/23/2004 9/20/2004 2/17/2005 7/17/ /14/2005 5/13/ /10/2006 3/9/2007 8/6/2007 1/3/2008 6/1/ /29/2008 3/28/2009 8/25/ Date Figure 6: Sediment loading at Ranch Rudolph based on data collected by the USGS Observed flow data for the complete period of record ( ) were used to develop a flow frequency distribution for this location of the Boardman River (Figure 7). This was annualized for input to SIAM (by dividing the total number of observed days for each flow interval by 365) to represent the range of flows that would be experienced over the course of a typical year (Figure 7). 10

23 40% 35% % Frequency (%) 25% 20% 15% Days/Yr 10% 40 5% 20 0% Discharge (cfs) 0 Figure 7: Flow-frequency analysis for USGS gauge upstream of Brown Bridge Pond As the period of observed record did not include any significant design events, a frequency and duration analysis was conducted on data from the nearby USGS stream gauge at Mayfield, just downstream of Brown Bridge dam (Figure 2). Since Brown Bridge dam is operated as a run-of-river impoundment, it was assumed the dam had no significant flow attenuation effects. Analysis of flow data, scaled by watershed area for the two gauges, revealed the water yield per unit drainage area was significantly different between the two. These results suggest differences in baseflow regime for the two gauges and that watershed area alone does not explain differences in yield. A review of the gauge data and precipitation records indicated a 10% decrease in average annual precipitation between the period of record for the two gauges (Figure 8). Consequently, simple scaling of the flow data was not sufficient to synthesize a greater range and duration of flows for the Ranch Rudolph gauge. An analysis of flow statistics revealed the two gauges had similar frequency and distribution characteristics when flows were scaled by total range and median flows. Consequently, the frequency distribution and duration values for the Mayfield gauge were used to extend the flow data at Ranch Rudolph to include up to the 5-year recurrence interval event. Given the relatively short 11

24 duration of observed data at Ranch Rudolph, it was not feasible to include larger design events, such as a 10-year recurrence interval event Average Annual Precip = 30" Average Annual Precip = 27" Annual Yield (in) Annual Precip (in) Annual Precip Water Yield at Mayfield Water Yield at Ranch Rudolph Figure 8: Precipitation and Water Yield for the Boardman River Sediment Data and Analyses Sediment rating curves were developed from observed sediment transport data gathered at the USGS gauging station at Ranch Rudolph (USGS, 2008) (Figure 9). The data provided sediment transport by particle size class. Individual sediment transport curves were developed for each size class. The smallest size classes of bed material, mm and mm, were not sampled by the bedload sampler; rather, they were transported as suspended load. As suspended sediments were not quantified on a per size basis, the total suspended load was equally divided across the range of suspended sediment size classes. For larger flows, on the order of a 5-year event, sediment transport was estimated using the Sediment Transport Capacity tool of HEC-RAS. Such flows represented a small fraction of the observed flow duration record, accounting for less than 0.2% of the observed flow duration. The results of sediment yield using sediment transport data gathered at the gauging station and the transport data estimated from HEC-RAS for the larger flows are shown in Figure

25 SUSPENDED BEDLOAD TOTAL Sediment Load (tons/day) y = 7E-06x R 2 = y = x R 2 = Discharge (cfs) Figure 9: Sediment loading data from USGS data collection at gauging station upstream of Brown Bridge Pond at Ranch Rudolph (Regression equations for sediment loading rates based on data collected from the USGS at gauge number (USGS, 2008)) Average Annual Sediment Yield (tons/acre) Figure 10: Average annual sediment yield based on observed daily flow at Ranch Rudolph gauge 13

26 Regression analysis of bedload and suspended load against discharge was used with flow duration analysis to estimate annual sediment loading rates. Particle size distributions of bedload sediment exhibited no significant trends with discharge, see Figure 11. Consequently, an average bedload distribution was used to develop daily and annualized bedload data for SIAM. This loading information was scaled based on watershed size for additional tributaries to estimate sediment loading from tributary sources since no other sediment monitoring data was available. Washload (particles < mm) comprised over 70% of total observed sediment transport data, see Figure 12. However, none of the sediment samples from river and reservoir sampling were analyzed for organic matter (USGS, 2008). Consequently, the washload and reservoir sediments contain an unknown amount of organic matter that is at least 10% by mass and likely greater than 50% by mass, based on field observations of extensive organic deposits in reservoir deltas, organic debris in the Boardman River, water quality sampling at the mouth of the Boardman River at Grand Traverse Harbor, and published literature (Riedel, et. al., 2007; MDEQ, 2005; Riedel and Vose, 2002; Cummings, 1984). 100% 90% 80% % of Total Bedload 70% 60% 50% 40% 30% 8-16 mm 4-8 mm 2-4 mm 1-2 mm.5-1 mm mm mm 20% 10% 0% Average Instantaneous Discharge (cfs) Figure 11: Observed bed load grain size data (Adapted from USGS, 2008) 14

27 100% 90% 80% % of Total Sediment Load 70% 60% 50% 40% 30% 8-16 mm 4-8 mm 2-4 mm 1-2 mm.5-1 mm mm mm <.125mm 20% 10% 0% Average Instantaneous Discharge (cfs) Figure 12: Observed total sediment grain size data (Adapted from USGS, 2008) Sediment Loading from Major Tributaries No sediment loading data or relationships were available for major tributaries to the Boardman River. During field reconnaissance, Jaxon, East, and Swainston Creek were identified as potentially significant sources of bed material load (as sand) to the Boardman River. Contributions from other tributaries were assumed to be not significant based upon relatively small watershed areas, flat terrain, and the existence of impoundments that capture bed material load. Sediment contributions from East, Jaxon, and Swainston Creeks were estimated by scaling the sediment load rating curve from Ranch Rudolph by the ratio of watershed areas. Swainston Creek has a large impoundment upstream of its confluence with the Boardman River, so only the subwatershed area of Swainston Creek downstream of this impoundment was used to scale the sediment load rating curve from Ranch Rudolph. The flow frequency and sediment duration data from these tributaries served as upstream/tributary sources to their respective sediment reaches in SIAM. 15

28 Sediment Loading from Bank Erosion Information on soil physical characteristics for the eroding bank site was obtained from the USDA Web Soil Survey, see Figure 13. The soil in the eroding bank is a terrace soil remnant from previous glaciation and fluvial outwash materials. It is categorized as a Gladwin-Richter gravelly sandy loam with a bulk density of 1.58 g/cc. The makeup of the soils for all soil size categories was estimated based on the percent clay, silt, and sand data available from the soil report (USDA Web Soil Survey, websoilsurvey.nrcs.usda.gov). Values for particle size classes between the sand, silt, clay, and gravel thresholds were estimated based on the published distribution, see Figure 14. The characteristic particle distribution is 4.8% clay, 23.7% silt, 67% sand, and the remaining 4.5% is assumed as gravel. Figure 13: Soil map of bank erosion example. Soil type is GrA (Gladwin-Richter gravelly sandy loams, 0 to 2 percent slopes). (Image Source: USDA Web Soil Survey, websoilsurvey.nrcs.usda.gov). 16

29 Based on field observation and estimates from aerial imagery, the bank failure location shown in Figure 4 is approximately 50 meters (164 feet) long, ranges 3 to 5 meters (10 to 16 feet) high, and 1 to 8 meters (3 to 26 feet) into the bank. For estimation purposes, we used average dimensions measured from both field data and aerial imagery; 50 m long by 4 m high by 4 m deep. This is equivalent to 800 cubic meters (28,000 cubic feet) of eroded material, or 1,300 metric tons (1,400 tons) based upon published bulk density data. Gravels Sands Silts % Finer Grain Size (mm) Figure 14: Bank material estimation based on percent clay, silt, and sand values available from the web soil survey (NRCS Soil Survey Staff, USDA, 2009). Values of soil sizes between particle size classification thresholds were estimated using trend lines between known size classes. A table of loading rates for each size class required for a complete SIAM model was developed using the results from Figure 14 and the known bulk density of the soil from the soil survey (NRCS Soil Survey Staff, USDA, 2009). These data were incorporated into the SIAM model to examine the impact that bank erosion has on the sediment transport of the Boardman River. As the actual period over which the erosion occurred is not known, two separate scenarios were developed: bank erosion occurring over a two year period or over a five year period. 17

30 Sediment Trapping Efficiency of Dams The sediment trapping efficiency for each dam was estimated using a plug-flow over-topping assumption to determine the likelihood of sediment pass-through as a function of discharge, particle size, reservoir volume, and release characteristics. Settling velocity was estimated using modified Stoke s Law (Jury, et. al., 1991). Over the observed range of flows, the reservoirs had sufficient retention time to allow settling of all sediments coarser than the silt size fraction (4 um) (Table 3). Consequently, each dam behaves as a sediment sink for all bed materials in the Boardman River. Table 3: Retention time and theoretical minimum particle size for sediment retention by Boardman River Dams under three flow scenarios. Low Mean and High Mean represent average bounds from USGS flow data (scaled from Brown Bridge) and Peak is the largest event of record. Sabin Boardman Brown Bridge Rt (days) Dia (um) Rt (days) Dia (um) Rt (days) Dia (um) Low Mean High Mean Peak SIAM Implementation The SIAM modeling was developed for the Boardman River from 6,000 feet upstream of Brown Bridge Road above Brown Bridge dam and extends to the backwater of Boardman Lake. Figure 2 and Figure 19 show where the sediment reaches are located along the Boardman River. A description of each of the sediment reach locations is shown in Table 4. Table 4: Boardman River sediment reach descriptions Sediment Upstream Extent Downstream Extent Reach A USGS gauge Brown Bridge Pond backwater B Brown Bridge Pond backwater Brown Bridge Dam C Brown Bridge Dam Confluence with East Creek D Confluence with East Creek Confluence with Jaxon Creek E Confluence with Jaxon Creek Boardman (Keystone) Pond F Boardman (Keystone) Pond Boardman (Keystone) Dam G Boardman (Keystone) Dam Sabin Dam H Sabin Dam Boardman Lake Sediment reaches were selected based upon results of field observations in October 2008, interpretation of bed material data gathered by the USACE, and consultation with SIAM model authors (Charlie Little, Personal Communication, February 2009). Figure 3 shows the representative 18

31 particle size distribution for each sediment reach. Field observations, which looked at the potential contribution of sediments to the Boardman River, concluded that East Creek and Jaxon Creek would contribute significant sediment loads to the system. Swainston Creek would not carry significant sediment to the main stem of the Boardman River due to the presence of a mill dam in Mayfield near the confluence. Each reservoir was identified as a unique sediment reach and, based on the significance of sediment deposits and longitudinal profiles, the deltas at Brown Bridge and Boardman Pond were also identified as unique sediment reaches. Dams along the Boardman River act as sediment traps, capturing a significant amount of the sediment traveling through the Boardman River system since their construction. Brown Bridge Dam, the most upstream of the dams on the Boardman River, has the largest impoundment and captures most of the sediment from upstream. Brown Bridge Dam is a top-release dam so the majority of the sediments are able to settle out before reaching the dam. The dam was completed in 1922 (ECT, 2008a), so it is assumed that most of the sediments from upstream sources since that time have been captured within the impoundment. The original Boardman Dam facility was constructed in 1894, but a newer facility was completed in 1931 to raise the pool elevation (ECT, 2008a). This dam, like Brown Bridge, is also a top release dam, so most of the sediment from upstream is able to settle in the impoundment. Only the river and drainage area downstream of Brown Bridge Dam is assumed to be contributors of sediment to Boardman Pond as Brown Bridge Dam would capture most of the sediments from upstream sources. 2.4 Model Checking SIAM is a sediment budget model that constructs an annualized estimate of sediment balances along reaches of a river based upon derived frequency and duration hydrologic and sediment source data. SIAM does not conduct sediment transport, yield, or bed change computations. Consequently, SIAM cannot be calibrated, per se, to observed sedimentation data. Rather, the relative trends predicted by SIAM can be compared to observed data to check the annualized sediment budgets against observed incision and deposition patterns. The process of model checking was conducted in a manner similar to calibration by tuning model parameters such that model output and estimates of sediment volume going through the system best match observed trends in river sediment dynamics. 19

32 Over the life span of the Boardman and Brown Bridge dams, sediment has accumulated within the impoundments. Given the accumulation of organic matter, along with sediment in these deposits, the volume of these deposits represents an upper limit on total sediment accumulation since the impoundments were constructed. The volume of these deposits was estimated to provide a check against the sediment budget estimates from SIAM. The estimate for Boardman Pond was developed by comparing a historic map to current bathymetric data, while the estimate for Brown Bridge Pond was developed by measuring the size of the sediment delta from cross-section and longitudinal surveys. Both of these methods are described below. Volumetric analysis for Boardman Pond was completed using a combination of current data collected by the USACE and historic data available before the new Boardman Dam was installed in A historic survey map of Boardman Pond, circa 1916 (E.P. Waterman Co. Surveyors), was compared to current aerial photographs and bathymetric data to determine how the volume has changed (Figure 15). During field data collection, the USACE gathered several cross sectional profiles of the Boardman Pond and additional topographic data above the lowered pool level in These datasets were combined to create a current bathymetry of the pond; this data was then used to calculate volumes at various elevation transects through the pond to develop a storagevolume curve (Figure 16). Sediment density of 1.8 Mg/m 3 was assumed to convert the volume data to mass. 20

33 Figure 15: 1916 imagery of Keystone (Boardman) Pond over current (2006) aerial photo 21

34 Boarmand Pond Storage Curves Elevation (ft) Current Volume (ac*ft) Figure 16: Boardman Pond storage curves. Current volume data are based on available bottom surveys and topography data. Volumes for the pond in 1916 were calculated by estimating surface areas from the 1916 drawing by E.P. Waterman Co. Surveyors. The volume of sediment in the sediment wedge that forms the Brown Bridge delta was estimated by measuring the length, width, and thickness from the longitudinal profile and aerial imagery. This was also converted to mass using a bulk density of 1.8 Mg/m 3. The sedimentation volumes for the Boardman and Brown Bridge ponds were converted to equivalent depths of sediment across the pond areas and are shown in Table 5. This allowed for direct comparison of the observed data to values predicted by SIAM. The values were also converted to watershed denudation rate by dividing the volume of material by the contributing watershed area. For Boardman Pond, this excluded the watershed area above Brown Bridge because Brown Bridge is assumed to effectively trap the majority of all sediment entering it. The similarity of the estimate for denudation rates at each impoundment suggests this method of estimating sedimentation volume produces reliable results. However, these denudation rates are nearly an order of magnitude lower than typically reported for other Great Lakes watersheds in this region (Riedel, et. al., 2009). One explanation for this may be 22

35 that this portion of the Boardman River Watershed has always been forested land use whereas most of the other Great Lakes watersheds have undergone significant conversion to agricultural land use. Table 5: Estimates of volume loss for Boardman and Brown Bridge Ponds Impoundment Volume of Sediment Mass of Sediment Sedimentation Denudation Rate ac-ft tons in/yr mm/yr Brown Bridge 1,240 3,027, Boardman 710 1,700, The base SIAM run used a default washload threshold diameter of mm, very fine sand. With the base run of the model, the results were predicting widespread sediment deficits in Boardman River including a very large deficit below Brown Bridge Dam. These results suggested the Boardman River would likely be undergoing widespread channel incision, yet field reconnaissance and inspections of old bridge footings and hydraulic structures indicated the riverbed is stable and actively connected to its floodplain. The following modifications were made to tune the SIAM model such that the results best matched the observed conditions in the Boardman River and the three impoundments: 1. Washload maximum diameter for sediment reach C, below Brown Bridge Dam, was increased to 2 mm, very coarse sand, to reduce the sediment deficit prediction. 2. Washload maximum diameter for the three dams in the model was reduced to make it impossible for larger classes of sediments to pass through the dams. 3. SIAM runs were conducted with maximum washload diameter for river reaches of 0.25 mm, 0.5 mm, and 1.0 mm to represent the range of sediment size classes that are in the threshold range of being a significant fraction of bed material (Figure 3). 23

36

37 3.0 RESULTS The results from SIAM are reported in annualized sediment balance values, by particle size and source, for each sediment reach. We converted these values to equivalent depths of incision or deposition based upon the length and average width of each reach and an assumed sediment density of 1.8 Mg/m3. This removes scale effects and allows for an apples to apples comparison across reaches. Figure 17 shows the results of the three different scenarios. The 0.25 mm washload threshold scenario had the most imbalance of sediment across the reaches and indicated widespread supply limitations along the Boardman River; suggesting channel scour and bed incision on the order of 2 to 6 inches per year. As the size class for washload was increased to 0.5 mm and 1 mm, the predicted supply limitations for the reaches declined and most were within a two inches/yr of zero, or balance. All scenarios showed a 5 in/yr sediment shortage in the Boardman River below Brown Bridge Dam. 4 Annualized Sediment Deposition/Incision Results from SIAM Modeling Results for a Range of Maximum Washload Diameters Annualized Sediment Deposition/Incision (in/yr) A B_Delta B C D E F_Delta F G H Sediment Reach Figure 17: Estimated magnitudes of deposition/erosion for each reach as per SIAM modeling. Series name indicates maximum washload diameter used for all reaches with the exception of reaches B, F, G, and C. 24

38 Supply excess was predicted for Brown Bridge and Boardman Dams (reaches B and F), conversely all scenarios showed sediment shortage, or 1.3 in equivalent incision per year at Sabin Dam (reach G). The values from Figure 17 are reported in Table 6. The sedimentation results from the 0.5 mm washload scenarios for Brown Bridge (sediment reach B) and Boardman River (sediment reach F) best compare to observed field conditions and sedimentation rates reported in Table 5. Table 6: SIAM modeling results of deposition/incision by sediment reach Estimated Sediment Sediment Deposition/Incision (in/yr) Reach A B_Delta B C D E F_Delta F G H The additional scenario examining the relative impact that bank erosion has on the volume of sediment being deposited into Boardman Pond was applied to the 0.5 mm washload scenario, since this scenario best matched observed conditions. The results, Figure 18 and Table 7, show that while the eroding bank used for volume estimation was large, the relative impact to the Boardman delta and reservoir is small (<0.03 inches/year). 25

39 2.00 Annualize Sediment Deposition/Incision (in/yr) Delta Pond No Erosion Bank Erodes Over 5 Years Bank Erodes Over 2 Years Figure 18: Results of scenarios examining impact of bank erosion on sediment volume in Boardman Pond. Bank erosion was estimated based on the dimensions of a large eroding bank upstream of the impoundment. Table 7: Results examining impact of bank erosion on sediment volumes in Boardman Pond; differences in values were only equivalent to differences in depth of hundredths of an inch per year. Impact on Sediment Volumes in Boardman Pond from Bank Erosion Delta Pond Sediment Depth Sediment Depth (tons/yr) (in/yr) (tons/yr) (in/yr) No Erosion -8, , Bank Erodes Over 5 Years -8, , Bank Erodes Over 2 Years -8, ,

40

41 4.0 DISCUSSION 4.1 SIAM Results for Sediment Reaches To interpret the results of the SIAM modeling in Figure 17, it is helpful to identify a stable reference condition and interpret each reach relative to that. For example, the results of field survey and review of USGS gauge records showed the Boardman River is stable at Ranch Rudolph; there is no evidence of active channel incision or aggradation, the floodplain is hydraulically connected to the channel, and there is no evidence of bed or water level profile changes in the USGS gauge data or at the bridge located at the crossing site. This section of river therefore, has been selected to serve as a stable reference condition. As the results for each SIAM scenario predicted a sediment shortage for this reach, that shortage should be interpreted as within the natural range of variability for a dynamically stable, or graded channel. Sediments move through the reach, and the reach may experience scour and deposition, but over long-term conditions, the net bed change is zero. Given the 0.25mm and 0.5mm washload conditions best represented the observed channel sediment and transport processes, any scour values between 4 to 1 inches/yr, respectively, actually reflect the natural channel system at this site. Similarly, the results for the reach at Sabin Dam, reach G, suggest a sediment supply limitation that on their own, could be interpreted as potential area for scour. However, Sabin Pond is a sediment sink. SIAM shows a supply limitation here because this reach is immediately located below Boardman Dam which captures virtually all sediment entering it and because SIAM only accounts for sediment balance, SIAM predicts a shortage for the Sabin Reach to balance the sediment budget. Within the context of these results, SIAM predicted sediment budgets for most reaches of the river were balanced, less than 2 inches of deposition or incision predicted per year. SIAM also predicted sediment budget results for the reaches that were consistent with changes at each reach and sediment transport theory. For example, sediment sources from tributaries were shown to increase sediment supply in downstream reaches and increases in bed slope from reach D to E resulted in increased transport capacity relative to sediment supply resulting in greater sediment deficits. 27

42 The reach directly downstream of Brown Bridge Dam, Reach C, showed a large shortage of sediment due to the sediment trapping efficiency of Brown Bridge impoundment, equivalent to five inches of incision for each of the three scenarios. Conversely, field observations and stable gauge data from the USGS gauge downstream indicated that this reach of the river is stable despite being sediment-starved. This can be explained by the well-armored condition of this reach. The results for the delta at Brown Bridge pond, sediment reach B_delta, do indicate this delta may be subject to incision and scour. While these may seem counter intuitive because deltas are formed from sediment deposition, deltas build and evolve by pro-grading. This process occurs as coarse sediments deposit at the upstream end of the delta during lower flows periods. When large flow events occur, these sediments are remobilized, transported across the delta, and deposited when the river flow enters the relatively quiescent waters of Brown Bridge pond. Here, the sediments are redeposited in deeper waters on the face of the delta, building the delta farther and deeper into the impoundment, or pro-grading. Thus, the body of the delta is actually a transitory storage zone for sediments. The signature profile of pro-grading deltas consists of a series of longitudinal waves in the delta profile that represent various advances of the delta under large flow event that remobilize large portions of the delta. 4.2 SIAM Results with Bank Erosion The results of the scenario examining the bank erosion contribution to sediment loads in Boardman Pond showed the bank erosion had little impact relative to the total sediment loading rate. Under the 2-year bank erosion rate with the example scenario, the total annual volume of sediment was only increased by 0.3%. Although there were select locations along the river showing significant bank erosion, these were rare and mostly confined to the reach of the river from the Beitner Road crossing to the backwater of Boardman Pond. The SIAM modeling of the current conditions show that the river is relatively stable with some deposition in Brown Bridge and Boardman Ponds. Alteration of the structures may change the sediment dynamics of the system, especially given the amount of sediment that has accumulated in the impoundments since the dams were constructed in the early 1900s. The sediment deltas, or wedges, in Brown Bridge and Boardman Ponds would be exposed to increased shear and likely to 28