One-Dimensional Modeling of Sedimentation Impacts for the Mississippi River at the West Bay Diversion ()
1. Introduction
West Bay Sediment Diversion Project (WBSD) is located on the west bank of the Mississippi River in Plaquemines Parish, Louisiana, 24803.15 ft (7.56 km) above Head of Passes, Figure 1 [1]. The project included the excavation of an uncontrolled diversion channel through the rightdescending bank of the Mississippi River. Construction was completed in November 2003. The project objective is to restore and maintain approximately 9830 acres (3978 ha) of fresh water to brackish marsh in the West Bay area by diverting both fresh water and sediment from the Mississippi River over the 20-year project life [2]. Thus, helping to alleviate the rapid erosion, which is on the order of 15.4 - 29.7 mi2/year (40 - 77 km2/year) [3,4] in coastal Louisiana.
Along the West Bay reach of the Mississippi River, the Pilottown Anchorage Area (PAA) is parallel to the navigation channel. The PAA is a US Coast Guard designated safe harbor outside the federally maintained navigation channel. The area is located along the right descending bank of the river from River Miles (RM) 6.7 to 1.5 (River Kilometer (RK) 10.8 to 2.4). Concerns about increased sediment deposition and subsequent increased dredging in the PAA and navigation channel prompted the Coastal Wetlands Planning, Protection and Restoration Act (CWPPRA) Task Force to authorize this study to evaluate the impacts of the WBSD. In response to, the Corps of Engineers’ Engineer Research and Development Center, Coastal and Hydraulics Laboratory (ERDC-CHL) developed a work plan that included 4 primary tasks: comprehensive channel geometry, discharge, suspended sediment, and bed material data collection program; a detailed geomorphic assessment; onedimensional (1-D) modeling which is the focus of this paper; and multi-dimensional modeling of the reach and WBSD. The multi-prong study further explores the impacts of diversions on the Lower Mississippi River which are not well understood [5].
A diversion of water without an appropriate amount of diverted sediment increases the potential for induced sediment deposition in the main channel [6-8]. If the diverted sediment-to-water ratio is less than that of the main channel, then a disproportionate amount of water is being diverted relative to sediment. By reducing the sediment transport capacity in the main channel without a corresponding reduction in sediment load will result in downstream deposition along the main channel [8]. Ref. [8], though not including WBSD, found that increases in upstream divert flows will increase maintenance dredging, 30,000 - 70,000 cu yd annually, in PAA. This is a critical
Figure 1. West Bay diversion project location map.
issue on the Mississippi River, where increased sediment deposition has an adverse impact on both commercial navigation and flood control.
2. WBSD History
The WBSD was first initiated with the excavation of an uncontrolled diversion channel through the right descending bank of the Mississippi River. The initial channel was excavated during the fall of 2003. This channel was constructed 24.9 ft (7.6 m) deep by 194.9 ft (59.4 m) wide using a hydraulic cutterhead dredge. The channel was designed to convey an average discharge of 20,000 cfs (570 cm) at the 50 percent duration stage at the Venice, LA gage. However, measured discharge in 2004 and 2005 indicated that the excavated channel passed only approximately 14,000 cfs (400 cm). A second phase of excavation planned to expand the channel conveyance to 50,000 cfs (1400 cm). This second phase excavation has not been constructed. However, the channel has been naturally enlarging since the initial construction in 2003, but has not reached the planned capacity. Measured discharge in 2007 and 2008 indicated that the diversion had almost doubled in capacity to approximately 27,000 cfs (760 cm).
Even during planning, sponsors realized the diversion could induce shoaling in the main navigation channel of the Mississippi River and the adjacent PAA. The US Army Corps of Engineers’ ongoing Operations and Maintenance Program is responsible for dredging of the main navigation channel. Additional dredging of the PAA would be an added feature to this program and would be a cost to the WBSD. After detailed negotiations with the navigation industry, an agreement for maintaining the PAA and navigation channel was developed and executed. The Cost Sharing Agreement executed between the State of Louisiana and the Corps of Engineers and the budget approved by the CWPPRA Task Force in 2002 state:
Included as a Project feature is the maintenance of the outermost (eastern) 250-ft-wide strip of the PAA and the entire width of the adjoining access area between this strip of the PAA and the Mississippi River navigation channel. Advanced maintenance of the PAA area shall be undertaken to account for the anticipated shoaling induced by the Project.
Thus channel maintenance is a direct project cost through the project life, 2023. However, initial advance maintenance dredging was conducted in the PAA in 2003; bathymetric surveys indicated a shoaling tendency prior to the opening of the WBSD. Subsequent maintenance dredging was conducted in both 2006 and 2009. The Task Force wanted to know the percentage of shoaling being caused by the diversion and the percentage being caused by other effects. ERDC-CHL developed a multitask work plan to address the shoaling issue.
3. 1-Dimensional Analysis
Using the HEC-6T numerical model software package, 1-D model, estimates the long term river responses to the diversion and the upstream sediment boundary conditions for the multi-dimensional models.
3.1. Model Background
The Engineering Research and Development Center (ERDC) conducted an investigation with the HEC-6T 1-D model. The effort established the usability and potential impact of the WBSD on dredging above head of passes, and evaluated the sensitivity of the model to key input parameters. An initial effort in studying the West Bay Diversion is documented in [9]. The ERDC model is based on the validated Vicksburg District, MVK, regional scale model. Changes from the MVK model to the ERDC model are discussed here and [9] along with key aspects critical for the model description. For a complete account of the MVK regional model see Copeland and Lombard (2009).
3.2. HEC-6T Model
The HEC-6T software is an enhanced version of HEC-6. HEC-6 is “a 1-D movable boundary open channel flow numerical model designed to simulate and predict changes in river profiles resulting from scour and/or deposition over moderate time periods, typically years” [10]. Model input requirements include: channel geometry, subsidence rates by cross-section, boundary conditions, bed material gradations, distributary outflow and sediment concentration, water temperature, and user specified sediment transport functions [10].
Flow conditions are specified by a series of sequential steady state flow discharges where water surface elevations at each cross-section are calculated with the standard step method, Method II [11,12]. Thus, from the user defined hydrograph HEC-6T calculates velocity and depths. Then, in a decoupled manner at each timestep, the calculated parameters (depth, velocity, and discharge) are then applied to determine the sediment transport potential. For a complete description of the governing equation see HEC-6 user manual [10,13]. The computed transport potential is compared to the available sediment supply in the water column and the river bed to determine bed erosion or deposition. Finally, these bathymetric changes are applied within the movable bed limits and the next flow condition is calculated repeating the process.
HEC-6T offers four capabilities needed for the evaluation of WBSD. First, HEC-6T allows for long-term simulations. For the WBSD evaluation, 50 year simulations were conducted to describe a broad range of potential flow events. Secondly, the model has the ability to simulate dredging activities. Dredging in both the navigation channel and in the PAA is required for the WBSD study. For the ERDC Phase II model, modifications were made to the code by MBH to allow multiple dredging templates at any cross-section, so adjacent sites can be dredged concurrently or at different times for varying widths and depths at the same cross-section. Additionally, HEC-6T allows for the diversion of both water and sediment, and calculates the downstream impacts of the diverted flux. Finally, it directly accounts for subsidence and sea level rise, important factors in the Gulf region for a long term simulation.
The primary disadvantage is that HEC-6T is a 1-D model which uses average hydraulic and sediment parameters to simulate 3-dimensional processes. HEC-6T includes no provision for specifying either a lateral distribution of sediment load or a bed material gradation across a cross-section. Additionally, HEC-6T does not consider salinity or the impacts or organics on fine sediment transport. Furthermore, in HEC-6T the standard procedure for deposition and scour is to move each cross section point, within the movable bed limits, an equal amount (the area that is shifted vertically during each time step due to sediment movement i.e. deposition or scour). For ERDC Phase I & II models, the $GR 3 option was selected that preferentially deposits sediment within the dredging template before deposition is distributed over the rest of the moveable bed portion of the crosssection. This prevents the artificial building of levees along the dredged channel, but does not necessarily distribute the sediment laterally in a realistic manner in all situations.
3.3. Modeling Approach
Two scenarios, each running the same 50 yr hydrograph, represent conditions with and without the WBSD. Comparisons of the two scenarios provide a means to identify both temporal and spatial changes in the sedimentation rates for both with and without WBSD alternatives.
The MVK model is part of a regional model being developed by the Mississippi River and Tributaries (MR & T) Project to identify long term channel maintenance sites within the Lower Mississippi River. Since the MVK model was developed for regional use, modifications were made for the WBSD evaluation, ERDC model modifications include:
a) Additional cross-sections downstream of Belle Chase with the highest density of cross sections within the PAA (River Mile, RM 1.5 to RM 6.7).
b) Subsidence and sea level rise rates were estimated and incorporated into the model.
c) Implementing the ERDC multi-dimensional model study and field data collection effort to refine sediment diversion ratios, flow diversion, sediment concentration, and bed material gradation. If needed, MVK Model values were modified.
d) Code modifications for multiple dredging templates were made to represent dredging in the Navigation channel and the PAA.
e) A typical discharge hydrograph which provides a plausible range of future flows is selected and duplicated as needed to create a projected 50 year hydrograph and its corresponding tailwater elevations.
For validation, the hydrograph prior to WBSD, 1991- 2002, was simulated. The MVK model was validated for the same time period, so for every change in the ERDC model comparisons were made to the MVK model to verify validation. The checks were primarily in the form of water surface elevations, dredging comparisons, and sediment load. Then sensitivity testing evaluated the impacts of varying sediment diversion ratios, sea level rise, subsidence, and sediment transport functions.
3.4. ERDC HEC-6T Model Input
The ERDC model extends from Vicksburg, MS to the Gulf. The primary focus was on the Belle Chasse, LA RM 75 to Head of Passes RM 0 reach. Key aspects of the MVK model were changed/modified to re-focus the model to the study area. With all changes to model input a congruent model validation was maintained.
3.4.1. Cross-Sections
The model provided by MVK extends about 455 miles from Vicksburg, Mississippi RM 437 .3 to Pilots’ Station in Southwest Pass at RM -18.0. The extended model allows for sediment adjustments prior to entering the study area, thus reducing bias from the inputted sediment load at Vicksburg. Model cross-sections are derived from the 1992 Mississippi River comprehensive hydrographic survey. The MVK model originally contained 201 crosssections, but the ERDC modifications added 28 crosssections between Belle Chasse and Head of Passes to better define the channel geometry within the study reach. The greatest increase in cross-section density occurred from Venice at RM 10.6 to Head of Passes RM 0, which includes the PAA. The average cross-section spacing through the PAA reach is 0.42 mi (0.68 km).
Within the Head of Passes (RM 0) to Venice (RM 10.6) reach, the ERDC model contains 19 cross sections which provide an average cross section spacing of 0.56 miles. The Pilottown PAA extends from RM 1.5 to RM 6.7. Through that reach, the ERDC model contains 12 cross sections. Eight of those sections are located downstream of the WBSD. The average cross section spacing through the PAA reach is 0.43 miles. The data for all cross-sections added to the model were obtained directly from the 1992 comprehensive hydrographic survey.
3.4.2. Boundary Conditions
Model computed sediment loads, deposition and erosion locations, and trends can vary if larger floods or drier periods occur more frequently than contained in the typical hydrograph. For water discharge, a typical average daily flow hydrograph is constructed. This hydrograph includes the 25-year period from 1 January 1984 to 31 December 2008. The period contains several higher flow years (1984, 1991, 1997, 2005, and 2008) as well as several lower water years (1988, 2000, and 2007). The highest flow in the hydrograph occurred during 2008, which approached the 50 year frequency flow. The 25-year hydrograph is simply repeated to create the 50-year typical hydrograph used for the simulations. The fifty year downstream water surface elevations are developed from 8:00 am daily stage data at Grand Isle East Point and match the same time period.
In south Louisiana, both subsidence and sea level rise are significant. Reported subsidence rates along the lower Mississippi River vary from different sources. The ERDC model subsidence rates were derived from NOAA Technical Report NOS/NGS 50 [14]. Subsidence rates vary from 0.87 in/year (22 mm/year) at RM 22.0 to 0.12 in/year (3 mm/year) at RM 306.00. The adopted subsidence rate from RM 16.0 (upstream of Venice, Louisiana) to the downstream end of the model is 0.63 in/year (16 mm/year). This rate equates to approximately 2.6 ft (0.8 m) of subsidence over the 50 year simulations. Subsidence rates in NOAA Technical Report NOS/NGS 50 were computed with a eustatic sea level rise of 0.05 in/ year (1.25 mm/year) at Grand Isle. The daily stages at the downstream boundary are increased at this rate for the 50 year simulations. Finally, incoming sediment loads are specified at the Vicksburg, MS gage.
3.4.3. Sediment Transport
The evaluation of transport capacity is calculated with a specified transport function. For the ERDC modeling effort, the Toffaleti function was applied. The Toffaleti equation was derived based on field data from the Lower Mississippi at Talbert Landing, Atchafalaya Rivers, five other river locations, and flume data from four data sets [15]. Data was collected over a broad range of flows for twelve years on the Mississippi River [16-18]. Other river data is from Mississippi River at St. Louis [19], Rio Grande at Bernalillo [20], Middle Loup [21], Niobrara [22]. The data included depths ranging from 0.98 ft - 49.2 ft (0.3 m - 15 m) with fine to medium sands [15]. The flume data was taken by [23-26], and USACE Waterways Experiment Station. Flume data was collected in flume widths ranging from 0.25 m - 2.4 m, flow depths ranging from 0.16 ft - 1.97 ft (0.05 m - 0.6 m), and sediment sizes of 0.01 in - 0.04 in (0.3 mm - 0.93 mm) [15]. The Toffaleti function was applied in this study since its main derivation was from large rivers.
While HEC-6T does not provide for the direct impact of salinity in the sediment transport functions, this impact can be approximated by varying the silt and clay shear threshold deposition coefficients. For the MVK model, the deposition coefficients for both silt and clay were increased downstream of Venice and the coefficient for clay was further increased in Southwest Pass to account for the effects of salinity on sediment deposition. The model allows for varying the threshold coefficients by reach but does not allow for varying the coefficients with discharge or stage. The salinity throughout the PAA varies greatly with discharge. During low flow, the salinity is much higher than during high flow periods.
Initial bed material gradations in the MVK model were derived from particle size distribution of bed sediments collected along the thalweg of the Mississippi River by Nordin and Queen in 1989 [27] (Copeland and Lombard 2009). One hundred seventy-six (176) samples were collected between Vicksburg, MS and Head of Passes. Of those samples, 25 were collected between Belle Chasse and Head of Passes [27]. Nordin did not collect any samples in Southwest Pass. Bed material samples were collected ERDC-CHL effort from RM 19.6 through Southwest Pass and comparisons made to the Nordin Data/ MVK model. Where vartiations occurred modifications were made to the ERDC models.
3.4.4. Diversions
In the HEC-6T model, the percentage of flow leaving the river through diversions compared to the flow in the river immediately upstream of that diversion is an input parameter. Flow distribution measurements were taken by ERDC at Baptiste Collette Bayou, Grand Pass, WBSD, Cubits Gap, Mississippi River upstream of Baptiste Collette Bayou, Mississippi River immediately upstream of WBSD, Mississippi River immediately downstream of WBSD, and in various small outlets in the bank of the Mississippi River between Venice and Head of Passes. Review of the diversion data, both from MVN and ERDC, indicates that for most flow conditions, Baptiste Collette and Grand Pass each diverts approximately 10 to 13 percent of the Mississippi River flow at Venice. Cubits Gap diverts approximately 13 to 18 percent of the flow and WBSD diverts approximately 5 percent of the flow. Figure 2 shows the flow distribution at WBSD by year. This plot shows the increase in flow over time as the WBSD has enlarged. For the ERDC model evaluation, the flow distribution at WBSD was set at the current rate of approximately 7 percent of the Mississippi River flow at Venice.
Table 1 provides the locations of the diversions contained in the ERDC model. The Davis Pond Diversion, WBSD, and Fort St. Philip Diversion were added to the ERDC model. For the diversions that were included, flow discharge through each diversion was modeled as a percentage of the discharge in the Mississippi River upstream from that diversion. When available, the percentages were estimated from measured data. When no measured data was available, the percentage of flow in the diversions was calculated (Copeland and Lombard 2009). For diversions added in ERDC Phase II the diversion ratios were estimated from a combination of ERDC field data and multi-dimensional model data.