LWI Region 3 Modeling Overview

LWI Region 3 Modeling Overview

Table of Contents

1 2

Region 3 Description

6 8

LWI Model Development Phases

3 Hydrologic and Hydraulic Modeling Methodology

10 10

3.1 Analysis of Record for Calibration (AORC)

3.1.1 Data Sources and Specifications

10

3.1.2 Development Process

10

3.1.3 Converting AORC to Gridded Precipitation for Modeling

11

3.2 Land Cover Classification

12 12 14 15 16 16 18 18 20 20 20 22 25 28 28 30 34 34 34 34 36 36

3.2.1 Remote Sensing and Land Cover Classification

3.3 LiDAR

3.3.3 Terrain Modifications for Modeling

3.4 Survey Planning

3.4.1 LWI Region 3 Survey Scope and Methodology

3.5 Soil Data Processing and Integration

3.5.1 Soil Data - Parameter Selection and Processing

3.6 Watershed-Specific Description

3.6.1 Boeuf

a. Hydrologic Model (HEC-HMS)

b. Hydraulic Model Setup (HEC-RAS 2D) c. Hydraulic Model Calibration and Validation

3.6.2 Macon

a. Hydrologic Model (HEC-HMS) b. Hydraulic Model (HEC-RAS)

3.6.3 Tensas

a. Hydraulic Model (HEC-RAS) b. HEC-RAS Boundary Options

c. HEC-RAS Model Calibration and Validation

3.6.4 Cocodrie

a. Hydraulic Model (HEC-RAS)

2

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

This document references the final reports for the Boeuf, Cocodrie, Macon, and Tensas Basins. You can access these reports through the ‘LEARN MORE’ links, which correspond to the following document references: B_*** for Boeuf, C_*** for Cocodrie, M_*** for Macon, and T_*** for Tensas.

4

Precipitation and Frequency Analysis 4.1 Precipitation and Frequency Analysis:

38 38

4.1.1 Atlas 14 Precipitation

38

4.1.2 Aerial Reduction Factor (ARF) Development Using MetVue

40

4.2 Watershed-Specific Description

44 44 44 44 45 46 46 47 48 48 48 50 54 54

4.2.1 Boeuf

a. Hydraulic Model - Design Storm Analysis

b. Rating Curve Development c. Calibration – Gage Selection

4.2.2 Macon

a. Model and Boundary Condition Update

b. Gage Analysis HEC-SSP

4.2.3 Tensas

c. Rating Curve Development d. Gage Analysis HEC-SSP

4.2.4 Cocodrie

5

Consequence Modeling

5.1 Flood Impact Assessment Using Go Consequence Software

5.1.1 Data Collected

55

6

Proof of Concept 6.1 City of Monroe

58 58

6.1.1 Boeuf Watershed – Young's Bayou Flood Mitigation Scenario

58

6.2 Bayou Cocodrie

60 60

6.2.2 Concept 1: City of Vidalia Pump station

6.2.3 Concept 2: Canal and Lake Sediment Removal

60

7

Future Use of Region 3 Models 7.1 Breakout Model Development

62 62

8

EnDMC

64

3

Louisiana Watershed Initiative (LWI) Region 3 Watershed Map LWI Region 3 Mapbook

4

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

5

1 Region 3 Description

For the Louisiana Watershed Initiative, WSP performed hydrologic and hydraulic studies in modeling Region 3. Within Louisiana, this Region is bounded by the Mississippi River mainstem levee to the east, the Ouachita River to the west, the Red River levee to the south, and the Arkansas State border to the north. The Region includes four watersheds—Boeuf River, Bayou Cocodrie, Tensas River, and Bayou Macon (Figure 1-1 and Figure 1-2).

Although no longer a part of the Mississippi River floodplain, the topography of this Region still exhibits its characteristics, such as previous meander belts of the Mississippi River, oxbow lakes, geologic scarring, and alluvial features like natural levees. The Region is mostly flat, with low lying areas and interconnected streams, bayous, and canals. Backwater from the Ouachita River greatly influences water surface levels in the Boeuf River and Tensas River Watersheds. Bayou Cocodrie Watershed, which is bound by levees on all sides, acts like a bowl with minimal natural drainage. The Tensas Cocodrie Pump Station is the only outflow from the basin. As a result, frequent, widespread, and unconfined flooding characterize Region 3. Most of the Region is rural, with vast agricultural lands and wildlife refuges. The City of Monroe, located in the Ouachita Parish in the Boeuf River Watershed, is the largest population center in the Region Ouachita Parish suffered major flooding in March 2016, leading the Department of Housing and Urban Development to identify it as a “most impacted and distressed” area.

Figure 1-2: LWI Region 3 relative to the Ouachita River basin.

6

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

Figure 1-1: LWI Region 3 spans four Hydrologic Unit Code (HUC) 8s and covers portions of Arkansas and northeastern Louisiana.

Major rivers in the Region include Boeuf River, Bayou Lafourche, Turkey Creek, and Big Rive (Boeuf River Watershed); Bayou Macon and Joes Bayou (Bayou Macon Watershed); Tensas River and Tensas Bayou (Tensas River Watershed); and Bayou Cocodrie in the Bayou Cocodrie Watershed. Besides the flat topography of the Region, the presence of levee and pump systems, dams, and diversion structures result in complex hydrologic and hydraulic conditions. WSP modeled this system using the following software and key inputs: { Hydrologic Engineering Center – Hydrologic Modeling System (HEC-HMS) (v10.1) for the hydrologic study of the Arkansas portion of Boeuf River and Bayou Macon Watersheds

{ Hydrologic Engineering Center – River Analysis System (HEC-RAS) (v6.3.1) for two-dimensional (2D) rain-on-mesh (ROM) studies of all four watersheds in Louisiana { Light detection and ranging (LiDAR ) terrain dataset provided by the U.S. Geological Survey (USGS) { High Water Marks (Ouachita Parish, March 2016 event), USGS, and U.S. Army Corps of Engineers (USACE) gage data for calibration of models and verification of model results { Survey collection of over 600 channel cross sections and 1,350 hydraulic structures with data incorporated into the models

7

Region 3 Description

2 LWI Model Development Phases

WSP developed hydrologic and hydraulic models for the Louisiana Watershed Initiative over three phases:

1. Data Collection & Modeling Methodology Development

3. Design Storm Events & Consequence Assessment

2. Model Development & Calibration

1. Data collection and modeling methodology development: During the first phase, existing model and supporting data was collected, including the following:

{ Previous watershed studies: USACE Corps Watershed Model System model for the Ouachita River, Federal Emergency Management Agency (FEMA) studies, and inundation maps from the March 2016 event { Existing channel/hydraulic structure survey: channel modification plans from the Tensas River Levee District, pump operation manuals, hydraulic structure data from past studies, and GIS-based survey data for the City of Monroe { Historical flood information: Newspaper accounts of past events, FEMA claims data, Louisiana Watershed Resiliency (LaWRS) data

{ Observed data: USACE and USGS stage and flow datasets and high water marks { Landuse and soils data: National Land Cover Database (NLCD) and Natural Resources Conservation Service (NRCS) soils datasets for each parish in Region 3 { Development of modeling methodologies: detailed reports describing watershed characteristics and appropriate modeling methodologies (one-dimensional [1D] vs. 2D, HEC HMS, ROM, etc.)

8

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

2. Model development and calibration: The two major components of the second phase included the following: { Channel and hydraulic structure survey: Survey data was collected for all four watersheds, including a channel cross section survey and bridge, culvert, levee outlets, diversions, dams and weirs. Pump station operation information was also obtained, and field trips were conducted to better understand the Region’s drainage patterns. { Model development and calibration: Watershed wide models were developed and calibrated to a selection of the following rainfall events, as appropriate for each watershed: • September 2008 (Hurricane Gustav): short event • May 2009: short event • September-November 2009: multiple days and backwater event • March 2016: historic event • February-May 2018: multiple days and backwater event • May 2019: short, backwater event

• September 2019: short, low flow event • September-October 2020: multiple days • January 2013: short event

3. Design storm events and consequence assessment:

The calibrated models developed in the second phase for each watershed were used to evaluate 35 design storm events of multiple frequencies and durations. Consequence Assessment was performed using Go Consequence for each of the 35 events.

9

LWI Model Development Phases

3 Hydrologic and Hydraulic Modeling Methodology

3.1 ANALYSIS OF RECORD FOR CALIBRATION (AORC) The AORC is a high-resolution, gridded meteorological dataset developed primarily for hydrologic modeling and analysis across the United States. Maintained by the National Weather Service Office of Water Prediction, this dataset provides precipitation and other atmospheric variables on a standardized Standard Hydrologic Grid (SHG). The AORC is widely used in water resource management, flood forecasting, and environmental modeling. AORC precipitation data can be incorporated into hydrologic models as a meteorological boundary condition, typically through the RAS unsteady flow file in HEC-DSS format.

3.1.1 Data Sources and Specifications AORC integrates multiple data sources to ensure accuracy and spatial/temporal coverage: { Next Generation Radar Data (NEXRAD) (National Oceanic and Atmospheric Administration [NOAA]) { Gauge-based precipitation data (National Weather Service, USGS) { Satellite observations { Reanalysis products (e.g., National Centers for Environmental Prediction/National Center for Atmospheric Research Reanalysis)

SPECIFICATIONS:

SPATIAL RESOLUTION: Approximately 4 kilometers TEMPORAL RESOLUTION: Hourly

TIME SPAN : Typically from 1979 to near present

3.1.2 Development Process The AORC dataset is produced through a multi-step process: 1. Data collection: combines radar, satellite, ground-based gauges, and numerical model data 2. Bias correction: adjusts data using ground-truth measurements (e.g., rain gauges) 3. Spatial interpolation: creates continuous, gridded fields 4. Temporal aggregation: compiles data into hourly intervals 5. Validation: compares with independent datasets for model calibration and quality assurance

10

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

3.1.3 Converting AORC to Gridded Precipitation for Modeling To use AORC data in modeling applications like HEC-HMS or HEC-RAS , it must be converted into the appropriate format: { The Vortex Importer utility is used to convert AORC data into HEC-DSS format. { Detailed instructions are available via the HEC documentation portal: { Figure 3-1 displays real-time historical precipitation grids. { Figure 3-2 shows accumulated rainfall over the duration of a modeled storm. { Supporting video animates the passage of the storm across the basin, helping visualize the storm's development and movement.

Creating Gridded Boundary Conditions for HEC-HMS

Figure 3-1: Hourly AORC Precipitation Grid over HUC-8 Cocodrie Basin. Rainfall depth is 1.13 inches at indicated location on January 10, 2013 at 4 a.m.

Figure 3-2: Accumulated AORC Precipitation Grid over Macon Basin during 9/23/2020 and 10/10/2020 event.

11

Hydrologic and Hydraulic Modeling Methodology

3.2 LAND COVER CLASSIFICATION 3.2.1 Remote Sensing and Land Cover Classification

To improve model precision in the Boeuf River Watershed, WSP conducted high-resolution land cover mapping using National Agriculture Imagery Program (NAIP) aerial imagery (1-meter resolution, 4-band including near infrared). Mapping allowed the project team to distinguish between vegetation types, impervious surfaces, and water features with greater detail than coarser datasets like the NLCD ( Figure 3-3).

Figure 3-3: Boeuf land cover.

12

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

Machine learning techniques were used on NAIP imagery to classify key land cover types, such as trees, water, buildings, and crops, supplemented with road, wetland, and building footprint data from national sources. For parts of the watershed in Arkansas, where 2D ROM modeling was implemented, the team used the 2019 NLCD dataset due to lack of NAIP coverage (Figure 3-4). The Normalized Difference Vegetation Index was calculated using the red and near-infrared bands to separate vegetated and nonvegetated features, providing a strong basis for supervised classification.

Training samples were collected across the watershed, and a maximum likelihood algorithm with image segmentation was used to classify the imagery into meaningful land use categories. LEARN MORE: B_2.1.3 + This enhanced land cover layer informed Manning’s roughness distribution in the 2D model, with spatial variation applied across cell faces for improved accuracy.

NLCD

Remote Sensing

Figure 3-4: Comparison between land cover dataset developed from remote sensing (left) and the NLCD (right).

13

Hydrologic and Hydraulic Modeling Methodology

3.3.1 LiDAR Data Summary for Region 3 Modeling OVERVIEW 3.3 LiDAR LiDAR acquisition, DEM correction, bathymetric enhancements, modification.

REGION 3 LIDAR DATA { Acquired in 2021–2022 in two phases: • Catahoula-Concordia 1 and 2 (Figure 3-5) • Northeastern 1, 2, and 4 ( Figure 3-6) { Acquisition was not temporally consistent , requiring blending of datasets. LEARN MORE: T_APPENDIX 4 + { The northeastern corner of Louisiana remains without finalized USGS-approved LiDAR (Figure 3-7).

{ LiDAR captures detailed three-dimensional (3D) elevation data using laser pulses from aircraft or drones. { Raw data is processed to remove noise, align sensors, and filter vegetation and structures. { Digital elevation models (DEMs) are created from ground points and serve as the base for hydraulic and hydrologic modeling.

USGS 3DEP Lidar Explorer

• Preliminary data did not meet USGS quality standards. • Key watersheds— Boeuf, Macon, and Tensas — depend on this data. • Extensive corrections were needed to improve data usability for modeling.

Figure 3-5: Extent of the Catahoula-Concordia LiDAR data relevant to modeling Region 3.

Figure 3-6: Extent of the Northeast LiDAR data relevant to modeling Region 3.

3.3.2 Key Data Issues

{ Limited hydro-flattening. { LiDAR processing agency only flattened rivers wider than 100 feet (ft). { Smaller channels were excluded, causing elevated or sloped features in the DEM. INCONSISTENT APPLICATION { Even qualifying streams had variable width and depth flattening. { Inconsistencies resulted in terrain discontinuities affecting flow paths ( Figure 3-8). HIGH WATER ACQUISITION

Figure 3-7: USGS LiDAR product, Catahoula-Concordia and Northeastern data relevant to modeling Region 3.

{ Some data was captured during flooding events. { Elevated water surfaces obscured true ground features, complicating channel and floodplain delineation ( Figure 3-9).

14

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

3.3.3 Terrain Modifications for Modeling { Multiple iterations were needed to improve terrain accuracy: • Integrated bathymetry for stream channels and lakes (Figure 3-10 and Figure 3-11). • Addressed abrupt elevation changes affecting flow conveyance. TERRAIN INCONSISTENCIES { Differences existed between observed channel widths and DEM representations. { Sudden width variations along rivers introduced modeling challenges. FINAL TERRAIN IS AN APPROXIMATION { Channel shapes and slopes between surveyed sections were interpolated. { Terrain remains suitable for 2D HEC-RAS simulations but may not fully reflect actual ground conditions.

Figure 3-8: Significant oscillation in channel elevations along Youngs Bayou requiring hydro-flattening.

Figure 3-9: Example of stark changes in channel topography within the Region.

Figure 3-11: Terrain modification necessary for better representation of oxbow lakes and meander scars storage in topography within Region 3.

Figure 3-10: Example of channel width in DEM ~3 times width in imagery within Region 3 due to collection of LiDAR during or after high-flow event.

15

Hydrologic and Hydraulic Modeling Methodology

3.4 SURVEY PLANNING Survey planning, structure analysis, and automation tools, survey incorporation. 3.4.1 LWI Region 3 Survey Scope and Methodology OVERVIEW { Surveys were conducted in Region 3 along priority streams, targeting bridges, culverts, weirs, dams, levee outlets, and control structures. { Cross-sections were collected to determine stream gradients and adjust LiDAR in areas with standing water for accurate channel representation. DATA COLLECTION { Key survey locations included stream confluences and tops of detailed stream sections to support hydraulic model interpolation.

PLANNING AND FORMATS { WSP developed survey guidance and provided spatial files (ESRI and KMZ) identifying structure and cross section locations. { KMZ files include: • Point file: structure points with required data, such as road profile and upstream face. • Line file: cross section lines to be surveyed bank-to bank. { Deliverables include text files (NAVD88 elevations, Point ID, Structure ID), MicroStation drawings, geo tagged photos, and shapefiles.

{ Pump stations were modeled using data from USACE and local engineers (via City of Monroe manuals) (Figure 3-12).

Figure 3-12: Typical pumping station plan.

16

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

STRUCTURE-SPECIFIC REQUIREMENTS Bridges/culverts:

Dams/weirs: Data includes top elevation, upstream/downstream sections, drawdown, spillway, and outlet pipe details. Required photos: upstream/downstream faces, channels, outlet, and dam top.

Control gates: Surveys capture gate type, opening size, invert elevations, crest elevation, and typical operation rates. Required photos: upstream/ downstream faces and channels.

ADDITIONAL NOTES Structure classifications based on imagery are verified in the field. Digital data entry enables seamless model integration by WSP. Surveys include upstream face, road top profile, bent details, guardrails, low chord elevations, and photos. Each structure point includes type (BR, CU, Weir, XS), road name, and span lengths (Figure 3-13 and Figure 3-14).

Figure 3-13: Bridge survey collection points.

Figure 3-14: Culvert survey collection points.

17

Hydrologic and Hydraulic Modeling Methodology

3.5 SOIL DATA PROCESSING AND INTEGRATION 3.5.1 Soil Data - Parameter Selection and Processing For all Region 3 watersheds, soil parameters were selected based on the Deficit and Constant Loss method, as outlined in LWI's June 2021 Modeling Guidance. This approach simulates how soil absorbs and retains water before contributing to runoff ( Figure 3-15).

THE KEY PARAMETERS USED ARE: Initial Deficit: Represents the soil’s moisture deficit at the start of a storm event. To simulate realistic antecedent conditions, models are initialized three days after a prior rainfall, allowing the soil to drain to field capacity. The value varies by event and is typically expressed as a percentage of the maximum soil deficit. Maximum Deficit: Represents the soil’s total water holding capacity. Parameter was calculated using effective porosity minus wilting point storage over the assumed active soil depth. Active Soil Depth: Refers to the depth of soil that actively participates in runoff processes. Based on rooting depth research, an initial depth of 24 inches (in)was used, adjustable during calibration.

Hydraulic Conductivity: Defines how quickly water moves through saturated soil. Regionwide, values were assigned based on Rawls (1982) soil data. This parameter set helps ensure a consistent and regionally appropriate representation of infiltration behavior across all hydrologic models developed in Region 3. Table 3-1 presents the initial and maximum soil deficits along with hydraulic conductivity values. The calculations are shown in the report section. The initial deficit serves as a placeholder and is adjusted as a percentage of the maximum deficit, depending on the antecedent soil saturation for each specific storm event.

Table 3-1 Soil parameters used in the model. Initial deficit varies by event based on antecedent conditions.

HYDRAULIC CONDUCTIVITY (INCHES/ HOUR)

TEXTURE

MAXIMUM DEFICIT (INCHES)

4.64 1.18 0.43 0.13 0.26 0.04 0.04 0.02 0.01 0.01 0.063

9.768 9.048 8.472 8.232 8.904 5.952 6.336 6.264 5.496

Sand

Loamy Sand Sandy Loam

Loam

Silt Loam

Sandy Clay Loam

Clay Loam

Silty Clay Loam

Silty Clay

4.92 4.32

Clay

Made Land

0

0

Water

43.3 0.02

9.12

Gravel Pit

5.496

Sand

Initial Deficit is used a % of Max deficit, dependent on the modeled historical storm event

18

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

Figure 3-15: Soil data within Boeuf Basin.

19

Hydrologic and Hydraulic Modeling Methodology

BOEUF

MACON TENSAS

COCODRIE

3.6 WATERSHED-SPECIFIC DESCRIPTION 3.6.1 Boeuf a. Hydrologic Model (HEC-HMS) Boeuf Watershed – HEC-HMS Modeling Summary

Overview { HEC-HMS was used to model the Arkansas headwaters of the Boeuf Watershed, providing inflow data for the Boeuf 2D ROM HEC-RAS model. { Detailed modeling methodology and results are documented in the Boeuf Report.

LEARN MORE: B_HMS_APPENDIX_1 +

WATERSHED CHARACTERISTICS

MODELED AREA: 572 square miles (sq. mi.).

CONFIGURATION: 70 subbasins and 80 reaches; see (Figure 3-16).

LEARN MORE: B_HMS_APPENDIX_2.1.1 +

Figure 3-16: Boeuf River-Arkansas HEC-HMS delineation and Eudora Gage location.

Rainfall-Runoff Transformation

Reach Routing Methods

Loss Method { Approach: Deficit and Constant method to simulate initial abstraction and infiltration losses. { Detailed procedures available in the Soil Data–Parameter Selection section and: LEARN MORE: B_HMS_ APPENDIX_2.1.4 +

Baseflow Representation { Method: Linear Reservoir used to simulate subsurface water contribution post-infiltration. { See the link below for more information: LEARN MORE: B_HMS_ APPENDIX_2.1.6 +

METHOD: ModClark transform used to convert precipitation to runoff hydrographs at each subbasin outlet. APPROACH: Accounts for subbasin specific characteristics. LEARN MORE: B_HMS_ APPENDIX_2.1.5 +

MODPULS METHOD : Applied to areas with significant overbank storage; based on 1D HEC-RAS hydraulics. MUSKINGUM-CUNGE METHOD : Used for channelized

reaches; routing based on channel slope, roughness, and length. LEARN MORE: B_HMS_ APPENDIX_2.1.7 +

20

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

Model Calibration and Validation { Calibration events: Dec 2014, Mar 2016, Feb 2018, May 2019, Aug 2019, Sep 2020, Oct 2020, and Aug 2022 { Validation events: Nov 2015, Feb 2020, and Jun 2021 { Parameters calibrated: • Subbasin time of concentration • Storage coefficient • Antecedent moisture conditions • Routing parameters (e.g., Manning’s n) { Comparison of calibrated vs. uncalibrated outputs shown in Figure 3-17; full details below: LEARN MORE: B_HMS_APPENDIX_3 + Recurrence Interval Analysis { 35 storm events simulated using Atlas 14 precipitation data Atlas 14 Precipitation { Model outputs validated against Bulletin 17C Expected Moments Algorithm peak flow estimates at the Eudora gage (Figure 3-18) { Example outlet hydrographs shown in Figure 3-18.

Figure 3-17: Comparison of model results for May 2019 calibration event at watershed outlet before and after calibration.

Figure 3-18: 100yr_24hr Model Outlet Hydrograph versus EMA Peak Flow estimate and confidence limits.

Figure 3-19: Comparison of results at HMS model outlet of all 24-hour duration events, varying recurrence interval.

21

Hydrologic and Hydraulic Modeling Methodology

BOEUF

MACON TENSAS

COCODRIE

b. Hydraulic Model Setup (HEC-RAS 2D) Model Overview and 2D Mesh Design The hydraulic model of the Boeuf River Watershed was developed using a fully integrated HEC-RAS 2D ROM approach, replacing the previously segmented 1D/2D hybrid model architecture. This decision was based on calibration findings that showed improved hydraulic behavior and interaction when using a seamless 2D setup. As shown in Figure 3-20 , the model domain includes an additional 170 sq. mi. added upstream in Arkansas. The modeled domain covers approximately 2,000 sq. mi. in Louisiana, with added refinement in Arkansas during calibration. The base mesh resolution in rural areas was kept coarse to optimize runtime but refined to 50–100 ft in urban and channelized areas, such as the City of Monroe, and along narrow bayous. These refinements allowed the model to capture localized flood behavior and flow through control structures with greater precision. Breaklines were used extensively along levees, roads, channels, and high ground to ensure the mesh honored important hydraulic features and gradient changes, preventing computational shortcuts across bends and preserving realistic flow paths. The Arkansas portion of the watershed was modeled with coarser mesh and necessary refinement along the Boeuf River to simulate continuous flow across the state line and eliminate artificial boundaries. This portion was later merged into the Louisiana ROM model during calibration.

Figure 3-20: Additional Arkansas portion added to 2D ROM model domain.

Hydraulic Structures and Control Features A total of 657 major hydraulic structures were included, covering:

Structures were modeled using Storage Area/2D connections with surveyed data, including deck elevations, culvert diameters, and spillway profiles where available. Lateral structures and flap gates were added at strategic locations along levees and embankments. Key structures, such as Gunby Dam, were transitioned from static outflow boundaries to dynamic weirs within the ROM system to improve calibration of flow splits. The Monroe urban drainage system was captured through detailed modeling of its 11 pump stations.

{ 280 bridges { 258 culverts { 8 dams/spillways { 52 weirs/diversions { 48 gated levee outlets { 11 stormwater pump stations in Monroe

22

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

Each station was linked to a storage sump and included operational logic such as pump start/stop triggers based on interior and exterior water levels. The Rochelle Avenue station alone accounted for ~60,000 gallons per minute (GPM) of pumping capacity. Table 3-2 summarizes the pump station specifications used in the model, including flow capacities and horsepower extracted from as-built documents and GIS records. Figure 3-21 depicts the locations of these pump stations. These parameters were critical in accurately representing pump operations during storm events and were used throughout model setup and calibration.

Table 3-2 City of Monroe pump station specifications used in the model, including flow capacities and horsepower extracted from as-built documents and GIS records.

PUMP STATION NAME

ADDRESS

UNITS

HORSEPOWER

FLOW (GPM)

901 Stubbs Avenue

2

350

30000 60000

Stubbs

901 Rochelle Avenue 3

1250 HP (Diesel)

Rochelle Ave

2151 Island Drive 2003 Lamy Lane

2 3

Pump 1: 150

Pump 1: 20,000

N 10th Street W

350

35000

Lamy Lane

Pump 1&2: 60,000

Pump 1&2: 500 HP (Electric) Pump 3: 625 HP (Diesel)

2505 Oliver Road

3

Pope Westminster

Pump 3: 80,000

7 Olive Street

4 3

400 200

50000 27000

PineSt

518 Oregon Trail

OregonTrail

2708 Hawes Street

1

600

60000

Hawes St

Pump 1: 30 Pump 2: 60 Pump3: Diesel

Pump 1: 1,600 Pump 2: 7,000 Pump3: 15,000

220 Plum Street

3

Plum St

110 Allen Ave

2

200

15000

Allen Ave

Pump 1: 150 Pump 2: 200

Pump 1: 20,000 Pump 2: 35,000

Pump1_P1 - MarquetteSt

2151 Island Drive

2

Boundary Conditions { Upstream boundaries were applied using hydrographs derived from the HEC-HMS model, ensuring accurate inflow from hydrologic sources in Arkansas and neighboring basins. { Downstream stage boundaries were based on observed stage hydrographs at the Ouachita River (e.g., Columbia L&D, Harrisonburg, West Monroe), allowing propagation of backwater effects into the Boeuf system. For higher stages of the Ouachita River, a series of gates along the levee system (shown in Figure 3-21 ) close, preventing backwater from entering the City of Monroe. Figure 3-21 shows the applied boundary conditions. { Lateral inflows were added during calibration to simulate spill over from adjacent basins, particularly from Bayou Bartholomew in extreme events like March 2016. { Former internal boundaries between model groups (e.g., BF-West, BF-South) were replaced with seamless 2D mesh connectivity to ensure mass continuity.

23

Hydrologic and Hydraulic Modeling Methodology

BOEUF

MACON TENSAS

COCODRIE

Updated Structural Configurations

Wham Dam Simulated in pre- and post-2018

configurations to reflect changes in structure and functionality following its reconstruction. The pre-2018 configuration had a higher weir elevation of 83.06 ft and a narrower width of 88 ft, while the post-2018 reconstruction featured a lowered weir crest at 65.48 ft and a widened spillway of 110 ft. These geometry changes allowed water to overtop at a lower elevation and across a broader span, influencing the timing and magnitude of downstream peaks. Refer to Figure 3-22 for a visual comparison of the dam geometry before and after reconstruction. Irwin Lake Weir Located downstream of the Boeuf confluence on Bayou Lafourche; was modeled with its current geometry and routing logic. Figure 3-22 highlights the Irwin Lake Dam, which plays a crucial role in diverging flow on the Boeuf and Bayou Lafourche Rivers. Rochelle Detention Basin { Added to the model post-2018 as part of Monroe’s flood mitigation improvements.

Figure 3-21: Boundary conditions on Ouachita and Bayou DeSiard, shown along the City of Monroe and with pump stations.

This storage area delayed stormwater surface flooding and was linked to the Rochelle Avenue pump.

{ Maintaining multiple geometry versions across different hydrologic events allowed for improved model calibration and assessment of mitigation effectiveness.

Figure 3-22: HEC-RAS representation of Wham Dam showing pre- and post-2018 geometry changes, including updated weir crest elevation and spillway width.

24

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

c. Hydraulic Model Calibration and Validation Regional Calibration Approach Calibration was performed regionally using the previously grouped models (e.g., BF-East, BF-West, BF-South) and later integrated into one fully connected model to assess system-wide performance. Figure 3-24 shows the full model domain for calibration. Key calibration regions included:

• The Boeuf-Bayou Lafourche split at Gunby Dam • Urban Monroe (internal drainage and pumping systems)

Figure 3-23: Highlighting the Irwin Lake Dam that plays a crucial role in diverging flow on the Boeuf and Bayou Lafourche Rivers.

• Upper Boeuf River near the AR/LA Stateline • Boeuf River and Bayou Lafourche diversion

Additional inflows from Region 2 (Bartholomew/ Ouachita) were introduced to capture real-world, cross basin effects. Geometry Refinement During Calibration { All 1D elements were removed and replaced with 2D ROM components to avoid loss of interaction. { Culverts and conveyance paths were added or adjusted based on overtopping discrepancies between modeled results and known high water mark locations. { Manning’s roughness coefficients (n values) were tuned using land cover data and further adjusted to match timing and volume at gages. Final values ranged from 0.035 in channels to up to 0.15 in wooded floodplain areas. LEARN MORE: B_5.6 + Initial Condition Setup For each calibration event, three initial soil saturation conditions were available: 20%, 50%, and 80%, corresponding to dry, moderate, and wet antecedent conditions respectively. These conditions affected the initial deficit in the infiltration model. Table 3-3 provides the maximum and initial soil deficits used for infiltration modeling. LEARN MORE: B_4.6 +

Figure 3-24: Full model domain for Boeuf Watershed calibration showing key streams and gages.

{ For Silt Loam soils, this resulted in initial deficits of 7.12,” 4.45,” and 1.78” for each scenario. { Initial water surface elevations were set using baseflow conditions or warm-up periods to reach steady states before simulation start.

25

Hydrologic and Hydraulic Modeling Methodology

BOEUF

MACON TENSAS

COCODRIE

Table 3-3 Maximum and initial soil deficits used for infiltration modeling under varying antecedent saturation scenarios (20%, 50%, and 80%) for each soil type.

20% SATURATED INITIAL DEFICIT

50% SATURATED INITIAL DEFICIT

80% SATURATED INITIAL DEFICIT

SOIL CLASSIFICATION MAX DEFICIT

8.9

7.12

4.45

1.78

Silt Loam

0

0

0

0

Water

4.92 8.47 8.23 9.05 6.26 9.77 6.34 5.5

3.94 6.78 6.59 7.24 5.01 7.81 5.07 4.4

2.46 4.24 4.12 2.75 4.52 3.13 4.88 3.17

0.98 1.69 1.65 1.81 1.25 1.95 1.27 1.1

Clay

Silty Clay

Loam

Loamy Sand

Sandy Loam

Silty Clay Loam

Sand

Clay Loam

Table 3-4 Summary of calibration events with associated rainfall intensity, backwater conditions, and initial soil saturation assumptions.

ORDER

EVENT

CHARACTERISTICS

1 2 3 4 5 6

Aug-19

Low flow event. Dry initial condition

Aug-22 Intense rainfall event. Saturated initial conditions

Mar-16 Dec-14 Sep-08 Nov-15

Intense rainfall as well as a big backwater event. Saturated initial conditions

Headwater influence

Small rainfall events. 80% wet initial conditions Small rainfall events. 50% wet initial conditions

The Model Was Calibrated Using Observed Data from the Following Gages:

BAYOU LAFOURCHE: Crew Lake and Alto (stage) BOEUF RIVER: Girard, Alto (USACE), Fort Necessity (stage and flow) BOEUF STATELINE (EUDORA): Stage

OAK GROVE AND SLIGO: Stage only

LEARN MORE: B_4.6 +

Figure 3-25: Bayou Lafourche at Crew Lake calibrated to the August 2022 event (located western side of the Boeuf Watershed).

26

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

Bayou Lafourche at Crew Lake and Girard at Boeuf River represented the western watershed, while Oak Grove covered the eastern side. Calibration of both regions was essential for accurate results at Fort Necessity, where flows from both converge before discharging into the Ouachita River. Calibration outputs for these areas are shown in Figure 3-25 (Crew Lake, August 2022), Figure 3-26 (Crew Lake, March 2016), and Figure 3-27 (Big Colewa Bayou, March 2016). ROOT MEAN SQUARE ERROR (RMSE) FOR STAGE: <5% in nearly all cases PERCENT BIAS (PBIAS): ±3%, with slight conservative bias (high estimates) CORRELATION COEFFICIENTS (r): 0.92–0.99 at key sites CREW LAKE: RMSE ~2.2%, PBIAS ~1.2%, r ≈ 0.989 Big Creek at Sligo presented poor correlation due to unreliable gage data and was excluded from overall performance metrics. Validation against 19 high water mark points in Monroe showed over 70% of modeled elevations within 1.0 ft of observed values, with several within ±0.2 ft. Spatial comparison of flood extents confirmed accurate ponding and inundation. Spatial comparison of flood extents confirmed accurate ponding and inundation.

Figure 3-26: Bayou Lafourche at Crew Lake calibrated to the March 2016 event (located western side of the Boeuf Watershed).

Figure 3-27: Big Colewa Bayou calibrated to the March 2016 event (On the eastern side of the Boeuf Watershed).

Boundary Condition Enhancements and Final Validation { Adjustments included adding weir-based lateral inflow from Bayou Bartholomew to simulate levee overtopping. { Boundary inputs from Ouachita River were extended temporally and aligned with observed hydrographs to capture backwater delays. { Refined pump and gate rules simulated operational shutdowns when external water levels (e.g., Ouachita) exceeded internal stage, preventing unrealistic reverse flows. { Final validation runs (e.g., May 2009 flood) showed strong alignment with observed hydrograph peaks and recession limbs, demonstrating model robustness.

LEARN MORE: B_4.6 +

27

Hydrologic and Hydraulic Modeling Methodology

BOEUF

MACON TENSAS

COCODRIE

3.6.2 Macon a. Hydrologic Model (HEC-HMS) HEC-HMS was used to model the Arkansas headwater portion of the Bayou Macon Watershed, and the results were used as an inflow boundary condition for the Bayou Macon 2D ROM HEC-RAS model.

LEARN MORE (M, APPENDIX 1).

Macon Arkansas Basin

BAYOU MACON ARKANSAS HEC-HMS MODELED AREA: 492 sq. mi. MODEL INFORMATION: 39 subbasins and 24 reaches. Figure 3-28 depicts the location of the watershed, the elements that make up the model, and the relevant gages. Rain-to-Runoff Transform The ModClark Transform method was used as previously discussed in Rainfall-Runoff Transformation . More information can be found in the Macon report: LEARN MORE: M, APPENDIX 1, 1.1.5 + Reach Routing For reach routing, the Modified Puls (ModPuls) method was used as previously discussed in Reach Routing Methods. More information can be found in the Macon report: LEARN MORE: M, APPENDIX 1, 1.1.7 + Loss Method For the soil losses, the Deficit and Constant method was used as previously discussed in Loss Method. More information can be found in the Macon report: LEARN MORE: M, APPENDIX 1, 1.1.4 +

Figure 3-28: Location of the Bayou Macon Arkansas model with outflow gage locations.

Groundwater The Linear Reservoir Baseflow method was used as previously discussed in Baseflow Representation . More information can be found in the Macon report: LEARN MORE: M, APPENDIX 1, 1.1.6 +

28

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

Lake Chicot and Lake Macon Modeling The Bayou Macon Arkansas HEC-HMS model incorporates two lakes, Lake Chicot and Macon Lake, that operate based on a rule system defined by water level and outlet control structures. HEC-HMS version 4.11 incorporates the use of rule-based reservoir operations. These rules allow for the modeler to set target maximum and minimum flow values to achieve target storage patterns for the lake. Calibration For the calibration process in Bayou Macon Arkansas, events were selected based on historical data and community accounts. Calibration was done through adjusting model parameters and the lake system. More details about the calibration process can be found in the Macon report.

LEARN MORE: M, APPENDIX 1, 1.2.1 +

EVENTS The events chosen for calibration were the same events used in the calibration process for the Boeuf Arkansas Watershed Section 3.6.1.

PARAMETERS

MODEL PARAMETERS

• Subbasin Time of Concentration • Storage Coefficient

• Antecedent moisture conditions for loss calculations • Routing parameters, such as Manning’s Coefficient

LAKE SYSTEM (SEE FIGURE 3-29) { Using the Water Control Manual for the Chicot Pumping Plant, it was determined that the pumping plant at Rowdy Bend diverts flow from Macon Lake to the Mississippi River depending on water levels. { Water exits Macon Lake via: • The Rowdy pump system into the Mississippi River • Following the Connerly Bayou Dam Spillway to the gravity gates and ogee spillway at Connerly Weir to exit downstream to Lake Chicot • Combination of these methods

Figure 3-29: Lake system represented in the model.

Recurrence Interval Modeling { Atlas 14-derived rainfall was used to derive key storms for recurrence interval modeling. See Storm Generation (Section 4.11) for additional information.

{ Results were validated by applying subbasin ARF to Atlas 14 subbasin average precipitation depth durations and applied through HMS hyetographs, ensuring the model results were reasonable when compared to the recorded gage data.

29

Hydrologic and Hydraulic Modeling Methodology

BOEUF

MACON TENSAS

COCODRIE

b. Hydraulic Model (HEC-RAS) Model Overview and 2D Mesh Design { The hydraulic modeling of the Bayou Macon Watershed was performed using the HEC-RAS 2D ROM approach due to its flat topography and complex flow patterns throughout the basin. { Starting at Eudora, Arkansas, the model domain covers approximately 42 sq. mi. in Arkansas and 481 sq. mi. in Louisiana. { Mesh resolution was kept coarse at 1,000 ft in rural and flat areas and refined to 100 ft in urban areas like Delhi, Epps, and Oak Grove. { Breaklines were used extensively along roads, channels, dams, and high ground to preserve important hydraulic features and realistic flow paths. Survey Data Incorporation Hydraulic Structures { A total of 55 bridges, 77 culverts, 12 weirs, 20 control structures, and 105 cross-sections were surveyed along priority streams. { The number of structures surveyed differed from the actual modeled structures due to boundary updates and HEC-RAS model refinements during calibration. { Structures were modeled using Storage Area/2D connections with surveyed data, including deck elevations, culvert diameters, and spillway profiles, where available. Figure 3-30 shows the modeled structures in the Macon HEC-RAS model. LEARN MORE: M, 2.7 & M, 2.7.1.2.3

Terrain Modifications { LiDAR data received during model setup was flown during high water levels, causing inaccurate bathymetry in some cases. { Survey data was used to interpolate channel depth, width, and slopes between structures and cross sections. Figure 3-31 shows channel bathymetry before and after survey incorporation. { Terrain was refined through several iterations to best represent conveyance and storage during several storm events. { Changes in terrain data were evaluated across multiple events to verify accuracy. { Buildings on terrain were raised 20 ft, and hydroconnectors were used to create openings for unsurveyed structures. Boundary Conditions { The model's northern boundary at Eudora, Arkansas, uses a flow hydrograph from the USGS gage to account for inflows from the Arkansas portion of the Macon Watershed. { The downstream boundary at the Macon and Tensas River outlet uses the Fool River Stage hydrograph to represent significant backwater impacts during historical events. { Normal depth is set at the Joes Bayou outlet below Swan Lake. { The remaining boundary condition lines use rating curves characterized using results from the downstream Tensas Watershed RAS model. Figure 3-32 shows the rating curve developed for Sutt Bayou. LEARN MORE: M, 2.4.2.3 +

30

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

Figure 3-30: Modeled bridge (top) and culvert (bottom) in the Bayou Macon HEC-RAS model.

Figure 3-31: Channel bathymetry before and after fitting to survey.

Model Calibration and Validation { Six historical storm events were used for calibration, chosen for their significant observed data coverage and relevance to current conditions. { Base water levels at the beginning of storm were assigned using the Initial Condition (IC) Points indicating observed stages at their locations. { Soils were identified as either "wet" or "dry" based on observed river levels and prior precipitation. Table 3-5 shows these scenarios that were represented using different initial deficit values, established through iterations for optimal model response during calibration. LEARN MORE: M, 2.7.3.2 & LEARN MORE: M, 2.7.3.3

31

Hydrologic and Hydraulic Modeling Methodology

BOEUF

MACON TENSAS

COCODRIE

Table 3-5 Soil properties used during model calibration and validation.

HYDRAULIC CONDUCTIVITY (INCHES/HOUR)

DRY SOIL - 40% SATURATED INITIAL DEFICIT (INCHES)*

WET SOIL - 100% SATURATED INITIAL DEFICIT (INCHES)*

MAXIMUM DEFICIT (INCHES)

TEXTURE

0.020

3.298

0.000

5.496

Silty Clay

0.010

2.923

0.000

4.872

Clay

0.000

0.000

0.000

0.000

Water

0.300

5.299

0.000

8.832

Silt Loam

0.040

3.787

0.000

6.312

Silty Clay Loam

0.100

4.982

0.000

8.304

Loam

1.200

5.501

0.000

9.168

Loamy Sand

0.040

3.845

0.000

6.408

Clay Loam

{ The encroachment of channel and overbank land cover was reassessed and corrected around Kilbourne and Como gages. { The channel Manning’s n values were revised throughout the major streams and were eventually set to 0.045 during calibration. { Overbank imperviousness from upstream to downstream Como was raised to 80%. { The model was validated using two historical storms. Model results showed strong agreement with observed data in terms of shape and timing. Figure 3-33 shows a calibration example Near Delhi. Table 3-6 details statistical metrics for the Spring 2016 and Summer 2022 storms—a calibration and a validation event, respectively. Table 3-6 Summary of gage calibration and validation statistics.

EVENT

GAGE

RMSE (%)

RSR

PBIAS (%)

r

Kilbourne

1.76%

0.21

0.01%

0.99

Near Delhi

1.94%

0.27

0.51%

0.99

Spring 2016

South Delhi

1.63%

0.27

0.67%

0.99

Como

2.70%

0.31

1.30%

0.97

Kilbourne

2.22%

0.38

1.72%

0.98

Near Delhi

1.98%

0.21

1.31%

0.99

Summer 2022

South Delhi

1.68%

0.22

0.10%

0.98

Como

5.95%

0.35

3.33%

0.97

32

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE

Limitations Key issues were pointed out in Tensas Calibration Limitations (Section 3.3).

LEARN MORE: M_5.2.1 +

Figure 3-32: Rating curve developed for the boundary condition on Sutt Bayou.

Figure 3-33: Simulated and observed stage at the Near Delhi Gage for the Spring 2016 Calibration Event

33

Hydrologic and Hydraulic Modeling Methodology

BOEUF

MACON TENSAS

COCODRIE

3.6.3 Tensas a. Hydraulic Model (HEC-RAS) Topographic and Bathymetric Survey, Channel and Lake Bathymetry

LiDAR data cannot penetrate water surfaces, leading to missing or inaccurate channel and lakebed elevations and resulting in elevation spikes in the raw DEM that distorted flow and storage representation. These issues were corrected through DEM modifications, including channel burning and hydro-flattening. Where available, survey data was used to adjust DEM shapes; otherwise, aerial imagery and land use data were used to estimate channel widths and lake bathymetry, especially in oxbow lakes where storage plays a key role. Accurate bathymetry and terrain representation were essential for calibrating the Tensas River Watershed model, ensuring peak stage timing at gage locations matched observed data. Figure 3-34 and Figure 3-35 show Boggy Lake before and after bathymetric updates.

Figure 3-34: Terrain at Boggy Lake before lake bathymetry update.

Figure 3-35: Terrain at Boggy Lake After Lake Bathymetry Update

b. HEC-RAS Boundary Options Outflows The Tensas Watershed discharges near the confluence with the Ouachita River, which causes significant backwater effects that impact flooding just as much as local rainfall. To model this influence, a stage hydrograph derived from observed data at Clayton, Harrisonburg, and Jonesville gages was applied at the watershed outlet. Figure 3-36 illustrates the downstream confluence.

c. HEC-RAS Model Calibration and Validation Calibration was performed using historical storm event data from USGS and USACE gages across the watershed (e.g., Transylvania, Tendal, Newlight, and Clayton). Model adjustments were made to align simulated results with observed data. Accurate calibration depended on realistic terrain storage, boundary conditions, and hydraulic structures like weirs. Figure 3-37 shows a calibration example at Southeast of Tendal; Table 3-7 details Fall 2008 storm metrics.

LEARN MORE: T_2.4.3.1 +

LEARN MORE: T_2.7.3 +

34

LWI REGION 3 TRAINING

LOUISIANA WATERSHED INITIATIVE