Abstract

Urban channel modification is increasingly employed to integrate green stormwater infrastructure (GSI) with public space, ecological restoration, and climate adaptation objectives. However, the system-scale hydrologic and water quality implications of such interventions—particularly the interaction between local treatment features and altered hydraulic connectivity—remain insufficiently understood. This study uses EPA SWMM to evaluate the performance of sediment forebays (SFBs) integrated into a modified urban “Savannah Channel” under three design storms (1-inch, 1-year, and 10-year), compared to a control scenario representing existing conditions.

Results indicate that individual sediment forebays can achieve high local total suspended solids (TSS) removal efficiencies (>95% for through-flow units and near-complete retention for offline units). However, system-scale responses varied by storm magnitude, with channel modification and sediment forebays substantially reducing sediment export during frequent and moderate storms while providing partial mitigation during extreme events.

These findings highlight the importance of right-sizing and strategic placement of sediment forebays and demonstrate the value of SWMM modeling as an iterative, pre-design tool that facilitates knowledge transfer between civil engineering and landscape architecture, informing spatial design decisions early in the design process.

1 Introduction

1.1 Urban stormwater and first-flush dynamics

Urban stormwater in dryland cities presents a unique challenge due to the interaction of highly episodic rainfall with long antecedent dry periods. During extended dry weather, organic debris, vehicular particulates, atmospheric dust, and street-level sediments accumulate on impervious surfaces (Burton and Pitt 2001; Li et al. 2008). When an intense rainfall event finally arrives, especially in the early minutes of the storm, these accumulated solids are rapidly mobilized, producing a “first flush” event in which a disproportionate share of the pollutant load is transported in the initial runoff pulse.

Sediment plays a central role in this process. Total Suspended Solids (TSS) represent not only a physical pollutant that smothers habitat and increases turbidity but are also a carrier for other contaminants. Particle-bound forms of phosphorus, non-polar organics, and heavy metals such as lead (Pb), copper (Cu), and zinc (Zn) are known to sorb strongly to fine particulates and organic fractions (Deletic 1998; Kayhanian and Stenstrom 2020). Studies have shown that particles smaller than 100 µm, despite representing only a small percentage of total mass, can account for a majority of metal loading in urban runoff. Sediment capture is therefore both a direct and indirect water quality objective (Sansalone and Buchberger 1997; Burton and Pitt 2001).

1.2 Green stormwater infrastructure in dryland watersheds

Green Stormwater Infrastructure (GSI) is typically presented as a decentralized complement to conventional gray systems, using vegetated systems such as swales, detention basins, and constructed wetlands to capture sediment, increase residence time, and promote infiltration (USEPA 2016). In semi-arid regions, however, the hydrology, sediment characteristics, and vegetation dynamics differ markedly from humid systems where much of the GSI literature and design guidance originates (Meerow et al. 2021).

In this context, channel-scale design strategies may play a significant role. The “Savannah Channel” concept is best understood as a site-specific application of alternative stormwater channel design strategies, developed by the authors as the subject of a Master of Landscape Architecture (MLA) thesis project. It integrates geomorphic channel shaping with vegetated benches and infiltration zones to create a wide, shallow, low-energy flow path that encourages sediment deposition, increases hydraulic residence time, and supports drought-tolerant vegetation. Paired with strategic inlet pretreatment, such systems may improve sediment capture performance without requiring continuous standing water typical of micropools or wetlands in humid settings.

1.3 Sediment forebays as inlet pre-treatment

Sediment forebays are small upstream settling basins located at the points where stormwater enters a larger treatment feature (McNett 2010; Wang et al. 2023). Their core functions are:

1. Energy dissipation,
2. Coarse sediment capture, and
3. Maintenance efficiency through localized cleanout.

While numerous stormwater manuals recommend sizing forebays as a percentage of permanent pool volume, fewer provide guidance for dryland contexts, where first-flush events are the dominant sediment delivery mechanism and fine particulates may play an outsized role (Figure 1).

Figure 1 Savannah channel cross-sections and plan view.

Several previous studies, such as Flindall and Basran (2001) and Gregory (2014), have employed SWMM to explore the performance of sediment forebays (SFB) from a water quality perspective. The present study examines how SFBs installed at each inlet of a large stormwater impoundment area can influence sediment loading and peak flow dynamics under design storm events typical of West Texas.

1.4 Study objective

The objective of this study is to evaluate, using EPA SWMM, how sediment forebays integrated within a modified stormwater channel influence system-scale hydrologic and water quality responses across storm magnitudes. The study deploys three comparative scenarios:

1. Control (“Do Nothing”). Existing condition with conventional detention, no channel modifications or forebays.
2. Savannah Channel regrading without SFBs. Channel modification to a Savannah Channel form without any upstream SFBs. Modifications balance the need for stormwater capture and the desire for elevated areas where various programmatic features are deployed (elevated paths and platforms for pocket parks, playgrounds, etc).
3. Savannah Channel regrading with SFBs. Same channel modification, with sediment forebays installed near each inlet. Each forebay fills to specified height and overspills into the larger Savannah channel.

By comparing scenarios and storm sizes, this modeling isolates the contributions of channel-scale design versus inlet-focused pretreatment. While speculative in nature, the goal is to demonstrate how SWMM can supplement the design development and community engagement process with rigorous and actionable insights on the likely performance of proposed interventions (Figure 2). 

Figure 2 Proposed greenspace interventions within the Savannah Channel (cross-sections and plan views).

In Figure 2, Notation 1 refers to the placement of Sediment Forebays (SFB), which are located around the periphery of the detention area, often near the intersection of perpendicular roads. These zones are intended to fully impound the runoff from smaller storms, protecting downstream receiving waters from harmful pollutant loads. Notation 2 refers to a transitional buffer zone of planted areas and pedestrian paths that would be subject to periodic inundation during larger rain events. Notation 3 refers to the flat area at the bottom of the channel that serves as open lawn or flexible green space when the channel is dry while providing significant storage volume for larger storm events when needed.

2 Methods and Materials

2.1 Study area

The study area consists of approximately 325 acres of residential and commercial land in the urbanizing southern periphery of Lubbock, Texas. The watershed comprises two residential neighborhoods (Oakmont Estates and Vintage Township) and a burgeoning commercial district to the west. Impervious cover ranges from 38% (low-density residential) to 75% (commercial parcels). Soils are dominated by Amarillo fine sandy loam, which exhibits moderate infiltration when uncompacted but low infiltration in disturbed urban conditions. The region has a semi-arid climate with long dry periods and occasional intense rainfall, making it representative of dryland stormwater patterns (Figure 3).

Figure 3 Multi-scale overview of the study area and its regional/urban context. Lubbock, TX is a growing city in the semi-arid, southern high plains region of the US.

The detention basin examined in this study is intended to function as a receiving facility for all runoff from the watershed. In the control condition, the basin is a relatively conventional form with a primary outlet and limited internal hydraulic complexity. The modified channel design introduces increased hydraulic complexity, a range of dual-use areas (open lawns and public amphitheater spaces that double as stormwater bioretention zones when needed). All system-wide hydraulic response is assessed from a single outfall point “Quaker_Out”, representing a 4-barrell box culvert that conveys excess flows from Savannah channel underneath a vehicular road (Quaker Ave.) and eastward through an unpaved trapezoidal channel to a nearby ~22-acre playa lake, (Lake “5056”). Playa lakes are internally drained, ephemeral depressions with no surface outflow, making them particularly sensitive to sediment accumulation and water quality degradation. Upstream interventions and channel modifications are therefore intended to proactively protect the further deterioration of this at-risk playa lake while also catalyzing multifunctional greenspace benefits to the upstream communities of Vintage Township and Oakmont Estates. 

This region is characterized by its many playa lakes (shallow depressions that formed in an otherwise flat topography over the course of thousands of years and serve as ephemeral water features). Today, The Savannah Channel receives runoff from the communities of Vintage Township and Oakmont Estates, with more development underway in the area.  The channel is currently considered among many residents to be an unprogrammed eyesore, signifying an opportunity to propose alternative landscape designs that increase programmatic uses within and around this space while still serving critical stormwater management functions.

2.2 SWMM configuration

The model includes 115 subcatchments and 17 outfalls. One outfall (“Quaker_Out”) represents the primary downstream outfall that serves as the locus for system-wide assessment of water quality and quantity (Figure 4), and the remaining 16 represent boundary conditions of peripheral roads, allowing excess flows to leave the model if needed. Twelve storage nodes (representing sediment forebays) capture flows entering from the major (streets) and minor (buried culverts) drainage networks into the Savannah Channel. The upstream overland flow network (2D mesh) is comprised of 35,145 conduits and 14,744 junctions. Three mesh boundary types were defined: 1) Blocks, 2) Streets, and 3) Savannah Channel, with varying spatial resolutions, roughness coefficients, and seepage rates. These inputs define the attributes that are subsequently imparted to the conduit objects that make up the 2D mesh. Subcatchments were delineated at the block scale using publicly available parcel boundaries from the City of Lubbock. Subcatchment outlets are set to flow into nodal junctions in the street network representing natural low points in the closest arterial road. Routing used dynamic wave simulation. Infiltration was modeled with the Horton method using soil parameters for Amarillo fine sandy loam.

Figure 4 Overview of the EPA SWMM model configuration illustrating (left) delineated subcatchments draining to a two-dimensional (2D) Thiessen polygon mesh representing the Blocks, Streets, and the Savannah Channel corridor.

The spatial placement of sediment forebay storage nodes (SFBs) at select street inlets (center) along the channel edge, and (right) the hydraulic linkage from inlet nodes through conduits to storage units and outlet weirs that return treated flow to the 2D mesh. Minor outfalls distributed along the channel boundary and the primary downstream outfall at Quaker Street were used to evaluate local and system-scale hydrologic and water quality responses. The inset highlights representative weir geometry and parameterization used to control forebay overflow.

Invert elevations for all junctions were set using a lidar-derived Digital Elevation Model at a 3 cm spatial resolution. Channel modifications and grading design were accomplished using standard site grading tools in Autodesk Civil 3D 2024 (Autodesk 2024). Using these surface creation tools, the lidar-derived surface representing an “existing ground” (EG) condition was modified to produce a desired (“proposed”) channel morphology (Figure 5). The process employed subtractive and additive designs to create a series of SFBs while also increasing the spatial utility of the Savannah channel and directing the flow path in a more refined and intentional manner as it flowed from northwest to southeast of the study area (Figure 5). Using C3D offers designers and engineers a series of built-in functions for calculating the balance of proposed cut and fill. While minimizing the need for excessive cut was not considered within the scope of this speculative project proposal, it has been suggested that balancing cut and fill operations is ideal wherever feasible (with all cut material used elsewhere on site as a bermed feature to minimize the expense of hauling away unused soil). Proposed surfaces were exported as GeoTIFF raster files and were used to assign invert elevations of 2D nodes and junctions within the Savannah channel.

Figure 5 Unmodified (left) and modified (right) channel designs.

Proposed channel modifications include deepened primary detention areas in the northern and southern extents and providing excavated SFB features at each inlet point to the system. Several areas are raised/leveled “platforms” for new play and active areas within the channel. Dry weather base flow is guided throughout a meandering channel toward the outfall at the southeastern end of the site (red triangle).

2.3 Scenario structure

Three design scenarios were evaluated to compare baseline performance with progressively more active treatment strategies.

Scenario A represents the existing condition (“control”). The detention basin retains its conventional geometry, and all inflows enter without inlet pretreatment.

In Scenario B, the basin was reconfigured using the Savannah Channel design, which creates a broad, shallow channel with vegetated benches. This is intended to reduce flow velocity, increase residence time, and promote sediment deposition naturally along the flow path.

Finally, Scenario C includes the Savannah Channel geometry with the installation of sediment forebays at each inlet point where stormwater enters the basin. Forebays were created as storage nodes in the model and sized to provide localized storage, energy dissipation, and coarse sediment capture prior to entry into the main channel flow.

Storage nodes in Scenario C were set to a depth to 5 ft, applying a common tabular storage curve, and setting the crest elevation of the weir objects draining the storage nodes to 3.5 ft. Conduits immediately upstream of storage nodes (inlets) were set with a 2 ft “outlet offset” and flap gate option set to “YES” to minimize backflow effects during the filling and ponding of the SFB nodes (Figure 6).

Figure 6 Profile view of a typical inlet/storage/weir configuration within SWMM. Water moves through a 2D overland mesh network and into dedicated inlet conduits.

These conduits feed storage nodes (representing SFBs). Storage nodes are set with a specific treatment expression that governs the removal of TSS (see Equation 1) and weir objects regulate the overflow dynamics of the SFBs as ponding depths in excess of the weir crest elevation (3.5 ft) are reached. “Spillover” flows are directed back to the 2D mesh network and eventually to a final downstream assessment Outfall (Quaker_Out). Twelve storage units/weir objects were deployed in the study and allowed for the localized assessment of water quality performance.

Scenario B was created by setting the depth of the storage unit, and the weir crest elevation of the components mentioned above, to “0”. Scenario A also contained these features with null values, but with node elevations taken from the existing rather than the proposed DEM surface.

Storm events

Each scenario was modeled under three SCS Type II (24H) design storms:

• 1-inch convective event (intense short duration, first-flush emphasis)
• 1-year, 24-hour storm (2.19 in)
• 10-year, 24-hour storm (4.22 in)

For each storm, antecedent conditions assumed full pollutant build-up to represent a conservative first-flush scenario.

2.4 TSS modeling approach

TSS were modeled as a proxy for sediment-bound contaminants, reflecting the pollutant profile most relevant to dryland stormwater systems.

Surface build-up

To capture land-use differences in pollutant generation, the following TSS build-up rates were assigned:

• 20 lb/ac—low-density residential
• 30 lb/ac—higher-density residential
• 60 lb/ac—commercial areas

These values fall within reported ranges for urban stormwater models and reflect greater vehicular load and imperviousness in commercial parcels.

Build-up was assumed to reach maximum capacity before each storm event.

Wash-off

Wash-off was simulated using SWMM’s exponential function:

• Residential: coefficient = 2.0, exponent = 0.52
• Commercial: coefficient = 2.5, exponent = 0.60

These differentiate the rate at which accumulated sediment becomes mobilized given the smoother surfaces and minimal storage on commercial parcels.

Forebay TSS treatment

Sediment forebays were modeled as SWMM storage units with tabular depth–area relationships based on a 3:1 side slope geometry and a maximum ponding depth of approximately 5 ft. Each forebay receives inflow via a dedicated conduit and discharges back to the channel via a weir set at the design crest elevation of 3.5 ft.

TSS treatment within storage units was represented using an exponential residence-time-based attenuation formulation (Equation 1):

(1)

Where:

Cout and Cin	=	outlet and inlet concentrations, respectively,
HRT	=	hydraulic residence time within the storage unit,
D	=	instantaneous water depth, and
k	=	lumped removal coefficient informed by sediment settling behaviour reported in the literature (Clary et al. 2020; Johnson 2007).

This formulation increases removal with residence time and reduces removal under deeper ponding conditions, capturing depth- and residence-time-dependent sediment attenuation without relying on global pollutant decay terms.

In SWMM syntax, this equation is interpreted and input as the following treatment expression within each storage node (Equation 2):

(2)

The coefficient (1.18) represents a conservative, lumped settling parameter corresponding to effective particle settling velocities reported for fine urban sediments, adjusted to reflect residence time in hours and depth in feet. Sensitivity testing indicated that this value produced plausible removal without over-predicting system-scale attenuation.

Forebay removal efficiency (local)

Event-scale TSS removal for each sediment forebay i was quantified using a mass balance between paired inlet conduits and outlet weirs (Equation 3):

(3)

Where:

Min,i and Mout,i

=

total TSS masses entering and leaving forebay i over the storm event.

Forebay-level removal efficiency was computed using Equation 4:

(4)

Additionally, for each of the design storms, forebays were individually classified as:

• No-spill (offline retention),
• Spill (through-flow treatment), and
• Under-fed (negligible inflow).

System scale performance

System-scale performance was evaluated at the watershed outfall (Quaker_Out) using event-integrated hydrologic and water quality metrics:

(5)

(6)

Where:

MTSS	=	total TSS mass exported during the event,
Vrunoff	=	total runoff volume during the event,
Ci	=	TSS concentration at time step i,
Qi	=	flow rate at time step i,
	=	duration of time step i,
i	=	model time-step index; and
N	=	total number of model time steps included in the event.

Appropriate unit conversions were applied during event integration to express TSS mass and runoff volume in the reporting units used below. Comparative system-scale performance among scenarios was assessed using differences in total exported mass, runoff volume, and peak discharge. TSS were modeled as a lumped parameter, recognizing that suspended solids encompass a range of particle sizes with differing settling characteristics; modeling multiple particle classes is beyond the scope of this study but represents an important area for future work.

2.5 Model scope and limitations

Formal calibration using observed flow or water quality data was not performed due to the speculative nature of the design and lack of site-specific monitoring data; instead, model parameters were selected based on literature values and regional soil characteristics to support comparative scenario analysis.

This study is intended as a comparative, pre-design modeling exercise rather than a calibrated prediction of site-specific performance. Model inputs were selected to represent plausible conditions for dryland urban stormwater systems, and results are interpreted in a relative sense across scenarios. Sediment forebay sizing was held constant to isolate the influence of placement and hydraulic connectivity rather than to optimize individual structure dimensions. Additional work is needed to evaluate sensitivity to forebay sizing, particle size distribution, and long-term hydrologic variability.

3 Results

3.1 System scale response at Quaker outfall

System-scale hydrologic and water quality responses at the Quaker_Out varied across storm magnitudes and design scenarios, with updated model outputs indicating more nuanced differences than previously observed (Table 1). Under the 1-inch storm, both modified channel scenarios reduced total runoff volume and peak discharge relative to existing conditions, reflecting increased localized detention and delayed routing within the redesigned channel corridor. TSS export was also reduced relative to the control scenario, with the greatest reduction observed when sediment forebays were included.

Table 1 System-scale hydrologic and TSS metrics at Quaker_Out for the 1-inch design storm.

Scenario	Storm	Peak Flow (cfs)	Total Vol (MG)	Total TSS (kg)	Δ TSS vs Control (%)	Δ TSS vs NoSFB (%)	Δ Vol vs Control (%)	Δ Peak vs Control (%)
A (Control)	1-inch	0.909	0.0334	3.83	–	–	–	–
B (No SFB)	1-inch	0.633	0.0275	2.94	-23%	–	-18%	–
C (with SFB)	1-inch	0.431	0.0178	1.85	-52%	-37%	-47%	-32%
A (Control)	1-year	36.89	3.2344	113.75	–	–	–	–
B (No SFB)	1-year	9.86	1.1923	58.04	-49%	–	-63%	–
C (with SFB)	1-year	7.84	0.9043	33.01	-71%	-43%	-72%	-26%
A (Control)	10-year	115.6	7.7482	124.89	–	–	–	–
B (No SFB)	10-year	117.9	9.6722	249.70	50%*	–	20%*	–
C (with SFB)	10-year	114.2	9.2792	183.77	47%*	-26%	17%*	-3%

NOTE: *Denotes negative removal efficiency (increase in flow or TSS) between the proposed design modifications vs. control (unmodified channel as built).

Specifically, the modified channel without sediment forebays reduced total runoff volume by approximately 18% and TSS export by 23% compared to existing conditions. The addition of sediment forebays further reduced TSS export by approximately 37% relative to the modified channel alone, corresponding to an overall reduction of more than 50% relative to the control scenario. Peak discharge was similarly reduced under the forebay scenario, indicating that local storage and controlled overflow influenced both hydrologic and water quality response during small storms.

For the 1-year storm, the modified channel configurations again reduced peak discharge, runoff volume, and TSS export relative to existing conditions. Compared to the control scenario, the modified channel without forebays reduced total TSS export by approximately 49%, while the inclusion of sediment forebays achieved a reduction exceeding 70%. Runoff volume and peak discharge exhibited similar proportional reductions, suggesting that channel regrading and distributed storage altered system-scale routing behaviour in a manner that attenuated both flow and sediment delivery during moderate storm events.

The 10-year storm produced a different response pattern. While peak discharge under both modified channel scenarios remained comparable to existing conditions, total runoff volume and TSS export increased modestly relative to the control scenario. The modified channel without forebays increased TSS export by approximately 50%, while the forebay scenario reduced this increase by approximately 26%, indicating partial mitigation of sediment loading under large storm conditions. These results suggest that during extreme events, system-scale storage capacity and conveyance dominate watershed response, limiting the relative influence of localized treatment features.

Across all storm magnitudes, sediment forebays consistently reduced TSS export relative to the modified channel without forebays. The magnitude of this benefit was greatest during smaller and moderate storms, where detention and residence time effects were most pronounced, and diminished during the 10-year event, where widespread inundation and high conveyance reduced interception efficiency. The observed increase in runoff volume and sediment export during the 10-year event reflects a shift in system behaviour under extreme conditions, where basin-scale storage is exceeded and flow routing dominates performance, rather than indicating model inconsistency.

3.2 Temporal flow and sediment dynamics at Quaker_Out

Comparative hydrographs and pollutographs further illustrate how channel modification and sediment forebays influence both the timing and magnitude of flow and sediment export (Figure 7). Under the 1-inch storm, the forebay scenario exhibited delayed runoff onset, reduced peak discharge, and lower TSS concentrations relative to both the existing channel and the modified channel without forebays. TSS concentrations in the forebay scenario rose more gradually and declined earlier, reflecting sediment capture and extended residence times within storage units.

Figure 7 Comparative hydrographs and pollutographs at the Quaker_Out for the (a) 1-inch, (b) 1-year, and (c) 10-year storms.

Time series illustrate differences in flow rate and TSS concentration across scenarios, highlighting the influence of channel modification and sediment forebays on peak magnitude, timing, and recession behaviour.

For the 1-year storm, hydrographs show substantial attenuation of peak flow under both modified channel scenarios, with the forebay configuration producing the lowest peak discharge and the most sustained recession limb. Pollutographs reveal a similar pattern, with reduced peak TSS concentrations and a dampened first-flush response under the forebay scenario compared to the control and no-forebay cases.

During the 10-year storm, hydrograph differences among scenarios were less pronounced at peak flow, consistent with the convergence of system-scale conveyance under extreme rainfall. However, pollutographs indicate that sediment forebays moderated peak TSS concentrations and reduced cumulative sediment export relative to the modified channel without forebays, even as overall loads increased compared to existing conditions.

Taken together, these temporal analyses reinforce trends observed in event-integrated metrics, demonstrating that sediment forebays influence not only total exported mass but also the timing and shape of flow and sediment delivery to downstream receiving waters.

3.3 Local sediment forebay performance (1-inch storm)

Analysis of individual sediment forebays revealed consistently high local TSS removal during the 1-inch storm, despite limited system-scale reductions in exported load. Most forebays did not overtop their outlet weirs during the event and therefore operated as offline retention features. In these cases, nearly all sediment mass that entered the forebay was retained locally, resulting in effective removal efficiencies approaching 100% for the intercepted flow.

A smaller subset of forebays overtopped during the 1-inch storm and discharged treated overflow back to the channel. Among these through-flow units, event-scale TSS removal efficiencies exceeded 96%, indicating effective attenuation under conditions of limited overflow and relatively long hydraulic residence times (Table 2). These results confirm that the modeled residence-time-based treatment formulation performed as intended at the scale of individual storage units.

Table 2 Event-scale sediment forebay performance for the 1-inch storm.

SFB ID	Inlet TSS Load (kg)	Spill Classification	Outlet TSS Load (kg)	TSS Removed (kg)	Removal Efficiency (%)
114TH_N	28.5	Spillover	0.53	28.0021	98%
114TH_W	27.8	No Spill	–	–	100%
115TH	32.9	Spillover	0.23	32.7024	99%
117TH	107.3	Spillover	4.63	102.7110	96%
118TH	6.5	No Spill	–	–	–
120th	86.8	Spillover	0.64	86.1628	99%
120TH_N	0.0	No Spill	–	–	–
121st	28.5	No Spill	–	–	100%
HEB_n	281.6	Spillover	0.51	279.671	100%
HEB_S	27.8	Spillover	0.46	27.2906	98%
SAV_N	109.5	Spillover	6.81	102.7270	94%
SAV_S	5.1	No Spill	–	–	100%

Despite strong local performance, the influence of sediment forebays on system-scale sediment export was constrained by hydraulic connectivity. Several forebays received minimal inflow during the 1-inch storm, while substantial portions of watershed runoff bypassed the forebay network entirely and were routed directly to the modified channel and downstream outfall. As a result, high removal efficiencies within individual forebays did not always translate into commensurate reductions in total TSS export at the watershed outlet.

The observed variability in forebay engagement underscores the importance of evaluating treatment features in relation to storm magnitude and flow routing rather than treatment efficiency alone. During small storms, many forebays function primarily as localized detention and sediment traps, whereas their system-scale influence depends on the proportion of watershed runoff they intercept. This distinction highlights the need to assess sediment forebays as components of a connected hydrologic network, rather than as isolated treatment units.

Five (out of twelve) SFBs did not overtop during the 1-inch storm event and therefore retained all sediment that entered them. Forebays that overtopped exhibited high removal efficiencies, confirming effective local treatment despite limited system-scale influence. Time series of water depth within representative sediment forebay storage units illustrates contrasting operational behaviour during the 1-inch storm event (Figure 8). Two SFBs (114^th and 115^th) received inflows, but ponding remains below the outlet weir crest and functions as an offline detention feature (no-spill), retaining intercepted runoff and sediment locally. Another SFB (114^th N) received significant inflows and reaches the weir crest elevation (3.5 ft) and discharges treated overflow back to the channel (spillover), activating the modeled residence-time-based sediment removal process. Horizontal dashed lines indicate the outlet weir crest elevation. Together, these examples demonstrate how storm magnitude and hydraulic connectivity govern whether individual forebays operate as localized storage or through-flow treatment features.

Figure 8 Storage depth time series for representative sediment forebays during the 1-inch storm, illustrating no-spill (offline detention) and spill (through-flow treatment) behaviour relative to the outlet weir crest elevation (3.5 ft).

The goal of SFBs to arrest the flow of water and pollutants proactively to reduce stress on downstream receiving waters, with ponded stormwater encouraged to slowly infiltrate and evapotranspirate in-situ over an extended period. Together, these findings could illustrate a possible scale mismatch between local treatment performance and watershed-scale outcomes, providing a basis for the discussion of right-sizing, strategic placement, and interdisciplinary design implications that follow.

4 Discussion

4.1 Scale mismatch between local treatment and system response

The results reveal a consistent scale mismatch: sediment forebays function effectively at the local scale, yet system-scale TSS export remains governed by hydraulic connectivity and storage distribution. During small storms, many forebays operate as offline retention features, while substantial runoff bypasses the forebay network entirely. While this study focused on a single forebay sizing strategy to isolate system-scale interactions, future analyses should explicitly evaluate alternative forebay dimensions to assess sensitivity to storage capacity and optimize treatment performance across storm magnitudes.

4.2 Right-sizing and strategic placement through pre-design modeling

Rather than representing optimized designs, the sediment forebays evaluated in this study were intentionally sized uniformly to support comparative analysis across scenarios. The results suggest that forebay performance is highly sensitive to both placement and hydraulic connectivity, and that structure sizing may further influence system-scale outcomes. Future work should evaluate alternative sizing strategies to identify configurations that balance sediment capture with system storage capacity.

4.3 Cross-disciplinary knowledge transfer

This study demonstrates how SWMM can serve as a shared analytical platform between civil engineering and landscape architecture. Modeling allows designers and engineers to operate with a shared language, testing how forebays activate across storm magnitudes and to explore trade-offs between spatial design objectives and hydrologic performance before a project is ultimately constructed (Figure 9).

Figure 9 Schematic overview of the workflow involved in the study. 

These feedback loops between modeling, design, and assessment offers many opportunities for input and insights on the likely performance of a given design at an early stage in the process, before costly mistakes or inefficiencies are implemented in the final, constructed project.

Channel regrading, programmatic design, and public space objectives shape hydrologic connectivity, while modeling results inform spatial design decisions. Iterative pre-design modeling enables collaborative refinement before formal design commitments are made. This study aligns with and reinforces various calls in recent years by theorists such as Felson and Pickett (2005) as well as Grose (2017) to engage in a more robust and interdisciplinary process based on performative design exploration.

Acknowledgments

The authors gratefully acknowledge Computational Hydraulics International (CHI) for providing educational access to PCSWMM software in support of graduate research conducted by Zahid Hasan. Hasan’s scholarship work informed the initial development of the modeling framework and design concepts further refined and extended in this study.

References

1. Autodesk. 2024. Civil 3D [software]. www.autodesk.com
2. Burton, G.A. and R. Pitt. 2001. "Stormwater Effects Handbook: A Toolbox for Watershed Managers, Scientists, and Engineers." CRC Press. https://doi.org/10.1201/9781420036244
3. Clary, J., J. Jones, M. Leisenring, P. Hobson, and E. Strecker. 2020. "International Stormwater BMP Database: 2020 Summary Statistics." The Water Research Foundation: Alexandria, VA, USA. https://www.waterrf.org/resource/international-stormwater-bmp-database-2020-summary-statistics
4. Deletic, A. 1998. "The first flush load of urban surface runoff." Water Research 32 (8): 2462–2470. https://doi.org/10.1016/S0043-1354(97)00470-3
5. Felson, A.J. and S.T.A. Pickett. 2005. "Designed experiments: new approaches to studying urban ecosystems." Frontiers in Ecology and the Environment 3 (10): 549–556. https://doi.org/10.1890/1540-9295(2005)003[0549:DENATS]2.0.CO;2
6. Flindall, R. and D. Basran. 2001. "Design of a Constructed Wetland for the Treatment of Glycol Contaminated Stormwater." Journal of Water Management Modeling R207-14. https://doi.org/10.14796/jwmm.r207-14
7. Gregory, M. 2014. "Stormwater Pond Sediment Loading and Accumulation Analysis." Journal of Water Management Modeling C378. https://doi.org/10.14796/jwmm.c378
8. Grose, M. 2017. Constructed Ecologies: Critical Reflections on Ecology with Design. Routledge, pp 208. United Kingdom.
9. Johnson, J.L. 2007. "Evaluation of Stormwater Wetland and Wet Pond Forebay Design and Stormwater Wetland Pollutant Removal Efficiency." North Carolina State University Theses, Biological and Agricultural Engineering. Raleigh, NC, USA. https://repository.lib.ncsu.edu/handle/1840.16/1590
10. Kayhanian, M. and M.K. Stenstrom. 2020. "First Flush Phenomenon and Its Application for Stormwater Runoff Management." Journal of Water and Wastewater 31, 6. https://doi.org/10.22093/wwj.2020.230456.3020
11. Li, M-H, M.E. Barrett, R. Pavitra, O. Francisco, and H.C. Landphair. 2008. "Documenting Stormwater Quality on Texas Highways and Adjacent Vegetated Roadsides." Journal of Environmental Engineering 134 (1): 48–59. https://doi.org/10.1061/(ASCE)0733-9372(2008)134:1(48)
12. McNett, J.K. 2010. "Proposing a New Method of Stormwater BMP Assessment and Evaluating the Toxicity of Forebay Sediments." North Carolina State University Theses, Biological and Agricultural Engineering. https://repository.lib.ncsu.edu/handle/1840.16/6269
13. Meerow, S., M. Natarajan, and D. Krantz. 2021. "Green infrastructure performance in arid and semi-arid urban environments." Urban Water Journal 18 (4): 275–285. https://doi.org/10.1080/1573062X.2021.1877741
14. Sansalone, J.J. and S.G. Buchberger. 1997. "Partitioning and first flush of metals in urban roadway storm water." Journal of Environmental Engineering 123 (2): 134–143. https://doi.org/10.1061/(ASCE)0733-9372(1997)123:2(134)
15. USEPA. 2016. "What is Green Infrastructure." Available at: https://www.epa.gov/green-infrastructure
16. Wang, H., Y. Tai, L. Huang, C. Yang, and H. Jing. 2023. "Analyzing Water and Sediment Flow Patterns in Circular Forebays of Sediment-Laden Rivers." Sustainability 15 (24): 16941. https://doi.org/10.3390/su152416941