Behaviors of Highway Culverts Subjected to Flooding: A Comprehensive Review

Abstract

Highway culverts are essential components of transportation infrastructure, designed to convey water beneath highways and protect embankments from flooding. However, extreme flood events often impose hydraulic loads, overtopping, and debris accumulation that can trigger erosion, scour, blockage, and in severe cases, catastrophic washout. This paper presents a comprehensive review of highway culvert behavior under flooding conditions, integrating insights from hydraulics, geotechnical engineering, and structural performance. The review is organized around four themes: (1) types of flooding and their interactions with culverts; (2) hydraulic performance during flood events; (3) common failure modes, including scour, debris blockage, and structural instability; and (4) mitigation strategies to enhance resilience. Advances in hydraulic modeling, including 1D, 2D, 3D, and CFD approaches, are summarized, with attention to their accuracy, applicability limits, and validation needs. Representative experimental, numerical, and empirical studies are grouped by common properties to highlight key findings and constraints. Finally, emerging research opportunities are discussed, including the need for quantitative relationships between culvert geometry and flood intensity, methods to assess structural capacity loss during flooding, and the integration of artificial intelligence and computer vision for rapid post-flood inspection. This synthesis establishes a foundation for more robust evaluation, design, and maintenance strategies, supporting the long-term resilience of highway culverts in an era of increasingly frequent and severe floods.

1. Introduction

Highway culverts are typically classified as single or multiple pipes, box culverts, and pipe-arches [1,2]. Many highways are located along or near rivers [3] making them vulnerable to flooding [4,5,6,7,8,9,10,11,12]. During flood events, culverts often experience high flow rates, which can lead to significant damage [8,13]. As a result, highway culverts are frequently considered in flood mitigation plans, with attention given to their type and performance under flood conditions [14,15,16,17]. Culverts play a critical role in maintaining the natural flow of watercourses intersecting highways, helping to prevent flooding and localized damage to civil infrastructure, the environment, and nearby properties [18,19,20,21,22,23].

Properly designed culverts accommodate the design discharge without causing excessive upstream water elevation that could overtop highways, embankments, or adjacent structures [3,16,24,25]. Design discharge is typically determined through statistical analyses of historical rainfall or flood data, or by referencing peak discharges in nearby or similar catchments.

However, when extreme rainfall or flood events exceed the design discharge, overtopping may occur [1,26,27,28,29,30]. This can result in large volumes of water flowing over the highway embankment, leading to severe road damage and, in extreme cases, a complete highway failure. Additional factors such as flood duration, flow velocity, debris accumulation, and upstream backwater levels further influence a highway’s vulnerability to flooding [31,32]. Floods are among the most destructive natural disasters, causing widespread damage to civil infrastructure and posing serious risks to public safety and economic stability [26,33,34,35,36,37,38,39,40,41,42,43,44,45]. With the increasing frequency and severity of flood events due to climate change, understanding their impacts on highway culverts has become increasingly important [31,46]. Extensive research has been conducted on the flooding impacts on highway culverts, primarily focusing on (1) scour and debris blockage caused by flooding and (2) the hydraulic characteristics of culverts under flood conditions. The structural behaviors of highway culverts have not been quantitatively evaluated in previous studies.

The objective of this review paper is to summarize the state-of-the-art research on behaviors of highway culverts under flooding, with the goal of helping researchers and engineers better quantify flood-induced damage to highway culverts. This review covers the following key aspects: (1) types of flooding and its relationship with highway culverts; (2) hydraulic performance of highway culverts during flood events; (3) failure modes of culverts under flooding; and (4) mitigation strategies. In addition, future research directions are proposed to support the development of effective methods for assessing culvert damage caused by flooding.

2. Methodology

Flooding refers to the overflow of water that submerges land, causing a significant rise in water levels. Recent studies predict that the frequency of flooding will increase in many regions over the next two decades [29,47]. Elevated water levels combined with strong currents can erode foundation soils and lead to catastrophic structural failures, even during large-scale flood events with relatively low flow velocities. To better understand these dynamics, empirical correlations and flood-related scenarios have been developed and analyzed [1,48,49,50]. This comprehensive review draws upon 115 sources, including peer-reviewed articles, government reports, and documents from non-governmental organizations, gathered from academic databases such as Google Scholar, Web of Science, and ScienceDirect. The literature search was guided by keywords such as flood, flooding events, highway culverts, flood damage, flood mitigation techniques, highway embankment flooding to thoroughly investigate the impacts of flooding on highway culverts.

3. Types of Flooding and Its Relationship with Highway Culverts

Flooding is among the most severe natural hazards worldwide, affecting both developed and developing nations, with intensifying impacts in recent decades [1,51,52,53,54,55,56,57,58,59,60,61,62]. Culverts are structures that channel water beneath infrastructure such as roads or driveways and are typically installed at natural or artificial channel crossings along roadway embankments or drainage channels. Their design depends on several factors, including the roadway or driveway profile, the geometry and roughness of both the road and culvert, upstream watershed characteristics, and the magnitude of flow [1]. Floodway flows through culverts may exhibit either free-flow or submerged-flow behaviors [31].

In free-flow conditions, discharge is primarily controlled by the upstream water level, while in submerged conditions, both headwater and tailwater levels influence discharge [63]. Culverts provide significant advantages in floodway design by mitigating the rise in upstream water levels caused by embankments and reducing water ponding. They also elevate tailwater levels, which in turn decreases the need for extensive downstream batter protection. While high velocities over roadway shoulders can lead to erosion or scour at the downstream toe, this risk diminishes as flow velocity increases and separates from the embankment batter. Notably, flow velocities along the downstream batter are generally lower than those under free-flow conditions. Floodway discharge is typically estimated using open channel flow analysis that incorporates tailwater height and approach velocity, with the floodway length measured between sag points in the roadway profile.

Based on the failure modes of transportation facilities during and after flood events, Johnston et al. [64] classified flood events into four types, as illustrated in Figure 1:

(1)

Offset Head Floods. Offset head development occurs when highway embankments act as temporary dams, as shown in Figure 1a. In this scenario, water accumulates on one side of the embankment, with only partial drainage through the culvert. This can lead to internal soil erosion within the embankment. The combination of water head differences and internal erosion may result in partial or complete embankment failure.

(2)

Overtopping Floods. Overtopping floods occur when floodwater flows over the top of the embankment, completely submerging the highway embankment and culverts, as shown in Figure 1b. Erosion of the downstream slope due to overtopping is widely recognized as a major cause of highway culvert failure [65].

(3)

Basal Floods. Basal floods develop with or without soil saturation, as shown in Figure 1c. The presence of shallow water at the slope toe of a highway embankment raises water levels within the slope, reducing soil strength and increasing the risk of slope failure.

(4)

Above Slope Floods. Above-slope floods develop above cuttings along highway embankments, as shown in Figure 1d. Piping development and seepage outflow can lead to localized slope failures [66].

The associated failure mechanisms include internal soil erosion, rapid drawdown, scouring, piping, and surface erosion. The failure modes of highway embankments during flood events depend on various factors, including but not limited to soil type, flow rate, flood duration, culvert size, and highway slope dimensions.

4. Hydraulic Performance of Culverts

Flow characteristics at highway culverts depend on several factors, including culvert geometry and size, tailwater depth, embankment slope, drainage area, and rainfall intensity [23,67,68,69]. The geometry of the culvert and the configuration of the upstream and downstream channels strongly influence flow conditions at both ends of the structure [70]. Depending on these variables, culvert hydraulics may operate under inlet control, where headwater depth, culvert geometry, and inlet configuration dominate, or under outlet control, where tailwater elevation, barrel roughness, and culvert length are more critical. The resulting flow can be either pressurized or free-surface and may occur in subcritical, critical, or supercritical regimes, each of which dictates velocity distribution, energy dissipation, and scour potential. At the culvert entrance, the flow may converge, diverge, or align parallel with the culvert axis, while variations in boundary conditions—such as tailwater fluctuations—often produce transitions and flow disturbances [1]. In addition, the approach flow can be either skewed or perpendicular depending on the alignment between the culvert axis and the approach channel. Recognizing these flow types and criteria is essential for interpreting hydraulic performance and for comparing experimental, empirical, and numerical studies on culvert hydraulics.

Hydraulic modeling employs physically based numerical methods aimed at understanding river flow behaviors [71,72,73,74,75,76,77]. Modeling techniques vary based on flow complexity and temporal variations which can be categorized as one-dimensional (1-D), two-dimensional (2-D), or three-dimensional (3-D) [46,78,79]. The 1-D approach assumes that velocity components perpendicular to the primary flow direction are negligible compared to the longitudinal velocity [80]. Two-dimensional modeling techniques are routinely applied to nonuniform, unsteady free-surface flows in alluvial channels and estuaries to investigate flow velocities, distributions, and flow patterns [81]. Water surface profiles developed from 1-D and 2-D models agree with a computed-weighted mean of observed water surface profiles. Significant flow phenomena can be identified in 2-D modeling that may be masked in 1-D output. A guidance statement on the use of modeling techniques has been developed, along with selection criteria, recommendations on model applicability, and discussion of each approach’s strengths and weaknesses [82]. Abegaz et al. [50] summarized four widely adopted hydraulic models for flood simulation along with their procedure and required parameters. Computational fluid dynamics (CFD) has been widely used to investigate culvert failures under extreme flooding, utilizing various software programs such as OpenFOAM (v2506), ANSYS Fluent (2024R1), and Flow 3D (2025R1). These numerical models, which couple the Volume of Fluid (VOF) method with the Navier–Stokes equations, play a crucial role in simulating culvert behavior during flooding [83,84]. Flow 3D has proven effective in simulating local scouring at culvert outlets, as demonstrated by Hien and Chien [85]. Streftaris et al. [86] employed numerical modeling to simulate debris accumulation at a trash screen near a culvert outlet, linking blockage probability to the surrounding environment. Sorourian et al. [67,75] examined the effects of partial inlet blockage on scour characteristics at culvert outlets, concluding that the blockage ratio is the primary factor influencing scour depth. Taha et al. [76] used Flow 3D to model partial blockage conditions in box culverts and found that internal blockages had a limited impact on scour depth. They also developed an empirical formula correlating the blockage ratio with the submergence ratio and Froude number.

Table 1. Comparison of 1D, 2D, 3D, and CFD models for culvert hydraulics, with advantages, limitations, and typical applications.

Model Type	Advantages	Limitations	Typical Applications/Real-Life Relevance
1D models (e.g., HEC-RAS, HY-8)	Simple setup, fast computation, minimal data needs; widely accepted for regulatory/design use	Cannot capture lateral or vertical flow variations; limited in complex geometries	Preliminary culvert sizing; estimating water surface profiles; regulatory floodplain mapping
2D models (e.g., SRH-2D, MIKE21)	Captures lateral velocity distributions, secondary flows, and inundation extent; better for irregular terrain	Higher data and computational requirements; resolution limited by grid size	Floodplain mapping near culverts; evaluating overtopping, flow around embankments; sediment transport
3D models (e.g., FLOW-3D, OpenFOAM RANS/LES)	Resolves vertical structures, turbulence, vortices, and nappe flows; suitable for local hydraulics and scour	Computationally intensive; requires detailed boundary conditions and calibration	Outlet scour prediction; debris blockage effects; detailed velocity/turbulence distribution at culvert entrances/exits
CFD (Computational Fluid Dynamics) (RANS, LES, DNS)	Most flexible and detailed; can simulate multiphase flows, debris, sediment transport, and complex geometries; allows sensitivity testing	Very high computational cost; accuracy conditional on mesh, turbulence closure, and V&V; requires expertise	Research studies; design of culverts under complex flow regimes; evaluation of extreme events (e.g., debris impact, dam-break scenarios)

provides a comparative summary of 1D, 2D, 3D, and CFD modeling approaches commonly applied in culvert hydraulics [87]. Each method offers distinct advantages and limitations depending on the complexity of the flow conditions, data availability, and computational resources. While 1D models are widely used for preliminary design and regulatory assessments due to their simplicity and efficiency, they cannot capture lateral or vertical flow variations. In contrast, 2D models improve the representation of velocity distributions and floodplain dynamics but require greater data resolution. For more localized analyses, 3D models provide detailed insights into turbulence, vortices, and scour processes at culvert inlets and outlets, though at a higher computational cost. CFD approaches extend this capability further, enabling multiphase flow simulations, sediment transport, and debris interactions, but their accuracy depends heavily on mesh quality, turbulence modeling, and verification and validation practices. Together, these modeling frameworks illustrate a hierarchy of tools that can be selected according to the problem scale, desired accuracy, and available resources.

CFD can provide powerful insights into culvert hydraulics, but its accuracy is highly dependent on modeling choices and numerical practices. Achieving reliable results requires careful attention to turbulence model selection, mesh quality (including skewness, non-orthogonality, and aspect ratio), near-wall treatments, time-stepping strategies, and rigorous verification and validation (V&V) through grid-independence and sensitivity analyses. The key “accuracy levers” for developing more trustworthy culvert simulations are summarized in

Table 2. Key factors influencing CFD accuracy in culvert simulations and recommended best practices for improving model reliability.

Factor	Influence on Accuracy	Recommended Practice
Turbulence model selection	Governs prediction of separation, mixing, and energy dissipation	Match model to flow regime (e.g., k–ε/k–ω SST for bulk flows; LES/scale-resolving for complex vortices)
Mesh quality and resolution	Poor skewness, non-orthogonality, or high aspect ratios can cause instability and error	Use smooth grading, low skewness, refined cells near jets, vortices, and scour regions
Near-wall treatment	Inaccurate wall modeling leads to errors in boundary shear stress and roughness representation	Apply wall functions or low-Re models consistent with culvert roughness/ks or Manning equivalent
Time-stepping strategy	Large time steps miss unsteady features; unstable CFL numbers cause divergence	Choose Δt to satisfy CFL < 1 (transient runs); ensure temporal refinement for hydrographs
Boundary condition realism	Unrealistic inflow, tailwater, or blockage setup distorts hydraulics and scour response	Implement measured or physically consistent hydrographs, tailwater levels, debris scenarios
Verification & validation (V&V)	Without V&V, numerical results may appear converged but be physically inaccurate	Perform grid-independence tests, sensitivity analyses, and compare against lab/field data

. By systematically addressing these factors, CFD-based culvert studies can achieve greater accuracy and reliability, allowing numerical results to serve as credible complements to experimental observations and empirical equations.

Table 3. Representative studies grouped by common properties, with key results, validity limits, and accuracy notes.

Group	Study (Method)	Setup/Conditions	Principal Hydraulic Result	Reported/Explicit Validity Limits	Notes on Numerical/Experimental Accuracy
Outlet scour & blockage	Sorourian et al. [67,75,88] (experiments)	Box culverts; unsteady & steady flows; partial inlet blockage	Blockage ratio is the primary control on scour depth; peak scour on rising limb of hydrograph	Specific to tested box culvert geometries and blockage ratios; tailwater and hydrograph shapes as tested	Experimental; qualitative accuracy inherent to measurements; used to benchmark later models
	Taha et al. [76] (CFD + correlation)	Box culverts; varying blockage & submergence; Flow-3D (VOF)	Internal blockages showed limited impact on scour depth; developed empirical relation linking blockage ratio, submergence ratio, and Froude number	Valid within studied ranges of blockage, submergence, and Froude number	CFD calibrated to experiments (case-dependent); emphasizes parameter-space limits rather than global generality
	Tan et al. [89] (experiments)	Various culvert outlet conditions	Linear relation between maximum scour depth and hydraulic radius for unblocked & partially blocked cases	Valid for shapes/flows tested; extrapolation beyond ranges cautioned	Lab data provide direct envelopes for design-stage checks
	Liriano et al. [90] (experiments)	Turbulent jets at culvert outlets	Mean velocity identified as key parameter for outlet scour initiation/growth	Jet and tailwater conditions as tested	Experimental accuracy; informs velocity-based design thresholds
	Abt et al. [91,92,93] (empirical/experiments)	Multiple culvert shapes and slopes	Slope strongly affects outlet velocity/discharge & transport capacity; shape has limited effect on outlet scour	Ranges of slopes and shapes investigated	Empirical equations commonly cited for preliminary sizing
Debris & blockage risk	Streftaris et al. [86] (numerical/probabilistic)	Trash screens near culvert outlets	Blockage probability linked to surrounding environment/vegetation and flow	Dataset and site conditions used in the model	Statistical fit quality reported within study scope
Local scour & dam-break loads	Hien & Chien [85] (2-D/3-D numerics)	Dam-break flows on structures; Flow-3D	Demonstrated capability to reproduce impact forces and local scour patterns in fast transients	Dam-break scenarios and geometries tested	Numerical comparisons to measurements within case studies
3-D flow for scour estimation	Olsen & Kjellesvig [83] (3-D CFD)	Complex 3-D culvert/channel flows	Early demonstration that 3-D numerics can estimate maximum local scour depth	Limited by then-available turbulence/mesh strategies; concept still relevant	Highlights need for modern turbulence & mesh best practices
Model dimensionality guidance	Deal (KDOT) [81]; West [82] (guidance/benchmarks)	1-D vs. Two-dimensional river/canal hydraulics	2-D captures secondary circulations masked in 1-D; water surface profiles comparable where assumptions hold	Valid where quasi-1-D assumptions (small transverse velocities) apply	Field/model comparisons; choose dimension to match physics

summarizes representative studies on culvert hydraulics and scour, grouped according to common research themes. The table highlights key findings, the specific conditions and validity limits under which each study was conducted, and the degree of accuracy reported in experimental or numerical approaches. This structured synthesis emphasizes how blockage ratio, culvert slope, tailwater conditions, and flow dimensionality emerge as recurring determinants of hydraulic performance and scour behavior. At the same time, the table illustrates the conditional accuracy of CFD-based studies, where reliability depends on turbulence modeling, grid quality, and validation against experimental benchmarks. By organizing the literature in this way, the review provides a concise reference that connects study-specific insights to broader design and modeling implications.

Increasing global temperatures have escalated the severity and frequency of storm events, often overwhelming culverts and washing out roadways [94,95]. Predicting hydrographs for large storms or multiple storm centers challenges traditional hydrologic models, which tend to forecast unrealistically broad, flat hydrograph peaks [67,75]. Consequently, stormwater management must address more frequent extreme storm intensities and volumes [96]. As a key decision point in drainage design, accurate hydrograph modeling is essential. Various hydrologic methods, such as Soil Conservation Service Curve Number SCS CN, Horton, Green-Ampt, and CN-Decay, offer prospective solutions for hydrographs. Although mainly focusing on urban conditions, methods like inverse-distance weighting (IDW), antecedent moisture condition (AMC), and storm mean retention (SMR) also contribute to the hydrograph prediction [97]. Pipe culverts are among the most prevalent drainage structures in transportation infrastructures. Under design stormwater conditions, they are rarely overtopped and exhibit straightforward flow characteristics. Consequently, the existing hydraulic models for pipe culverts under overtopping stormwater run-off are not available, which highlights the urgent need to develop appropriate hydraulic models for highway culverts under overtopping conditions.

5. Failure Modes of Culverts During Flooding

Culverts play a critical role in water management, transferring surface water from one side of a highway embankment to the other [18,98]. The design and construction of culverts adopt multidisciplinary approaches, combining hydraulics, structural engineering, geotechnical engineering, and policy analysis, to accommodate anticipated hazards such as flooding [1]. Flooding may initiate with an overtopping hazard or capacity exceedance, but it can ultimately induce structural damage through soil scour, hydraulic uplift, or debris impact [31]. A comprehensive understanding of the behaviors of highway culverts under flood conditions, including hydrologic, hydraulic, and structural responses, can guide the assessment and mitigation of these hazards at both design and operation stages.

While comprehensive work has been conducted on the serviceability and ultimate behaviors of highway culverts for normal loading conditions, few studies have addressed culvert behavior under flooding hazards, partly due to the complexity of related phenomena and the scarcity of experimental and numerical data [18]. A multidisciplinary review integrating hydraulics, structural engineering, geotechnics, and policy is presented to elucidate the behavior of highway culverts in flooding events. Highway culverts must accommodate high discharge flows and withstand earth and hydrostatic pressures; structural components are designed accordingly, with top slabs bearing vehicular loads and abutments resisting the remaining stresses [31]. When overwhelmed by floodwater, various failure modes compromise structural integrity [63].

Flooding events produce significant hydraulic loads on culvert structures, which can lead to catastrophic failures [99,100]. These loads induce a variety of damaging mechanisms at multiple scales that can affect the structural performance of the culverts [63]. The main failure mechanisms for highway culverts subjected to flood loading were classified into damage due to debris accumulation, scour/erosion, and structural failure; and loss of functionality due to clogging, loss of structural stability, and overtopping [31]. The flow conditions during flooding may also exacerbate existing issues such as sliding and uplift [101]. Such failure modes reveal the vulnerable areas of the highway culvert system and which areas to target for mitigation and improvement. The forces exerted on the culvert system during a flood depend on the flow regime, duration, and release type. Similarly, soil-culvert interaction during flood loading depends on the soil type, density, and degree of saturation. Greene [102] summarized that the failure of floodway structures under extreme flooding is primarily caused by four factors: scour (soil erosion), washout, structural failures, and debris blockage. AASHTO [103] classified flood forces into five types: hydrostatic force (F_h), buoyant force (F_b), drag force (F_d), lift force (F_l), and overturning moment (M_o). Figure 2 shows the forces on a flood-affected culvert.

• Hydrostatic force (F_h) results from the difference in water levels upstream and downstream of a flood-affected culvert.
• Buoyant force (F_b) is equal to the weight of the water displaced by submerged culverts.
• Drag force (F_d) is generated by the pressure of flowing water in the direction of the flow.
• Lift force (F_l) is caused by the pressure of flowing water acting perpendicular to the water surface.
• Overturning moment (M_o) is the moment created by the imbalance of the previously mentioned forces.

These forces can be calculated as follows which were originally proposed for flood-affected bridges:

$$F_h= {\displaystyle \frac{1}{2}}{\rho }_{w}ghA$$

(1)

$$F_b={\rho }_{w}gV$$

(2)

$$F_d={\displaystyle \frac{1}{2}}{\rho }_w{C}_d{A}_d{v}^2$$

(3)

$$F_l={\displaystyle \frac{1}{2}}{\rho }_w{C}_l{A}_l{v}^2$$

(4)

$$M_o={\displaystyle \frac{1}{2}}{\rho }_w{C}_{M}L{W}^2{v}^2$$

(5)

where ρ_w is water density; g is gravitational acceleration; h is depth of water, A is projected area of the submerged portion of the bridge; A_d is projected area of the submerged portion upon which drag force is acting and normal to the flow direction; A_l is projected are of the submerged portion upon which lift force acts and parallel to the flow direction; V is volume of the submerged portion of the bridge; v is average flow velocity; L is the length of the bridge; C_d, C_l, and C_M are coefficients for drag force, lift force, and overturning moment, respectively.

Scour (soil erosion) is widely recognized as one of the major causes of culvert failure during flooding [104]. Extensive research has been conducted to predict scour depth at culvert outlets caused by flooding. Based on field test results, Ruff and Abt [105] proposed an empirical formula for predicting scour depth at culvert outlets. Lriano et al. [90] experimentally investigated the effects of turbulent flow on scour at culvert outlets. They evaluated the influence of turbulence intensity, mean velocities, near-bed bursting structures, and Reynolds stresses, concluding that mean velocity is the key parameter for scouring at the culvert outlet. Ali and Lim [106] discovered the influence of tailwater depth on scour behavior. Abt et al. [91,92,93] found that the culvert shape has a limited effect on the outlet scour but the culvert slope is a key parameter in estimating culvert flow velocity, discharge capacity, and sediment transport capacity. Papanicolaou et al. [107] examined the mechanism of outlet scour and assessed the impact of blockage on scour hole geometry, finding that blockage is a dominant factor in scour development at culvert outlets. Tan et al. [89] analyzed existing scour experiments considering outlet conditions, and culvert shapes. Their findings indicated a linear relationship between maximum scour depth and hydraulic radius under both unblocked and partially blocked conditions. Similarly, Sorourian et al. [88] experimentally studied the scouring process in partially blocked and unblocked culverts, concluding that maximum scour depth occurs during the rising limb of all hydrograph stages. Taha et al. [76] investigated the effects of various inlet blockage and submergence ratios on culvert efficiency and scour depth under different flow rates. They found that, at a given flow rate, maximum scour depth decreases as the submergence ratio increases. However, for blocked culverts, maximum scour depth increases with higher blockage ratios and Froude numbers. Jaeger and Lucke [108] investigated debris transport behavior in a natural channel and found that stream depth and width determined the mobility of debris and vegetation is the main factor in debris accumulation [18].

Culvert failure modes can be induced by various causes. This may stem from hydrologic or hydraulic factors like surcharges, blockages, floods, or buoyancy, and structural issues such as debris, scour, settling, uplift, or soil bearing failures [31].

Loss of surrounding medium from washout or scour destabilizes structures, while hydraulic blockages affect both the structure and its soil cover. Incoming flow is tied to upstream water levels, and outflow rates depend on downstream conditions. Buoyancy can reduce effective weight, increasing failure risk. Structural issues arise from three failure modes: sliding, bearing, and rotational. Sliding involves soil bearing capacity and foundation contact; bearing failure occurs when surface soil cannot support the load; and rotational failure pertains to heaving or punching beneath the foundation, reliant on soil shear strength [18].

The development of scour and the reduction in soil strength during flooding can significantly impact the load-bearing capacity of highway culverts. Additionally, traffic loading during flooding may further exacerbate the failure of flood-affected culverts. However, to the best of the authors’ knowledge, no studies have specifically examined the effect of traffic loading on the stability of highway culverts. Some research has been conducted on the impact of flooding on railway embankments. For instance, determining a critical train speed during flooding is essential to limit ground vibrations and prevent material weakening [109]. Jiang et al. [110,111] and Bian et al. [112] investigated the effects of high-speed trains on embankment slopes under flooding through full-scale tests, concluding that internal soil erosion within the embankment was a primary cause of failure.

6. Mitigation Strategies

Ahmari et al. [113] reviewed the design methods adopted by state DOTs to account for flooding impacts on bridges. Most state DOTs across the U.S. adopt a design flood with a return period of 50 to 500 years for bridge design, along with a minimum vertical clearance of 1 ft (0.3 m). Mississippi DOT sets a higher standard, requiring a minimum freeboard of 3 ft (0.9 m). To mitigate flood-related bridge failures, common design strategies include the use of riprap and reinforced foundations, while shear keys are implemented to counteract lateral forces induced by flooding. The FEMA [114] identified the primary causes of highway culvert failures as insufficient capacity, high-velocity flows, and debris impact leading to plugging. The major failure modes include erosion and scour, inundation and washout, and debris accumulation. This handbook [114] offers various options to mitigate different failure modes as shown in

Table 4. Mitigation Solutions of Highway Culvert Failures Caused by Flooding (Modified from [114]). (✓ denotes the current solution option is effective to mitigate this specific problem).

Solutions	Options	Effective to Mitigate Problem(s)
		Erosion and Scour	Inundation and Washout	Debris Impacts and Plugging
Increase drainage capacity	Increase ditch capacity	✓	✓	✓
	Replace a culvert with a box or arch culvert	✓	✓	✓
	Replace a culvert with a bridge	✓	✓	✓
	Add pipe culverts	✓	✓	✓
Reduce embankment erosion	Shape culvert entrance	✓	✓
	Construct a cutoff wall	✓	✓
	Install appropriate culvert end sections	✓	✓
	Install lining in the ditch	✓	✓
	Install check dams	✓	✓
	Construct an energy dissipater	✓	✓
Improve alignment	Realign culvert	✓	✓	✓
	Install approach berms	✓	✓
	Install flow diverters	✓	✓
	Install additional culverts	✓	✓
	Realign the stream channel	✓	✓
Reduce obstructions	Install an entrance debris barrier	✓		✓
	Install a sediment catch basin upstream			✓
	Install a relief culvert	✓	✓	✓
Relocate or replace with a water crossing	Relocate culvert	✓	✓
	Add a low water crossing	✓
	Add a high-water overflow crossing	✓

. It summarizes engineering strategies to mitigate culvert failures under flood conditions by linking specific options to the primary hazards they address, including erosion and scour, inundation and washout, and debris impacts/plugging. Increasing drainage capacity (e.g., larger culverts, additional culverts, or bridges) provides the broadest protection, reducing all three hazards by improving conveyance and reducing overtopping. Erosion-control measures such as cutoff walls, energy dissipaters, and end sections primarily mitigate scour and washout. Alignment improvements (e.g., realignment, berms, flow diverters) enhance hydraulic efficiency and reduce localized stresses, addressing erosion and washout, with some benefits against debris plugging. Debris-control measures, including entrance barriers, sediment catch basins, and relief culverts, directly reduce blockage risks while also supporting capacity. Finally, relocation or alternative crossings (low-water or overflow crossings) offer additional resilience where culverts are insufficient. Together, these solutions highlight that a combination of approaches is typically required to mitigate the full range of culvert failure mechanisms.

7. Future Research Directions

This review synthesizes a range of studies that collectively highlight the multifaceted challenges and innovative solutions related to the behaviors of highway culverts subjected to flooding. Based on this review paper, the authors would suggest the following topics for future studies on flooding impacts on highway culverts:

(1)

Failure modes in relation to highway culvert configuration and flooding intensity. Although a range of failure modes for highway culverts has been identified in the literature, there is no established methodology to systematically correlate these failure modes with specific culvert configurations and flooding intensities. This gap prevents the development of targeted preventative measures and mitigation strategies. Establishing predictive relationships that integrate culvert geometry, structural characteristics, and hydrologic loading conditions would improve vulnerability assessment and enable proactive flood risk management.

(2)

Quantifying structural capacity reduction in highway culverts due to flooding. Because highway culverts are embedded within the ground, assessing the reduction in their structural capacity under flooding conditions remains a significant challenge. Existing research has primarily emphasized hydraulic performance and failure modes, while limited attention has been given to quantifying structural degradation induced by flooding. The absence of reliable methods for such quantification constrains accurate risk evaluation and may lead to unrecognized safety hazards. Developing systematic approaches to measure structural capacity reduction during and after major flood events would fill this critical research gap and support timely safety warnings and effective mitigation strategies.

(3)

Advanced and rapid highway culvert inspection leveraging AI and computer vision [115]. Flood-induced external loads on highway culverts can cause cracks, voids, and internal soil erosion, all of which contribute to the structural vulnerability of highway culverts. Current inspection practices, however, are primarily designed for routine maintenance and are not equipped to provide rapid evaluations following flooding events. Consequently, there is no effective methodology available for emergency inspections of culverts after extreme flooding events. Emerging artificial intelligence and computer vision technologies present promising solutions to this gap, offering the potential for automated, rapid, and high-resolution inspections that can enhance post-flood response and improve infrastructure resilience.

(4)

3D printing technology might be adopted to develop flood-resistant highway culverts to minimize the flooding impacts and thus to stabilize the entire highway embankment. Portable self-standing flood barriers could be a mitigation measure to lower the flood energy, which can be 3D printed as well.

8. Conclusions

Highway culverts are critical to the functionality and resilience of transportation systems, yet they remain highly vulnerable to extreme flooding. This review highlights the complex interplay of hydraulic, geotechnical, and structural processes that govern culvert performance under flood loading. The findings indicate that while traditional design practices improve serviceability under typical conditions, they often fail to capture the dynamic responses associated with overtopping, scour, debris accumulation, and structural instability during severe flood events.

Key conclusions from this review are as follows:

• The relative position of upstream water levels and culvert inlets plays a decisive role in determining failure modes, especially under submerged conditions.
• Overtopping remains the most critical and least understood flooding scenario, with current hydraulic models unable to simulate such conditions accurately.
• Common failure mechanisms, erosion, debris-induced clogging, and structural washout, have been widely documented but are not yet systematically correlated with specific culvert geometries or flooding intensities.
• Existing mitigation strategies, such as drainage capacity enhancement, erosion control, and debris management, are primarily qualitative and lack standardized methods for quantifying flood-induced damage.

Future research should focus on bridging these gaps through: (a) predictive models that link culvert geometry and flood intensity to specific failure modes; (b) reliable methods to quantify structural capacity reduction under flood loading; and (c) advanced inspection technologies, such as AI and computer vision, for rapid post-flood assessment. By combining hydraulic analysis, soil-fluid-structure interaction, and modern advanced monitoring techniques, the resilience of highway culverts can be substantially improved, reducing risks to public safety and transportation infrastructure in the face of intensifying flood hazards.