Coupled Compound Flood Modeling: Integrating Fluvial Catchment Runoff, Pluvial Storm Deluges, and ADCIRC Coastal Surge

Coastal estuaries, deltas, and maritime urban centers represent the most ecologically dynamic and economically vital regions on Earth. However, they also occupy the crosshairs of a complex, devastating hydrodynamic phenomenon: Compound Flooding. Compound flooding occurs when multiple distinct flood-generating mechanisms coincide in space and time or follow in rapid chronological succession, driving water levels far higher than the statistical sum of the individual isolated hazards.

Historically, civil engineering practices and regulatory flood hazard mapping (such as FEMA National Flood Insurance Program Flood Insurance Studies) evaluated flood hazards in complete isolation:

• Hydrologists evaluated fluvial flooding (riverine discharge overtopping mainstem banks) using watershed runoff models like HEC-HMS and 1D/2D river routing engines like HEC-RAS, assuming static, low downstream ocean tide boundaries.
• Coastal oceanographers evaluated oceanic flooding (storm surge, wind wave setup, and astronomical king tides) using coastal ocean circulation models like ADCIRC and SWAN, assuming static or zero upland river discharge boundaries.
• Municipal stormwater engineers modeled pluvial flooding (localized urban precipitation ponding) assuming unhindered, free-outfall drainage into receiving tidal estuaries.

In tropical cyclones, atmospheric river deluges, and severe nor'easter winter storms, these three drivers strike simultaneously. Hurricane winds drive catastrophic ocean surge into the mouth of an estuary, elevating the coastal tailwater elevation by three to five meters. Concurrently, torrential rain deluges dump inches of precipitation per hour across the urban core, while massive upland river hydrographs race downstream toward the coast. When the downstream ocean surge physically blocks the river discharge from escaping into the sea, an extreme backwater effect surges inland, producing catastrophic inundation across low-lying coastal plains.

In late 2026, coastal hydraulic engineering has achieved a monumental breakthrough through Fully Coupled Multi-Physics Compound Flood Modeling. By dynamically linking upland hydrological runoff engines, 2D hydrodynamic overland flood routing, and the ADCIRC+SWAN coastal storm surge and wave action model across a continuous, seamless computational grid, engineers can simulate compound flood dynamics with uncompromising physical fidelity.

This comprehensive technical treatise examines the governing mathematics, boundary-coupling architectures, tidal backwater physics, and joint probability statistics required to model compound coastal flood hazards.

flowchart TD
  Meteo["Meteorological Forcing\nNOAA GFS / HRRR Atmospheric Pressure P_s(x,y,t)\n10-Meter Wind Velocity Vector W_{10}(x,y,t)"]
  ADCIRC["ADCIRC Ocean Solver\nGeneralized Wave Continuity Equation (GWCE)\nSea surface elevation ζ(x,y,t) & currents (U, V)"]
  SWAN["SWAN Wave Model\nAction Density Spectrum Balance N(σ, θ)\nRadiation stress tensors S_{xx}, S_{xy}, S_{yy} & wave setup ∇·S"]

  Meteo --> ADCIRC
  Meteo --> SWAN
  ADCIRC <-->|Two-Way Dynamic Coupling| SWAN

The Generalized Wave Continuity Equation (GWCE)

ADCIRC discretizes the ocean basin across an unstructured triangular finite element mesh stretching from the deep Atlantic or Pacific Ocean onto continental shelves and coastal floodplains. To ensure numerical stability without artificial damping, ADCIRC replaces standard continuity with the Generalized Wave Continuity Equation (GWCE):

(∂^2 ζ) / (∂ t^2) + τ_0 (∂ ζ) / (∂ t) + ∇ · J - U H ∇ τ_0 = 0

Where:

• zeta(x, y, t) is free-surface elevation above the geoid datum (NAVD88).
• H = zeta + h is total instantaneous water column depth.
• tau_0 is the numerical damping parameter that suppresses spurious high-frequency spatial noise.
• J represents depth-integrated flux vectors incorporating Coriolis accelerations, atmospheric surface pressure gradients, astronomical tidal potential, and wind shear stress:

τ_s = ρ_{air} C_d |W_{10}| W_{10}

Two-Way Wave-Current Coupling with SWAN

Wind waves generated by oceanic gales do not merely ride on top of storm surge; they actively pump additional water onto the coast through wave radiation stress gradients (nabla · S). Every 10 to 30 minutes, ADCIRC passes updated sea surface elevations zeta and current vectors $(U, V)$ to SWAN, while SWAN computes the radiation stress divergence and transfers it back to ADCIRC as an additional momentum driving force, resolving the total coastal water level with centimeter-level precision.

3. Dynamic Boundary Coupling Topologies: One-Way, Loose, and Tight Coupling

Integrating ADCIRC ocean surge with 2D river and urban watershed engines requires careful selection of coupling architecture:

The Seamless Continuous Mesh Advantage

In late 2026, leading modeling platforms (such as FloodSim Compound, HEC-RAS 2D with coastal boundaries, and unstructured ADCIRC+SWAN) favor the Seamless Continuous Unstructured Mesh approach.

By employing adaptive triangular, quadrilateral, and polygonal elements:

• In the deep Atlantic Ocean, elements are large (15 km across), where simple long-wave physics dominate.
• As the mesh enters coastal barrier islands and estuarine inlets, elements refine to 50 meters.
• Across urban riverbanks, drainage ditches, and coastal city streets, elements refine down to `1.5 to 3.0 meters`.

Because there is zero artificial boundary dividing "ocean" from "river," numerical reflections and conservation errors are eliminated. The model naturally computes how massive river discharges push back against oncoming storm surges, capturing exact boundary stagnation points without manual parameter tuning.

4. Estuarine Tidal Backwater Physics and Drainage Paralysis

To understand why compound flooding inflicts catastrophic damage on urban drainage, one must analyze the hydraulics of Gravity-Fed Stormwater Outfalls.

The Physics of Backwater Choking

Under fair-weather conditions, urban stormwater conduits discharge into tidal estuaries via gravity outfalls equipped with flap gates to prevent tidal ingress during high tide. When an extreme storm surge elevates the estuarine water level above the conduit crown (H_{surge} > z_{invert} + D), two destructive hydraulic conditions occur:

1. Flap Gate Closure and System Paralysis: The hydrostatic ocean pressure forces the outfall flap gate shut. Internal rainfall runoff collected in the pipe network has nowhere to go. Within minutes, the underground pipe network pressurizes, causing street-level catch basins to overflow and ponding stormwater to depths exceeding a meter—even if the physical ocean surge never breaches the coastal seawall!
2. Submerged Orifice Discharge Restriction: If the outfall is un-gated, the ocean water level imposes a massive downstream tailwater head (h_{tw}), choking the outfall discharge capacity:

Q = C_d A √(2g (h_{upstream) - h_{tw})}

As h_{tw} approaches h_{upstream}, discharge drops toward zero (Q → 0), completely paralyzing urban drainage networks.

5. Joint Probability and Bivariate Copula Statistics in Compound Risk

Evaluating compound flood risk requires moving beyond traditional single-variable return periods (e.g., "The 100-Year Flood"). A 100-year compound flood event is rarely caused by a 100-year rainfall event coinciding with a 100-year storm surge; rather, it is frequently produced by the coincidence of a moderate 10-year storm surge and a moderate 10-year rainfall deluge, whose joint occurrence creates a catastrophic 100-year flood depth.

Modeling Non-Linear Dependency via Archimedean Copulas

Because storm surge and extreme rainfall are both driven by the same synoptic meteorological systems (hurricanes and tropical depressions), they exhibit strong tail dependence. Standard Pearson linear correlation coefficients fail to capture this non-linear clustering.

Hydraulic risk engineers utilize the Gumbel-Hougaard Copula, which exhibits upper tail dependence (lambda_U > 0):

C_θ(u, v) = exp ≤ft( - ≤ft[ (-ln u)^θ + (-ln v)^θ ]^{1/θ} )

Where:

• u = F_X(x), v = F_Y(y) represent the uniform marginal probability transformations.
• $\theta \ge 1$ represents the copula dependence parameter (calibrated via Kendall's rank correlation tau = 1 - theta^{-1}).

By sampling synthetic event pairs (Q_i, zeta_i) from the fitted copula, engineers execute Monte Carlo ensembles through coupled 2D hydrodynamic solvers, constructing Bivariate Return Period "AND" Curves (T_{AND}) that quantify true compound risk with statistical rigor.

6. Comprehensive Compound Flood Architecture Comparison Matrix

The table below contrasts isolated modeling approaches with the fully coupled compound simulation framework engineered by FloodSim:

7. Real-World Engineering Case Study: Hurricane Inundation in a Tidal River Delta

To demonstrate the critical necessity of coupled compound modeling, FloodSim executed a post-event compound hydrodynamic validation study for a 120-square-kilometer coastal river delta housing an active commercial deep-water port, industrial chemical terminals, and 185,000 residents.

The Compound Dynamics Unveiled

The coupled simulation revealed that the peak disaster was not caused by the ocean surge alone:

1. The Hydrodynamic Stagnation Point: At Hour 11 of the storm, the inward-surging ocean surge met the outward-surging 1,800 m³/s river discharge wave at River Mile 8.4. The opposing momentum vectors created a stationary stagnation boundary where flow velocity dropped to zero (u ≈ 0).
2. Backwater Elevation Surge: Unable to escape to sea, the river discharge backed up behind the surge dam, forcing water levels at an inland industrial park to reach `+5.27 meters NAVD88`—a full 1.42 meters higher than the ocean storm surge measured at the coast!
3. Infrastructure Defense Remediation: Using the coupled model, port engineers designed an active 2.2-meter setback levee and a high-capacity 80 m^3/s axial-flow storm pump station at the estuarine confluence, successfully safeguarding over $1.2 billion in petrochemical processing assets against future Category 3+ hurricane events.

8. Five-Phase Engineering Protocol for Executing Compound Flood Audits

Civil and coastal engineering firms auditing estuarine flood risk in 2026 should adhere to this five-phase execution protocol:

1. Establish Seamless Geodetic Elevation Datums: Harmonize all bathymetric soundings (NOAA hydrographic surveys) and upland LiDAR topography onto a single continuous vertical datum (e.g., NAVD88 with updated 2026 epoch geoid models). Never mix local Mean Lower Low Water (MLLW) with terrestrial datums.
2. Construct an Unstructured Multi-Scale Mesh: Design a continuous mesh spanning from the deep continental shelf to urban street centers, refining node density along barrier islands, navigation channels, and levee crests.
3. Execute Two-Way Coastal Wave-Surge Coupling: Ensure ADCIRC passes instantaneous water levels to SWAN every 15 minutes, feeding wave radiation stresses back into the hydrodynamic momentum solver.
4. Implement Direct-on-Mesh Hydrologic Precipitation: Instead of applying lumped river hydrographs at upstream boundaries, apply spatially distributed NEXRAD radar precipitation directly onto the 2D computational cells, accounting for dynamic soil infiltration losses via Green-Ampt or Horton equations.
5. Calibrate Joint Bivariate Probability: Fit marginal extreme value distributions to historical stream gauge and tide gauge records, utilizing Archimedean copulas to establish rigorous multi-hazard return period boundaries.

Through the disciplined integration of oceanic circulation modeling, shallow water hydrodynamics, and multivariate copula statistics, coupled compound flood modeling empowers coastal communities to anticipate, withstand, and adapt to the most extreme multi-hazard storms of the 21st century.