Coupling the National Water Model to the Coastal Ocean
Project Team
Project Lead/Scientific PI: Brian Blanton, UNC Chapel Hill
Transition PI: Rick Luettich, UNC Chapel Hill
Other Investigators: Clint Dawson, Jason Fleming
Partners: Edward Myers, Sergey Vinogradov (NOAA/NOS), E. Clark, Hassan Mashriqui (NOAA/NWS: Model), C. Massey (USACE-ERDC), Debra Hernandez (SECOORA)
NOAA/NOS Technical Points of Contact: Saeed Moghimi, Audra Luscher, Carolyn Lindley, Lianyuan Zheng
NOAA/NWI Technical Point of Contact: Trey Flowers
Project Overview and Results
Hurricanes Matthew (2016), Harvey (2017) and Irma (2017) caused unprecedented levels of coastal flooding, due to the occurrence of both substantial coastal storm surge and heavy precipitation-related hydrologic/inland flooding propagating toward the coast. The resulting “compound” floods remind us that prediction of water related hazards remains challenging. This project will evaluate different strategies for linking NOAA’s National Water Model with NOAA’s coastal Extratropical Surge and Tide Operational Forecast System (ESTOFS) and Hurricane Surge On-demand Forecast System (HSOFS), both of which utilize the ADCIRC modeling system.
The overall goal of this testbed project is to provide a new solution for compound flood predictions in the operational modeling capability for NOAA.
The primary objectives of this testbed project are:
- Adapt and evaluate ADCIRC as a middle model for linking NOAA’s CONUS scale National Water Model (NWM) and the Coastal Zone as modeled by NOAA’s Extratropical Surge and Tide Operational Forecast System (ESTOFS) / Hurricane Surge On-demand Forecast System (HSOFS).
- Develop appropriate data streams and procedures to use the Deltares Delft3D-Flow (D-Flow) model and the US Army Corps of Engineers Gridded Surface/Subsurface Hydrologic Analysis (GSSHA) model for linking the NWM with the ESTOFS / HSOFS.
- Develop appropriate data streams and procedures to use the Deltares Delft3D-Flow (D-Flow) model and the US Army Corps of Engineers Gridded Surface/Subsurface Hydrologic Analysis (GSSHA) model for linking the NWM with the ESTOFS / HSOFS.
- Evaluate the accuracy and performance of the three coupling options developed in years 1 and 2 and transition one or more strategies to NOAA units responsible for operating the coupled system.
Model Descriptions
ADCIRC:
ADCIRC is a widely used coastal circulation and storm surge model (see a partial list of ADCIRC publications at http://adcirc.org) that solves the 2D or 3D shallow water equations on a triangle-based unstructured grid to enable coverage of large domains and complex geometry, topography and bathymetry in the Coastal Zone. Forcing includes tides, meteorological fields and rivers, and it simulates wetting and drying of areas subject to transient flooding. It has been highly parallelized using domain decomposition for efficient application on high performance computers (Dietrich et al, 2012). Following Hurricane Katrina (2005), ADCIRC was coupled with the unstructured grid version of the SWAN wave model (Ziljema, 2010) to create a modeling system often called ADCIRC+SWAN (or often just ADCIRC) that provides a seamless solution to tides, surge and waves (Bunya et al, 2010; Dietrich et al, 2010; Dietrich et al, 2011). ADCIRC has also been coupled to the US Army Corps of Engineers ST-WAVE model via the Coastal Storm Modeling System (CSTORM-MS, Massey et al, 2011) and is currently being coupling to the unstructured version of its WaveWatch III wave model (A. van der Westhuysen, NOAA NWS/EMC personal communication). ADCIRC has been used to provide surge and wave conditions to delineate coastal flood hazards for the FEMA National Flood Insurance Program from New England to Texas, to evaluate flood hazards at coastal nuclear power plants for the Nuclear Regulatory Commission, to design coastal flood reduction systems (e.g., levees, flood walls) around major coastal cities including New Orleans (USACE, 2012), Houston/Galveston and New York City (Bloomberg, 2013). It is one of NOAA’s core coastal models and is the basis of NOAA’s V-DATUM program as well as the ESTOFS and HSOFS operational models. A previous pilot project associated with NOAA’s CI-Flow program demonstrated that output from NOAA’s Distributed Hydrology Model (HL-RDHM) could be used as a real time input to ADCIRC in eastern North Carolina (Van Cooten et al, 2011; Dresback et al, 2013) although the ADCIRC triangular grids restricted the allowable model time step and therefore the effectiveness of the application. By overcoming this limitation, ADCIRC may be an attractive hydrodynamic solution in the Middle Zone since it would simply be an extension of (and thus seamlessly integrate with) the operational ESTOFS / HSOFS models.
D-Flow:
NOAA’s Office of Water Prediction has recently begun working with the D-Flow Flexible Mesh (FM) model (H. Mashriqui, NOAA OWP, personal communication). D-Flow is the hydrodynamic component of the DELFT-3D hydro-morphological modeling suite used to simulate non-steady flows and transport in 1D, 2D or 3D, (Deltares, 2014). Forcing includes tides meteorological fields and rivers, and it simulates the wetting and drying for flooding computations. The original D-Flow solves the shallow water equations on either a rectilinear or curvilinear, boundary fitted mesh while the more recently developed D-Flow FM uses an unstructured, flexible mesh of triangles, quadrilaterals and / or hexagons. D-Flow FM has a finite volume numerical algorithm with a higher-order advection treatment that is better suited for advection dominated supercritical flows, bores and dam breaks than ADCIRC’s finite element algorithm. D-Flow FM is parallelized for high performance computing, using a domain decomposition approach similar to ADCIRC and using OpenMP. D-Flow FM may therefore be an attractive option in the Middle Zone provided an effective coupling can be developed with ESTOFS / HSOFS. The overall DELFT-3D package has been widely applied in coastal, river and estuarine areas for water flows sediment transport, waves, water quality, morphological developments and ecology. In 2011, the D-Flow curvilinear grid model was open-sourced. D-Flow FM is currently limited to a small group of development partners, but Deltares anticipates having D-Flow FM fully open source in late 2018. Given that NOAA is already using 2 D-Flow FM, the actual open source release date should not constrain its use in this COMT project.
GSSHA:
GSSHA (Gridded Surface/Subsurface Hydrologic Analysis - Downer and Ogden, 2004) is a two-dimensional, physically-based watershed model that simulates surface water and groundwater hydrology, erosion and sediment transport. It has been developed at the US Army Engineer Research and Development Center Hydraulic Modeling Branch and is approved for use for determining hazards from hydrologic flooding for the FEMA National Flood Insurance Program. GSSHA solves the diffusive wave approximation to the shallow water equations, that neglects transient terms in the momentum equation. A finite volume numerical solver is used on a regular square grid spatial discretization of the watershed with elevation data extracted from a digital elevation model. GSSHA can represent various hydraulic structures such as weirs, culverts, detention basins, etc, and includes a vector channel algorithm allowing channels to flow in any direction independent from the grid resolution. GSSHA models a large number of processes including precipitation distribution, snowfall accumulation and melting, infiltration, evapotranspiration, surface runoff routing, channel flow routing, unsaturated zone flow, saturated groundwater flow, overland sediment erosion, transport and deposition, and channel routing of sediments. GSSHA provides the capability to explicitly model precipitation input and hydrologic processes in the Middle Zone, versus simply routing water from the Upland or Coastal Zone across the Middle Zone by ADCIRC and D-Flow. Due to the low slope topography along much of the US coast (particularly the Gulf and SE Atlantic coasts) the width of the Middle Zone may be significant, and therefore including hydrologic processes in this zone may be important. On the other hand, the use of the diffusive wave approximation in rivers subject to tidal influence and upstream surge propagation is probably not adequate, thereby requiring a combination of ADCIRC / D-FLOW and GSSHA in the Middle Zone. Including GSSHA in this COMT project provides the opportunity to determine the data flows required to include GSSHA or a similar model of the Middle Zone hydrology and evaluate its value for inclusion as part of the Middle Zone solution.
References
Bloomberg, M. (2013). A stronger, more resilient New York. City of New York, PlaNYC Report. 445p.
Bunya, S., J. Dietrich, J. Westerink, B. Ebersole, J. Smith, J. Atkinson, R. Jensen, D. Resio, R. Luettich, C. Dawson, V. Cardone, A. Cox, M. Powell, H. Westerink, and H. Roberts. (2010). A high resolution coupled riverine flow, tide, wind, wind wave and storm surge model for Southern Louisiana and Mississippi: Part I–model development and validation. Mon. Weather Rev., 138, 345–377.
Downer, C., and Ogden, F. (2004). GSSHA: A model for simulating diverse streamflow generating processes, J. Hydrol. Engrg., 9(3), 161-174.
Massey, T. C., T. V. Wamsley, and M. A. Cialone. (2011). Coastal storm modeling-system integration. In: Solutions to Coastal Disasters 2011, pp. 99–108.
Van Cooten, S., Kelleher, K., Howard, K., Zhang, J., Gourley, J., Kain, J., Nemunaitis-Monroe, K., Flamig, Z., Moser, H., Arthur, A., Langston, C., Kolar, R., Hong, Y., Dresback, K., Tromble, E., Vergara, H., Luettich, R., Blanton, B., et al. (2011). The CI-FLOW project: A system for total water level prediction from the summit to the sea. Bull. Amer. Meteor., 92(11), 1427–1442.
Zijlema, M. (2010). Computation of wind-wave spectra in coastal waters with swan on unstructured grids. Coastal Engineering, 57(3), 267–277.