118
a number of high-profile incidents in the urban estuary including the ditching of a
commercial flight in the Hudson River in 2009 and urban inundation caused by
Hurricane Sandy in 2012 (Blumberg et al. 2015). The urban inundation model,
operating on a 3.1 m grid and using high-resolution LIDAR topography provides
guidance for flood prevention intervention in preparation for future such events.
6.3 Coastal Ocean Wave Models
In contrast to hydrodynamic models which operate in a four-dimensional space/
time frame (x, y, z, t), wave models operate in a five-dimensional space/time frame
(x, y, t, θ, σ) where θ is the propagation angle and σ is the frequency (Booij et al.
1999; Mellor et al. 2008), a feature that increases the computational demand by one
degree of freedom over that of hydrodynamic models. Third-generation wave models in use today incorporate varying wind fields (assimilated from satellite observations and/or operational wind models); resolve nonlinear wave-wave interactions,
which are responsible for energy transfer across the wave spectrum; and explicitly
model energy dissipation through bottom friction, bottom-induce wave breaking,
and whitecapping. Such spectral wave models portray well the wave fields in the
deep ocean. In nearshore waters where both waves and currents interact with convoluted topography and bathymetry model, performance depends greatly on the precision and resolution allowed by the digital representations of physical features
(stored as shapefiles for geographic information system (GIS) software) and of the
rugosity and consistency of these coastal features. These elements are crucial to
accurate representation of wave evolution and coastal inundation. Bottom types are
parameterized in terms of their frictional interactions with waves and currents
through so-called Manning roughness coefficients (Benitez and Mercado 2015) or
other empirical roughness parameterizations. Attenuation of wavefields with subgrid scale natural features such as floating sea ice must be similarly parameterized
according to the myriad sea ice types. Accurate estimates for such parameterization
are critical for successful modeling of wave dissipation in high rugosity environments such as grassy wetlands, the dense root structure of mangrove forests, and
coral reefs (Lowe et al. 2005).
Experienced mariners are well aware that opposing wave and current fields make
for choppy seas, a clear indication that accurate numerical representation of the
wave field is incomplete without reference to ocean currents, especially nearshore
where topography-induced jets occur. Such representation is achieved using fully
coupled hydrodynamic and wave models requiring simultaneous runs of both models. As a result, it is only recently, as computational capacity has increased, that
fully coupled models on unstructured grids have been achieved (Xie et al. 2016;
Chen et al. 2013). These are largely run to provide forensic analysis of past storms
or to develop storm surge atlases which are catalogs or look-up list for storm inundation preparedness. Inundation hazard databases are prepared carrying out multiple model runs with simulated hurricanes of different magnitude and size at different
6 Numerical Models for Operational Ocean Observing
a number of high-profile incidents in the urban estuary including the ditching of a
commercial flight in the Hudson River in 2009 and urban inundation caused by
Hurricane Sandy in 2012 (Blumberg et al. 2015). The urban inundation model,
operating on a 3.1 m grid and using high-resolution LIDAR topography provides
guidance for flood prevention intervention in preparation for future such events.
6.3 Coastal Ocean Wave Models
In contrast to hydrodynamic models which operate in a four-dimensional space/
time frame (x, y, z, t), wave models operate in a five-dimensional space/time frame
(x, y, t, θ, σ) where θ is the propagation angle and σ is the frequency (Booij et al.
1999; Mellor et al. 2008), a feature that increases the computational demand by one
degree of freedom over that of hydrodynamic models. Third-generation wave models in use today incorporate varying wind fields (assimilated from satellite observations and/or operational wind models); resolve nonlinear wave-wave interactions,
which are responsible for energy transfer across the wave spectrum; and explicitly
model energy dissipation through bottom friction, bottom-induce wave breaking,
and whitecapping. Such spectral wave models portray well the wave fields in the
deep ocean. In nearshore waters where both waves and currents interact with convoluted topography and bathymetry model, performance depends greatly on the precision and resolution allowed by the digital representations of physical features
(stored as shapefiles for geographic information system (GIS) software) and of the
rugosity and consistency of these coastal features. These elements are crucial to
accurate representation of wave evolution and coastal inundation. Bottom types are
parameterized in terms of their frictional interactions with waves and currents
through so-called Manning roughness coefficients (Benitez and Mercado 2015) or
other empirical roughness parameterizations. Attenuation of wavefields with subgrid scale natural features such as floating sea ice must be similarly parameterized
according to the myriad sea ice types. Accurate estimates for such parameterization
are critical for successful modeling of wave dissipation in high rugosity environments such as grassy wetlands, the dense root structure of mangrove forests, and
coral reefs (Lowe et al. 2005).
Experienced mariners are well aware that opposing wave and current fields make
for choppy seas, a clear indication that accurate numerical representation of the
wave field is incomplete without reference to ocean currents, especially nearshore
where topography-induced jets occur. Such representation is achieved using fully
coupled hydrodynamic and wave models requiring simultaneous runs of both models. As a result, it is only recently, as computational capacity has increased, that
fully coupled models on unstructured grids have been achieved (Xie et al. 2016;
Chen et al. 2013). These are largely run to provide forensic analysis of past storms
or to develop storm surge atlases which are catalogs or look-up list for storm inundation preparedness. Inundation hazard databases are prepared carrying out multiple model runs with simulated hurricanes of different magnitude and size at different
6 Numerical Models for Operational Ocean Observing