Session of the wmo-ioc joint Technical Commission for Oceanography and Marine Meteorology (jcomm) agreed that it would be logical to transform the wmo wave Programme into the jcomm wind Wave and Storm Surge Programme
Download 1.37 Mb.
|
- Navigate this page:
- Model Setup
- Tidal boundaries
IntroductionThis section deals with the impacts of various input parameters for the storm surge model both in the initial setup, tuning and validation phase and in the operational phase, and with possible output parameters and their form. Oceanographical data, which is needed in the setup of the model, consists of bathymetry and tidal forcing data at the open boundary. The location of the open boundary might even be determined by the availability of tidal forcing. Tide gauge data will be needed for tuning the model, validation and verification, and possibly data assimilation, together with harmonic analysis astronomical tides. Satellite altimeter data can be very useful for tuning and calibrating the model, but their availability is generally, especially if the model is not too large, too low to be of much practical use in the operational running of the model. Any available current data might also prove useful for tuning the model. Meteorological input to the model consists of surface 10 m winds or surface stress and mean sea level pressure. This Section will elaborate on spatial resolution of the different available meteorological sources and input time-interval. Hydrological inputs consist of river runoff and precipitation data. Ice input is needed for modeling in areas where this is relevant. For all inputs considerations of accuracy and timeliness will be addressed. Output parameters which could be considered are total water level and surge as either fields or time series only for high/low tides. This section will comment on how to deal with the surge-part of the total water level.
The first stage in the setup of the model is the definition of the area. For that purpose one needs geometry and bathymetry data for the area under consideration which correspond to the required resolution of the storm surge model. The open boundary of a 2D storm surge model is best located in water deeper than 200 m and requires available tidal forcing, either as currents or as levels. For shoreline data, the GSHHS dataset [Wessel and Smith, 1996] with a typical resolution of about 200 m is available in the public domain and can be downloaded from the University of Hawaii at http://www.soest.hawaii.edu/wessel/gshhs/gshhs.html. Bathymetry data can come from sea charts, satellite altimetry maps, but also from local measurement campaigns. The General Bathymetric Chart of the Oceans (GEBCO) project, which operates under the joint auspices of the Intergovernmental Oceanographic Commission (IOC) of UNESCO and of the International Hydrographic Organization (IHO), provides global bathymetry data sets for the world's oceans. Their One Minute Grid is available for download from the British Oceanographic Data Centre (BODC) at http://www.bodc.ac.uk/data/online_delivery/gebco. Furthermore, NOAA's National Geophysical Data Center (NGDC) has an extensive collection with a resolution of 2 min in the public domain which can be downloaded from http://www.ngdc.noaa.gov/mgg/bathymetry/relief.html." Local data sources are typical national hydrographical institutes, which may resort under the navy, or research institutes. When using bathymetry data one should verify the purpose for which the data were originally collected and used. This might influence the usefulness for storm surge modelling. An important use of sea charts e.g. is to prevent ships from running aground and therefore these charts tend to give a minimum depth rather than an average depth, which is needed for modelling water movement. Such deficiencies can be corrected in the tuning phase to get proper astronomical tides inside the model domain.
The open boundary condition proposed by Davies and Flather (1978) is applied to the difference between the interior value and the prescribed value of the tide and the storm surge, so that the perturbation leaving the domain follows the expression  (4.1)
where q is the current normal to the boundary;  is the speed of the external gravity waves leaving the model domain;  and   are the water level and the depth-averaged current of the tide prescribed in the open boundary, respectively; h is the height calculated in the interior point adjacent to the boundary and hs and qs are the height and current associated with the storm surge entering the domain obtained from other sources such as a coarser grid model, inverse barometer effect, etc. The tide is prescribed as a combination of several harmonic constituents, as follows:
 (4.2)
where n is the prescribed number of tidal constituents, F is the nodal factor, a is the speed of the constituent, t is the time starting from the beginning of the prediction in Greenwich meridian time (GMT), V0u is the equilibrium tide at the Greenwich meridian, p is 1 for diurnal components and 2 for semidiurnal,  is the amplitude of the component and G is the phase (expressed as epoch) of the tidal constituent. The values of F and V0u depend on time and, according to usual practice, may be tabulated for every year, as well as their correction for month and day. These parameters should be calculated for every prescribed day/time of a run to establish the phase of the selected tidal components. For further details on tidal analysis, the work of Schureman (1958) may be referred.
In most cases, the constituents for the tidal current at the open boundaries are not known. The linear approach,  can be used, where the sign corresponds to the normal component of the speed of a perturbation that enters into the domain.
The constituents used may be the few most representative of the local tidal regime. The values of amplitude and epoch are determined by some local observations. At open sea boundaries amplitude and co-tidal fields from global tidal models or other sources can be interpolated and introduced in Equation (4.2). Blending this information with data from any oceanographic station located close to the boundary is appropriate. In other cases, such as semienclosed seas or estuaries, an adequate interpolation of coastal values at the open boundary, by following known patterns or prescribing a Kelvin wave shape, may be suitable. Tidal constants can be calculated at selected points after an adequate spin-up period, i.e., the time needed by the tidal waves to establish a steady state. The spin-up period depends on the local tidal regime and the area covered by the model domain. In a general case, about a month run is sufficient for the time series needed to obtain the tidal constituents (Foreman, 1979). It is strongly recommended that the tidal part of the model (i.e., the storm surge model applied to the tide only) should be calibrated with the available analysed tidal data. The tidal constituents and their distribution throughout the open boundaries may be used in the calibration process, together with other model parameters with their own range of uncertainty, such as bottom friction or some critical depth for land flooding and drying in high resolution coastal models. The modelled tidal wave may be more sensitive to the values at a particular boundary, depending on the incidence of the external wave. Occasionally, differences between different bathymetric sources may be sufficient to produce significant modifications in speed and direction of long gravity waves. A "best" solution can then be found to fit the observations, if needed after re-visiting the bathymetric information. The correct simulation of the tide will ensure the adequacy of the model for the representation of the local hydrodynamical processes and hence the storm surge. Tidal constituents which are used in the British model CS3X (POL, 2007) are Q1, O1, P1, K1, MNS2, 2N2, µ2, N2,  ,  , M2, L2, T2, S2, K2, MSN2, 2SM2, MO3, MK3, MN4, M4, MS4, M6, 2MS6, 2MK6, and 3M2S2. The Dutch model DCSM (KNMI, 2007) uses M2, S2, N2, K2, O1, K1, Q1, P1, and L2.
|