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:
- Using the Model
- 4.3.1.1 Meteorological inputs for tropical cyclone case
- 4.3.1.2 Interpolation issues for numerical models
Data for tuning and calibrationOnce the model has been set up and tidal boundaries have been established, it has to be tuned and calibrated to make it fit for the intended use. A first step is to verify that the astronomical tide which comes in through the open boundaries is propagated correctly through the model. Data which are useful here are time series of observed water levels for quiet periods where the weather was calm and no significant surges can be expected. But also the tidal constituents which are determined by a harmonic analysis of long time series can be used. The model can then be run without meteorological input for a certain period to establish the model equivalents of these constituents or the time series for the quiet periods. Discrepancies between the model and observations can lead to adjustments of the bathymetry or the astronomical tide at the open boundaries. Satellite altimeter data might prove useful to get an overview of the spatial structure of the astronomical tide and help to identify regions where adjustments are necessary. Calibration for surges is best done for periods in which there is much wind and should cover both positive and negative surges. Here, apart from the time series of observed water levels, also meteorological input is needed, preferably comparable to that which will be used for the forecast mode in the operational phase of the model. Any available current data are also useful in this phase for comparison with the model.
Meteorological input for (2D) storm surge forecasting usually consists of (near) surface wind (stress) and sea level pressure. For more sophisticated air-sea interaction schemes [this should have been dealt with in chapter 2], also other parameters might come into play, like temperature and wind at other levels, see Makin (2003), Flather et al (1991), Mastenbroek et al (1993). Accurate meteorological fields, especially wind fields, are crucial, since the magnitude of storm surges is at least proportional to the square of the wind speed. A natural source for these parameters is the atmospheric models, which are used for weather forecasting. Global models, like those from ECMWF, UKMO or NOAA/NWS have horizontal resolutions of 25–100 km up to 240–360 hours ahead, which makes them suitable for extra-tropical storms. However, for tropical storms and also for high-resolution water level and current forecasts, these models are often not sufficient because the structure of tropical cyclones and other relevant details are not sufficiently resolved. There, limited area models, as operated by many of the National Meteorological Services, are a better option. Limited area models nowadays generally have resolutions in the order of 10 km up to 48 hours, and the next generation is aiming at 1 km up to 24 hours. There are several ways to use the meteorological data for storm surge forecasting. If used in a numerical model, the meteorological fields generally have to be interpolated from the grid and the time steps of the atmospheric model those of the storm surge model. Questions to address are the type of interpolation in space and time, handling of land-sea boundaries. Other methods use empirical formulas and need other, derived, parameters like the maximum wind and pressure in the centre of a tropical cyclone. 4.3.1.1 Meteorological inputs for tropical cyclone caseBecause tropical cyclones are usually smaller but stronger than extratropical cyclones, the modeling of tropical storm surge requires higher-resolution (say, 5 to 10km) wind fields that resolve the most intense part of a tropical cyclone. State-of-the-art mesoscale NWP models can provide such high-resolution fields but they are currently available only in limited number of meteorological centers, since they are computationally intense and require an advanced data assimilation scheme that can generate realistic tropical cyclone fields. Therefore, simple empirical formulas of tropical cyclone pressure or wind are often used to create forcing fields for tropical storm surge models. As a tropical cyclone has a nearly axisymmetric structure, its distribution can be reasonably represented with an empirical formula for the radial distribution of wind or pressure. Once the distribution of either wind or pressure is determined, the other can be calculated by the gradient wind balance:  , (4.3)
where v is the speed of gradient wind at the distance r from the storm center, f the Coriolis parameter,  the density of air and p the atmospheric pressure at sea level, respectively.
Many researchers have suggested empirical formulas for the sea level pressure of a tropical cyclone as functions of r, the radial distance from the storm center, and the following forms are often used: Fujita (1952):  (4.4)
Myers (1954): Holland (1980):
These values can be obtained from the tropical cyclone advisories and/or bulletins issued by the Regional Specialized Meteorological Centres (RSMCs) and Tropical Cyclone Warning Centres (TCWCs) assigned by WMO, through the Internet and GTS. For detail on the bulletins, refer to http://www.wmo.ch/web/www/tcp/Advisories-RSMCs.html. The above methods are also useful for ensemble storm surge forecasts in case of tropical cyclones. Even when an ensemble NWP system is not available, one can easily create an ensemble of forcing fields with one of the formulas by changing the track or intensity of the target tropical cyclone. 4.3.1.2 Interpolation issues for numerical modelsWhen using output from a numerical weather prediction (NWP) model as input for a storm surge model, several interpolation issues have to be addressed. Generally, the grid of the NWP model is different from that of the storm surge model and the meteorological fields are not available on all the time steps of the storm surge model. The wind interpolation can be done in two different ways: on wind components or on wind intensity and direction. The first method leads to a loss of energy that can be important in case of strong wind rotation (Figure 4.1). More details can be found in Skandrani and Daniel (2001). Interpolation of the winds on force and direction, rather than on zonal and meridian components minimize the energy loss and seems to be preferable in most meteorological conditions. 
Figure 4.1: Two interpolation methods: on vector components (left) or on intensity (right) For interpolation in time there are two choices. Generally, a linear interpolation between the two nearest points in time for which the meteorological fields are available can be applied. But another possibility is to keep the meteorological forcing constant over a period of time around the NWP output step. An advantage of that could be that sharp gradients in the fields are better conserved. Interpolation in space is usually done in bi-linear way. Care should be taken with the interpolation of the wind field in the vicinity of land-sea boundaries. As the wind significantly decreases over land, the wind for a sea point in the storm surge model should not be (directly) interpolated from NWP points over land. Especially if the grid of the NWP model is coarser than that of the storm surge model, this might play an important role in the coastal zone. Solutions of this involve changing the coefficients of the interpolation near the coast with the aid of the land-sea masks of both the atmospheric and the storm surge model, e.g. setting the weights of the land points to 0 and renormailzing the remaining weights, or less radical solutions which still include something like 20% of the wind over land.
|
(4.5)
(4.6)
and 
are respectively the pressure at the storm center 
is the pressure drop defined as 
, and 
, A and B are empirical constants that control the storm size. Usually, B has a value between 1 and 2.5.
(4.7)