source: production/pt_hedland_2006/report/computational_setup.tex @ 3365

To set up a model for the tsunami scenario, a study area is first 
determined. Preliminary investigations have indicated the point
at which the output from MOST is the input to ANUGA is
sufficient at between the 50m and 100m bathymetric contour line.
In constrast to the Onslow study,
we choose the 50m contour line as the seaward boundary due primarily
to computational constraints.
\footnote{ 
Preliminary investigations indicate that MOST and ANUGA compare
well at the 50m and 100m contour line. In addition, the resolution for
the MOST modelling indicate that it can theoretically model 
tsunamis with a wavelength of 20-30km, and the wavelength of
the tsunami wave at the boundary is approximately 20km. A much
higher model resolution will be used in developing the probabilistic
models for further studies.}.  Historical runup heights are 
of the order of 10m and we would expect that a tsunami wave
would penetrate no higher for this scenario.
Current computation requirements define a coastline
extent of around 100km. Therefore, the study area of around ? km$^2$ 
covers approximately 100km of
coastline and extends offshore to the 50m contour line and inshore to 
approximately 10m elevation. 

The finite volume technique relies on the construction of a triangular mesh which covers the study region. This mesh can be altered to suit the needs of the scenario in question. The mesh can be refined in areas of interest, particularly in the coastal region where the complex behaviour is likely to occur. In setting up the model, the user defines the area of the triangular cells in each region of interest\footnote{Note that the cell 
area will be the maximum cell area within the defined region and that each 
cell in the region does not necessarily have the same area.}.
The area should not be too small as to exceed realistic computational time, and not too great as to inadequately capture important behaviour. There are no gains in choosing the area to be less than the supporting data.
Figure \ref{fig:pt_hedland_area} shows the study area and where further mesh refinement has been made. For each region, a maximum triangular cell area is defined and its associated lateral accuracy.
With these cell areas, the study area consists of 401939 triangles
in which water levels and momentums are tracked through time. The lateral accuracy refers to the distance at which we are confident in stating a region is inundated. Therefore we can only be confident in the calculated inundation extent in the Port Hedland town centre to within 30m.

\begin{figure}[hbt]

  \centerline{ \includegraphics[width=100mm, height=75mm]
             {../report_figures/pt_hedland_data_poly.png}}

  \caption{Study area for the Port Hedland scenario highlighting four regions of increased refinement. 
Region 1: Surrounds Port Hedland town centre with a cell area of 500 m$^2$ (lateral accuracy 30m).
Region 2: Surrounds the coastal region with a cell area of 50000 m$^2$ (lateral accuracy 220m).
Region 3: Water depths to the 50m contour line (approximately) with a cell area of 250000 m$^2$ (later accuracy 700m).
}
  \label{fig:pt_hedland_area}
\end{figure}

\begin{figure}[hbt]

  \centerline{ \includegraphics[width=100mm, height=75mm]
              {../report_figures/mesh.jpg}}

  \caption{Computational mesh for Port Hedland study area where the
cell areas increase in resolution; 500 m$^2$, 50000 m$^2$ and 250000 m$^2$.}
  \label{fig:mesh_pt_hedland}
\end{figure}

The final item to be addressed to complete the model setup is the 
definition of the boundary condition. As 
discussed in Section \ref{sec:tsunamiscenario}, a Mw 9 event provides
the tsunami source. The resultant tsunami wave is made up of a series
of waves with different amplitudes which is affected by the energy
and style of the event as well as the bathymetry whilst it travels
from its source to Port Hedland. The amplitude and velocity of each of these
waves are then provided to ANUGA as boundary conditions and propagated
inshore.