next up previous contents
Next: Organisation of input files Up: Preparation and sampling of Previous: Preparation and sampling of

Digitised coastline, contour and sounding data

The data files for the demonstration case are normally to be found in directory trigrid/demodata on the drive where TRIGRID has been installed. It is recommended that all the files in this directory should be copied to a backup diskette, in case you accidentally overwrite them while working through the demonstration case. It is best to create a separate working directory and copy to it the following files from trigrid/demodata :


All the coastline and bathymetric data required to prepare a depth grid in the demonstration case are contained in the following three files used as input to the SAMPLER:

bnd.dig - digitised coastlines in DIGIT format

cont.dig - digitised contours in DIGIT format

sound.nod - depth soundings in NODE format

Partial listings of the files holding the digitized boundary and contour data, in DIGIT format, are shown in Part III.

The guidelines followed in producing these files apply to all TRIGRID applications and are as follows :

  1. All boundaries are digitised counterclockwise.
  2. If possible, boundaries are digitised at several times the resolution required eventually for the depth or model grid boundaries.
  3. The outer boundary of the domain, made up alternately of stretches of land and open sea boundary, is digitised as an entity, so that the digitised locations of all points on the outer boundary appear in consecutive data blocks in the digitised boundary data file bnd.dig, with the first block corresponding to a stretch of land boundary.
  4. Valid depth information is required for some distance beyond the intended locations of the model's open sea boundaries. Consider the lower boundary in Figure 2.1 for instance. Line MN is to constitute an open sea boundary for the model, but depth contours have been digitised not only inside the intended model domain but also for the area MNBC outside. The outer boundary actually digitised, A-S-M-B-C-N-P-D-E-Q-R-F, embraces this extension, as well as similar extensions beyond the other intended open boundaries P-Q and R-S. The amount of extension needed is related to the size of triangles to be produced later near the open boundaries of the model grid. This is discussed in more detail in a later section dealing with creation of interior nodes for the model grid; here it is enough to say that the depth grid should extend an appreciable distance beyond the intended final location of the open boundary.
  5. Note that in digitizing the outer boundary, a small gap has been left at A (Figure 2.1), i.e. the boundary is not quite closed. Similarly, digitization of the island stops just short of the initial point G. This is done to avoid getting coincident or nearly coincident boundary nodes, which can cause difficulties in the gridding program GRIDIT.
  6. Contours should preferably be digitised at fairly high resolution, so that a fairly wide choice of sub-sampling interval is possible when using the SAMPLER. It is not necessary to digitise whole contours or to use closed contours. Where two contours come very close together, only one should be digitised, to avoid having nearly coincident nodes for the depth grid after the contours are subsampled. For the same reason, any individual sounding supplied (see (7) below) should not lie very close to any of the digitised contours used. Program NODER has an option (TOO-CLOSE) for detecting nearly coincident nodes, which can cause problems with the triangulation program GRIDIT if not eliminated.
  7. The depth contours chosen for digitization should cover the model domain fairly evenly; large areas should not be left blank, otherwise sampling of the contours does not provide an adequate number of nodes to form a satisfactory depth grid, unless supplemented with suitably placed depth soundings. Contours should extend some distance beyond the intended positions of the model grid open boundaries (see (4) above).
  8. Wherever convenient or necessary, for instance, over part of a chart without depth contours, a file consisting of a list of depth soundings, e.g. sound.nod in the demonstration case, can be added to the other depth information read in by the SAMPLER.
  9. As an option, all the depth values can be furnished as a sounding file. This appropriate for the case where depth is provided as gridded data from some database program, for instance.

next up previous contents
Next: Organisation of input files Up: Preparation and sampling of Previous: Preparation and sampling of

Channel Consulting Ltd.