114 



The convergent procedure often works well for bathymetric data. It uses 

 multiple data points as controls for calculating the values at nearby nodes. 

 The values are blended with a distance-weighting technique such that close 

 points have more influence over the node than distant points. Several itera- 

 tions are made, with the first being crude and including many points, and the 

 final being confined to the closest points. The least-squares method produces 

 a plane that fits across several points near the node. Once the plane has been 

 calculated, the Z-value at the node is easily computed. The reader must con- 

 sult software manuals to learn the intricacies of how these and other algo- 

 rithms have been implemented. 



Another important parameter that must be chosen is the gridding 

 increment. This is partly determined by the algorithm chosen and also by the 

 data spacing. For example, if survey lines are far apart, there is little purpose 

 in specifying closely spaced nodes because of the low confidence that can be 

 assigned to the nodes located far from soundings. In contrast, when the 

 original data are closely spaced, large X- and Y-increments result in an 

 artificially smoothed surface because too many data points influence each 

 node. Some programs, such as CPS-3, can automatically calculate increments 

 that produce good results for a wide variety of survey patterns. 



Applications and display of gridded data 



Contouring of an area is one of the most common applications of mapping 

 software (Figure 46). Not only is this faster than hand-contouring, but the 

 results are uniform in style across the area and precision (i.e. repeatability) is 

 vastly superior. 



The power of mapping programs is best demonstrated when analyzing 

 different surveys. If at all possible, the different data sets should be gridded 

 with the same algorithms and parameters in order that the results be as 

 comparable as possible. Difficulty arises if earlier surveys contain data much 

 sparser than later surveys. Under these circumstances, it is probably best if 

 the optimum grid is chosen for each data set; the grid produced for the 

 densely sampled survey should not be compromised just to maintain uni- 

 formity with an earlier survey. A simple application is to plot a suitable 

 contour to demonstrate the growth over time of a feature like a shoal 

 (Figure 47). Computation of volumetric changes over time is another applica- 

 tion (Figure 48). This can graphically demonstrate how shoals develop or 

 channels migrate. 



Volumetric data can be used to estimate growth rates of features like 

 shoals. As an example, using all 18 of the 1,000-ft squares shown in 

 Figure 47, the overall change in volume of the East Pass ebb-tidal shoal 

 between 1967 and 1990 was only 19 percent (Figure 49). Although the shoal 

 had clearly grown to the southwest, the minor overall increase in volume 

 suggests that considerable sand may have eroded from the inner portions of 

 the shoal. In contrast, when plotting the change in volume of nine selected 



Chapter 5 Analysis and Interpretation of Coastal Data 



