A amp (f h 6j = nodal amplification factor for HARBD computational 

 frequency f t and direction n 



N T = number of HARBD computational wave periods 



N D = number of HARBD computational wave directions 



Finite element grids 



The finite element numerical grid depicting existing conditions at Kahului 

 Harbor was created using WES's finite element grid development software 

 (Turner and Baptista 1993) (Figure 38). The grid covers the entire Kahului 

 Harbor area and extends somewhat seaward into Kahului Bay. The land 

 boundary was digitized from an aerial photograph. Grid element size is based on 

 the criterion of 6 elements per wavelength (the minimum recommended resolu- 

 tion with HARBD) for a 10-sec wave in 15-ft water depth. Depths for virtually 

 all areas of interest exceed 15ft. For the longer period waves, the grid gives a 

 high degree of resolution. Grid characteristics are summarized in Table 11. 



The radius of the seaward semicircle is 2,307 ft. This is equivalent to 2.9 and 

 9.7 wavelengths for the longest and shortest short wave periods considered, 

 assuming a representative water depth of 35 ft. The semicircle size and location 

 were chosen to include both breakwaters and the immediate nearshore area. The 

 semicircle extends sufficiently far seaward to cover the most important nearshore 

 bathymetry while rriamtaining a reasonable number of grid elements. 



Bathymetric data were obtained from National Oceanic and Atmospheric 

 Aclministration hydrographic chart 19342 and WES bathymetric survey data. 

 Digitized depths were transferred onto the finite element grid using the WES grid 

 software package. A contour plot of bathymetry is given in Figure 39. 



Reflection coefficients K r are needed for all solid boundaries. For the short 

 wave tests, K r values were estimated from existing Corps of Engineers guidance, 

 photos, and field notes from a recent site visit by WES personnel, and past 

 experience. The solid boundary was divided into 13 zones and a reflection 

 coefficient was estimated for each zone (Figure 40). Reflection coefficients 

 ranged from 0.2 for the shallow sandy beach along the southwest shore of the 

 existing harbor to 0.5 for all pier areas and 0.9 for the grouted revetment along 

 the western side of Pier 2. Additional parameter values used in the numerical 

 model are summarized in Table 12. 



Different parameters are used for the long wave tests. The reflection 

 coefficient was set to 1.0 for all boundaries, since long waves generally reflect 

 very well from a coastal boundary. Long waves are more affected by bottom 

 friction than short waves, so a value of p greater than zero is appropriate. The 

 value of P is best determined by calibration with field data, as discussed in the 

 following section. A value of p=0.032 was selected. This and other parameters 

 are summarized in Table 12. 



52 



Chapter 4 Numerical Model 



