"loose" sand in the case of uniform well-rounded sand grains, slightly 

 loose for uniform subangular sand, and the loosest possible for well- 

 graded subangular sand. 



However, it is probable that sand along the shoreline should be more 

 on the dense side, rather than on the loose side because of the compact- 

 ing effects of water soaking and wave action. Moreover, it is evident 

 that the weight density will vary with the grain-size distribution and 

 grain shape (Sowers and Sowers, 1970). Thus, the constant a' = 0.60 is 

 an assumption not likely to be generally true. 



Unit volume is equal to the reciprocal of the unit weight. Since 

 uniform sands may vary 24 percent in unit weight (Table), the same 

 variation may occur in unit volume. Since most independent, local field 

 estimates of longshore transport are based on surveys of sand volumes, 

 it is possible that the energy flux prediction can be significantly 

 affected by variation in unit volume of the sand. 



For example, Caldwell's (1956) and Komar's (1969) data are from 

 surveys of the nearshore zone subject to wave action, and these data are 

 19 of the 23 data points used to establish the SPM longshore transport 

 curve (Figure 4-37 in SPM). If the sand settled out in quieter waters, 

 the same number of sand grains might be expected to yield larger surveyed 

 volumes. This possibility is consistent with the data from Channel 

 Islands Harbor (Bruno and Gable, 1977) and Santa Barbara (Galvin and 

 Vitale, 1977), California, which do plot above the SPM curve (Fig. 1). 

 (However, even a 24-percent decrease in unit volume would only bring 

 these California data points about 20 percent closer to the SPM curve on 

 that log- log plot.) 



III. PRESENT USE OF IMMERSED WEIGHT CALCULATION 



The immersed weight formulation has been strongly recommended by 

 some for longshore transport prediction. As shown by equation (3), the 

 immersed weight rate equals the volume rate multiplied by two sand- 

 related parameters, a 1 (eq. 8) and y' (eq. 4). 



However, the present use of In with the SPM design curve (eq. 2) 

 implies a constant unit weight which probably was lacking in the under- 

 lying data. The existing design curve in SPM is based on three sets of 

 field data for which a' and even y^ are not available. One set meas- 

 ures short-term volume changes in the high tide surf zone (Komar, 1969); 

 the second set measures longer term variations in the littoral zone 

 (Caldwell, 1956); and the third set measures pumping rates of probable 

 carbonate sand (Watts, 1953). Thus, there is a good deal of uncertainty 

 in a' and Yg for all three sets, and a 1 in particular is likely to 

 be different in each set of data. Therefore, it is probable that all 

 three sets of data involve slightly different unit weights of sand. 



Those studies that use Ijj, have assumed a Yg for quartz sand and 

 a' = 0.6 to compute 1^. This is permissible when other data are lacking. 



10 



