688 EIGHTH PACIFIC SCIENCE CONGRESS 



of the rivers are governed by the major water sources. Lowland rivers 

 peak during the winter from rainfall. Those rivers having mountainous 

 areas as their principal watershed peak during the late spring from 

 snow melt. Others may have both peaks. Hydroelectric developments 

 on some rivers tend to flatten the peaks and maintain a more uniform 

 runoff. Discharge varies greatly with flood stages being as much as 150 

 to 200 times the minimum flow. 



The population of the area is over 1.5 million, the principal sea- 

 board cities being Seattle, Tacoma, Bremerton, Port Townsend, Everett, 

 Olympia, and Shelton. Virtually all sewage from the municipal systems 

 is discharged untreated into the Sound. While industrialization of the 

 area is not heavy, some industries, particularly those manufacturing 

 pulp and paper, are potential pollution hazards because of the large 

 volume and high oxygen demand of the wastes. 



Construction of the Model 

 It is common practice to make tidal models as large as is consistent 

 with space and facilities available and cost. The horizontal scale is or- 

 dinarily determined by these empirical factors. The vertical scale 

 is governed by the practical requirements for suitable time scale and 

 water depth in critical areas and by the theories of dynamic and kine- 

 matic similitude. Space available limited the present model to a hori- 

 zontal scale of 1:40,000. The vertical scale selected was 1:1,152 or 

 1/16-inch per fathom, giving a distortion of 34.6:1. With these scales, 

 the model is approximately 7 feet by 15 feet and has a maximum water 

 depth of 10 inches with l-i/^ to 2 inches over the critical sills. As a 

 consequence of the distortion, some channels are deeper than they are 

 wide. The time scale from the Froude modeling law is 1:1,178 or 3.05 

 seconds per hour, permitting a year's tides to be observed in 7.4 hours. 

 With this choice of scales, the Reynold's number in the model will be 

 in the turbulent region over the critical sills, for most of the tidal 

 cycle. That the motion is actually turbulent is easily verified by in- 

 serting dye into the model and watching its rapid diffusion. The ap- 

 propriate criterion for turbulence in a fluid with stable stratification 



is given by the Richardson number ^ , j^' f , • With Froude modeling, 

 ^ ^ p(du/dzy 



the salinity scale must be unity to give the same Richardson number 

 in model and prototype. With a reduced salinity scale, the mixing in 

 the model could be expected to be increased. 



The customary method of constructing model basins consists of 

 moulding the contours to conform to templates fixed in position within 

 the model area. This technique was found unsatisfactory for reproduc- 



