period of the 2.5-min basin tests, only the midflow speed could be measured 

 and subsequently used as the independent variable. The different types of 

 equations are as indicated below: 



Linear: F - aV + b (11) 



Exponential: F - ae (12) 



Power: F - aV b (13) 



Power Threshold: F - a(V - V.) b (14) 



where 



F = sand flux, in g/cm 2 /nri.n 



a = empirically determined coefficient, with dimensions 

 consistent with variables in equation 



V = either midflow (V mid ) or bottom flow (V bot ) speed in 

 cm/sec 



b = empirically determined coefficient, with dimensions 

 consistent with variables in equation 



V* = threshold midflow (28 cm/sec) or bottom (18 cm/sec) flow 

 speed for sand movement, determined experimentally, 

 in cm/sec 



85. Table 6 presents the empirically determined coefficients and 

 squared correlation coefficients resulting from regression analyses for each 

 type of fit for each type of nozzle and basin test. Both the power fit and 

 threshold power fit with the midflow speed had identical average squared 

 correlation coefficients (r 2 = 0.94). Because most modern sediment transport 

 formulae incorporate a threshold flow speed or shear velocity, the threshold 

 power fit was chosen to represent the data and for use in further analysis. 

 The term "threshold flow speed" V* used in this analysis refers to the 

 measured flow speed either in the middle or bottom (approximately 0.5 cm above 

 the bed) of the flow corresponding to incipient motion. The threshold flow 

 speed is related, but not equal, to the critical shear velocity U* c , the 

 shear velocity (discussed previously in the section entitled "Flow condi- 

 tions") at incipient motion. Midflow as well as bottom and threshold flow 

 speeds were used in this analysis rather than corresponding values of shear 



57 



