Unfortunately, correlations between the two methods were not consistently 

 high. All correlation values for the July 1991 experiment were quite low 

 (< 0.002). Some correlation values for the January 1992 experiment were 

 high (near 1.0), but many of these predicted transport in the wrong direction 

 (thus, they had high anti-correlations). White (1994) discusses these problems 

 in more detail and suggests several possible explanations. White (1994) also 

 notes a dismrbing lack of agreement between the various SPM (1984) formula 

 (Equations 6-9) predictions. These issues are the subject of ongoing research. 

 However, because of the disagreement in these results, the OBS data were not 

 used further to determine the average yearly longshore transport rate at the 

 Colorado River. 



Research was conducted to investigate one possible reason for the poor 

 correlations between the two procedures. Sled 1 was usually positioned just 

 outside the breaker line. This sled contained a pressure sensor and three 

 current meters. Though not designed for this purpose, the gages were 

 combined to form a directional wave gage (a Puv gage). Adjustments were 

 made in the data processing because the sensors were not co-located and thus 

 contained small phase shifts in their signals, and realistic directional spectra 

 were obtained for the January 1992 data. These data were then considered to 

 be measured wave values at the breaker line. Equation 6 was used to obtain an 

 additional set of transport rate values. These data were compared with the OBS 

 transport data and with the shoaled 10-m-deep wave gage data using the linear 

 regression method discussed above. The results are shown in Table 15. 



Table 1 5 



Regression and Correlation of Breaker Gage Data with OBS and 



10-m Gage Data for January 1992 



Data Sources 



m 



Slope of 



Regression 



b 

 Intercept 

 (m^/hour) 



r 

 Correlation 

 Coefficient 



X is the breaker gage data 

 y is the OBS data 



-1.9268 



-38.7602 



-0.9397 



X is the breaker gage data 

 y is the shoaled 10 m data 



0.5869 



11.5309 



0.8735 



Number of Spectra, n = 14. 



While both correlations are high, the correlation between the OBS data and 

 the breaker line wave gage is negative. These results suggest that the problem is 

 not a result of shoaling the wave data. However, there are insufficient data to 

 draw firm conclusions. 



Chapter 5 Prediction of Sediment Transport Rates 



49 



