with a maximum of nine segments allowed per cross-section. The 

 IFG-4 predicted and measured velocities and depths within each 

 segment were averaged in order to obtain data comparable to 

 the WSP output. The water surface elevations and segment velo- 

 cities and depths generated by the two models were statistically 

 compared to the measured values using t-tests for paired data 

 (Table 29) . Only four svibreaches are compared since the WSP 

 model was not applied to the Madison #3 subreach. Cross-section 

 #7 of the Beaverhead subreach was also eliminated from the 

 analysis since the WSP model could not be calibrated to the 

 field data for this cross-section. 



Table 29 . Statistical comparison of the reliability of the predictions 

 of the water surface elevations and the segment velocities 

 and depths generated by the IFG-4 and WSP hydraulic simu- 

 lation models for subreaches of the Madison, Beaverhead, 

 Gallatin and Big Hole Rivers. The number of observations 

 is in parenthesis. 



f 



IFG-4 



WSP 



Subreach 



WSE 



Madison (#1) 0(15) 

 Beaverhead(#2)0(18) 

 Gallatin (#2) 0(21) 

 Big Hole(#l) 0(24) 



X = Predicted values are significantly different (P ^.05) from the 

 measured values. 



= Predicted values are not significantly different (P>. 05) from 

 the measured values. 



83 



