A quadratic regression was run with catch rates plotted against the 

 average weekly discharge using the model: 



Y = aQ+a,x + a„x2 



where ag, a^ , and ^2 ^'"^ constants 



This regression produces a hyperbola. 



The graphs for the 1975 sturgeon run .depicted in figure 15, show that 

 the optimum discharge for shovel nose sturgeon was between 65 m^/sec (2300 cfs) 

 and 68 m^/sec (2400 cfs). Plotting discharge frequency for 1975 reveals that 

 the most sturgeon were taken in the interval 34 m^/sec (1200 cfs) to 45 m3/sec 

 (1600 cfs). The two discharge figures differ because the optimum is based on 

 catch rates at predicted flows from the regression model and the second 

 interval on the total number of fish caught at each flow. Sampling bias 

 produced by varying section lengths resulted in the difference between the 

 two discharge figures. 



In 1976, the optimum discharges for the linear and trip catch rates were 

 30.90 m3/sec (1090 cfs) and 34.91 m^/sec (1233 cfs), respectively (figure 16). 

 The majority (56.5 percent) of the sturgeon in the discharge frequency 

 histogram were taken at discharges between 11.3 m3/sec (400 cfs) and 22.7 

 m3/sec (800 cfs). Runoff patterns for the years 1975 and 1976 differed, 

 with 1975 considered a high water year. No sturgeon were taken either year 

 when flows fell below 8.50 m3/sec (300 cfs). Sampling efficiency decreased 

 noticeably when flows exceeded 62.29 m3/sec (2200 cfs) because the sampling 

 equipment was more difficult to operate in higher flows. 



Based on the quadratic regressions, it appears that a minimum flow of 

 8.50 m3/sec (300 cfs) is necessary to provide shovelnose sturgeon passage. 

 The upper limit was not estimated due to the decreased sampling efficiency 

 at greater flows. The quadratic regression was utilized because, for any 

 environmental condition, there is typically an intermediate range, with 

 less favorable conditions above and below (Ricker 1975). For example, 

 there may theoretically be flows too low or too high for successful migration; 

 the optimum is intermediate. 



Temperature 



The linear and trip catch rates (dependent variables) of sturgeon were 

 compared with the five-day average minimum and maximum water temperatures 

 (independent variables) using the quadratic regression model described 

 above. The graphs for the 1975 shovelnose run (figure 17) show an optimum 

 minimum temperature of 18.3°C (65°F) for the linear catch rate and 16.9°C 

 (62.5 F) for the trip catch rate. The optimum maximum temperatures for the 

 linear and trip catch rates were 20.0OC (680F) and 18.30C (60.5OF), 

 respectively. In 1976 the minimum linear and trip catch rate averages 



40 



