Effects of Turbine Operating Efficiency on Smolt Passage Survival John R. Skalski University of Washington and Dilip Mathur Normandeau Associates History of Presentation • Turbine Passage Survival Workshop, Portland, Oregon, 14-15 June 2000 • Skalski, Mathur, and Heisey, 2002, North American Journal of Fisheries Management 22:1193-2000 • New input from 2002 McNary Dam studies, US Army Corps of Engineers Purpose of Talk Review of existing data on relationships between smolt survival and turbine operating efficiency using: 1. Bell (1981) report 2. Analyses of planned experiments – Lower Granite Dam 1995 – Wanapum Dam 1996 – Rocky Reach Dam 1997 – Bonneville Dam 2000 – McNary Dam 2002 3. Meta-analysis across multiple projects Background 1994 Columbia River Basin Fish and Wildlife Program 5.6D.1 “Operate turbine units within 1 percent of peak operating efficiency from April through August of each year, and especially during peak migration periods.” Background (con’t) 1995 FCRPS Biological Opinion (IV) • “Improvement in the operational control of turbine units, allowing operation within one percent of peak efficiency at all eight mainstem federal dams on the Snake and Columbia Rivers.” • “Turbine survival is directly related to turbine efficiency.” • “ . But the precise benefits of increased turbine efficiency . are unknown.” Background (con’t) 2000 FCRPS Final Biological Opinion Action 58: “The Corps and BPA . shall operate all turbine units at FCRPS dams for optimum fish passage survival. The Corps and BPA will operate turbines within 1% of peak efficiency during juvenile and adult migration seasons . .” 10.5.1.8 Monitor Turbine Efficiency “The BPA and the Corps shall provide an annual summary report detailing compliance with the 1% peak efficiency turbine operation guideline . .” Historical Information Bell (1981) Report Big Cliff Dam Study (1964, 1966) Evaluated Kaplan turbine unit • 6 blades • 164 rpm • Head at 90 ft Study design • Release at 1 location Fish • Chinook salmon (mean length 101 mm) Bell (1981) Kaplan Turbine at Big Cliff Dam Regression Year Efficiency Range r 2 P-value 1964 32.5% - 96% 0.254 0.017* 1966 32.5% - 96% 0.020 0.258 Combined 32.5% - 96% 0.112 0.003* Survival vs. Efficiency Plot Big Cliff (1964 and 1966) Sˆ Efficiency Bell (1981) “There does not seem to be a smooth ascending and descending curve following the efficiency line of the turbines as might have been expected.” “The data offer some support, however, . the best points of machine efficiency should give the best points of fish passage survival.” Bell (1981) •Did not actually measure turbine efficiency – Instead used % wicket-gate opening as surrogate •Hence, no actual data on efficiency-survival relationship Site-Specific Studies Columbia Basin • Lower Granite, 1995 • Wanapum, 1996 • Rocky Reach, 1997 • Bonneville, 2000 • McNary, 2002 Lower Granite Dam Study (1995) Evaluated turbine unit 4 •Kaplan • 6 blades •90 rpm • Head at 98 ft Study design • 3 turbine operating levels – Low end of +1% –Normal +1% – Cavitation mode • 1 mid-release location Fish • Spring chinook salmon (mean length 150 mm) Survival Plot: Lower Granite Dam (1995) Peak efficiency 1.00 0.99 l 0.98 a v i 0.97 v r u 0.96 S d 0.95 te a 0.94 m 0.93 ti s E 0.92 0.91 0.90 130001350014000 15000 16000 17000 18000 19000 20000 Discharge (cfs) Wanapum Dam Study (1996) Evaluated turbine unit 9 • Kaplan with adjustable blades • 5 blades • 87.5 rpm • Head at 74.5 ft Study design • 4 turbine operating levels • 2 release locations – 10 ft and 30 ft below intake ceiling Fish • Coho salmon (mean length 154 mm) Survival Plot: Wanapum Dam 10 Feet 30 Feet Efficiency Peak efficiency 0.99 99% l a 0.97 97% v i v cy r 0.95 95% en Su 0.93 93% ci d fi e f t a 0.91 91% E m % i t 0.89 89% Es 0.87 87% 0.85 85% 7000 9000 11000 13000 15000 17000 Discharge (cfs) Rocky Reach Dam Study (1997) Evaluated turbine unit •Kaplan • 6 blades •90 rpm • Head at 92 ft Study design • 3 turbine operating levels • 2 release locations – 10 ft and 30 ft below intake ceiling Fish • Spring chinook salmon (mean length 184 mm) Survival Plot: Rocky Reach Dam Peak efficiency 10 Feet 30 Feet Efficiency 1 95% 0.99 94% l a 0.98 v i v 93% cy r 0.97 u en S ci i d 0.96 92% f f te a 0.95 E m 91% % ti s 0.94 E 90% 0.93 0.92 89% 6000 8000 10000 12000 14000 16000 Discharge (cfs) Bonneville Dam Study (2000) Evaluated turbine unit 5 • Kaplan with adjustable blades • 5 blades •75 rpm • Head at 57 ft Study design • 4 turbine operating levels • 3 release locations – Hub, Tip, and Mid-blade Fish • Spring chinook salmon (mean length 155 mm) Survival Plot: Bonneville Dam (2000) Tip Mid Efficiency Peak efficiency 1.00 90% 0.99 90% l a 0.98 89% iv y v 0.97 c r 0.96 89% en Su ci d 0.95 88% i f e f t a 0.94 88% E m % i 0.93 t 87% Es 0.92 0.91 87% 0.90 86% 5000 6200 7000 9000 1050011000 12000 13000 Discharge (cfs) Survival Plot: Bonneville Dam (2000) Hub Efficiency Peak efficiency 1.02 90% l 90% a 1.00 v i 89% v r 0.98 cy u 89% n e S ci d 0.96 88% fi e f t a 88% 0.94 E m % i t 87% Es 0.92 87% 0.90 86% 5000 6200 7000 9000 1050011000 12000 13000 Discharge (cfs) McNary Dam Study (2002) Evaluated turbine unit 9 • Kaplan with adjustable blades • Six blades • 85.7 rpm • Head at 75 ft Study design • 4 turbine operating levels in April • 2 levels repeated in May • 1 release location below screen Fish • Spring chinook salmon (mean length 155 mm April, 140 mm May) Survival Plot: McNary Dam (2000) April May Efficiency Peak Efficiency 1.00 1.00 l a v 0.98 i 0.95 v r cy u 0.96 S en d 0.90 ci i te f 0.94 f a E m ti 0.85 s 0.92 E 0.90 0.80 7 8 9 10 11 12 13 14 15 16 17 18 Discharge (Kcfs) Multi-Project Analysis Across Country • Meta-analysis across: − 51 trials − 17 turbine units − 13 projects • All Kaplan turbines Multi-Project Analysis • Assess correlation between turbine passage survival and – Efficiency – Mean length – Number of blades – Speed (rpm) –Head – Strike probability Results of Regression Analysis Factor P-Value r 2 Efficiency 0.2933 0.0298 Mean Length 0.0149* 0.1173 No. of Blades 0.2498 0.0275 Speed 0.1021 0.0547 Head 0.2412 0.0285 Strike Probability 0.0157* 0.1157 Scatterplot: Survival vs. Efficiency non-salmonid salmonid 102 100 98 l a v i 96 v r 94 Su d te 92 a m ti 90 Es 88 2 86 r = 0.0311 84 20% 30% 40% 50% 60% 70% 80% 90% 100% 110% % Efficiency Conclusions • Bell (1981): Did not measure turbine efficiency despite wide-spread interpretation • Lower Granite (1995): Peak survival ≠ peak efficiency • Wanapum (1996): Peak survival ≠ peak efficiency • Rocky Reach (1997): Peak survival ≠ peak efficiency • Bonneville (2000): Possible peak efficiency relationship • McNary (2002): Peak survival ≠ peak efficiency – In 8 of 10 curves, peak survival not at peak efficiency • Meta-analysis: No efficiency relationship, r2 = 0.030 Conclusions (cont’d) • Turbine efficiency curves for Kaplans have broad shallow shapes –+1% rule encompasses wide range of discharges – Range generally includes maximum survival • Peak survival usually not at peak efficiency • Difference between: Peak Survival and Up to 3.2% Survival at Peak Efficiency Conclusions (cont’d) • Potential benefits from operating turbines at peak survival • Rule needs to be examined for new generation of turbines.