154 



3. Are there risks to migrating smolts and returning adults associated with 

 high levels of dissolved nitrogen resulting from spill? 



Absolutely! I believe that uurroiit dissolved gas levels are directly and indirectly 

 killing salmon smelts in numbers greater tl\an tlie incremental survival benefit from spill. 

 However, I doubt diat tiie smolts typically die directly from gas bubble disease per se; 

 rather, gas bubble disease-weakened smolts are probably eaten rapidly by predators, Gas 

 bubble disease signs are likely to be found in only 1 % of tiie smolts or less at any given 

 time, because they are continuously removed from the river by predators and digested. 

 Thus, predators remove evidence that a supersaturation problem exists, much to the 

 frustration and confusion of all concerned, and we are asked to demonstrate and display 

 that which no longer exists. Additionally, this syndrome makes it esseptiallv impossible 

 to accurately monitor the impact of gas levels from dmns. 



EPA and several states adopted a dissolved gas standard not to exceed 1 10 % of 

 barometric pressure which is roughly equivalent to 75 mm Hg or 1.5 psi above barometric 

 pres-ture. 'fhis criterion was supported by the National Academy of Sciences, and tiie 

 NMFS's Panel on Gas Bubble Disease. Oregon and Washington granted a variance that 

 allows gas levels in the river to average 120 % in the spill and 1 15 % well down river, but 

 in fact, extensive area exceed these levels, and tlie Corps have been sent letters of non 

 compliance from Oregon and Washington. 



I approach the risk analysis for .smolts differently than state fishery- agencies. First 

 I estimated tiie proximate portion of the smolt population that spill could benefit at a 

 hypothetical dam, assuming no mortality' related to spill or gas supersaturation (equation 1 

 below), I assumed tliat spill would increase non turbine passage or fish passage efficiency 

 (FPE) from a conservative low of 60 % to a typical goal of 80 % (equation 2 below), 

 altiiough a rise from 70 % to 80 % would be more typical. I also assimied tliat smolt 

 survival in turbines would be 85 "/o (altliough recent studies support higher survival) and 

 that smolt survival via non turbine routes would be 98%, botii being ft-aditional values. 

 Therefore: 



smolt survival at 60 % FPE = (0.85 X 0.4U) + (0,98 X 0.60) = 92.8 %; (1 ) 



and smolt survival at 80 % FPE - (0.85 X 0.20) + (0,98 X 0.80) = 95.4 <>b. (2) 



The potential benefit of spill at 80 % FPE is equation (2) minus (1) or 95.4% - 

 92.8 % = 2.6 % higher survival per day (3). Thus an 80% FPE might result in 2.6 % 

 higher survival, while a 70'5b FPE would only result in 1,3 "'o higher smolt survival (but 70 

 "/o FPE usually doesn't require spill to achieve it). I believe this calculation illustrates the 

 problem and applies to most Columbia River danis. 



Next I estimated ttie potential adverse impact of gas supersatiiration. Put simply, 

 what is the risk tliat gas supersaturation would kill 3 "o of the smolts per da>, either 

 directly or indirectly, and therefore destroy the potential benefit of spill? To assess the 

 above risk, at least t\vo additional conditions must be known: 



