Modeling the Effects of Supersaturated Dissolved Gas on Resident Aquatic Biota in the Main-Stem Snake and Columbia Rivers

Abstract

Dissolved-gas levels in the Columbia and Snake rivers during the spring freshet often exceed 110% of saturation, the maximum level permitted by the U.S. Environmental Protection Agency. The highest levels of supersaturation result from high springtime river flows and turbine outages, conditions over which there is little control and that cause high volumes of water passing over the dams and through spillways instead of through turbines. During the spring freshets of 1994–1997, we surveyed nonsalmonid fishes and invertebrates for signs of gas bubble disease (GBD) and conducted holding experiments in three river reaches where gas saturation commonly exceeds 120%. We developed a dissolved-gas exposure index for nonsalmonid fishes sampled at specific times and locations; mean daily total dissolved-gas saturation (TDGS) was ranked and then summed over a 7-d period. We analyzed observations of 39,924 nonsalmonid fishes in an iterative process, which led to development of a mathematical equivalence model for TDGS duration and level of exposure that was strongly correlated with prevalence of GBD signs (R² = 0.79). We believe this simple model is a reliable predictor of external GBD signs resulting from prolonged exposure to supersaturated dissolved gas in the Columbia and Snake rivers. When TDGS levels were below 120%, GBD signs in fish were rare; when TDGS levels exceeded 120%, the model reliably predicted the extent to which fish displayed external GBD signs. In addition, we attempted to evaluate GBD-related mortality over the 4 years of net-pen holding experiments. However, due to the high variability we observed in these evaluations and the paucity of dead fish recovered from these rivers, we concluded that an accurate model relating TDGS to mortality could not be developed using these methods.