Lower Columbia River (CLBMON #42[A]) and Kootenay River Fish Stranding Assessments: Annual summary

Author: Golder Associates Ltd

The main objective of the monitoring program was to collect fish stranding data to assess the impact of flow reductions from Hugh L. Keenleyside Dam/Arrow Lakes Generating Station (HLK/ALH) on native fish species of the lower Columbia and Kootenay rivers. The program assessed fish stranding at pre-determined sites (Appendix A) between HLK/ALH and the Canada/USA border. Secondary objectives addressed in this study included: 1) determining whether physical habitat manipulation (e.g., re-contouring the shoreline) reduced incidences of fish stranding in the lower Columbia River; and 2) reducing (through risk management strategies) the number of occurrences when stranding crews need to be deployed during flow reductions (BC Hydro 2007).

MQ1: Is there a ramping rate (fast vs. slow, day vs. night) for flow reductions from HLK/ALH that reduces the number of fish stranded (interstitially and pool) per flow reduction event in the summer and winter?
Ramping rates within the range of variability experienced in the previous years of study were not a statistically significant predictor of fish stranding in the Columbia and Kootenay river systems (Golder 2007, Golder/Poisson 2010, Golder 2016). Given these results the ramping rate component of this hypothesis is not rejected. Previous studies indicate that time of day (day vs night) was not a significant variable for stranding risk (Golder and Poisson 2010). However, this finding is based on limited data. Time of day ramping studies were not conducted during the present study. The time of day component of this hypothesis can not be rejected and must be deferred until additional time of day ramping experiments are conducted. Additional ramping experiments are outside of the scope of the present study, therefore this component of the hypothesis will not be addressed.

MQ2: Does wetted history (length of time the habitat has been wetted prior to the flow reduction) influence the number of fish stranded (interstitially and pool) per flow reduction event for flow reductions from HLK/ALH?
A significant increase in the number of stranded fish was observed after a 10-day wetted history, although the effect size (proportion of the population affected and the response to wetted histories of variable lengths greater than 10 days) has not been accurately quantified (Golder and Poisson 2010). Previous studies suggest that this hypothesis can be rejected, however Golder recommends that the hypothesis be deferred and the feasibility of using River2D models from Golder (2013) in Year 11 should be considered. These models allow for quantification of fluctuations in river stage at select stranding sites and can be used to calculate wetted history.

MQ3: Can a conditioning flow (temporary, one step, flow reduction of approximately 2 hours to the final target dam discharge that occurs prior to the final flow change) from HLK/ALH reduce the stranding rate of fish?
This hypothesis cannot be rejected at this time and must be deferred due to limited data (Golder 2007, Golder/Poisson 2010). Conditioning flow studies were not conducted during the present study. For a definitive answer to this management question an experimental conditioning flow study including manipulation of flows with significant time between replicates would be required. Experimental studies would be both costly and may lead to significant mortality as indicated by previous studies (Poisson and Golder 2010). Abandonment of this mitigation method should be considered if there is no desire for conducting conditioning flow experiments.

MQ4: Can physical habitat works (i.e., re-contouring) reduce the incidence of fish stranding in high risk areas?
Previous analysis (Golder and Poisson 2010, Irvine et al. 2014) and the results of recent re-contouring efforts suggest that this hypothesis can be rejected; however Golder recommends the hypothesis should be deferred to Year 11 of the current study to allow for a statistical analysis of stranding assessments on all re-contoured sites. The effect size (the proportion of the population or the relative number of fish not stranded as a result of the physical habitat works) has not been estimated for previously re-contoured sites due to limited data. Additional stranding assessments at varied discharge volumes at recently re-contoured sites (Lions Head and Ft. Shepherd Launch [RUB]) is recommended.

MQ5: Does the continued collection of stranding data, and upgrading of the lower Columbia River stranding protocol, limit the number of occurrences when stranding crews need to be deployed due to flow reductions from HLK/ALH?
Based on 10 years of data collection this hypothesis is rejected. Continued collection of stranding data and updating the Lower Columbia River Fish Stranding Database has not decreased the number of stranding events where crews were deployed. During the previous 8 years, 84% of HLK/ALH reduction events initiated stranding assessments. During the present study (1 April 2016 to 1 April 2017), 100% of HLK/ALH reduction events initiated a stranding assessment.

MQ1: Is there a ramping rate (fast vs. slow, day vs. night) for flow reductions from HLK/ALH that reduces the number of fish stranded (interstitially and pool) per flow reduction event in the summer and winter?
Ramping rates within the range of variability experienced in the previous years of study were not a statistically significant predictor of fish stranding in the Columbia and Kootenay river systems (Golder 2007, Golder/Poisson 2010, Golder 2016). Given these results the ramping rate component of this hypothesis is not rejected. Previous studies indicate that time of day (day vs night) was not a significant variable for stranding risk (Golder and Poisson 2010). However, this finding is based on limited data. Time of day ramping studies were not conducted during the present study. The time of day component of this hypothesis can not be rejected and must be deferred until additional time of day ramping experiments are conducted. Additional ramping experiments are outside of the scope of the present study, therefore this component of the hypothesis will not be addressed.

MQ2: Does wetted history (length of time the habitat has been wetted prior to the flow reduction) influence the number of fish stranded (interstitially and pool) per flow reduction event for flow reductions from HLK/ALH?
A significant increase in the number of stranded fish was observed after a 10-day wetted history, although the effect size (proportion of the population affected and the response to wetted histories of variable lengths greater than 10 days) has not been accurately quantified (Golder and Poisson 2010). Previous studies suggest that this hypothesis can be rejected, however Golder recommends that the hypothesis be deferred and the feasibility of using River2D models from Golder (2013) in Year 11 should be considered. These models allow for quantification of fluctuations in river stage at select stranding sites and can be used to calculate wetted history.

MQ3: Can a conditioning flow (temporary, one step, flow reduction of approximately 2 hours to the final target dam discharge that occurs prior to the final flow change) from HLK/ALH reduce the stranding rate of fish?
This hypothesis cannot be rejected at this time and must be deferred due to limited data (Golder 2007, Golder/Poisson 2010). Conditioning flow studies were not conducted during the present study. For a definitive answer to this management question an experimental conditioning flow study including manipulation of flows with significant time between replicates would be required. Experimental studies would be both costly and may lead to significant mortality as indicated by previous studies (Poisson and Golder 2010). Abandonment of this mitigation method should be considered if there is no desire for conducting conditioning flow experiments.

MQ4: Can physical habitat works (i.e., re-contouring) reduce the incidence of fish stranding in high risk areas?
Previous analysis (Golder and Poisson 2010, Irvine et al. 2014) and the results of recent re-contouring efforts suggest that this hypothesis can be rejected; however Golder recommends the hypothesis should be deferred to Year 11 of the current study to allow for a statistical analysis of stranding assessments on all re-contoured sites. The effect size (the proportion of the population or the relative number of fish not stranded as a result of the physical habitat works) has not been estimated for previously re-contoured sites due to limited data. Additional stranding assessments at varied discharge volumes at recently re-contoured sites (Lions Head and Ft. Shepherd Launch [RUB]) is recommended.

MQ5: Does the continued collection of stranding data, and upgrading of the lower Columbia River stranding protocol, limit the number of occurrences when stranding crews need to be deployed due to flow reductions from HLK/ALH?
Based on 10 years of data collection this hypothesis is rejected. Continued collection of stranding data and updating the Lower Columbia River Fish Stranding Database has not decreased the number of stranding events where crews were deployed. During the previous 8 years, 84% of HLK/ALH reduction events initiated stranding assessments. During the present study (1 April 2016 to 1 April 2017), 100% of HLK/ALH reduction events initiated a stranding assessment.