Date of Award

Fall 2018

Document Type


Degree Name

Master of Science (MS) in Biology




Construction of dams dramatically alters the hydrology of rivers, influencing a wide variety of water quality characteristics directly impacting fish species. These physical alterations will interact with increasing temperatures, due to climate change, reducing thermal habitat for cold-water fish species. Rainbow Trout are sensitive to changes in their habitat and prefer water temperatures < 20 °C and DO levels ≥ 5.0 mg/L. The overarching goal of this study is to document Rainbow Trout movements in an environment that experiences oxygen stress and high water temperatures not meeting preferred levels. Lake Spokane is a 38.62 km long reservoir on the Spokane River located between Long Lake and Nine Mile Falls Dam in Washington State. Long Lake has three reaches, riverine, transitional, and lacustrine, and experiences low DO and high temperatures. The objectives of this project were to 1: Determine summer habitat use of Rainbow Trout from tracking data that provides location, depth, and temperature. 2: Compare habitat use and variation between two sequential years. 3: Use water quality and tracking data to predict future habitat characteristics Rainbow Trout will select for, based on past use. Hypotheses were 1: Tagged Rainbow Trout will leave the lower lacustrine zone during the summer period and select locations within the upper lacustrine and transitional zone. 2: Rainbow Trout will inhabit the upper hypolimnion and lower metalimnion for cooler water temperatures and higher DO during the summer period. To determine summer salmonid habitat utilization 20 and 25 triploid Rainbow Trout were acoustically tagged and released in Lake Spokane in the spring of 2017 and 2018 respectively. Weekly tracking events followed a pre-determined grid, and were conducted to determine location, water column depth, and water temperature for each tagged fish. Locations of the Rainbow Trout were then compared to the water quality (DO, pH, and temperature) from six to ten sampling locations. Parameters were collected on a bi-weekly basis from June to September, and monthly for May and October. Data analysis consisted of using tracking date to make monthly kernel density maps in Geographic Information Systems and a combination of tracking data and water quality data for a range bagging model in Program R. Monthly kernel density maps showed locations selected during two summers (2017 and 2018). The probability of presence of Rainbow Trout will be determined from water quality data and Rainbow Trout depth and temperature selections. Range bagging is the presence-only model in Program R that was used to store, average, and predict frequency of selections of depth, temperature, DO, and pH. Kernel density maps for 2017 and 2018 revealed similar patterns of selected locations. Early in the tracking season (July and April respectively) revealed Rainbow Trout primarily selecting areas in the mid-range of the lacustrine reach, with a few individuals within the mid-range of the transitional reach. As the tracking season progressed into late July, selections in 2018 were within the transitional reach unlike the previous year. August selections were nearly identical with Rainbow Trout selecting locations throughout the entire reservoir, except for the riverine reach. As temperatures cooled in September Rainbow Trout were found to school in similar areas throughout the mid-range of the lacustrine and transitional reaches. October was the only month that revealed different patterns of use between years. In 2017 the Rainbow Trout remained in small schools of fish in the two reaches, but in 2018 the fish exhibited the same pattern as August of both years. Depth selections in 2017 were more variable earlier in the season between 0 and 16 m. After the August 4th tracking event, depth selections averaged less than 6 m. Water temperatures were correlated with depths, fish that were shallow experienced warmer temperatures as high as 23.6 °C. As temperatures cooled, temperatures at the surface were as low as 8.4 °C. Depth selections in 2018 were more consistent across the entire season with two patterns being revealed. Either Rainbow Trout selected depths less than 6 m or were found between 6 and 15.6 m. Therefore Rainbow Trout experienced warmer temperatures of 18.0 to 20.4 °C during the peak of the season. Deeper fish were found at temperatures at, or below, 19.6 °C. When temperatures cooled, fish were found at 14.8 °C. Range bagging for both years revealed that Rainbow Trout were more likely to select intermediate water temperatures within the range of 0 to 30 °C and shallower depths (< 16 m) in the range of 0 to 20 m. With those selections falling between 10 and 22 °C and depths of 5 m or less. The predicted DO and temperature range bagging revealed that fish were going to select warmer temperatures (nearing 25 °C) at a wide range of DO between 4.0 and 13.0 mg/L. However, the 2018 model showed fewer Rainbow Trout selecting DO as low as 4.0 mg/L, falling between 5.0 and 11.0 mg/L. The final range bagging models for predicted DO and predicted pH revealed a wide range of DO being selected in 2017 with a higher probability of presence between 7.5 and 13.0 mg/L for 2018. Similarly, the predicted pH was over a wider range between 4 and 9 for 2017, but fell between 6 and 9 for 2018. The most surprising conclusions of this study were that Rainbow Trout were selecting depths within the epilimnion and upper metalimnion. Therefore were found at temperatures bordering the upper temperature limit at 20 °C and above. Furthermore, these fish selected locations throughout much of the reservoir during the warmest period of the season in August, the time period expected for Rainbow Trout to find cold-water refuge. Literature suggests that there were more location selections due to increased movement to locate coldwater refuge, perhaps not finding it, and surviving until temperatures cooled. Although they did not appear to be selecting the riverine section, as shown by the data, there were individuals that were not during the 2018 season, and then were found in the upper transitional reach. The lack of data for location selections within the riverine reach are believed to be due to the limitations of acoustic telemetry. Therefore, tracking adaptations would need to be made for future studies. The range bagging model could be improved with using water quality data set over a finer scale from a prediction model developed by Portland State University. For example, the probability of presence of Rainbow Trout were shown at DO levels as low as 4.0 mg/L, this could not occur because it is the level of asphyxiation. In addition to improving the predictive power of the range bagging model and changing tracking methods, a diet and growth study could be conducted to understand Rainbow Trout health. Although Rainbow Trout in this study were shown to have survived selecting warmer water temperatures, particularly in August, these fish may be thermally stressed and attempting to locate cold-water refuge. These fish may be pushed towards the surface due to lower DO levels near 5.0 mg/L. Although this is still within the suitable limit for Rainbow Trout to survive, it is not at 9.0 mg/L and above which are suggested when temperatures are ≥ 15.0 °C to reduce stress. Therefore, even though these selections are made, there is still concern for increasing water temperatures and low DO with climate change. Studies of Rainbow Trout responses to climate change and habitat selections and understand behavior under, what had traditionally been believed to be, marginal conditions.

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Included in

Biology Commons