Quantifying the Behavioral Response of Spawning Chum Salmon to Elevated Discharges from Bonneville Dam, Columbia River : Annual Report 2005-2006.

In unimpounded rivers, Pacific salmon (Oncorhynchus spp.) typically spawn under relatively stable stream flows, with exceptions occurring during periodic precipitation events. In contrast, hydroelectric development has often resulted in an artificial hydrograph characterized by rapid changes in discharge and tailwater elevation that occur on a daily, or even an hourly basis, due to power generation (Cushman 1985; Moog 1993). Consequently, populations of Pacific salmon that are known to spawn in main-stem habitats below hydroelectric dams face the risks of changing habitat suitability, potential redd dewatering, and uncertain spawning success (Hamilton and Buell 1976; Chapman et al. 1986; Dauble et al. 1999; Garland et al. 2003; Connor and Pflug 2004; McMichael et al. 2005). Although the direct effects of a variable hydrograph, such as redd dewatering are apparent, specific effects on spawning behavior remain largely unexplored. Chum salmon (O. keta) that spawn below Bonneville Dam on the Columbia River are particularly vulnerable to the effects of water level fluctuations. Although chum salmon generally spawn in smaller tributaries (Johnson et al. 1997), many fish spawn in main-stem habitats below Bonneville Dam near Ives Island (Tomaro et al. 2007; Figure 1). The primary spawning area near Ives Island is shallow and sensitive to changes in water level caused by hydroelectric power generation at Bonneville Dam. In the past, fluctuating water levels have dewatered redds and changed the amount of available spawning habitat (Garland et al. 2003). To minimize these effects, fishery managers attempt to maintain a stable tailwater elevation at Bonneville Dam of 3.5 m (above mean sea level) during spawning, which ensures adequate water is provided to the primary chum salmon spawning area below the mouth of Hamilton Creek (Figure 1). Given the uncertainty of winter precipitation and water supply, this strategy has been effective at restricting spawning to a specific riverbed elevation and providing minimum spawning flows that have the greatest chance of being maintained through egg incubation and fry emergence. However, managing the lower Columbia River for a stable tailwater elevation does not provide much operational flexibility at Bonneville Dam, which has little storage capacity. When river discharges increase due to rain events, the traditional approach has been to pass excess water at night to maintain stable tailwater elevations during the daytime. The underlying assumption of this strategy, referred to as reverse load following, is that fish do not spawn at night. However, Tiffan et al. (2005) showed that this assumption is false by documenting nighttime spawning by chum salmon in the Ives Island area. Similarly, McMichael et al. (2005) reported nighttime spawning by Chinook salmon (O. tshawytscha) in the Columbia River, indicating that diel spawning may be a common occurrence in Pacific salmon. During the latter portion of the chum spawning period in December 2003 and 2004, discharges from Bonneville Dam increased from an average of 3,398 m3/s (tailwater elevation {approx} 3.5 m above mean sea level) during the day to over 5,664 m3/s (tailwater elevation {approx} 5.1 m) at night, with peak discharges of 7,080 m{sup 3}/s (tailwater elevation {approx} 6.1 m). This caused concern among fishery managers regarding the potential effects of these high discharges on this population of spawning chum salmon, which is listed under the Endangered Species Act (National Oceanic and Atmospheric Administration 1999). We hypothesized that increased water velocities associated with elevated tailwaters might alter chum salmon spawning behavior if water velocities at redd locations increased beyond the range of suitability (>0.8 m/s; Salo 1991). In 2005, we investigated the movement and behavioral responses of spawning chum salmon at Ives Island to increased tailwater elevations at Bonneville Dam. We used acoustic telemetry to determine if the higher velocities associated with increased tailwater elevations caused fish to leave their redds. We related the duration fish were away from redds and the distances moved to water velocities estimated from a two-dimensional hydrodynamic model. Finally, we described specific changes in spawning behavior (e.g., nest digging; swimming activity) during elevated-tailwater tests using a dual-frequency identification sonar (DIDSON)