The influence of stream flow on survival of emigrating juvenile (smolts) Pacific salmon Oncorhynchus spp. and steelhead trout O. mykiss is of key management interest. However, few studies have quantified flow effects on smolt migration survival, and available information does not indicate a consistent flow-survival relationship within the typical range of flows under management control. It is hypothesized that smolt migration and dam passage survival are positively correlated with stream flow because higher flows increase migration rates, potentially reducing exposure to predation, and reduce delays in reservoirs. However, available empirical data are somewhat equivocal concerning the influence of flow on smolt survival and the underlying mechanisms driving this relationship. Stream flow effects on survival of emigrating anadromous salmonids in the Yakima Basin have concerned water users and fisheries managers for over 20 years, and previous studies do not provide sufficient information at the resolution necessary to inform water operations, which typically occur on a small spatiotemporal scale. Using a series of controlled flow releases from 2012-2014, combined with radio telemetry, we quantified the relationship between flow and smolt survival from Roza Dam 208 km downstream to the Yakima River mouth, as well as for specific routes of passage at Roza Dam. A novel multistate mark-recapture model accounted for weekly variation in flow conditions experienced by radio-tagged fish. Groups of fish were captured and radio-tagged at Roza Dam and released at two locations, upstream at the Big Pines Campground (river kilometer [rkm] 211) and downstream in the Roza Dam tailrace (rkm 208). A total of 904 hatchery-origin yearling Chinook salmon O. tshawytscha were captured in the Roza Dam fish bypass, radio-tagged and released upstream of Roza Dam. Two hundred thirty seven fish were released in the tailrace of Roza Dam. Fish released in the tailrace of Roza Dam were tagged concurrently with fish released upstream of the dam using identical tagging methods. Tagging and release events were conducted to target a range of flow conditions indicative of flows observed during the typical migration period (March-May) for juvenile spring Chinook salmon in the Yakima River. Three, five and four separate upstream releases were conducted in 2012, 2013, and 2014 respectively, and at least 43 fish were released alive on each occasion. The release sample sizes in 2014 were much larger (~130) compared to previous years for the purpose of increasing precision of survival estimates across the range of flows tested. Migration movements of radio-tagged spring Chinook salmon smolts were monitored with an array of telemetry receiver stations (fixed sites) that extended 208 rkm downstream from the forebay of Roza Dam to the mouth of the Yakima River. Fixed monitoring sites included the forebay of Roza Dam (rkm 208), the tailrace of Roza Dam (rkm 207.9), the mouth of Wenas Creek (rkm 199.2), the mouth of the Naches River (two sites, rkm 189.4), Sunnyside Dam (two sites, rkm 169.1), Prosser Dam (rkm 77.2), and the mouth of the Yakima River (two sites, rkm2 3). This array segregated the study area into four discrete reaches in which survival of tagged fish was estimated. Aerial and underwater antennas were also used to monitor tagged fish at Roza Dam. Aerial antennas were located in the forebay, on the East gate, on the West gate, and in the tailrace of Roza Dam. Underwater antennas were located in the fish bypass, upstream of the East gate, and upstream of the West gate to collect route-specific passage data for tagged fish. Additional years of data collection and analysis could alter or improve our understanding of the influence of flow and other environmental factors on smolt survival in the Yakima River. Nevertheless, during 2012-2014, yearling hatchery Chinook salmon smolt emigration survival was significantly associated with stream flow in the