AbstractPost‐spawning migrations of sea trout (Salmo truttaL.) in two northern Swedish river systems, Vindelälven and Piteälven, were studied. Telemetry was used to evaluate spawning locations, post‐spawning mortality, overwintering survival/habitat selection/location, partitioned natural mortality from dam passage mortality and overall survival from post‐spawning to the sea entry. Fifty‐eight sea trout (LT = 50–86 cm) were radio‐tagged at the beginning of their spawning migration from June to September in 2003 and 2004, and their spawning in September and October each year was monitored. In total, 91% (n = 53) survived spawning and were thus defined as kelts. Of these, 92% (n = 49) overwintered under ice‐cover in deep, slow flowing sections of the rivers. Overwintering mortality was low, on average 8%. The main seaward migration occurred during May and June and was initiated when ambient water temperatures exceeded 4–6°C, independent of the start of the spring flood when flows rose to 1000 m3 s−1. The highest downstream migration speed was approximately 25 km 24 h−1and migrations mainly took place at daytime. During seaward migration, the kelts encountered hydropower stations that lacked bypass systems for fish passage, which delayed their downstream migrations. Passage mortality at the power stations was 69 and 25% in the Vindelälven and Piteälven, respectively. These losses, combined with the naturally high winter survivals of kelts at the unregulated river parts, underline the importance of developing safe passage routes for kelts at hydropower facilities for conserving sea trout populations in regulated rivers. Copyright © 2008 John Wiley & Sons, Ltd.