1. Smallmouth bass (Micropterus dolomieu) have been widely introduced to fresh waters throughout the world to promote recreational fishing opportunities. In the Pacific Northwest (U.S.A.), upstream range expansions of predatory bass, especially into subyearling salmon-rearing grounds, are of increasing conservation concern, yet have received little scientific inquiry. Understanding the habitat characteristics that influence bass distribution and the timing and extent of bass and salmon overlap will facilitate the development of management strategies that mitigate potential ecological impacts of bass.

2. We employed a spatially continuous sampling design to determine the extent of bass and subyearling Chinook salmon (Oncorhynchus tshawytscha) sympatry in the North Fork John Day River (NFJDR), a free-flowing river system in the Columbia River Basin that contains an upstream expanding population of non-native bass. Extensive (i.e. 53 km) surveys were conducted over 2 years and during an early and late summer period of each year, because these seasons provide a strong contrast in the river’s water temperature and flow condition. Classification and regression trees were applied to determine the primary habitat correlates of bass abundance at reach and channel-unit scales.

3. Our study revealed that bass seasonally occupy up to 22% of the length of the mainstem NFJDR where subyearling Chinook salmon occur, and the primary period of sympatry between these species was in the early summer and not during peak water temperatures in late summer. Where these species co-occurred, bass occupied 60–76% of channel units used by subyearling Chinook salmon in the early summer and 28–46% of the channel units they occupied in the late summer. Because these rearing salmon were well below the gape limitation of bass, this overlap could result in either direct predation or sublethal effects of bass on subyearling Chinook salmon. The upstream extent of bass increased 10–23 km (2009 and 2010, respectively) as stream temperatures seasonally warmed, but subyearling Chinook salmon were also found farther upstream during this time.

4. Our multiscale analysis suggests that bass were selecting habitat based on antecedent thermal history at a broad scale, and if satisfactory temperature conditions were met, mesoscale habitat features (i.e. channel-unit type and depth) played an additional role in determining bass abundance. The upstream extent of bass in the late summer corresponded to a high-gradient geomorphic discontinuity in the NFJDR, which probably hindered further upstream movements of bass. The habitat determinants and upstream extent of bass were largely consistent across years, despite marked differences in the magnitude and timing of spring peak flows prior to bass spawning.

5. The overriding influence of water temperature on smallmouth bass distribution suggests that managers may be able limit future upstream range expansions of bass into salmon-rearing habitat by concentrating on restoration activities that mitigate climate- or land-use-related stream warming. These management activities could be prioritised to capitalise on survival bottlenecks in the life history of bass and spatially focused on landscape knick points such as high-gradient discontinuities to discourage further upstream movements of bass.