Patterns of recent brook trout invasion in bull trout streams in relation to habitat, source connectivity, biotic resistance, and disturbance

Knowledge of which habitats are at risk of invasion by nonnative species is essential to conservation, but a changing climate often complicates assessments of where future invasions may occur. Nonnative brook trout Salvelinus fontinalis threaten cold-adapted native salmonids throughout western North America, and the extent of future invasions in coldwater streams is recognized as an area of uncertainty with important consequences. We addressed this uncertainty using spatial-stream-network (SSN) modeling to develop a species distribution model (SDM) for brook trout in Idaho, creating high-resolution prediction maps of suitable (i.e., invadable) habitat under baseline (1990–2015) and future (2080s) climate scenarios, and querying predictions within natal bull trout S. confluentus habitat patches (stream–subwatershed scale) for each scenario. Results indicated that brook trout will experience a net loss of suitable habitat in Idaho, but a large increase of suitable habitat in cold streams. The total proportion of bull trout habitat invadable by brook trout increased from 0.33 during the baseline scenario to 0.44 during the future scenario—a result of upstream expansions of brook trout habitat that were partially offset by contractions in bull trout habitat. Changes in the proportion of invadable habitat varied widely among bull trout patches. Between scenarios, the largest increases in suitable habitat occurred in small, low-gradient, unconfined streams with mean August water temperatures of 7–9 °C during the baseline period. A relatively small proportion of patches contained the majority of invadable habitat in both scenarios, particularly patches in “U-shaped” glacial valleys. A changing climate will broadly increase the threat that brook trout pose to bull trout, but the degree of change experienced by individual populations will vary with local context. The high-resolution prediction maps created by our SSN model will help conservation planners identify and prioritize control efforts on the small proportion of brook trout populations that most threaten bull trout.