Complexity of addressing cumulative effects that vary in space and time, especially for species occupying large ranges, makes conservation and recovery of populations difficult. In Alberta, declines of all three native stream trout species led to them being listed as species at risk. We developed a novel, semi-quantitative cumulative effects modelling process to quantify threats using stressor-response curves with a single common response scale, wherein inputs were determined for each population, and outputs were used to create population-specific recovery action hypotheses to inform management. Using a case study of bull trout recovery in Rocky Creek, Alberta, we tested these hypotheses using a before–after control-impacted design. Recovery actions positively affected bull trout, and the modelling approach provided insight into threats (sedimentation and angling effort) that most likely limited the population.