Timing and magnitude of climatic extremes differentially elevate mortality but enhance recovery in a fish population.

The countervailing effects of disturbances (e.g., high mortality and enhanced recovery) on population dynamics can occur through demographic processes under rapidly increasing climatic extremes. Across an extreme-event gradient, we mechanistically demonstrated how dramatic changes in streamflow have affected the population persistence of endangered salmon in monsoonal Taiwan over a three-decade period. Our modeling indicated that the dynamics of the age-structured population were attributed to demographic processes, in which extensive mortality was characterized as a function of climatic extremes and vulnerability in the young stage of fish. In the stochastic simulations, we found that the extensive mortality and high proportion of large fish resulted from extreme flooding, which caused high values of postimpact population recovery. Our empirical evidence suggests that the magnitudes and timing of disturbance can explain the population persistence when facing climatic extremes and thereby challenges the understanding of the mechanistic drivers of these countervailing phenomena under changing environmental conditions.