The relation between age-0 rainbow trout (Oncorhynchus mykiss) abundance and winter discharge in a regulated river

We identified and experimentally tested a discharge–abundance relation that predicted, based on the mean river discharge in the second half of winter (15 January – 31 March), the spring abundance of age-0 rainbow trout (Oncorhynchus mykiss) in a section of the Henrys Fork of the Snake River, Idaho, with complex bank habitat. We also considered a competing hypothesis in which autumn abundance determined spring abundance. We established that large abundances of age-0 trout were present in autumn (34 000 – 81 000) and lower abundances remained in spring (8000 – 15 000). Winter loss of age-0 trout was initiated in January. Spring abundance in 1996–1998 was related to autumn abundance (r2 > 0.99) and mean discharge in the second half of winter (17.1–22.8 m3·s–1; r2 > 0.99) but not mean discharge in the first half of winter (15.1–21.1 m3·s–1; r2 = 0.11). We experimentally maintained a high discharge (20–21 m3·s–1) in the second half of winter in 1999 to test model predictions. Autumn abundance failed to predict spring abundance (observed = 11 109; predicted = 6822; 95% prediction interval = 4669–8975). However, the discharge–abundance model accurately predicted spring abundance (predicted = 11 980; 95% prediction interval = 10 728 – 13 231). Higher discharge in the second half of winter may have provided more bank habitat at a critical time for survival.