Juvenile fish can strand in pools or in interstitial spaces when the water elevation drops in regulated rivers due to flow reductions. Three years of summer and winter experiments on the Columbia and Kootenay Rivers (Canada) assessed the effect of the rate of change in water level (ramping rate) on the probability of pool and interstitial stranding for juvenile (<100 mm) fish. The factors of wetted history of the site, time of day, natural fish density and the occurrence of a conditioning reduction prior to the operational reduction were also examined for their effect on stranding. Experimental net pens were constructed to test these factors in situ in the varial zones of the two rivers. Linear models with plausible additive combinations of the potential explanatory factors and a null model were fitted to the logistically transformed data and ranked using the second-order Akaike Information Criterion (AIC c ). The null model was the top ranked model for the interstitial stranding analyses, highlighting that none of the factors tested were significant variables in predicting the probability of stranding. Natural fish density, wetted history of the site, ramping rate and the presence of a conditioning reduction were variables included in the top three ranking models for the pool stranding analyses. Probability of pool stranding in summer was reduced by the occurrence of a conditioning reduction prior to the operational reduction. Higher natural fish density, longer periods of wetted history and higher ramping rates all led to higher probabilities of pool stranding.