We used a model of the pelagic ecosystem in the eastern tropical Pacific Ocean to explore how climate variation at El Nino - Southern Oscillation (ENSO) scales might affect animals at middle and upper trophic levels. We developed two physical-forcing scenarios: (1) physical effects on phytoplankton biomass and (2) simultaneous physical effects on phytoplankton biomass and predator recruitment. We simulated the effects of climate-anomaly pulses, climate cycles, and global warming. Pulses caused oscillations to propagate through the ecosystem; cycles affected the shapes of these oscillations; and warming caused trends. We concluded that biomass trajectories of single populations at middle and upper trophic levels cannot be used to detect bottom-up effects, that direct physical effects on predator recruitment can be the dominant source of interannual variability in pelagic ecosystems, that such direct effects may dampen top-down control by fisheries, and that predictions about the effects of climate change may be misleading if fishing mortality is not considered. Predictions from ecosystem models are sensitive to the relative strengths of indirect and direct physical effects on middle and upper trophic levels.