As marine food webs are increasingly altered by climate change and human exploitation, it is often desirable to estimate the amount of prey at one trophic level that is needed to sustain higher trophic levels. Critical to such estimates is the tightness of trophic coupling, or the fraction of prey that are effectively available and eaten by predators. For example, small diving seabirds are limited in dive depth, so that even abundant prey like krill are unavailable unless they come near the surface. Thus, physiological constraints on these predators can make the realized patch structure of available prey quite different from that of the entire prey population. Availability of pelagic prey is also determined by prey behavior, with diel vertical movements that can vary temporally and spatially with bathymetry, light levels, and tides. Overall prey availability may vary predictably with these factors at large scales (fronts) -- however, smaller-scale features (eddies) that concentrate prey into profitable densities are far less predictable, and may not be detected without costly exploratory dives. As a result, much of the prey is never exploited. In this research, I will link models of the foraging energetics of auklets to variations in 3-dimensional dispersion of krill prey over a time series of months and years. In particular, I will explore mechanisms and develop predictive models of how physiological limits on prey availability can control and weaken trophic coupling. Resulting concepts and models will have broad utility in foodweb approaches to marine ecosystem management, for both quantifying prey stocks needed to support top predators, and determining the location and extent of viable foraging habitat. collapse