Marine ecosystems are being altered increasingly by climate change and fisheries. To manage such systems, we often need to estimate the amount of prey that must be left to sustain higher predators such as mammals and birds. Critical to such estimates is knowing the fraction of prey stocks that are actually available to predators. For example, small seabirds can only dive so deep, so that prey such as krill are unavailable to them unless the krill come near the surface. As a result, sustaining the birds requires a much larger population of krill than the birds actually eat. On a large scale, movements of krill to the surface depend on currents such as those caused by upwelling; but at smaller scales at which seabirds search for food, spatial patterns of krill are poorly understood. In this research, I am developing computer models that indicate the ability of a small seabird (Cassin's Auklet) to feed profitably based on the dispersion of krill prey as determined by at-sea surveys. These birds have experienced widespread breeding failures in recent years all along the Pacific Coast, apparently because of food shortages. My results will help to quantify prey stocks needed to support top predators, and to determine the location and extent of key foraging areas that need protection.