No description available.

I studied the onshore migrations of Atlantic cod (Gadus morhua) in the northern Gulf of St. Lawrence to test the predictability of the distributions and catches of a migratory predatory fish relative to physical oceanographic processes and the distributions and abundances of its prey. To survey daily fish distributions and abundances by species, hydroacoustic signal discriminant methods were developed which classified schools of cod, capelin (Mallotus villosus), and mackerel (Scomber scombrus) with 93% success. Physical oceanographic processes along the 150 km of coastline studied were characterized by Ekman-type upwellings and downwellings forced by alongshore winds at periods $>$3 d. Physical elements of this system (principally currents and temperature) were associated with cross-shore movements of cod and accounted for as much as 83% of the variance in mean trap catch. Cod and capelin formed coherent spatial distributions when capelin were congregated where sea temperatures were most favorable to cod (1 to 5$ sp circ$C) and mean capelin densities were above 100/10$ sp5$m$ sp3$. Coherences were in phase at larger scales ($>$10 km) but became increasingly out of phase at smaller scales. Cod distributions fit the predictions of an "ideal" distribution (allometric constant = 0.67) when two constraints were put on the relationship: (1) capelin densities above 100/10$ sp5$m$ sp3$, and (2) temperatures between 1 and 9$ sp circ$C.