Testing the shallow water refuge hypothesis in flatfish nurseries

The 'shallow water refuge hypothesis' (SWRH) holds that predation upon juvenile fish and crustaceans is reduced in shallow water because larger predatory fish increase in abundance with depth. We tested predictions of this hypothesis in 2 Kodiak Island flatfish nurseries, Pillar Creek Cove and Holiday Beach, using baited camera, tethering and video sled techniques. Baited camera deployments during 2007 indicated that predators increased with depth in both nurseries. In the same year, tethering indicated increased mortality of juvenile flatfish with increased depth. In contrast, video sled data from 2003 and 2004 indicated that predator abundance increased with depth at Holiday Beach, but not at Pillar Creek Cove. At Holiday Beach, juvenile abundance decreased rapidly with depth, while at Pillar Creek Cove abundance increased slightly with depth. Thus, predator-predation distribution over 3 yr at Holiday Beach was consistent with the SWRH, while at Pillar Creek Cove it was more variable. Although the SWRH is normally considered in the context of waters <2 m in depth, our results suggest that it may have efficacy for deeper water, with the relative depth distribution of predators and suitable juvenile habitat being more relevant than absolute depth.