Inferring Marine Distribution of Adult Atlantic Salmon Using a Hidden Markov Model
Atlantic salmon populations are currently experiencing substantial declines throughout the species' distribution range and for many stocks the reduction is linked to decreased marine survival. Despite this, little is known about the migratory behavior and spatial distribution of individuals at sea. In the presented study, migration of post-spawned adult salmon from the Miramichi River, Canada, was described using Pop-up satellite tags (PSATs) in combination with acoustic telemetry. Individual migrations were modeled using a Hidden Markov Model (HMM), specifically designed to geolocate salmon. Reconstruction of individual tracks revealed utilization of various parts of the Gulf of St. Lawrence initially after ocean entry. All fish with track lengths exceeding 50 days migrated through the Strait of Belle Isle and in to the Labrador Sea. Here, fish exhibited little overlap in spatial distributions and the reconstructed tracks revealed that fish utilized distinctly different oceanographic environments. This high individual variability in habitat use indicates that the growth and survival of the Miramichi salmon population depends on multiple ecological habitats while at sea.