Predictive models of the preferential distribution of demersal fish larvae in the southern part of the California Current

Abstract

Habitat characterization provides predictive information about the distribution of species and is useful for as­sessing habitat quality and population stability. Larval abundance of six frequent and abundant demersal species and the relationship of each with the environment were analysed through generalized additive models to determine their preferential distribution and predictive response to the environmental variables in the southern part of the California Current (25-31°N) between two periods of data collection: 1997-2000 and 2006-2010. Essentially, the main associated variables governing the distribution patterns were related to common and oceanographic characteristics of the water column (temperature and salinity at 50 m depth, dynamic height and degree of water column stratification); however, the set of variables and their ranges are usually species-specific. Species of northern distribution, Sebastes sp. and Citharichtys stigmaeus, were recorded mainly in newly emerged, relatively unstratified waters characterized by a shallow mixed layer and low temperatures. Low dynamic height values were the most significant predictor of larval distribution for Merluccius productus. Citharichthys xanthostigma and Symphurus atricaudus were widespread, distributed across the study area mainly in autumn in unstratified or stratified waters and at a shallow mixed layer. Particularly C. Xanthostigma and S. Lucioceps were related to high dynamic height val­ues, likely influenced by a coastal flow towards the pole, as evidenced by counter-currents.