Bayesian hierarchical spatio-temporal models to identify the European hake (Merluccius merluccius) recruits dynamic in the northern Iberian Peninsula

Abstract

Spatial management of commercial resources is becoming an effective measure to be broadly implemented in the European Seas. However, it is currently unconnected from the population dynamics and the temporal assessment. Indeed, it is known that species abundance can be influenced by the environmental features of its own habitat and/or by biotic process that are spatially structured (e.g. reproduction, predation, among others). Usually, this variability is assumed to be implicitly in the abundance trends used as inputs of the stock assessment models and it is not explicitly taken into account. Within this context, in this study we propose a novel methodological approach for an effective implementation of spatial and ecological knowledge that could help to embrace species spatial management providing a more holistic and ecosystem-based approach. As case study we used the European hake recruits (Merluccius merluccius) in the northern continental shelf of the Iberian Peninsula. Data collected during the scientific survey series “DEMERSALES” by the “Instituto Español de Oceanografía” (IEO) from 2005 to 2016 were analyzed using Bayesian hierarchical hurdle spatial-temporal models, considering as environmental variables Sea Surface Temperature, Sea Surface Salinity, bathymetry and rugosity of the seabed. The presented models allowed to solve data issues such as spatial dependence, temporal autocorrelation, high proportions of zeros and big size of samples. In addition, using the integrated nested Laplace approximation (INLA) method different types of spatiotemporal structures (persistent, opportunistic, and progressive) were compared in order to achieve the most robust fit for our data type. We argue that the analytical framework proposed in this study allowed (1) to assess which environmental factors influence hake recruits abundance in the northern continental shelf of the Iberian Peninsula, (2) to identify areas in which individuals of this functional group are more aggregated as well as their space-time fluctuations, and (3) to provide improved habitat-based standarization abundance indexes for stock assessment models