Development of a spatially explicit, individual-based model of marine fish early life history

Abstract

A growing body of evidence suggests that survival of the planktonic early life stages of marine fish varies spatially, often in concordance with mesoscale and larger circulation patterns and with spatial variation in biotic and abiotic factors. Environmental variation experienced by individuals may contribute to variance within the population. Detailed spatial information about life histories cannot be obtained however, by simply subjecting individuals to identical time series of environmental variates, or adding random noise to the mean values of those variates. In this paper, we present a numerical biophysical model to address this problem. The model combines a 3-dimensional hydrodynamic model with a probabilistic, life-stage model of young fish. The oceanographic model is an eddy-resolving, semi-spectral primitive equation model (SPEM) adapted for Shelikof Strait and the western Gulf of Alaska, USA. The biological component is an individual- based model (IBM) which follows each fish through its first 3 mo of life. The 2 models are coupled via a float-tracking algorithm, which adverts each fish through the model grid based on the time-varying velocity field derived from the SPEM model, and returns information to the IBM on any physical variables of interest (e.g. temperature and salinity) at the individual's new location. These variables are used to drive biological processes in the IBM. Simulation 1 demonstrates how a probabilistic IBM using Lagrangian temperature information derived from float tracking can yield significantly different biological results than the same model with constant temperature, or a model where all individuals are subjected to identically increasing time series of temperature. In Simulation 2 we compare the float tracks and growth of individuals using a scenario of constant depth for each life stage versus algorithms that include mechanisms producing variable depth. These mechanisms differ for each life stage (e.g. changes in depth due to development of eggs, or diel migration of feeding larvae). This simulation shows that inclusion of specific mechanisms which determine depth of individual eggs or larvae at particular periods of their development are important in determining the direction of horizontal advection, and the history of exposure to environmental variables for each individual. Simulation 3 compares modeled distributions of the early life stages with distributions derived from field surveys.