A COMPUTER-INTENSIVE ALGORITHM TO CREATE GROWTH TRANSFER MATRICES FOR USE IN AGE AND SIZE BASED SIMULATION MODELS

摘要

Commonly, age and size based simulation models introduce fish at the youngest age into bins (recruitment), then move (transfer) the fish through age and size bins as time passes and growth occurs in the simulation. The numbers of fish in each cohort decrease over time due to size specific natural and fishing mortality. This basic procedure quickly leads to a complex simulation requiring significant computer resources. Growth variability can be incorporated in this type of simulation model in two ways. The growth transfer rates can be computed every time step within the simulation or can be created once outside the simulation, assuming growth is independent of the number of animals at age or size. Computation of the growth transfer rates within the simulation assume a mean (expected) population growth curve and introduce variability to produce a range of fish sizes for any given age (Gutreuter and Anderson, 1985; Buijse et al., 1992; Shepherd and Idoine, 1993). This approach requires many computations per time step with a new transfer matrix value generated for each age-size combination. This can become a computational bottleneck within the simulation.