Stochastic model of the Isle Royale biome

A stochastic model employing and event-consequence technique is developed for a three-species interaction. If the system is in some state x1 at time t, the elapsed time to the next event is chosen by means of a Poisson distribution exp [-tauΣ/sub x/2g (x1,x2,t)]. The type of event (birth, death, predation, etc.) is then chosen from the distribution: g (x1,x2,t)/Σ/sub x/2g (x1,x2,t) where g is some non-negative function. Thus the classical system of differential equations is replaced by a set of transition probabilities P/sub tt+delta/,(x1→x2) <g (x1,x2,t) deltat+0(deltat). The set of states and transition probabilities incorporates the age structure of the three species (17 for the moose, 9 for the wolf, 1 for the plant). This discrete space-continuous time model is applied to the data from the Isle Royale National Park studies of Allen, Mech, etc. An estimate of the predictive ability of this simulation is made after considering the statistical uncertainty of the model and the model's goodness of fit.A code, called BIOTA1, exists in Fortran, which will process a general n-species stochastic model