PATCH OCCUPANCY MODELS OF METAPOPULATION DYNAMICS: EFFICIENT PARAMETER ESTIMATION USING IMPLICIT STATISTICAL INFERENCE

摘要: The practical value of a predictive metapopulation model is much affected by the amount data required for parameter estimation. Some models require information on population turnover events parameterization, whereas other models, such as incidence function that used in this study, can be parameterized with spatial patch occupancy. latter are more readily available. original method using pattern to parameterize and has been criticized involving potentially troublesome assumptions, independence habitat patches constant colonization probabilities. This study describes an improved estimation not these problems. proposed based Monte Carlo inference implicit statistical it adapted any stochastic occupancy dynamics. As additional advantage, new allows amplitude regional stochasticity. Tested simulated data, was found produce substantially accurate estimates than method. approach applied two empirical metapopulations, false heath fritillary butterfly Finland American pika at Bodie, California.