Risk analysis in fisheries and natural resource management

The Monte Carlo approach now pervades risk analysis; it is essential for calculating consequences of alternative actions, and even one of the more common methods of Bayesian integration is the Markov chain simulations, which are Monte Carlo methods (Gelman et al., 1995). The major methodological debate concerns the use of resampling or Bayesian methods to assign probabilities to alternative hypotheses. While resampling methods were very popular when formal risk analysis began, I believe that Bayesian methods are to be preferred, and will soon dominate the actual applications.