The perspective of quantitative science in the debate about environmental degradation
The hierarchical statistical model is an enormously powerful way to describe complex environmental phenomena. Its three generic components are: the data model, where a probability distribution is given for data given process and parameters; the process model, where a probability distribution is given for process given parameters; and a prior model, where a probability distribution is given for all parameters. The probability modeling of the parameters is optional, but regardless of this, the hierarchical approach allows one to model locally and infer globally. In practice, the inference is carried out via Markov chain Monte Carlo (or some other computational scheme), which allows relatively straightforward sampling from the posterior distribution of all unknowns given the data.