Search

Markov mixture models combine a Markov model for transitions between low and normal streamflow states with a mixture model blending two normal subpopulations. The models are particularly effective for generating synthetic streamflow records with long and severe droughts. Their use in a hypothetical planning problem illustrates the application of a set of modeling precepts.

Birth-death models can generate synthetic flow sequences that demonstrate differential persistence, i.e., sequences in which low flows show more persistence than high flows do. A simple phenomenological model helps justify the assumption of differential persistence. The use of a birth-death model, together with a set of modeling precepts, shows that differential persistence is a phenomenon of considerable descriptive importance but at least in some planning situations does not carry corresponding prescriptive importance.