Identification of Dominant Hydrological Mechanisms Using Bayesian Inference, Multiple Statistical Hypothesis Testing, and Flexible Models

In hydrological modeling, the identification of model mechanisms best suited for representing individual (physical) processes is major scientific and operational interest. We present a statistical hypothesis-testing perspective on this challenge contribute mechanism framework that combines: (i) Bayesian estimation posterior probabilities from given ensemble structures; (ii) test statistic defines “dominant” as more probable than all its alternatives observed data; (iii) flexible modeling to generate structures using combinations available mechanisms. The uncertainty in approximated bootstrap sampling ensemble. Synthetic experiments (with varying error magnitude multiple replicates) real data are conducted system FUSE (7 2–4 per process yielding 624 feasible structures) Leizarán catchment northern Spain. method reliable: it identifies correct dominant synthetic trials where an made. As data/model errors increase, power (identifiability) decreases, manifesting no identified dominant. case study results broadly consistent with analysis, 4 7 processes. Insights which most/least identifiable also reported. expected broader community efforts improving representation hydrology.