Rainfall Generation Revisited: Introducing CoSMoS‐2s and Advancing Copula‐Based Intermittent Time Series Modeling

Abstract What elements should a parsimonious model reproduce at single scale to precisely simulate rainfall many scales? We posit these are: (a) the probability of dry and linear correlation structure wet/dry sequence as proxy reproducing distribution spells, (b) marginal nonzero its structure. build two‐state model, CoSMoS‐2s, that explicitly reproduces is easily applicable any timescale. Additionally, paper: introduces Generalized Exponential ( ) system comprising six flexible distributions with desired properties describe facilitate time series generation; extends CoSMoS framework allow simulations negative correlations; (c) simplifies generation binary sequences by analytical approximations; (d) rank‐based CoSMoS‐2s preserves Spearman's correlations, has an formulation, also for infinite variance series, (e) copula‐based enabling intermittent times values having dependence copula, (f) offers conceptual generalizations modeling beyond, specific ideas future improvements extensions. The tested using four long hourly records; multiple scales including dry, characteristics rainfall, behavior extremes.