A new framework for a multi-site stochastic daily rainfall model: Coupling a univariate Markov chain model with a multi-site rainfall event model

Multi-site rainfall models are useful tools to provide synthetic realizations of spatially-correlated at multiple stations, which great importance for flood and drought risk assessment climate change impact analysis. Therefore, a good preservation various observed characteristics including time-series statistics event individual stations the inter-site correlations these is very crucial. To achieve this purpose, study aims develop multi-site stochastic daily model by coupling univariate Markov chain with (MSDRM-MCREM), based on our previously-developed single-site SDRM-MCREM. The in MSDRM-MCREM used generate occurrence time series extract simulated events continuous wet days. then constructed using Vine copulas simulate those that occur simultaneously durations, depths temporal patterns. Subsequently, was applied Changshangang River basin Zhejiang Province, East China its performance reproducing spatial evaluated three cases, i.e. simulations two, four stations. Results show except overestimation light rainfall, can well preserve (i.e. different percentiles, mean monthly standard deviations probabilities values days), extreme exceedance annual maximum 1-day, 3-day 5-day rainfall) cumulative spell, dry spell depth, patterns types depth-based classes) In addition, have also been maintained, groups, being slightly underestimated.