Stochastic modeling of hydrometeorological data is essential to predict the probability of occurrence of extremes such as droughts and floods that are likely to occur in future, especially because the events that have occurred in the historical past, may not repeat in terms of the magnitude or intensity or time sequence of occurrence. With the measured hydrometeorological data available being limited, it becomes challenging to model these natural processes that are stochastic in nature. At a fine time scale, such as daily or hourly, the evolving complex dynamics involved in these processes are to be captured by the stochastic model being fitted using the measured (empirical) data. An attempt is made in this research study to propose novel hybrid stochastic models for characterising the daily precipitation and the daily streamflows (river flows) at a single as well as multiple stations (gauging sites). The hybrid model proposed for daily precipitation involves three components: i) principal component streaming, a data streaming method to capture the dynamics involved in the process through an effective evolving clustering mechanism; ii) Markov chain based Monte-Carlo simulation to generate multiple realisations of the cluster sequences obtained from data streaming; and iii) non-parametric multi-site precipitation simulator that effectively resamples the spatial and temporal precipitation data corresponding to the cluster sequences generated by the Markov chain using a varying block resampling method that effectively reproduces both cross-station correlations and higher order moments. The hybrid model proposed for streamflow simulation involves: i) extraction of the daily streamflow dynamics through PC-streaming of the observed streamflow data that results in effective clustering of days across years that are similar; ii) Markov chain based Monte-Carlo simulation to generate multiple realisations of the cluster sequences obtained from data streaming that simulate the occurrences including the wet and dry state transitions; iii) a parametric streamflow simulator that generates the flow magnitudes corresponding to the length of alternate cycles of ascension and recession that indicate the wet and the dry sequences respectively, in which the flows corresponding to wet state are generated from a monthly varying four-parameter kappa distribution fitted for the ascension segment followed by a shuffling to maintain the rank order and the dry state flows are generated from a monthly two-parameter exponential decay model fitted for the recession segment. The daily precipitation and the daily streamflow data generated from the respective hybrid stochastic models proposed, will be subjected to validation tests for reproduction of the respective hydrological characteristics. Furthermore, the synthetic sequences generated from the proposed models will be used for the prediction of drought and flood characteristics for hydrologic design purposes.