Stochastic modeling of rainfall processes: A Markov chain-mixed exponential model for rainfalls in different climatic conditions

Watershed models simulating the physical process of runoff usually require daily or sub-daily rainfall time series data as input. However, even when rainfall records are available, they contain only limited and finite information regarding the historical rainfall pattern to adequately assess the response and reliability of a water resource system. This study is therefore concerned with the development of a stochastic rainfall model that can reliably generate many sequences of synthetic rainfall time series' that have similar properties to those of the observed data.;The 'MCME' model developed is based on a combination of the rainfall occurrence (described using a Markov Chain process) and the distribution of rainfall amounts on wet days (represented by the Mixed-Exponential probability function). Various optimization methods were tested to best calibrate the model's parameters and the model was then applied to daily rainfall data from 3 different regions across the globe (Dorval, Quebec, Sooke Reservoir in British Columbia and Roxas City in the Philippines) to assess the accuracy and suitability of the model for daily rainfall simulation. The feasibility of the MCME model was also assessed using hourly rainfall data available at Dorval Airport in Quebec (Canada).;In general, it was found that the proposed MCME model was able to adequately describe various statistical and physical properties of the daily and hourly rainfall processes considered. In addition, an innovative approach was proposed to combine the estimation of daily annual maximum precipitations (AMPS) by the MCME with those by the downscaled Global Circulation Models (GCMs). The combined model was found to able to provide AMP estimates that were comparable to the observed values at a local site. In particular, the suggested linkage between the MCME and downscaled-GCM outputs would be useful for various climate change impact studies involving rainfall extremes.