Generator of Synthetic Rainfall Time Series through Markov Hidden States

摘要

This paper presents a method for generating synthetic daily rainfall time-series using Markov hidden states that preserve the temporal, spatial and rainfall quantity correlation in a climatologic network station or in a basin divided into catchments. The method does not require that the rainfall time series must be normally distributed and the seasonal rainfall dependence is managed with a non-homogeneous Markov process. Because the computer implementation does not require a large amount of resources, it is fast and accurately produces synthetic time series. The hydrologic model for simulation and planning developed for the Lerma river basin in Mexico was fed with synthetic rainfall time series generated with this method.