Song Songbai

A measure of hydrological system complexity based on sample entropy

The principles of sample entropy algorithm and Lopes-Ruiz complexity are presented in this paper. Integrating sample entropy algorithm into Lopez-Ruiz complexity theory, the authors derive a novel method for studying the hydrological system complexity. The derived method is applied to the monthly mean runoff observation data for Linjiacun, Zhangjiashan and Zhangtou sites in the Weihe watershed area. The obtained results for these sites detect an increase in the complexity of runoff series in early 1970s and 1990s due to the human activities. The appropriate results obtained for the study area lead the authors to conclude that the proposed model provides a new method for analyzing the complexity in hydrological systems and identifying the influence of environmental changes on the runoff series.

Bayesian estimation of parameters for pearson III distribution

In this paper Bayesian method was used to estimate the parameters of Pearson III distribution, the posterior distribution was calculated by the Markov Chain Monte Carlo (MCMC) approach which avoiding the difficulty of boresome integration. Metropolis-Hasting (MH) algorithm, Delayed Rejection (DR) algorithm and Adaptive Metropolis (AM) algorithm are employed for parameter sampling at the site of interest, respectively, then EM (Ensemble Mean) and ES (Ensemble Spread) were applied to evaluate convergence of the parameter sampling. The results indicate that the AM sampling is much better than the other two algorithms in convergence rate.