River water quality modelling and simulation based on Markov Chain Monte Carlo computation and Bayesian inference model

Abstract

Hierarchical Bayesian methods are experiencing increased use for probabilistic ecological modelling. Influence of water quality indicators in the river water are studied. Bayesian inference through Markov Chain Monte Carlo (MCMC) algorithm is used as the basic model to assess the rate of water pollution using conjugate and non-informative priors. The algorithm used flow velocity, physico-chemical and biological parameters as the three model parameters. MCMC simulates a chain that converges on posterior parameter distributions, which can be regarded as a sample for posterior estimations. The results show the biological parameters have a negative impact on quality of water, whereas the quality is improved while considering the physico-chemical parameters and flow velocity. The Bayesian MCMC produces the posterior distributions which are heavily influenced by the priors along with given likelihood function. However, the simulation (MCMC) based estimates of posterior distributions may vary due to the use of a random number of generators in procedures.

Keywords:

• Bayesian inference
• likelihood
• MCMC
• probability distributions
• Water Quality Index (WQI)

Disclosure statement

No potential conflict of interest was reported by the authors.