Multi-point source identification of sudden water pollution accidents in surface waters based on differential evolution and Metropolis–Hastings–Markov Chain Monte Carlo

Yang, Haidong; Shao, Dongguo; Liu, Biyu; Jian-hua, Huang; Ye, Xianbao

doi:10.1007/s00477-015-1191-5

Cited by 47 publications

(20 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Essentially, the MCMC algorithm can be used to explore the state space of a random parameter using Markov chain mechanism and generate samples from the posterior distribution while spending most of the sampling steps in the highest density regions of the posterior state space [10,14]. In other words, it generates a point series as a chain, and the distribution of these points follows the posterior probability density function P(m|C,I).…”

Section: Markov Chain Monte Carlomentioning

confidence: 99%

“…Metropolis algorithm, a special case of Metropolis-Hastings [10,16,17], was adopted in this study. Its proposal distribution (p) satisfies symmetrical random sampling, q(m (*) |m (i) )=q(m (i) |m (*) ), and the acceptance probability (A) of Markov chain ranges from m (i) to m (*) is as follows.…”

Section: Metropolis Algorithmmentioning

confidence: 99%

“…In the past two decades, several deterministic and stochastic methods have been developed successively to facilitate the solution to the source identification problem for identifying the release locations of unknown sources, and estimate the release time and emission loads [2][3][4][5][6][7][8][9][10][11]. Among them, Bayesian approach has been applied widely for source identification in recent years as it has a number of distinctive attributes [10,11]. The method converts contamination source identification into the reiterative computation of posterior probability distribution of the source parameters and can quantify random errors in the concentration data and their uncertainties [5].…”

Section: Introductionmentioning

confidence: 99%

“…The method converts contamination source identification into the reiterative computation of posterior probability distribution of the source parameters and can quantify random errors in the concentration data and their uncertainties [5]. For example, Yang et al [10] constructed differential evolution algorithm based on Bayesian inference to identify multi-point sudden water pollution sources. Compared with the deterministic calculations, the Bayesian framework estimates the posterior distribution of historical contaminants in an expanded stochastic space with current measurement concentrations and thus restates the ill-posed inverse problem as a well-posed problem [8,12].…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Utilization of a Bayesian probabilistic inferential framework for contamination source identification in river environment

et al. 2018

View full text Add to dashboard Cite

In the environmental event of hazardous release into river, quick and accurate identification of the contamination source is important for emergence response. Generally, given a noisy and finite set of monitoring information, determining the source items (i.e. location, strength and release time) is an ill-posed inverse problem. In this study, a Markov chain Monte Carlo method combined with advection-dispersion equation (ADE) was proposed for the source identification of contamination event in river system based on a Bayesian probabilistic inferential framework. Case study with analytical solution for one-dimensional ADE showed that the proposed methodology was effective and the mean posterior errors for all source parameters were lower than 3%. Case simulation based on two-dimensional ADE with numerical solution obtained similar results and further demonstrated the utility of the proposed approach for source identification. We hope the study will provide a helpful guidance to develop approach for contamination event source identification to support environmental risk management of river system.

show abstract

Section: Markov Chain Monte Carlomentioning

confidence: 99%

Section: Metropolis Algorithmmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Utilization of a Bayesian probabilistic inferential framework for contamination source identification in river environment

et al. 2018

View full text Add to dashboard Cite

show abstract

“…In Sweden, several thousand accidents are reported annually to the Swedish Rescue Services (Ohlén and Larsson 2000 ). The probability of traffic accidents causing water pollution can be estimated by Bayesian Network analysis (Yang et al 2015 ; Tang et al 2016 ). Liquid spills sometimes leave the paved road and infiltrate into the surroundings.…”

Section: Introductionmentioning

confidence: 99%

Spread of Water-Borne Pollutants at Traffic Accidents on Roads

Olofsson

Rasul

Lundmark

2017

Water Air Soil Pollut

View full text Add to dashboard Cite

Traffic accidents sometimes lead to the spread of hazardous compounds to the environment. Accidental spills of hazardous compounds on roads in the vicinity of vulnerable objects such as water supplies pose a serious threat to water quality and have to be assessed. This study compared three different assessment methods, electrical resistivity measurements, analytical flow calculations, and 1D and 2D dynamic flow modeling, to describe rapid transport processes in the road shoulder and roadside verge after a major spill. The infiltration and flow paths of water-borne substances were described during simulated discharge of pollutants on different road types. Full-scale tracer tests using sodium chloride were carried out at nine different road locations in Sweden. Analysis of grain size distribution and infiltrometer tests were carried out at the road shoulder and verges. The pathways and travel times were traced using resistivity measurements and 3D inverse modeling. The resistivity measurements were compared to analytical flow calculations and 1D and 2D dynamic modeling. All measurement sites were highly heterogeneous, which caused preferential flow. Vertical flow velocities of 1.4–8.6 × 10−4 m/s were measured. The results of the analytical calculations and flow modeling were of the same order of magnitude. The measurements showed that almost all infiltration goes directly into the road embankment, hence the composition and structure of the built-up road must be considered. The non-destructive resistivity measurements and 3D modeling provided useful information for clarifying the infiltration and flow pattern of water-borne compounds from road runoff.

show abstract

An innovative framework for real-time monitoring of pollutant point sources in river networks

Moghaddam

Mazaheri

Samani

et al. 2022

Stoch Environ Res Risk Assess

View full text Add to dashboard Cite

Multi-point source identification of sudden water pollution accidents in surface waters based on differential evolution and Metropolis–Hastings–Markov Chain Monte Carlo

Cited by 47 publications

References 44 publications

Utilization of a Bayesian probabilistic inferential framework for contamination source identification in river environment

Utilization of a Bayesian probabilistic inferential framework for contamination source identification in river environment

Spread of Water-Borne Pollutants at Traffic Accidents on Roads

An innovative framework for real-time monitoring of pollutant point sources in river networks

Contact Info

Product

Resources

About