Evaluation of Bayesian source estimation methods with Prairie Grass observations and Gaussian plume model: A comparison of likelihood functions and distance measures

Source term estimation for atmospheric dispersion deals with estimation of the emission strength and location of an emitting source using all available information, including site description, meteorological data, concentration observations and prior information. In this paper, Bayesian methods for source term estimation are evaluated using Prairie Grass field observations. The methods include those that require the specification of the likelihood function and those which are likelihood free, also known as approximate Bayesian computation (ABC) methods. The performances of five different likelihood functions in the former and six different distance measures in the latter case are compared for each component of the source parameter vector based on Nemenyi test over all the 68 data sets available in the Prairie Grass field experiment. Several likelihood functions and distance measures are introduced to source term estimation for the first time. Also, ABC method is improved in many aspects. Results show that discrepancy measures which refer to likelihood functions and distance measures collectively have significant influence on source estimation. There is no single winning algorithm, but these methods can be used collectively to provide more robust estimates.

[1]  David B. Dunson,et al.  Bayesian Data Analysis , 2010 .

[2]  G Johannesson,et al.  Dynamic Bayesian Models via Monte Carlo - An Introduction with Examples - , 2004 .

[3]  Xiaoping Zheng,et al.  Inverse calculation approaches for source determination in hazardous chemical releases , 2011 .

[4]  M. Cheng,et al.  Bayesian treatment of a chemical mass balance receptor model with multiplicative error structure , 2009 .

[5]  John Salvatier,et al.  Probabilistic programming in Python using PyMC3 , 2016, PeerJ Comput. Sci..

[6]  Eugene Yee,et al.  Source reconstruction: a statistical mechanics perspective , 2012 .

[7]  John Geweke,et al.  Evaluating the accuracy of sampling-based approaches to the calculation of posterior moments , 1991 .

[8]  Javier Bajo,et al.  Resampling methods for particle filtering: identical distribution, a new method, and comparable study , 2015, Frontiers of Information Technology & Electronic Engineering.

[9]  Guido Cervone,et al.  Non-Darwinian evolution for the source detection of atmospheric releases , 2011 .

[10]  Lowry A. Harper,et al.  Deducing Ground-to-Air Emissions from Observed Trace Gas Concentrations: A Field Trial with Wind Disturbance , 2004 .

[11]  John Skilling,et al.  Data analysis : a Bayesian tutorial , 1996 .

[12]  K. S. Rao.  Uncertainty Analysis in Atmospheric Dispersion Modeling , 2005 .

[13]  Zaoxiao Zhang,et al.  Comparison and improvements of optimization methods for gas emission source identification , 2013 .

[14]  X. L. Zhang,et al.  Modified ensemble Kalman filter for nuclear accident atmospheric dispersion: prediction improved and source estimated. , 2014, Journal of hazardous materials.

[15]  Fue-Sang Lien,et al.  Bayesian inference for source determination with applications to a complex urban environment , 2007 .

[16]  Katherine von Stackelberg,et al.  Estimation of fugitive lead emission rates from secondary lead facilities using hierarchical Bayesian models. , 2005, Environmental science & technology.

[17]  Sarvesh Kumar Singh,et al.  A least-squares inversion technique for identification of a point release: Application to Fusion Field Trials 2007 , 2014 .

[18]  Inanc Senocak,et al.  Stochastic reconstruction of multiple source atmospheric contaminant dispersion events , 2013 .

[19]  Hyondong Oh,et al.  A review of source term estimation methods for atmospheric dispersion events using static or mobile sensors , 2017, Inf. Fusion.

[20]  Shinsuke Kato,et al.  Identification of Pollution Sources in Urban Areas Using Reverse Simulation with Reversed Time Marching Method , 2009 .

[21]  Mark M. Tanaka,et al.  Sequential Monte Carlo without likelihoods , 2007, Proceedings of the National Academy of Sciences.

[22]  Anna Wawrzynczak,et al.  Application of the Approximate Bayesian Computation methods in the stochastic estimation of atmospheric contamination parameters for mobile sources , 2016 .

[23]  E. Somersalo,et al.  Statistical and computational inverse problems , 2004 .

[24]  Juan M. Corchado,et al.  Fight sample degeneracy and impoverishment in particle filters: A review of intelligent approaches , 2013, Expert Syst. Appl..

[25]  H. Leon Harter,et al.  Tables of Range and Studentized Range , 1960 .

[26]  Donald D. Lucas,et al.  Impact of meteorological inflow uncertainty on tracer transport and source estimation in urban atmospheres , 2016 .

[27]  Tiancheng Li.  The Optimal Arbitrary-Proportional Finite-Set-Partitioning , 2014, ArXiv.

[28]  Branko Ristic,et al.  Localisation of a source of hazardous substance dispersion using binary measurements , 2016, 1601.02304.

[29]  Julien Cornebise,et al.  On optimality of kernels for approximate Bayesian computation using sequential Monte Carlo , 2011, Statistical applications in genetics and molecular biology.

[30]  Sarvesh Kumar Singh,et al.  Inverse modelling methods for identifying unknown releases in emergency scenarios: an overview , 2015 .

[31]  Keith D. Copsey,et al.  Approximate Bayesian Computation for Source Term Estimation , 2009 .

[32]  Janez Demsar,et al.  Statistical Comparisons of Classifiers over Multiple Data Sets , 2006, J. Mach. Learn. Res..

[33]  Wenguo Weng,et al.  Source identification for unsteady atmospheric dispersion of hazardous materials using Markov Chain Monte Carlo method , 2009 .

[34]  Jean-Michel Marin,et al.  Approximate Bayesian computational methods , 2011, Statistics and Computing.

[35]  Yan Wang,et al.  Stochastic source term estimation of HAZMAT releases: algorithms and uncertainty , 2015, ISCRAM.

[36]  Steve Zegelin,et al.  Atmospheric tomography: a Bayesian inversion technique for determining the rate and location of fugitive emissions. , 2012, Environmental science & technology.

[37]  Patrick Armand,et al.  An adaptive Bayesian inference algorithm to estimate the parameters of a hazardous atmospheric release , 2015 .

[38]  C C Drovandi,et al.  Estimation of Parameters for Macroparasite Population Evolution Using Approximate Bayesian Computation , 2011, Biometrics.

[39]  Yan Wang,et al.  Rao-Blackwell dimension reduction applied to hazardous source parameter estimation , 2017, Signal Process..

[40]  J. Stockie,et al.  Bayesian estimation of airborne fugitive emissions using a Gaussian plume model , 2016, 1602.09053.

[41]  Radford M. Neal.  Slice Sampling , 2003, The Annals of Statistics.

[42]  Yu Li,et al.  Sequential multi-nuclide emission rate estimation method based on gamma dose rate measurement for nuclear emergency management. , 2017, Journal of hazardous materials.

[43]  S. Walker.  Invited comment on the paper "Slice Sampling" by Radford Neal , 2003 .

[44]  Philip Jonathan,et al.  An improved algorithm for locating a gas source using inverse methods , 2007 .

[45]  K. S. Rao.  Source estimation methods for atmospheric dispersion , 2007 .

[46]  Nicolas W. Hengartner,et al.  Stochastic event reconstruction of atmospheric contaminant dispersion using Bayesian inference , 2008 .

[47]  Branko Ristic,et al.  Achievable accuracy in Gaussian plume parameter estimation using a network of binary sensors , 2015, Inf. Fusion.

[48]  Branko Ristic,et al.  Beyond the Kalman Filter: Particle Filters for Tracking Applications , 2004 .

[49]  Branko Ristic,et al.  Bayesian likelihood-free localisation of a biochemical source using multiple dispersion models , 2014, Signal Process..