A Stochastic Regression Approach to Analyzing Thermodynamic Uncertainty in Chemical Speciation Modeling

详细信息    查看全文

• 作者：Christopher L. Weber ; Jeanne M. VanBriesen ; and Mitchell S. Small
• 刊名：Environmental Science & Technology
• 出版年：2006
• 出版时间：June 15, 2006
• 年：2006
• 卷：40
• 期：12
• 页码：3872 - 3878
• 全文大小：515K
• 年卷期：v.40,no.12(June 15, 2006)
• ISSN：1520-5851

文摘

Chemical speciation modeling is a vital tool for assessingthe bioavailability of inorganic species, yet significantuncertainties in thermodynamic parameters and model formlimit its potential for decision-making. In this paper wepresent a novel method for the quantification of thermodynamic parameter uncertainty and ionic strength correctionmodel uncertainty using Bayesian Markov Chain MonteCarlo (MCMC) estimation methods. These methods allowfor the inclusion of correlation modeling, which has not beenpresent in previous work. The MCMC simulations areused to model a natural river water to determine theuncertainty in the calculated environmental speciation ofethylenediamenetetraacetate, a chelating agent thathas attracted considerable environmental interest. Theresults indicate that incorporating correlation among relatedthermodynamic parameters into the uncertainty model isnecessary to correctly quantify the overall system uncertainty.This result indicates the superiority of MCMC estimationmethods over traditional Monte Carlo methods when availabledata are used to estimate parameter uncertainty insystems with closely related model parameters.