Calculation of reaction parameter sensitivity coefficients in multicomponent subsurface transport models

Caroline L. Tebes-Stevens, Albert J. Valocchi

Research output: Contribution to journalArticlepeer-review


A significant source of uncertainty in multicomponent transport models is due to reaction parameter estimation error. In this paper, two codes are developed to compute reaction parameter sensitivity coefficients, which provide a quantitative measure of the impact of this uncertainty. One of the codes solves the full coupled system of sensitivity equations, while the other applies the operator splitting approach to decouple the sensitivity equations for each chemical component. CPU time comparisons demonstrate that operator splitting significantly reduces the simulation time for solving both the state and sensitivity equations. These comparisons also confirm that solution of the sensitivity equations is more efficient than a traditional direct perturbation sensitivity analysis. Sensitivity calculations for an example transport problem with cobalt and nitrilotriacetate (NTA) are used to gain insight into the relative significance of reaction processes and to rank individual reaction parameters in terms of importance. In this example, speciation reactions cause an indirect linking of the NTA biodegradation and the cobalt sorption reactions, making the NTA concentration nearly as sensitive to the sorption parameters as to the degradation parameters. (C) 2000 Elsevier Science Ltd. All rights reserved.

Original languageEnglish (US)
Pages (from-to)591-611
Number of pages21
JournalAdvances in Water Resources
Issue number6
StatePublished - May 20 2000


  • Biodegradation
  • Contaminant transport
  • Groundwater
  • Mathematical modeling
  • Operator splitting
  • Sensitivity analysis
  • Sorption

ASJC Scopus subject areas

  • Water Science and Technology


Dive into the research topics of 'Calculation of reaction parameter sensitivity coefficients in multicomponent subsurface transport models'. Together they form a unique fingerprint.

Cite this