Multispecies Reactive Transport---A Computer Mediated Distance Learning course by Walt W. McNab.

Updated: June 7, 2001.
Copyright © 2000 by Walt W. McNab, Concord, CA, U.S.A. 
All Rights Reserved.

Computer-Mediated Distance Learning

Course on

MULTISPECIES REACTIVE TRANSPORT IN GROUNDWATER

INSTRUCTOR:

WALT W McNAB

Dr Walt McNab

Lawrence Livermore National Laboratory

Livermore, California, U.S.A.

TOPIC D: SUMMARY AND FUTURE DIRECTIONS IN MODELING REACTIVE TRANSPORT

Lecture 2: Conclusions and the Path Forward

A Retrospective

Reactive transport models allow the contaminant hydrologist, geologist, or soil scientist to view chemically reactive aqueous systems in soil or rock as dynamic rather than static. With a purely static model, the questions we may pursue are limited in scope:

What are the pertinent components/species in a system?

What are the relevant reactions?

Are the aquifer and solid phase chemical compositions altered in a significant way by heterogeneous reactions?

With a dynamic model, we may broaden our list of questions to include:

What materials are flowing into and out of our system (i.e., an open system)?

What are the appropriate boundary conditions?

Over what time scales will heterogeneous chemical reactions exert a measurable influence on the solid phase chemistry? As we observed in Topic C (specifically in Lecture 2), this question need not involve explicit definitions of chemical kinetics; rather, it may depend in some cases only upon the limited solubilities of many different types of minerals.

Throughout this course, our strategy has been to understand reactive transport in a manner analogous to the construction of the models themselves; that is, a divide-and-conquer perspective emphasizing the compartmentalization of chemical processes into different volume elements linked by the transport equation. Weíve looked at the various types of chemical processes of concern ñ aqueous complexation, redox, mineral precipitation and dissolution, adsorption and ion exchange. Weíve used specific example applications ñ acid mine drainage (contaminant hydrology), ore formation (geology), and soil profile genesis (soil science) ñ to illustrate reactive transport from the perspective of different disciplines. However, the examples were only intended as a survey, some of the chemical issues of interest (e.g., ion exchange, reaction kinetics) were not addressed in the examples. Weíve left it up to the student to explore these and other topics independently.

Challenges and Opportunities in the Future

So what will reactive transport models do for us in the future? The capabilities of these models, and their list of applications, will continue to grow in response to several trends:

Enhanced computing power, both at the higher end for complex research applications as well as the lower end (e.g., desktop PCs) for solving practical problems.

An improved understanding of the coupling between flow and transport arising from field and laboratory studies as well as further experimentation with the models themselves.

An ever-increasing need to apply computer simulations to understand complex problems of interest to science and engineering.

Here are a few potential ìgrowth areasî for reactive transport simulation as we look ahead:

Improved definitions/concepts/databases. As alluded to above, laboratory and field experiments continue to provide insights into natural processes that hold implications for reactive transport modeling. Specific examples include improved models for reaction kinetics, identification of relevant species, and more accurate thermodynamic data. Each of these will improve the general applicability of reactive transport models to a broad assemblage of problems.

ìUsabilityî. Current developments in desktop computation emphasize the development of improved user interfaces to facilitate the operation of various kinds of software. With regard to groundwater models, this implies, for example, the replacement of character-based input files with graphical interfaces featuring pull-down menus, buttons, various levels of menu options, etc. Reactive transport models, with all of their complex requirements for input, stand to gain much from this trend, allowing non-experts easier access to these types of tools. Whether or not the use of reactive transport models by ìnon-expertsî is a desirable is a debatable issue, but you get the point!

Sensitivity analyses. Reactive transport models depend on large collections of input parameters, including definitions of the transport problem, solution chemistry, and thermodynamic and kinetic constraints. The impact that uncertainties in any of these quantities exert on the overall quality of the model output is sometimes difficult to ascertain a priori. Sensitivity analyses attempt to address this problem in a systematic manner. For example, for a given parameter (e.g., the rate constant of a particular reaction), a series of estimates, perhaps based on a probability distribution, can be used in the model instead of a single value, producing a series of simulations. This approach is commonly referred to as a Monte Carlo simulation. Because of the computational demands exerted by reactive transport models, application of Monte Carlo techniques has been problematic. However, increases in computational power will bring Monte Carlo analyses and other sensitivity analysis approaches into more common practice.

Complex flow coupling. As the dynamics of non-Darcian flow regimes (e.g., turbulent flow in rivers, streams, and through Karst terrain in the subsurface) are better understood and modeled, more sophisticated coupling between transport and chemistry may be realized in reactive transport models.

Parallel processing. The two-step approach used by most reactive transport models ñ solution of the transport equation for each component and the solution of the chemical speciation equations for each volume element, naturally lends itself to a parallel, rather than a sequential computation approach. That is, during each time step, the speciation of the local geochemistry within a given volume element is independent from all others. On a computer with multiple processors, each processor can solve the chemical speciation problem for a subset of volume elements over a given time step. This can lead to tremendous efficiency gains, greatly reducing the time need to reach a solution for an entire system, bringing within reach large-scale, enormously complex simulations for problems that would be too intractable to solve otherwise on single-processor computers.

These possibilities, and others you may think of, suggest that reactive transport modeling will expand into an ever more interesting field. Nevertheless, it is always important to keep in mind that a reactive transport model, like any science or engineering simulation, is only a cartoon of reality. It produces output that is only accurate to the same degree as the input data, and thus ultimately reflects our understanding of a given system (complex or simple). In this regard, the greatest challenge in applying reactive transport models is, and always will be, assessing how well the conceptual model of a system is defined, based on real field observations, and how well that conceptual model can be manifested in the model.

This brings this course to its end. I hope you have learned new material, new tools and new ideas which will henceforth be useful to you professionally. Not les important, I hope you ahve enjoyed the course.

You may e-mail me questions and comments.

Walt W. McNab
E-mail address: mcnab1@llnl.gov.