The saying, "out of sight, out of mind," doesn't apply to geochemists. For them, phenomena that occur underground are very much on their minds, and for good reason. Subsurface plumes of magma and the underground transport of nutrients make Earth what it is.

What happens underground: a schematic of the oxidation-reduction zones that can develop in an aquifer downstream from an organic-rich landfill. Closest to the landfill, methane is produced; then the moving plume passes through zones of sulfate reduction, iron reduction, and denitrification, becoming progressively oxidized by aerobic respiration and the influx of oxygenated water. Inset shows some of the chemical reactions that may occur among bacteria, dissolved organics, clays, and other minerals.

Unfortunately, studying the physical, biological, and chemical processes that drive the underground world is incredibly difficult. Often, geochemists' only glimpse of how magma forms or how rocks change under heat and pressure is by examining whatever bubbles up from deep underground or whatever erosion lays bare.

That's beginning to change. To learn more about underground phenomena, Berkeley Lab scientists are taking a computer-driven technique called reactive transport modeling — traditionally used as an engineering tool — and applying it to fundamental Earth Sciences research.

"We want a ringside seat to what is happening underground, and reactive transport modeling gives us that," says Carl Steefel, a geochemist in the Earth Sciences Division (ESD) who is among a group of scientists worldwide advancing the technique.

Reactive transport modeling enables researchers to create computer simulations of geochemical, microbiological, and physical processes that occur just below the surface, such as flowing contaminants and fluxes in nutrients, as well as those that happen deep underground, such as rock metamorphism and magma formation. It integrates the various mechanisms that drive underground processes: the way chemicals and microbes interact with each other, how minerals react with fluids, and the role of heat and pressure.

The technique isn't new. It is widely used as an engineering tool to help monitor the spread of contaminants at places like the Department of Energy's Hanford Site. But scientists like Steefel hope to use it not just to track underground phenomena but to gain fundamental insights into the subsurface processes that shape the planet. By integrating innovative research such as high-resolution images of mineral surfaces obtained at synchrotrons like Berkeley Lab's Advanced Light Source, they're working to transform reactive transport modeling into a powerful research tool.
 
In addition, they hope to use the technique to elucidate subsurface phenomena across a range of scales, from the microscopic scale to the field scale, which spans hundreds of square kilometers.

"The complex interplay of material flow, transport, and reactions at multiple spatial and time scales which characterize most Earth systems requires an integrated approach," says Steefel. "This is a particularly challenging task, since processes may play out at scales as small as the pore scale, and yet they affect the system's behavior at much larger scales."

Berkeley Lab is uniquely poised to advance this effort because it boasts state-of-the-art computing resources as well as geochemists, microbiologists, chemists, and computer scientists. For example, Steefel is part of a small group of geochemists at Berkeley Lab, including ESD staff scientists Eric Sonnenthal and Nic Spycher, who use computers to study the interplay between fluid transport — traditionally the province of physical hydrologists — and geochemistry.

Berkeley Lab is also home of the National Energy Research Scientific Computing Center (NERSC) and a new high-performance computer cluster in the Earth Sciences Division devoted to geochemistry. These resources allow scientists to develop computer simulations of complex natural systems and, importantly, link these simulations to laboratory and field observations.

"The Berkeley Lab group is highly qualified to lead the development of this field," says Steefel.

Reactive transport modeling's merits as a research tool are explored in a recent paper in the journal Earth and Planetary Science Letters by Steefel; Don DePaolo, head of ESD's Geochemistry Department and a professor of geochemistry in UC Berkeley's Department of Earth and Planetary Science; and Peter Lichtner of Los Alamos National Laboratory.

	High-resolution simulations of reactive transport in sediment reveal differences among nonreactive contaminants. Here nitrate is carried along by the subsurface flow of water, while cesium is slowed as it reacts with mineral surfaces. (The lower panels show effects on cesium transport of two different concentrations of sodium nitrate in the soil.) The simulations are conducted in physically diverse soil, which is why the "fingering" shows up.

Among its many applications, reactive transport modeling can be used to predict the transport of man-made contaminants that have the potential to threaten nearby water supplies and ecosystems. Specifically, reactive transport models are helping scientists determine the degree to which the spread of contaminants can be slowed by natural processes. If this doesn't work, it can also help scientists choose remediation strategies that can cost-effectively stem the flow of contaminants. The need for better transport models is underscored by the fact that some current models dramatically underestimate the extent to which dangerous chemicals spread.

On a larger scale, reactive transport models have been used to estimate the flux of elementals and nutrients in the atmosphere, ocean, and continents. Dating back to the 1970s, the technique has been used to describe how biogeochemical and transport processes at the interface between seawater and sediment regulate fluxes between these two important Earth reservoirs.

Based on this legacy, scientists now hope to apply the latest in computing power and basic research to some of today's most urgent problems. For example, estimating the transfer of carbon dioxide and methane between the ocean, atmosphere, soils, and vegetation can help scientists understand how each of these systems plays a role in regulating atmospheric carbon dioxide, which is a greenhouse gas.

"This could help elucidate carbon sequestration efforts," says Steefel.

Reactive transport modeling can also shed light on the extent to which rivers carry nutrients like nitrogen and phosphorous to coastal ocean waters, where they support a diverse array of biological activity.

"We need a good understanding of all of the processes, including transport and biogeochemical reactions, which can affect these fluxes," says Steefel. "The goal, just like in contaminants, is to have a sufficiently mechanistic model that works in several locales and scales."

Then there's the study of how magma forms deep inside the Earth and subsequently rises to the surface. As DePaolo explains, one of the best windows into the subsurface characteristics of magma is the study of lava that emanates from volcanoes. These analyses offer many valuable clues, but there are still many uncertainties. In addition, today's models rely on the material properties of partially molten rocks and the thermodynamics of melting rock, both of which are not fully understood.

"It is like studying stars," says DePaolo. "We know there is a lot going on in a star's interior, but all we can study is their light. Likewise, we know there is a lot going on with respect to magma, but all we can study is what comes out of the volcano — which has gone through many changes in its journey."

Because of these limitations, it is far from clear how magma forms and migrates to the surface. In the case of melting rock, for example, today's computer models don't describe how the rock's viscosity changes as it melts, and how this changes the way it flows.

Lava offers several clues to the hidden world of magma, but not enough for a full explanation of underground processes.

"These processes have never been captured together in the models," says DePaolo.

There is a better way. Reactive transport models that account for changing thermodynamic conditions along magma's flow path have the potential to pick up where current models leave off. Using them, scientists can infer how magma forms deep underground based on what is known about magma, such as its chemical concentrations and isotopic ratios.

"It is here that reactive transport models come into play," DePaolo says.

Additional information

• "Reactive transport modeling: An essential tool and a new research approach for the Earth sciences," by Carl I. Steefel, Donald J. DePaolo, and Peter C. Lichtner, appears in the 15 December 2005 issue of Earth and Planetary Science Letters (pdf)
• More about tracking contaminants and other environmental processes