Breaking New Ground for Clean Water:
ORNL's past, present and future

Groundwater, the potable water supply for over half of the U.S. population, is increasingly threatened by organic, inorganic and radioactive pollutants. Given the complexity and the inherent uncertainty of the subsurface, the need for research addressing subsurface characterization, transport, and remediation cannot be over emphasized. ORNL has been in a unique position to conduct this research because of the complex nature of the subsurface environment at the Oak Ridge Reservation, scientifc expertise, and the availability of experimental and computational resources. The following timeline is a sampling of research accomplishements related to subsurface science at ORNL during the past decade.

Aerial photograph of the Oak Ridge Reservation showing various waste area groupings (WAGS)--SWSA 6 (807-84).

New tracer techniques developed by ORNL researchers at Oak Ridge Reservation help understand complex subsurface transport processes occuring in heterogeneous, fractured porous media (5819-90).

3D FEMWATER, a three-dimensional finite element model is developed to simulate water flow through saturated-unsaturated media.

SWSA4

A FEMWATER simulation for Solid Waste Storage Area 4 (SWSA4) at the Oak Ridge Reservation.

FEMAIR, A finite-element model for simulating airflow through porous media is developed at ORNL to study novel remediation strategies such as in situ soil venting and vacuum extraction.

New breakthroughs for understanding scale dependent transport phenomena are enabled by Pedon scale experiments (5815-90).

LEWASTE, A three-dimensional transport model for simulating waste leaching in the subsurface is developed at ORNL.

P-LEWASTE, A parallel port of LEWASTE enables high resolution simulations for nuclear criticality studies.

The DOE PICS (Partnership in Computational Science) groundwater grandchallenge project was initiated to develop new algorithms and tools for parallel computation of subsurface flow and transport. The PICS consortium involves 9 institutions, 7 of which participated in the groundwater project.

PFEM, A Parallel port of 3D Femwater funded through PICS enables high resolution simulations on massively parallel computers such as the Intel Paragon (ORNL-DWG 93Z-6007).

The PICS groundwater contaminant code GCT 1.0 is released. GCT 1.0 simulates saturated/unsaturated flow and reactive transport using massively parallel archectures.

credit: Texas A&M University

GCT simulation showing migration of a high-density contaminant in a clayey aquifer.

MURF/MURT, multiregion flow and transport models are developed to accurately simulate flow and transport in fractured porous media such as the Karstic aquifers found at the Melton Branch site in the Oak Ridge Reservation.

credit: University of Texas at Austin

The PICS code GCT 1.2 and a graphical post processing tool EYE are released. At the left is an EYE visualization of a GCT 1.2 simulation of contaminent concentrations.

The DONIO library developed at ORNL enables 100 fold speedup of I/O in the PICS code GCT.
DOLIB/DONIO, contact: Ed D'Azevedo

credit: University of South Carolina PICS milestone: A 1.2 million node simulation is performed using GCT 1.2 and 128 processors on the Intel Paragon XPS/35. Simulation shows contaminant transport at Savannah River Site old burial ground.

credit: University of South Carolina A graphical user interface G3D (an integrated graphical user interface used for post-processing simulations, for preprocessing models and 3D grid generation) is developed by PICS.

High resolution simulations using PGREM3D with over 10 million cells on the Intel Paragon XPS/150 enable us to study the effectiveness of pump-and-treat remediation in heterogeneous aquifers using horizontal wells. contact: Kumar

PICS activities concluded. The PICS groundwater project resulted in over 50 publications, new algorithms and codes for accurate computation of ground water flow and transport, software libraries for parallel computation, visualization, and interactive computational tracking/steering.

Computational steering enables interactive simulation of groundwater remediation experiments.

A new U.S. Department of Energy Environmental Technology Partnership Initiative (ETPI) project "Influence of Coupled Processes on the Fate and Transport of Industrial Mixed Waste Plumes in Structured Media" is started. Part of this project involves developing an integrated high performance hydrobiogeochemistry code (HBGC) for predicting mixed waste migration that is influenced by coupled processes.
contact: Jack Gwo

High resolution transport simulations using PGREM3D transport code for TCE leaching at Portsmouth Gaseous Diffusion plant is used to study remediation using in situ chemical oxidation (ports.tiff).
contact: Kumar

Parallel hydrogeochemistry code simulations showing distributions of iron-carbonate and uranium-proton at the Melton Branch site at the Oak Ridge Reservation. contact: Jack Gwo

A new Natural and Accelerated Bioremediation (NABIR) project is started to study scale-dependent mass-transfer processes in structured porous media (nabir.tiff).

contact: Kumar	Genetic search algorithms are coupled to existing groundwater codes on massively parallel and workstation cluster platforms to solve (a) subsurface microbial activity zone detection (left) , and (b) subsurface fracture network characterization problems (below).
				contact: Jack Gwo

Computational and memory saving features implemented in the PGREM3D transport code enables solution of a 120 million degrees of freedom nonlinear multicomponent transport problem in less than 10 seconds per timestep on the 1024-processor Intel Paragon XPS/150.

An HBGC Simulation showing volume fraction of a secondary mineral "alunite" after 0.25 years of copper leaching. This simulation involved 7 species and 56 components.

As we look to the future, ORNL will continue to work on developing computational algorithms and simulation software that will not only address future modeling needs such as coupled multi-scale biohydrogeochemical transport processes, but also take advantage of emerging high performance computing environments such as distributed-shared memory systems and network distributed resources. As in the past, the computational work will go hand in hand with experimental work addressing issues such as the effects of soil/rock structures on biogeochemically enhanced transport and mass-transfer mechanisms.