LYNX Case Study (Environmental)

3D Geostatistical Modeling for Hanford Waste Site

Hanford Site History
The Hanford site encompasses 560 square miles within the Columbia River Basin in southeastern Washington. Hanford originally focused on plutonium production but has now shifted to environmental restoration and managing wastes generated by past operations.

At the current time Hanford also conducts advanced research and development for advanced reactors, energy technologies and waste disposal practices.

study area modelNumerical information was available in the form of boreholes which indicated the depth to the water table and the measured mount of tritium contamination, measured in nano Curies. The basic objective was to use the existing information to develop a preliminary site characterization model of the tritium plume. This would provide a better understanding of the distribution and location of the contaminant as well as create a visual representation of the subsurface. In addition, such characterization could provide an objective, defensible modeling method that could be used in optimizing future sampling programs and set the initial guidelines for remediation planning.

Site Characteristics
The Hanford site is in flat terrain underlain by a complex stratigraphy of alluvium and gravel, with granite bedrock. Topography slopes gradually southwest from 230 to 120 meters adjoining the river. Below surface the water table gradient slopes gradually in the same direction from elevations of 140 to 120 meters. This southeast gradient provides natural drainage into the river, allowing contaminant plumes to migrate down vertically into the water table and along gradient towards the river. Numerous monitoring wells and boreholes have been drilled to sample the location and extent of the tritium contaminants which varies from 0 to 700 nano Curies.

Application of 3D Geoscience Technology
tritium plumeA significant problem exists in that the extent of contamination and the amount of volume to be remediated dramatically effects the cost of remediation. Further complicating the process is the distribution and quantity of sample information which requires interpretive methods to extrapolate the spatial continuity of the contaminant.

The LYNX modeling technology was a logical choice to best characterize subsurface since it is designed specifically to deal with such spatial problems. It integrates spatial data management, geostatistical techniques, 3D modeling, volumetrics, engineering and visualization in one facility, thereby allowing precise representation and estimation of complex subsurface problems.

Site Information and Basic Objectives
Information made available for the study included a topographic map indicating the surface elevations over the area and a site plan showing the location of major features such as the river, the boundaries of the plants and generating stations.

Numerical information was available in the form of boreholes which indicated the depth to the water table and the measured mount of tritium contamination, measured in nano Curies. The basic objective was to use the existing information to develop a preliminary site characterization model of the tritium plume. This would provide a better understanding of the distribution and location of the contaminant as well as create a visual representation of the subsurface. In addition, such characterization could provide an objective, defensible modeling method that could be used in optimizing future sampling programs and set the initial guidelines for remediation planning.

Geostatistical Analysis
Initial analysis was performed on the tritium data using basic statistics, with the intention of determining data relationships and spatial behavior. Results indicated distinct lognormality forcing the need for log transformations. Upon transformation, the tritium data exhibited excellent spatial continuity up to 1500 meters, allowing a spherical semi-variogram model to be used. Further analysis of anisotropy was performed by selecting several different directions confirming a similar range of influence. A secondary, indicator analysis was carried out with three cut-off values of 1, 20, 80 nano Curies. Semi-variogram modeling confirmed the spatial continuity ranges of tritium and the extent of the tritium plume.

Contaminant Modeling
The spherical semi-variogram model was the basis for Kriging techniques used to estimate the contaminate distribution throughout the study area. 3D plumes were generated thereby allowing further inspection of plume migration, extents and behavior, making it possible to cut plans and sections at any orientation though the plume to detail contaminant ranges. By using enhanced visualization, 3D isosurfaces could be scrutinized showing the effect of plume limits depending on the threshold chosen.

Characterization Summary
90% confidence modelThe use of 3D Geoscience modeling has been used to effectively characterize the site and to provide the basis to further planning of remediation and sampling. Although not performed in this exercise, it is noteworthy that the development of the model not only facilitates the quantification of volumes and a measure of the certainty at any particular threshold or location, but it also provides the basis for cost effective, defensible sample planning and control by being able to display areas which have poor confidences and require additional sampling.

Finally, the development of a visual 3D model also provides an explicit visual picture of extent, degree and uncertainty, a vital tool that facilitates better analysis, understanding and expedient cleanup.