Introduction
McClellan Air Force Base, northwest of Sacramento
covers almost 3000 acres of intense activity in support of jet fighter
maintenance. For decades the base has been involved in use, storage and disposal
of hazardous materials. In the late 1970's groundwater contamination was
discovered with studies pointing to historical waste disposal practices as
the source. In some cases the prevalent VOC's were detected in the groundwater
to levels of 100 feet. In 1981, McClellan was incorporated into the new Air
Force IRP and in 1987, the base was placed on the NPL. In 1991 a new remediation
initiative began to reduce risk and accelerate the cleanup process. Under
this plan a prototype remedy for a base wide model of soil gas extraction
was developed.
The subjects of the pilot became three local areas on the base representing a mix of diverse contamination problems, including multiple plumes, erratic sampling and complex soil stratigraphy. VOC plumes from varying sources formed the basis of the examination. Soils were highly variable, undergoing abrupt lateral and vertical facies changes or pinch outs over short distances.
3D Geoscience Modeling
Technology
McClellan AFB used the LYNX 3D Geoscience Modeling
System to assist in the study. LYNX uses a volume modeling method to create
3D representations of subsurface geology, incorporating geostatistical techniques
to predict contaminant distributions.
Analytical and soil data, plus site plans were imported into GMS. Soil horizons were interpreted on cross sections and a 3D volume model of each was developed. The spatial variability of contaminant samples were analyzed to produce geostatistical models. These were then used to predict concentrations in the soil volumes by controlling the prediction of contaminant distribution within the layers. Volumetrics analysis was used to determine the soil volumes where the contaminant values exceeded threshold.
The study showed that the soil vapor extraction method would be the best remediation method. It also outlined a site selection methodology for SVE removal actions. In addition, the LYNX 3D volume modeling clearly high-lighted several other interesting benefits.
Handling Diverse Data
The ability to deal with large amounts of diverse
spatial and analytic information provided ways of reduction and analysis
not possible before.
An Objective Medium
An effective template was created providing
analytical credibility to regulators, managers and remediators.
Adaptable Prediction
The ability to model several unique contaminant
distributions within different soil types was an advantage. By analyzing
data within volumes, sample distributions could be treated independently,
catering to their own realistic spatial relationships.
Objective Sampling
By applying geostatistical analysis it was possible
to determine an objective future sampling strategy. Consistent semi-variogram
ranges allowed new drill patterns to be developed requiring less holes.
Visualization
The approach showed ways to develop a plume quickly
and inspect it visually, providing an extremely effective way of understanding,
and presenting complex problems.
Precise Volumes
The benefits of the 3D volume approach allowed
a precise quantification of each soil stratigraphy, despite its
complexities.
Risk Quantification
By using geostatistics, it was possible to develop
data specific models that specifically quantify how good, or how badly the
samples relate to each other, dealing with both sparse and clustered
data.
Optimized Remediation
The results highlighted that extraction wells
could be placed at a lower density than first projected. In addition, the
extraction wells could be specifically placed for optimum extraction
effectiveness.