VARIABILITY IN VADOSE ZONE GASEOUS PHASE
CONTAMINANT CONCENTRATIONS RESULTING FROM
DIURNAL AND PERIODIC BAROMETRIC PRESSURE CHANGES
Eric C. Miller
LMITCO
ABSTRACT
The measurement of subsurface soil gas concentrations is employed on a large scale at many DOE facilities and at other locations throughout the world to characterize and monitor volatile organic compound contamination. However, the quantitative significance of contaminant vapors in soil gas data is currently debated among scientists. A potential source of uncertainty, typically not considered in debating quantitative significance of soil gas data, is a phenomenon best described as "barometrically induced variability (BIV)" Barometrically Induced Variability is a phenomenon by which atmospheric pressure changes induce changes in measured concentrations of soil gas at depth in the subsurface. In general, these types of barometric effects are thought to be confined to the near surface (1m), while typical soil gas collection occurs at greater depths. However, recent evidence suggest that barometric effects may penetrate to depths of over 30 m, a typical depth for soil gas monitoring. The effect of subsurface pressure fluctuations on measured VOC concentrations in soil gas at approximately 30 m below land surface (bls) was examined at a test well at the Department of Energy's Idaho National Engineering and Environmental Laboratory (INEEL). Data from this research indicates that measured subsurface contaminant concentrations varied from significantly during time periods of several weeks. This research also provided data that suggests that barometric pressure changes are closely related to these variations in measured subsurface contaminant concentrations. An understanding of this relationship at a given site may have significant influence on the number of, and manner in which, subsurface soil gas data is collected. In addition, these pressure changes may need to be considered in the interpretation of soil gas data sets except for the long term (i.e.., years to decades) or where order of magnitude results are adequate.
INTRODUCTION
Soil gas monitoring is a technique that is utilized to characterize the nature and extent of volatile organic compound (VOC) contamination in the subsurface. Although soil-gas investigations have been used to successfully delineate volatile organic contamination in the vadose zone and saturated zone, scientist still debate the quantitative significance of contaminant vapors in soil gas (1). This debate typically focuses on uncertainties in soil gas data relating to factors such as soil moisture distribution, soil permeability, and fracture flow. Another potential source of uncertainty, typically not considered in debating the quantitative significance of soil gas data, is a phenomenon best described as "barometrically induced variability (BIV)". BIV is a phenomenon by which atmospheric pressure changes induce changes in measured concentrations of soil gas at depth in the subsurface. BIV is likely a special case of a more general class of barometric effects generally well understood by scientists in theory, but seldom considered when planning soil gas sampling or in interpreting soil gas data. This is likely due to the conventional thinking that the typical depth of atmospheric pressure influence is much less than one meter (2), while the majority of soil gas investigations occur at greater depths. Recent evidence however suggests that pressure influences may extend to depths of 25 m (3), a depth typical of soil gas investigations in regions with relatively thick vadose zone sequences.
Test Objective
The potential influence of atmospheric pressure fluctuations on measured VOC concentrations in soil gas at approximately 30 m below land surface (bls) was examined at a test well in the Idaho National Engineering and Environmental Laboratory (INEEL). Practical implications on soil gas sampling and soil gas data interpretation were also examined.
Atmospheric Pressure and the Subsurface
Temporal changes in atmospheric pressure are propagated into the subsurface creating pressure gradients in the vadose zone. The atmospheric pressure changes that induce these subsurface pressure gradients can be complex, involving air temperature and air density. Atmospheric pressure changes find their origin in two major sources. The first is pressure changes resulting from the diurnal rising and falling of temperature in the atmosphere known as the thermal tide. The second source of atmospheric pressure change is associated with the movement of regional high and low pressure systems over the earth's surface. These pressure systems are temporally random and their size, shape, and pressure gradient are seldom uniform (3). The daily rotation of the earth upon its axis facilitates the solar or thermal tide, which results in two high pressure and two low pressure events each day (4). The change in atmospheric pressure associated with the earth's thermal tide is generally on the order of 5 millibars or less. In addition to these daily atmospheric pressure changes, locations in the mid-latitudes (30 - 60 degrees) experience additional pressure changes facilitated by the movement of high and low pressure air which generally overshadow daily pressure variations (4). Unlike diurnal pressure changes, pressure changes associated with regional air masses are difficult to predict in terms of either time or magnitude. Typically these changes are of greater magnitude than those associated with the thermal tide and can vary as much as 30 millibars. Atmospheric pressure changes, induced by these two sources, are propagated into the subsurface at a rate closely related to the permeability of the vadose zone matrix. Because of the difference in the rates at which pressure changes are propagated in the atmosphere and into the vadose zone, there exists complex pressure gradients between the atmosphere and the soil, and within the soil itself.
MATERIALS AND METHODS
Over an approximate 19 day period, beginning in late December 1996, subsurface pressure data and downhole carbon tetrachloride concentration data were collected at a frequency of one data point per hour. These data sets were collected from a vadose zone monitoring port completed at a depth of approximately 30m below land surface (bls). The data collection frequency and time period for the test were chosen in order to obtain adequate data resolution and to provide reasonable assurance of experiencing both diurnal and periodic barometric pressure fluctuations respectively during the test period.
Equipment
Subsurface pressure data were measured with a Setra 270 (psia) pressure transducer with reported accuracy of one-tenth of a millibar. The pressure transducer was calibrated within several weeks of the test by the INEEL calibration laboratory. Carbon tetrachloride concentrations were automatically sampled and analyzed with a Bruel and Kjaer 1302 photoacoustic data analyzer. The photoacoustic analyzer and pressure transducer were plumbed into a vapor port near land surface via Swagelok fitting and Teflon tubing. Both data sets were collected and stored on a Wavetek 52A data logger housed in a weather resistant enclosure. The enclosure was not heated other than by instrument waste heat for the duration of the test. The Wavetech data logger was powered via 12V battery.
Well Completion
The well from which the data was collected consists of a 6" ID casing surrounded by eight, 3/8" stainless steel vapor sampling ports installed exterior to the main well casing to varying depths. These vapor ports were completed with a sand/bentonite/grout pack above and below the monitoring point which consists of the crimped and perforated end of the stainless steel tubing.
Subsurface Lithology
The lithology of the test area consists of approximately 6m of surface alluvium overlaying fractured basalt sequences of varying thickness and vesicularity. Basalt sequences are separated by sedimentary interbeds of varying thicknesses at approximately 10m, 35m, and 75m bls. Groundwater depth at the test area is approximately 185m bls.
RESULTS AND DISCUSSION
Over the approximate 19 day period of the test, carbon tetrachloride concentrations ranged from approximately 500 ppmv to approximately 900 ppmv.
Fig. 1. Carbon Tetrachloride concentration Changes Measured at 30m bls with 10% Error Bands Around the Average Carbon Tetrachloride Concentration Showing Analytical Uncertainty
During the same period subsurface pressure ranged from approximately 830 mb to approximately 855 mb showing evidence of both diurnal and larger scale periodic pressure changes.
Fig. 2. Subsurface Pressure Measured at 30m bls Showing Both Periodic and Diurnal Subsurface Pressure Changes.
These changes occurred in the absence of soil vapor extraction activities, which had ceased temporarily approximately 3 weeks prior to the test. A significant inverse relationship exists between measured carbon tetrachloride concentration and subsurface pressure.
Fig. 3. Carbon Tetrachloride concentration and Subsurface Pressure Measurements
Statistical analysis of the entire data set indicated that the linear regression coefficient was -0.53. Analysis of subset of the data, delineated by changes in trend direction for each data set, indicated a linear regression coefficient of -0.93. This difference is likely due to lags in the response of concentration changes to subsurface pressure changes.
Fig. 4. The Linear Correlation Coefficient of -0.93 Indicates that Greater Than 85% of the Total Variance in Carbon Tetrachloride concentration can be Accounted for by a Linear Relationship with Subsurface
One hypothesis that may explain such variability would include the concept of a discrete soil gas contamination layer undergoing vertical oscillations relative to some stationary sampling point under the influence of subsurface pressure changes. This and other hypotheses remain to be tested in future investigations. Depending upon the ultimate use for this type of subsurface data, the impact of this variation can be meaningful. For purposes of risk assessment, where risks are generally calculated to the nearest order of magnitude, such variations may be academic. However, where remediation goals are stated in terms of specific subsurface vapor concentrations or where subsurface data is used in short to mid-term trending, such changes may be meaningful.
CONCLUSIONS
Atmospheric pressure changes, and the subsurface pressure changes they induce, should be considered in both planning for soil vapor data sampling and in the interpretation of subsurface soil gas data.
ACKNOWLEDGEMENTS
Special thanks to Dr. Wayne Downs of Brigham Young University whose continuing association and work in passive vapor extraction utilizing barometric effects have served as the intellectual catalyst for this investigation. In addition, thanks to Jeff Sondrup and Joel Hubbell of the Idaho National Engineering and Environmental Laboratory for their intellectual efforts in supporting this work. This work was funded by the US Department of Energy under DOE Idaho Office contract DE-AC07-94ID13223
LITERATURE CITED