Environ. Sci. Technol. 2003, 37, 3141-3144
In-Situ Characterization of Soil-Water Content Using Gas-Phase Partitioning Tracer Tests: Field-Scale Evaluation J A S O N M . K E L L E R †,‡ A N D M A R K L . B R U S S E A U * ,†,§ Department of Soil, Water, & Environmental Science and Department of Hydrology & Water Resources, The University of Arizona, 429 Shantz Building, Tucson, Arizona 85721
Field-scale tests were performed to evaluate the effectiveness of the gas-phase partitioning tracer method for in-situ measurement of soil-water content. The tracer tests were conducted before and after a controlled infiltration event to evaluate performance at two water contents. Nonpartitioning (sulfur hexafluoride) and waterpartitioning (difluoromethane) tracers were injected into the test zone, and their effluent breakthrough curves were analyzed using the method of moments to calculate retardation factors for difluoromethane. Soil-water contents estimated using the tracer data were compared to soilwater contents obtained independently using gravimetric core analysis, neutron scattering, and bore-hole ground penetrating radar. For the test conducted under drier soil conditions, the soil-water content estimated from the tracer test was identical to the independently measured values of 8.6% (equivalent to water saturation of 23%). For the test conducted under wetter soil conditions, the tracer test derived soil-water content was 81% of the independently measured values of 12.2% (equivalent to water saturation of 32%). The reduced efficacy at the higher soil-water content may reflect the impact of advective and/ or diffusive mass transfer constraints on gas-phase transport. The results presented herein indicate that the partitioning tracer method is an effective technique to measure soil-water content at the field scale, especially for sites with moderate to low soil-water contents.
Introduction Soil-water content is often difficult to fully characterize at the field scale because of the spatial and temporal variability inherent to the subsurface. To date, characterization of soilwater content has emphasized the measurement of small sample volumes (e.g., point measurements). This includes the use of gravimetric analysis of soil cores, neutron scattering, electrical resistance, tensiometric data, and timedomain reflectometry. These methods provide soil-water content information at the scale of ∼10-1 m3. For larger, relatively heterogeneous systems, the collection of a sufficient * Corresponding author phone: (520)621-3244; fax: (520)621-6782; e-mail:
[email protected]. † Department of Soil, Water, & Environmental Science. ‡ Present address: Pacific Northwest National Laboratory, Richland, WA. § Department of Hydrology & Water Resources. 10.1021/es0340329 CCC: $25.00 Published on Web 06/07/2003
2003 American Chemical Society
number of point measurements to accurately characterize the spatial distribution of soil-water content may often be time- and cost-prohibitive. The gas-phase partitioning tracer method offers an alternative for in-situ characterization of soil-water content at a larger scale. The application of gasphase partitioning tracer tests for measuring soil-water contents has been recently evaluated for both laboratory and intermediate-scale systems. Brusseau et al. (1) conducted column experiments using helium as the nonreactive tracer and CO2 as the water-partitioning tracer. The soil-water content determined from the retardation of CO2 was similar to the gravimetrically measured value. Kim et al. (2) conducted column experiments using methylene chloride and chloroform as water-partitioning tracers. They found good correspondence between the tracer-derived soil-water contents and the gravimetrically measured values for a range of soil-water contents. The water-partitioning tracer method has been tested at the intermediate scale with a series of experiments conducted in a large (4.0 m deep, 2.5 m diameter) weighing lysimeter (3, 4). The lysimeter contains a homogeneous packing of fine sand and is instrumented with multiple devices with which to measure soil-water content. Experiments were conducted at three soil-water contents using SF6 as the nonreactive tracer and various halogenated methanes as water-partitioning tracers. For the first set of tests, soilwater contents of 0.04, 0.06, and 0.06 were estimated using the tracer data collected at the effluent sampling location with bromochlorodifluoromethane, dibromodifluoromethane, and trichlorofluoromethane, respectively, as the waterpartitioning tracers. The latter two values are identical to independently measured values of soil-water content obtained using gravimetric analysis of core samples and timedomain reflectometry. For the second set of tests, a soilwater content of 0.12 was estimated using the tracer data collected at the effluent sampling location with trichlorofluoromethane as the water-partitioning tracer. This value is 80% of the independently measured values obtained using time-domain reflectometry (0.15), neutron scattering (0.15), and conversion of soil-tension data (0.15). For the third set of tests, a soil-water content of 0.07 was estimated using the tracer data collected at the effluent sampling location with difluoromethane as the partitioning tracer. This value is identical to the independently measured values obtained using time-domain reflectometry, neutron scattering, and conversion of soil-tension data. Water-partitioning tracer tests have been conducted at the field scale as part of gas-phase partitioning tracer tests designed to characterize immiscible organic liquid contamination in vadose zone systems (5-7). However, the soilwater contents obtained in these tests were not compared to independently measured values to evaluate the efficacy of the method. The purpose of the study presented herein is to explicitly evaluate the field-scale performance of the water-partitioning tracer method. This will include comparing soil-water contents estimated using partitioning tracer data to soil-water contents measured with independent methods.
Materials and Methods The study was conducted at The University of Arizona West Campus Agricultural Center. The vadose zone at the site consists primarily of an unconsolidated, homogeneous sand with a 0.5-m-thick clay lense residing approximately 1 m below ground surface. The flow field for the tracer tests was generated using a single injection/extraction well couplet VOL. 37, NO. 14, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3141
TABLE 1. Experimental Conditions
test
flow rate (L min-1)
tracer pulse (h)
tracer pulsea (pore vol)
1 2
30.6 44.9
4.47 4.25
0.05 0.05
aCalculated
FIGURE 1. Plan view of the experimental site showing layout of wells and access tubes. spaced 7.3 m apart. The well casing is 5.1 cm o.d. schedule 40 PVC installed to a depth of 3.4 m below ground surface. The screened intervals are between 2.8 and 3.3 m below ground surface. Four access tubes (5.1 cm o.d. schedule 40 PVC) installed to a depth of 15.25 m exist at the site, allowing for independent measurements of soil-water content to be obtained using bore-hole ground penetrating radar and neutron scattering. A schematic of the site layout is provided in Figure 1. Sulfur hexafluoride (SF6) and difluoromethane (DFM) were used as the nonpartitioning and partitioning tracers, respectively. Difluoromethane has been used successfully as a water-partitioning tracer in previous studies (4, 6, 7, 11). Henry’s constants of 70 (8) and 132 (9) have been reported for SF6, while a value of 0.505 has been reported for DFM (10). The tracers were premixed at concentrations of 300 ppmv for SF6 and 20 000 ppmv for DFM in a balance of nitrogen and stored in two high-pressure gas cylinders (Spectra Gases, Branchburg, NJ). The tracer-free air used for injection was humidified by passing it through a water tower consisting of 66 L of water held within a schedule 40 PVC pipe (0.3 m o.d. and 1.8 m long). The air was humidified to prevent removal of water from the test domain. The tracer mixture injection bypassed the water tower to avoid retention of DFM. Gas flow was generated using a 5-hp air compressor coupled to the injection well and a 3/4-hp vacuum pump coupled to the extraction well. Pressure fluctuations due to cycling of the air compressor were dampened with a highprecision gas regulator (Moore Products Co., Spring House, PA) coupled to a 1/3-hp vacuum pump. A stainless steel gas line (12.7 mm diameter) was keyed into both well casings for either gas delivery or extraction. Each gas line was fitted with a needle-valved gas flow rotameter (Omega, Stamford, CT), a pressure transducer (Omega, Stamford, CT), and a stainless steel septum-injector nut (Valco Instruments Co. Inc., Houston, TX). Flow rates were maintained throughout the experiment by monitoring and adjusting the gas flow rotameters. The tracer pulse was injected once steady-state gas flow was established. Steady-state gas flow was determined to be established once pneumatic pressure within the wells remained stable. Following tracer injection, tracerfree air was injected for the remainder of the experiment. The tracer pulse width and flow rate are presented in Table 1 for each test. 3142
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 14, 2003
mass recovery (%) SF6
DFM
tracer swept vol (m3)
49.3 37.4
51.9 40.1
176.4 230.1
using SF6 mean travel time.
After completion of the first partitioning tracer test, a 66-h infiltration event was used to increase the soil-water content within the test zone. Water was evenly applied to the surface of the site through porous hoses connected to a manifold. The advance of the wetting front and redistribution of soilwater after cessation of the infiltration event were monitored, and the second partitioning tracer test was conducted after soil-water contents were stable. Influent and effluent gas samples were collected using the septum-injector nut sampling ports. Gas samples were collected in 80-mL evacuated canisters (Tracer Research Corp., Tucson, AZ). The samples were analyzed by Tracer Research Corporation (Tucson, AZ) using a Hewlett-Packard 5890 gas chromatograph, equipped with electron capture (SF6) and flame ionization (DFM) detectors. The quantifiable detection limits for SF6 and DFM are 0.1 and 80.5 µg/L, respectively. Breakthrough curves were constructed by plotting relative concentrations versus pore volumes. The time axis was converted to pore volumes by normalization with the mean travel time of the nonpartitioning tracer. The concentration axis was converted to relative concentrations by dividing by the input concentration. Retardation factors for DFM were determined by calculating the ratio of the mean travel times for DFM and SF6. The mean travel times were determined by moment analysis of the tracer breakthrough curves (12, 13). The moment analysis incorporated an exponential extrapolation of the elution tails, as described in prior work (e.g., ref 7). The retardation factors (R) obtained from the tracer tests are equated to the mass balance definition of the retardation factor, given as (assuming retention solely by soil-water):
R)1+
θw Sw )1+ θaKH (1 - Sw)KH
(1)
where θw (cm3 cm-3) is the volumetric water content, Sw (cm3 cm-3) is the water saturation, θa (cm3 cm-3) is the gas porosity, and KH (mol L-1 mol-1 L) is the Henry’s law constant. Four cores were retrieved from the location of the injection and extraction wells at depths of 2.6 and 2.9 m below ground surface for gravimetric analysis of soil-water content and soil bulk density. The samples were placed in airtight containers, weighed, oven-dried for 72 h at 105 °C, and weighed again. Gravimetric soil-water content was converted to volumetric soil-water content using the calculated average bulk density of 1.65 Mg m-3. Four bore-hole groundpenetrating radar measurements were taken from 2.75 to 3.5 m below ground surface using the B-D access tube pair, while four neutron probe measurements were taken over the same depth interval in both access tubes D and B, for a total of eight neutron probe measurements. The soil-water contents were averaged over the interval of measurement. Previous work at the site indicates that soil-water content exhibits a relatively uniform distribution within the test domain.
Results and Discussion Independently determined soil-water contents for both tracer tests are presented in Table 2. The soil-water content
TABLE 2. Independent Measurements of θw and Sw test 1 test 2
θw Sw θw Sw
GPR
neutron probe
cores
0.088 0.24 0.122 0.32
0.087 0.23 na na
0.082 0.22 na na
TABLE 3. Experimental Results and Tracer Determined Soil-Water Content and Soil-Water Saturation test 1 2
tracer SF6 DFM SF6 DFM
travel times (h) 96.1 152.3 85.4 146.1
retardation (R)
θw
Sw
1a
0.086
0.23
1.58 1a 1.71
0.099
0.26
aAssumed.
FIGURE 2. Breakthrough curves and tail extrapolation functions for test 1, conducted prior to a controlled infiltration event.
FIGURE 3. Breakthrough curves and tail extrapolation functions for test 2, conducted after a controlled infiltration event. for test 1 was determined using gravimetric analysis of soil cores, neutron scattering, and bore-hole ground-penetrating radar, while bore-hole ground-penetrating radar was used for the second test. All methods of independent measurement used in the first experiment are in reasonable agreement, providing a mean soil-water content of 0.086 cm3 cm-3. This is equivalent to a soil-water saturation of 0.23 cm3 cm-3. The independently measured soil-water content and soilwater saturation for test 2 were 0.122 and 0.32 cm3 cm-3, respectively. Tracer breakthrough curves and tail extrapolation functions for the first and second tests are presented in Figures 2 and 3, respectively. The breakthrough curves display sharp arrival waves and extensive elution-wave tailing. The tailing is likely a product of the flow field associated with the single injection-extraction well couplet and the influence of subsurface heterogeneities on flow. The arrival of DFM is delayed with respect to SF6, indicating retention of DFM. Mass recoveries were similar for the two tracers for both tests (see Table 2). The estimated tracer swept volumes are also reported in Table 2. Travel times, retardation factors, and tracer-derived soilwater contents and saturations are reported in Table 3. Comparison of the calculated DFM retardation factors for tests 1 (R ) 1.58) and 2 (R ) 1.71) demonstrates the
responsiveness of the partitioning tracer method to the increase in soil-water content associated with the infiltration event. A soil-water content of 0.086 cm3 cm-3 is estimated from the tracer test results obtained from the first test. This value is identical to the mean of the independently measured soil-water contents. A soil-water content of 0.099 cm3 cm-3 is obtained for test 2, which is 81% of the independently measured values. Interestingly, the results obtained from the field tests are similar to those obtained from the lysimeter experiments of Carlson et al. (4) discussed above. The soil-water contents obtained from the partitioning tracer tests were identical to the independently measured values for the tests conducted at the lower soil-water contents of 0.06 to 0.09. These water contents are equivalent to water saturations of approximately 13% and 23% for the lysimeter and field site, respectively. Conversely, the tracer-test derived soil-water contents were approximately 80% of the independently measured values for the tests conducted at the higher water contents, which are equivalent to water saturations of approximately 32%. The reduced efficacy observed for the tests conducted at the higher water saturation may be related to advective transport or mass transfer constraints associated with the higher soil-water content. For example, the larger water saturation reduces the air-phase relative permeability and may decrease the continuity of the gas phase for the higher soil-water content system. This could make it more difficult for the tracer pulse to contact all of the water in the system. In addition, it is possible that the thicker water films and pendular rings associated with the higher soil-water content may cause mass transfer of the partitioning tracer between the air and the water to be rate-limited by aqueous-phase diffusive constraints (e.g., ref 14). Theoretically, nonequilibrium conditions caused by rate-limited mass transfer do not influence the first temporal moment (travel time). Therefore, nonequilibrium conditions should not affect the calculated retardation of the partitioning tracer and the measured soil-water content. However, in practice, because of experimental constraints such as difficulty in measuring associated low-concentration tailing, rate-limited mass transfer can lead to underestimates of the first moment and therefore of the estimated soil-water content. The study reported herein provided the opportunity to evaluate the performance of the gas-phase water-partitioning tracer method at the field scale through comparison against independent measurements of soil-water content. On the basis of the results, the partitioning tracer method appears to provide reasonable estimates of soil-water content at the field scale and is responsive to changes in soil-water content. The method may be particularly appropriate for sites with extensive vadose zones, such as those found in arid and semiarid environments, for which the use of traditional methods would generally be cost-prohibitive. Implementation of tracer tests in conjunction with a limited number of point measurement devices may be a viable approach for applications requiring long-term monitoring of soil-water contents, such as those associated with vadose zone waste disposal facilities. While the point measurement devices serve as sentinels characterizing temporal changes in soil-water content, the VOL. 37, NO. 14, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
3143
partitioning tracer test would provide information regarding total volumes of water present at any given time. Further research is needed to evaluate the performance of the tracer method under a range of field conditions.
Acknowledgments Funding for this work was provided by the United States Department of Agriculture, National Research Initiative Program. We thank Lawrence Schenmeyer and Dr. Glen Thompson of Tracer Research Corp in Tucson, AZ, and Dale Rucker and Ty Ferre at the University of Arizona for their generous assistance.
Literature Cited (1) Brusseau, M. L.; Popoviu ´ ova´, J.; Silva, J. A. K. Characterizing Gas-Water Interfacial and Bulk-Water Partitioning for GasPhase Transport of Organic Contaminants in Unsaturated Porous Media. Environ. Sci. Technol. 1997, 31 (6), 1645-1649. (2) Kim, H.; Suresh, P.; Rao, C.; Annable, M. D. Gaseous Tracer Technique for Estimating Air-Water Interfacial Areas and Interface Mobility. Soil Sci. Soc. Am. J. 1999, 63, 1554-1560. (3) Nelson, N. T.; Brusseau, M. L.; Carlson, T. D.; Costanza, M. S.; Young, M. H.; Johnson, G. R.; Wierenga, P. J. A Gas-Phase Partitioning Tracer Method for the In Situ Measurement of SoilWater Content. Water Resour. Res. 1999, 35 (12), 3699-3707. (4) Carlson, T. D.; Costanza, M. S.; Keller, J.; Wierenga, P. J.; Brusseau, M. L. Intermediate-Scale Tests of the Gas-Phase Partitioning Tracer Method for Measuring Soil-Water Content. Soil Sci. Soc. Am. J. 2003, 67, 483-486. (5) Simon, M. A.; Brusseau, M. L.; Golding, R.; Cagnetta, P. J. Organic and Aqueous Partitioning Gas Tracer Field Experiment. In Platform Abstracts of the First International Conference on Remediation of Chlorinated and Recalcitrant Compounds; Monterey, CA, 1998.
3144
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 14, 2003
(6) Mariner, P. E.; Jin, M.; Studer, J. E.; Pope, G. A. The First Vadose Zone Partitioning Interwell Tracer Test for Nonaqueous Phase Liquid and Water Residual. Environ. Sci. Technol. 1999, 33 (16), 2825-2828. (7) Deeds, N. E.; Pope, G. A.; McKinney, D. C. Vadose Zone Characterization at a Contaminated Field Site Using Partitioning Interwell Tracer Technology. Environ. Sci. Technol. 1999, 33 (16), 2745-2751. (8) Wilson, R. D.; Mackay, D. M. Direct Detection of Residual Nonaqueous Phase Liquid as a Partitioning Tracer. Environ. Sci. Technol. 1995, 29 (5), 1255-1258. (9) Olschewski, A.; Fischer, U.; Hoffer, M.; Schulin, R. Sulfur Hexafluoride as a Gas Tracer in Soil Venting. Environ. Sci. Technol. 1995, 29 (1), 264-266. (10) Costanza-Robinson, M. S. Ph.D. Dissertation, University of Arizona, 2001. (11) Deeds, N. E.; McKinney, D. C.; Pope, G. A.; Whitley, G. A. J. Environ. Eng. 1999, 125, 630-633. (12) Skopp, J. Analysis of Solute Movement in Structured Soils. In ISSS Symposium on Water and Solute Movement in Heavy Clay Soils; Bouma, J., Raats, P. A. C., Eds.; International Institute for Land Reclamation and Improvement: Wageningen, The Netherlands, 1984; pp 220-228. (13) Jin, M.; Delshad, M.; Dwarakanath, V.; McKinney, D. C.; Pope, G. A.; Sepehrnoori, K.; Tilburg, C. E. Partitioning Tracer Test for Detection, Estimation, and Remediation Performance Assessment of Subsurface Nonaqueous Phase Liquids. Water Resour. Res. 1995, 31 (5), 1201-1211. (14) Popoviu ´ ova´, J.; Brusseau, M. L. Contaminant Mass Transfer During Gas-Phase Transport in Unsaturated Porous Media. Water Resour. Res. 1998, 34 (1), 83-92.
Received for review January 10, 2003. Revised manuscript received April 30, 2003. Accepted May 6, 2003. ES0340329
