ARTICLE pubs.acs.org/est
A Novel Method Using a Silicone Diffusion Membrane for Continuous 222 Rn Measurements for the Quantification of Groundwater Discharge to Streams and Rivers Harald Hofmann,*,†,‡ Benjamin S. Gilfedder,†,‡ and Ian Cartwright†,‡ † ‡
School of Geosciences, Monash University, Clayton, Victoria, Australia National Centre for Groundwater Research and Training, Australia ABSTRACT: 222Rn is a natural radionuclide that is commonly used as tracer to quantify groundwater discharge to streams, rivers, lakes, and coastal environments. The use of sporadic point measurements provides little information about short- to medium-term processes (hours to weeks) at the groundwatersurface water interface. Here we present a novel method for high-resolution autonomous, and continuous, measurement of 222Rn in rivers and streams using a silicone diffusion membrane system coupled to a solid-state radon-in-air detector (RAD7). In this system water is pumped through a silicone diffusion tube placed inside an outer air circuit tube that is connected to the detector. 222Rn diffuses from the water into the air loop, and the 222Rn activity in the air is measured. By optimizing the membrane tube length, wall thickness, and water flow rates through the membrane, it was possible to quantify radon variations over times scales of about 3 h. The detection limit for the entire system with 20 min counting was 18 Bq m3 at the 3σ level. Deployment of the system on a small urban stream showed that groundwater discharge is dynamic, with changes in 222 Rn activity doubling on the scale of hours in response to increased stream flow.
’ INTRODUCTION Documenting the exchange between surface water and groundwater is critical for understanding the hydrology of surface water bodies, such as rivers, lakes, and oceans. Simplification or even neglecting the connectivity between groundwater and surface water has led to many water allocation and water exploitation problems, particularly “double accounting” of water resources.1 Limited knowledge of the interconnectivity between these reservoirs, which may vary on hourly to yearly time scales in response to external forcing, such as precipitation, floods, drought, and human exploitation, is likely to have caused mismanagement. Indeed, the limited data available, most of which comes from coastal and estuarine environments, suggests that groundwater surface water exchange is highly dynamic and can vary on the scale of hours to days (e.g., refs 27). Sporadic geochemical sampling campaigns (e.g., monthly to half-yearly point samples) are unlikely to capture the full complexity of the groundwatersurface water system. To gain a better understanding of short- to mediumterm groundwatersurface water interactions, high-resolution physical and chemical data is needed that can be measured continuously over an extended period of time. 222Rn is an excellent tool for tracing short-term processes at the interface between groundwater and surface water. 222Rn is a radioactive gas produced in rocks and minerals by the decay of 226Ra, within the 238U decay chain. It accumulates in pore spaces after emanating from the surface of mineral grains through α-recoil.8 222Rn is sparingly soluble in water (Henry constant kH = 1.08 102 mol L1 atm1 9), has a short half-life (∼3.8 days), and, as a noble gas, is chemically and r 2011 American Chemical Society
biologically inert. It degasses, however, to the atmosphere from surface water bodies, and high 222Rn activities occur only in the vicinity of zones of groundwater input. This makes it an ideal natural tracer for identifying groundwater input to streams and rivers. 222Rn activities have been used extensively to constrain groundwater discharge to rivers,1015 lakes, 1618 coastal environments,6,1923 and glacial environments.24 It has also been used to a limited extent to trace recharge of aquifers from rivers via river bank infiltration25,26 and to study hyporheic exchange.12,27 Measurement of Radon in Water. 222Rn activities in water are traditionally measured by taking discrete point samples, followed by liquid scintillation counting (LSC) after liquidliquid extraction with a solvent, or by stripping the radon from the sample and measuring the evolved gas by a radon-in-air detector (e.g., Lucas scintillation cell, ionization chamber, or solid-state silicon detector). The Durridge RAD7 has been widely used in environmental sciences for the last 510 years. LSC is an expensive and time-consuming technique that is generally confined to the laboratory. Furthermore, the short half-life of 222Rn and low surface water activities require that samples are rapidly measured, which essentially precludes using 222Rn as a tracer in remote locations, where transport time to the laboratory is more than a Received: August 2, 2011 Accepted: September 2, 2011 Revised: September 1, 2011 Published: September 02, 2011 8915
dx.doi.org/10.1021/es202683z | Environ. Sci. Technol. 2011, 45, 8915–8921
Environmental Science & Technology
ARTICLE
diffusion membrane for 222Rn, despite its excellent properties for long-term deployment in the field (e.g., very robust, permeable to most gases, and inexpensive;40 Kies, personal communication, 2008).
Figure 1. Silicone radon diffusion “tube in a tube”. The PVC mantles the silicone tube and is connected in a closed loop to a desiccant unit and the RAD7 radon-in-air detector.
few days. Although portable LSC devices are available, they are complicated to operate in the field. Moreover, as natural processes can vary significantly in space and time on a variety of scales, continuous field-based monitoring has many advantages. For example, due to the current lack of robust long-term measurement techniques, there is very little known about how rapidly groundwater discharge to rivers varies. Continuous radon measurements in air and water have been undertaken since the 1950s,28 and continuous measurements in water have been increasingly reported over the last 10 years. A method to measure 222Rn continuously in water using an “air water exchanger” was developed by Friedmann and Hernegger,29 Surbeck,30 and Burnett et al.31 The airwater exchanger developed by Burnett et al. 31 was later modified by Dulaiova et al. 32 and Dimova et al.33 to increase the temporal resolution of the extraction device to about 1520 min. It has become a standard method for continuous measurements and has been applied in the field by Santos et al.,4 Santos et al.,6 Peterson et al.,34 and Stieglitz et al.,23 among others. Membranes, such as microporous polymer tubes, e.g., the Membrana Accurel (Membrana Inc.), have also been used to transfer dissolved 222Rn from water to an air circuit for measurement with a radon-in-air detector.22,3537 Although very efficient for the waterair exchange of 222Rn, these membranes clog easily with sediments and biofilms36,37 and are therefore less suitable for long-term autonomous field deployment. This study was driven by our need for a system that can measure 222Rn activities in streams and rivers continuously to better understand the variability of groundwatersurface water interaction. The system is required to work autonomously in the field for at least 34 weeks and thus must be simple, rugged, and reliable. Another requirement is that the system is responsive enough to capture changes in groundwatersurface water interactions over a few hours to days, for example, to quantify bank storage and return flow following high water events. The method described in this paper is based on 222Rn gas exchange through a silicone membrane installed within an airtight poly(vinyl chloride) (PVC) tube (Figure 1). Although Jonsson et al.38 described a system where silicone tubing was used in combination with a liquid scintillation counting device for 222Rn measurements in Icelandic bores, and Surbeck39 used a silicone tube to measure 222 Rn in karst springs, silicone tubing has rarely been applied as a
’ MATERIALS AND METHODS For the evaluation of the efficiency, response time, and accuracy of the new diffusion method, we have simultaneously compared the silicone diffusion membrane with an airwater exchanger based on the design of Burnett et al.31 The airwater exchanger sprays water at approximately 2.5 L min1 into the exchanging chamber, where radon in the water equilibrates with the chamber headspace. The 222Rn activity in the headspace is then measured in closed circuit using a RAD7 radon-in-air detector. Under ideal conditions, the gas airwater equilibrium for 222Rn should be reached after ∼4(V/F), where V is the air volume in the headspace plus detector and the tubing and F is the water flow rate through the airwater exchanger chamber. In our setup, this should take only 2 min; however, it is unlikely that the exchange is ideal and the results presented in the Results and Discussion section suggest that it takes about 25 min. The semipermeable silicone membrane diffusion tube (abbreviated TINT—tube in a tube) developed here for radon sampling, consists of a standard laboratory grade silicone tube (ROTH GmbH, Germany) installed inside a commercially available 13 mm inner diameter (i.d.) reinforced PVC pressure hose (Figure 1). Silicones are polymerized siloxanes, which are a mixture of organic and inorganic polymers of the generic chemical formula (R2SiO)n, where R represents an organic group, such as a methyl group. The silicone tube outer diameters (o.d.) range from 6 to 7 mm. The PVC hose is sealed at both ends with permanently installed tube connectors where the silicone tube enters and exits. During operation, sample water passes through the inner silicone tube and radon diffuses from the water through the membrane into the sheath air of the PVC tube. Two 8 mm tube connectors are installed in the PVC hose at each end to connect the sheath air with the solid-state radon detector in a closed circuit. The sheath air is circulated (∼1 L min1) counter flow to the water using the internal pump of the detector. The water temperature was measured using a T-type thermocouple (TC direct) with a Lascar data logger to compensate for the temperature dependence of the airwater partitioning coefficient (Ostwald solubility coefficient) for 222Rn. According to Meyer and Schweidler (1916, as cited in Weigel41), the Ostwald solubility coefficient α is related to the water temperature by the empirical relationship: α ¼ ð0:105Þð0:405e0:0502T Þ
ð1Þ
where T is the temperature in °C. At equilibrium the radon activity in water can then be calculated from the measured radon activities in the closed air loop by 222
Rnwater ¼ 222
222
Rnair α
ð2Þ
222
where Rnwater and Rnair are the activities of radon in water and air, respectively. Radon Detection System. 222Rn activities were measured using a Durridge RAD7 radon-in-air monitor, which uses a solid-state passivated ion-implanted silicon crystal held at a ground potential of ∼2200 V to count the decay of 222Rn progeny 218Po and 214 Po. The Si chip electrostatically attracts the positively charged 222 Rn daughter isotope 218Po+ (t1/2 = 3.05 min; E = 6.00 MeV) 8916
dx.doi.org/10.1021/es202683z |Environ. Sci. Technol. 2011, 45, 8915–8921
Environmental Science & Technology
ARTICLE
Figure 2. Setup of TINT optimization and TINTairwater exchanger comparison experiments.
and discriminates each α decay pulse from 218Po and its progeny into a pulse-height spectrum consisting of eight windows. This has the advantage that α counts from 218 Po and 214 Po decay can be separated. By using the 218Po channel, the RAD7 can quantify rapid changes in 222Rn activity (secular equilibrium between 222Rn and 218Po ∼15 min, 218Po and 214Po ∼3 h). During this study, the α counts from the A window (218Po) were used to achieve the maximum temporal resolution. Experimental Setup of the TINT. The most crucial variable for high-resolution continuous measurements is the time required for the TINT to attain airwater equilibrium in response to a change in aqueous 222Rn activities. To optimize the diffusion system for maximum response times, we varied the silicone tube length, the wall thickness, and water flow rate through the tube. Three tube lengths (2, 5, and 10 m) were compared simultaneously at a flow rate of 1 L min1 to determine the effect of the membrane surface area on the time required to reach equilibrium. Three 5 m silicone tubes, with wall thicknesses of 0.5, 1.0, and 1.5 mm, were compared simultaneously to determine if the more robust 1.5 mm tubes reduced the exchange rate significantly compared to the 0.5 mm tubing. Finally, three water flow rates were tested on a 10 m TINT with 1 mm wall thickness by adjusting the flow to 0.66, 1.3, and 2.5 L min1. All experiments were conducted under field conditions on the bank of the Avon river, southeast Victoria, Australia. Groundwater was constantly pumped into a 100 L water drum with a submersible impeller pump at a flow rate of 6 L min1 from 5 m bore depth. At this flow rate, the drum was constantly overflowing to ensure a continual renewal of groundwater. A second barrel was filled with river water that had low 222Rn activities (∼1 Bq L1). To simulate rapidly changing 222Rn activities in a river, and to observe the rate of the systems response to this change, we installed a three-way valve between the two barrels and the TINT/exchanger pumps (Figure 2). This allowed us to rapidly switch between water with high and low 222Rn activities, as may occur in a river during a storm event or in a bore if pumping induces surface water inflows. Water was pumped though the airwater exchanger at about 2.5 L min1, whereas the TINTs were generally run at a flow rate of 1 L min1. The TINTs were operated in closed circuit with both drums. All experiments were run using a 5 min counting time (TINT and exchanger), which resulted in a precision of ∼5% relative standard deviation (RSD) when the airwater exchange was at equilibrium.
Figure 3. RAD7 and TINT detection limit experiment.
Point Measurements. Discrete water samples were taken from the water drums throughout the experiment, and 222Rn activities were measured using the RAD7 combined with the commercial RAD H2O kit (Durridge Inc.) (250 mL sample bottles, 10 min integration time). The bottles were filled to the top and then closed using the RAD H2O kit bottle head. 222Rn was stripped from the water by purging the sample with air, which was cycled through the RAD7 in a closed loop. Five replicate measurements of 222Rn activities were made on each sample. A similar procedure was applied to river water samples, using a 500 mL sample bottle with 30 min integration time. The 500 mL sampling bottles had been previously cross-calibrated against the RAD H2O kit. Detection Limit. The background count rate of the RAD7 instrument was measured by running the system in closed circuit with an inline activated carbon filter, which quantitatively adsorbs all radon from the air stream. With this setup, the instrument background (detection limit) for 222Rn at 3σ was 8 Bq m3 using a 20 min counting time (Figure 3). The effect of the TINT on background noise was determined by passing water with low 222 Rn activities (14 Bq m3) thorough a 10 m 1 mm TINT 8917
dx.doi.org/10.1021/es202683z |Environ. Sci. Technol. 2011, 45, 8915–8921
Environmental Science & Technology
ARTICLE
Figure 4. Comparison between a 1 mm 10 m TINT diffusion membrane (black) and the airwater exchanger.
and using a variety of counting times. The detection limit was then estimated by the 3σ value of repeated measurements at this low level.42 Surprisingly, there was very little difference in the detection limit when using 20, 40, 60, or 120 min counting time, e.g., 3σ value of 13 Bq m3 for 2 h versus 18 Bq m3 for 20 min (Figure 3). This suggests that the best option for continuous measurements is to use a 20 min integration time to increase the temporal resolution of the data with little loss in detection probability.
’ RESULTS AND DISCUSSION Comparison with the AirWater Exchanger. The performance of the TINT was compared to the well-established air water exchanger and grab samples. For this experiment, we used a 10 m TINT with 1 mm wall thickness silicone tube. The flow rates for both the airwater exchanger and the TINT were 2.5 L min1. To simulate varying 222Rn activities, we changed the water source from river water to the bore water using the threeway valve (Figure 2). Once equilibrium was reached, we again changed from a high to low 222Rn source by changing the valve back to the river water reservoir. Figure 4 shows that both extraction systems react rapidly to a change in aqueous 222Rn activities. While the airwater exchanger reached equilibrium after 25 min, it took approximately 180 min for the TINT. Nonetheless, the equilibrium 222Rn activities for both the airwater exchanger and the TINT did not differ within the error of the measurements. The 222Rn activities of the point samples taken during the experiment also agreed well with both systems. The 222Rn activities of both point and airwater exchanger measurements increased slightly (from 18 000 to 20 000 Bq m3) over the course of the experiment, probably due to variations in 222Rn activities of the groundwater that is drawn from different parts of the aquifer. After 400 min, the source of water was changed back to river water (222Rn activity ∼1000 Bq m3) to produce rapidly decreasing 222Rn activities. 222Rn activities decreased to river water levels in ∼35 min using the airwater exchanger and ∼200 min using the TINT. The delayed response of the TINT is due to 222 Rn diffusion through the membrane, which, as expected, is slower than the direct airwater exchange of the exchanger.33 The response time of the TINT during periods of increasing and decreasing 222Rn concentrations is similar, between 200 and 300 min. TINT Optimization. The diffusion 222Rn across the silicone tube membrane may be influenced by a number of factors, such
Figure 5. Dependence of equilibrium time on silicone tube length.
as the surface area available for exchange, thickness of the tube wall, and the water flow rate through the tube. The surface area available for exchange is directly proportional to the tube length. The ratio between the surface area A of the silicone tube available for 222Rn exchange and the volume V in the PVC hose (TINT dead volume) does not change with increasing tube length, as A/V scales as 2/r. However, the A/V ratio increases with silicone tube length when the total dead volumes (RAD7750 mL, the desiccant tube ∼350 mL, connection tubing ∼50 mL) are included in the calculation of V. In our system the A/V increases by ∼300% when using a 10 m TINT rather than a 2 m TINT, including the dead volume of 1150 mL. To test the effect this has on the instrument response time, TINTs (wall thickness 1 mm) of 2, 5, and 10 m length were simultaneously compared. As can be seen in Figure 5, the time taken to reach equilibrium varied significantly between the three lengths. The 10 m tube had reached equilibrium after 200 min, the 5 m tube reached a similar equilibrium activity after 350 min, while the 2 m tube had not reached equilibrium when the test was stopped after 500 min. According to Fick’s first law of diffusion, the flux across a membrane is inversely proportional to its thickness. However, a thicker wall will add extra stability and robustness to the TINT design. Thus, if a reduction in the response time with increasing wall thickness is minimal, it may be beneficial to use a thicker silicone tube. To test this, we compared silicone tubing with 0.5, 1, and 1.5 mm wall thickness on a 5 m TINT at a flow rate of 1 L min1. As expected the TINT with 0.5 mm silicone tubing reached equilibrium first (∼250 min), followed by the 1 mm tube (300 min) (Figure 6). The TINT with 1.5 mm wall thickness had not reached equilibrium after 330 min. This tube with 1.5 mm wall thickness had a slightly smaller i.d. (4 mm) than the 0.5 and 1 mm TINTs (5 mm), which may affect the rate of 222Rn diffusion due to the lower surface area available for exchange. When we apply a correction factor proportional to the difference in surface area between the 1.5 mm and the other membranes (20%), the modeled response of the 1.5 mm tube is closer to that of the other tubes, but it is still slower at reaching equilibrium. By differentiating the linear sections of each curve in Figure 6, we find a very strong inverse correlation between the gradient and the wall thickness (r2 = 0.99). Overall, the 1.5 mm silicone tube is 8918
dx.doi.org/10.1021/es202683z |Environ. Sci. Technol. 2011, 45, 8915–8921
Environmental Science & Technology
ARTICLE
Figure 8. Continuous 222Rn, depth, and temperature measurements in an urban creek. Figure 6. Effect of silicone tube wall thickness on diffusion rate and time to equilibrium. The 1.5 mm corrected trend derives from applying a 20% correction factor applied to the 1.5 mm tube due to slightly different surface area available for radon exchange.
container”, i.e., the radon activity in the air circuit being measured. The exhalation rate is calculated from measured data by E¼
Figure 7. Effect of flow rate on radon equilibrium time in a 1 mm 10 m TINT.
prohibitively slow in reaching equilibrium, and so either 0.5 or 1 mm tubing should be used to achieve a satisfactory response. Finally, the effect of flow rate through the silicone tube was determined on a 10 m TINT. The time required to reach airwater equilibrium decreased from 350 to 250 min, as we increased the flow rate from 0.66 to 1.33 and finally to 2.5 L min1 (Figure 7). With the use of those data, it is possible to calculate the diffusion rate constant D for the silicone tube using a slightly modified form of eqs 1 and 2, presented in Jiranek and Svoboda:43 D¼d
E Csc Crc
ð3Þ
where D is the diffusion constant (m2 s1), E is the radon exhalation rate (Bq m2 s1), Csc is the radon activity in the “source” container, which in our case is the water passing through the silicone tube, and Crc is the radon activity in the “receiving
λV ðCrc ðt2 Þ Crc ðt1 Þeλðt2 t1 Þ Þ Að1 eλðt2 t1 Þ Þ
ð4Þ
where λ is the 222Rn decay constant (2.1 106 s1), V is the volume of the air circuit, A is the surface area of the membrane, converted to a sheet by A = 2πrL. t1 is the time at the start of the experiment, and t2 was the time required for the radon activity to reached 1 e1 (∼63%) of the final activity in the closed air circuit. Since we are comparing gas-phase activities with waterphase activities, the equations must be modified to account for the preferential partitioning of 222Rn into the gas phase. Thus, the Crc is converted into water-phase activity using eqs 1 and 2. Using these expressions, the diffusion coefficient varies from 4.6 1010 to 9 1010 m2 s1, increasing with the flow rate. It is thought that increased pressure in the tubes at higher flow rates slightly stretches the molecular lattice (SiO(CH3)2) of the silicone membrane, increasing the effective pore size of the material and, therefore, the diffusion rate. D for the lowest flow (and so the least deformation of the material structure) was similar to the value of 3.2 1010 m2 s1 found by Mamedov et al.44 for a silicone membrane at atmospheric pressure. Field Deployment. To test the suitability of the TINT for continuous field deployment, we installed an optimized system, consisting of a 10 m TINT with 1 mm wall thickness, on the bank of a small stream near Monash University (Scotchman’s creek). A water level logger (pressure sensor) was also placed in the stream to correlate changes in 222Rn with changes in stream depth. The system was run continuously for approximately 4 days. Even over this short time span, 222Rn activities doubled (from 290 to 680 Bq m3) (Figure 8). These variations are considerably larger than can be accounted for by degassing and decay and indicate the dynamic nature of groundwater surface water interactions. The system also has been deployed for 3 months on a research raft in a coastal wetland in Victoria without an external power source and with the replacement of the desiccant on monthly intervals. As with Scotchman’s creek, there are significant changes in 222Rn activity over this time scale (Figure 9). In summary, the results from the optimization experiments show that tube length, membrane wall thickness, and flow rate are all important factors influencing the response of silicone tube 8919
dx.doi.org/10.1021/es202683z |Environ. Sci. Technol. 2011, 45, 8915–8921
Environmental Science & Technology
ARTICLE
’ ACKNOWLEDGMENT This study was supported by the National Centre for Groundwater Research and Training P3 program. The National Centre for Groundwater Research and Training is an Australian Government initiative supported by the Australian Research Council and the National Water Commission. We thank Professor Antoine Kies (University of Luxembourg) for the discussions on the use of membranes for radon and CO2 extraction, in particular the use of silicone tubing. We also thank Chris Pierson from the Geoscience workshop for help in the construction of the TINT setup. Figure 9. Continuous wetlands in Victoria.
222
Rn using the TINT on a raft in coastal
membranes to 222Rn diffusion. When these parameters are optimized, it is possible to construct a continuous 222Rn monitoring system that can capture changes in 222Rn activities, which allow determining variations in groundwater discharge to streams and rivers on a variety of scales from a few hours to a few months. This could be particularly useful for investigating processes, such as bank return flow, the impacts of storm events on groundwater discharge and recharge, and changes in the regional groundwater input to rivers and streams. The TINT system requires minimal power consumption, with a small pump capable of delivering 1 L min1, drawing only a few hundred milliamps, which is essential for long-term deployments in remote areas. The minimum power consumption we were able to achieve included a Rotek Inc. rotary pump drawing significantly under 1000 mA. The system can be supplied with a regular solar panel (e.g., 80 W) and a 712 Ah battery as buffer. Furthermore, the consumption of desiccant is significantly reduced when using a membrane compared to an airwater exchanger, as less humidity enters the air circuit, which results in longer maintenance intervals of at least a few weeks. In contrast to the microporous polymer membranes, which are constructed of porous fibres, clogging by sediments is not a problem, even without prefiltration. In the experiments conducted here, the bore water had high sediment loads (the water was turbid and gray), but this did not have any effect on the response, even after long exposure. The silicone tube is also very flexible, in contrast to the rather brittle microporous polymer tubing, which allows it to be molded into convenient shapes. In our system the TINT is rolled into a loop so that the entire 10 m length only occupies a 40 40 15 cm3 volume. Moreover, there is little chance of water entering and destroying the sensitive detection system as may be the case with the airwater exchanger if an outlet is blocked or if the air return tubing develops a leak. Thus, although the TINT does have a slower response than the other methods, it is more suitable for long-term, autonomous measurement. For systems that vary rapidly, such as tidal areas or radon mapping, the airwater exchanger would be a more suitable system, but regular supervision is required. The first field data presented here demonstrates that groundwater exchange with rivers and wetlands is complex and that it varies on the scale of hours to days. Thus, a continual monitoring system is needed to understand the real complexity of the groundwatersurface water system.
’ AUTHOR INFORMATION Corresponding Author
*Phone: +61 (3) 9905 5786; fax: +61 (3) 9905 4903; e-mail:
[email protected].
’ REFERENCES (1) Sophocleous, M. Interactions between groundwater and surface water: The state of the science. Hydrogeol. J. 2002, 10, 52–67. (2) Taniguchi, M.; Iwakawa, H. Submarine groundwater discharge in Osaka Bay, Japan. Limnology 2004, 5, 25–32. (3) Crusius, J.; Koopmans, D.; Bratton, J.; Charette, M.; Kroeger, K.; Henderson, P.; Ryckman, L.; Halloran, K.; Colman, J. Submarine groundwater discharge to a small estuary estimated from radon salinity measurements and a box model. Biogeosciences 2005, 2, 141–157. (4) Santos, I. R.; Dimova, N.; Peterson, R. N.; Mwashote, B.; Chanton, J.; Burnett, W. C. Extended time series measurements of submarine groundwater discharge tracers (222Rn and CH4) at a coastal site in Florida. Mar. Chem. 2009, 113, 137–147. (5) de Weys, J.; Santos, I. R.; Eyre, B. D. Linking groundwater discharge to severe estuarine acidification during a flood in a modified wetland. Environ. Sci. Technol. 2011, 45, 3310–3316. (6) Santos, I. R.; Peterson, R. N.; Eyre, B. D.; Burnett, W. C. Significant lateral inputs of fresh groundwater into a stratified tropical estuary: Evidence from radon and radium isotopes. Mar. Chem. 2010, 121, 37–48. (7) Burnett, W. C.; Peterson, R. N.; Santos, I. R.; Hicks, R. W. Use of automated radon measurements for rapid assessment of groundwater flow into Florida streams. J. Hydrol. 2010, 380, 298–304. (8) Porcelli, D.; Swarzenski, P. The behavior of U- and Th-series nuclides in groundwater. In Uranium-Series Geochemistry; Bourdon, B., Henderson, G. M., Lundstrom, C. C., Turner, S. P., Eds.; Reviews in Mineralolgy and Geochemistry; Mineralogy Society of America, 2003. (9) Sander, R. Compilation of Henry‘s law constant for inorganic and organic species of potential importance in environmental chemistry, 1999. http://www.henrys-law.org. (10) Wanninkhof, R.; Mulholland, P.; Elwood, J. Gas exchange rates for a first order stream determined by deliberate and natural tracers. Water Resour. Res. 1990, 26, 1621–1630. (11) Cook, P. G.; Favreau, G.; Dighton, J.; Tickell, S. Determining natural groundwater influx to a tropical river using radon, chlorofluorocarbons and ionic environmental tracers. J. Hydrol. 2003, 277, 74–88. (12) Cook, P.; Lamontagne, S.; Berhane, D.; Clark, J. Quantifying groundwater discharge to Cockburn River, South-Eastern Australia, using dissolved gas tracers 222Rn and SF6. Water Resour. Res. 2006, 42, 1–12. (13) Mullinger, N.; Binley, A.; Pates, J.; Crook, N. Radon in Chalk streams: Spatial and temporal variation of groundwater sources in the Pang and Lambourn Catchment, UK. J. Hydrol. 2007, 339, 172–182. (14) Gleeson, T.; Novakowsji, K.; Cook, P.; Kyser, T. Constraining groundwater discharge in a large watershed: integrated isotopic, hydraulic, and thermal data from the Canadian Shield. Water Resour. Res. 2009, 45, W08402. (15) Kies, A.; Hofmann, H.; Tosheva, Z.; Hoffmann, L.; Pfister, L. Using 222Rn for hydrograph separation in a micro basin (Luxembourg). Ann. Geophys. 2005, 45, 101–107. (16) Kluge, T.; Ilmberger, J.; von Rohden, C.; Aeschbach-Hertig, W. Tracing and quantifying groundwater inflow into lakes using a simple method for radon-222 analysis. Hydrol. Earth Syst. Sci. 2007, 11, 1621–1631. (17) Schmidt, A.; Gibson, J.; Santos, I.; Schubert, M.; Tattrie, K.; Weiss, H. The contribution of groundwater discharge to the overall 8920
dx.doi.org/10.1021/es202683z |Environ. Sci. Technol. 2011, 45, 8915–8921
Environmental Science & Technology water budget of two typical boreal lakes in Alberta/Canada estimated from a radon mass balance. Hydrol. Earth Syst. Sci. 2010, 14, 79–89. (18) Dimova, N.; Burnett, W. Evaluation of groundwater discharge into small lakes based on the temporal distribution of radon-222. Limnol. Oceanogr. 2011, 56, 486–494. (19) Kim, G.; Hwang, D.-W. Tidal pumping of groundwater into the coastal ocean revealed from submarine 222Rn and CH4 monitoring. Geophys. Res. Lett. 2002, 29, 1678. (20) Lambert, M. J.; Burnett, W. C. Submarine groundwater discharge estimates at a Florida coastal site based on continuous radon measurements. Biogeochemistry 2003, 66, 55–73. (21) Burnett, W.; Dulaiova, H.; Stringer, C.; Peterson, R. Submarine groundwater: Its measurement and influence on the coastal zone. J. Coastal Res. 2004, SI39, 35–38. (22) Dulaiova, H.; Camilli, R.; Henderson, P. B.; Charette, M. A. Coupled radon, methane and nitrate sensors for large-scale assessment of groundwater discharge and non-point source pollution to coastal waters. J.Environ. Radioact. 2010, 101, 553–563. (23) Stieglitz, C.; Cook, P. G.; Burnett, W. C. Inferring coastal processes from regional-scale mapping of 222Radon and salinity: examples from the Great Barrier Reef, Australia. J. Environ. Radioact. 2010, 101, 544–552. (24) Kies, A.; Surbeck, H.; Tosheva, Z. Natural radioactive isotopes in glacier studies. In The Dynamics and Mass Budget of Arctic Glaciers— Extended Abstracts—Workshop and GLACIODYN Meeting, Pontresina (Switzerland); Oerlemans, J., Tijm-Reijmer, C., Eds.; IASC Working group on Arctic Glaciology, Institute for Marine and Atmospheric Research: Utrecht, The Netherlands, 2007; pp 5760. (25) Hoehn, E.; van Gunten, H. Radon in groundwater: A tool to access infiltration from surface water to aquifers. Water Resour. Res. 1989, 25, 1795–1803. (26) Dehnert, J.; Nestler, W.; Freyer, K.; Treutler, H.-C. Messung der infiltrationsgeschwindigkeit von oberfl€achenwasser mit hilfe des nat€urlichen isotops radon-222. Grundwasser 1999, 1, 19–30. (27) Lamontagne, S.; Cook, P. Estimation of hyporheic water residence time in situ using 222Rn disequilibrium. Limnol. Oceanogr.: Methods 2007, 5, 407–416. (28) George, A. C. World history of radon research and measurement from the early 1900’s to today. AIP Conf. Proc. 2008, 1034, 20–33. (29) Friedmann, H.; Hernegger, F. A Method for continuous measurement of radon in water of springs for earthquake prediction. Geophys. Res. Lett. 1978, 5, 565–568. (30) Surbeck, H. Radon monitoring in soils and water. Nucl. Tracks Radiat. Meas. 1993, 22, 463–468. (31) Burnett, W.; Kim, G.; Lane-Smith, D. A continuous monitor for assessment of 222Rn in the coastal ocean. J. Radioanal. Nucl. Chem. 2001, 249, 167–172. (32) Dulaiova, H.; Peterson, R.; Burnett, W.; Lane-Smith, D. A multi-detector continuous monitor for assessment of 222Rn in the coastal ocean. J. Radioanal. Nucl. Chem. 2005, 263, 361–365. (33) Dimova, N.; Burnett, W. C.; Lane-Smith, D. Improved automated analysis of radon (222Rn) and thoron (220Rn) in natural waters. Environ. Sci. Technol. 2009, 43, 8599–8603. (34) Peterson, R. N.; Santos, I. R.; Burnett, W. C. Evaluating groundwater discharge to tidal rivers based on a Rn-222 time-series approach. Estuarine, Coastal Shelf Sci. 2010, 86, 165–178. (35) Surbeck, H. A radon-in-water monitor based on fast gas transfer membranes. Proceedings of the International Conference on Technologically Enhanced Natural Radioactivity (TENR) Caused by NonUranium Mining, Szczyrk, Poland, 1996. (36) Schmidt, A.; Schlueter, M.; Melles, M.; Schubert, M. Continuous and discrete on-site detection of radon-222 in ground-and surface waters by means of an extraction module. Appl. Radiat. Isot. 2008, 66, 1939–1944. (37) Schubert, M.; Schmidt, A.; Paschke, A.; Lopez, A.; Balcazar, M. In situ determination of radon in surface water bodies by means of a hydrophobic membrane tubing. Radiat. Meas. 2008, 43, 111–120. (38) Jonsson, G.; Theodorsson, P.; Sigurdsson, K. Auto-Radon: A new automatic liquid scintillation system for monitoring radon in water
ARTICLE
and air. In LSC 2005, Advances in Liquid Scintillation Spectrometry; Chalupnik, S., Sch€onhofer, F., Noakes, J., Eds.; Arizona Board of Regents on behalf of the University of Arizona, 2006; pp 119123. (39) Surbeck, H. Dissolved gases as natural tracers in karst hydrogeology; radon and beyond. Proceedings of Multidisciplinary Approach to Karstwater Protection Strategy, UNESCO Course, 2005. (40) Surbeck, H.; Deflorin, O.; Kloos, O. Spatial and temporal variations in the Uranium series background in Alpine groundwater. In Uranium in the Environment, Mining Impact and Consequences; Merkel, B. J., Hasche-Berger, A., Eds.; Springer, 2006; pp 831839. (41) Weigel, F. Radon Chemiker Zeitung 1978, 102, 287–299. (42) Sanders, P.; Lippincott, R.; Eaton, A. Determining quantification level for regulatory purposes. J.—Am. Water Works Assoc. 1996, 88, 104–114. (43) Jiranek, M.; Svoboda, Z. Transient radon diffusion through radon-proof membranes: A new technique for more precise determination of the radon diffusion coeffient. Build. Environ. 2009, 44, 1–10. (44) Mamedov, F.; Cermak, P.; Smolek, K.; Stekl, I. Measurement of radon diffusion through shielding foils for the SuperNEMO experiment. J. Instrum. 2011, 6, C01068.
8921
dx.doi.org/10.1021/es202683z |Environ. Sci. Technol. 2011, 45, 8915–8921
