Derivation of travel time of limestone cave drip water using tritium/helium 3 dating method



[1] A method of deriving the travel time of drip water based on the tritium/helium 3 (3H/3He) dating method was demonstrated in limestone caves in Indonesia. The examined drip waters from two caves, which were taken using different methods, included 2.15 (± 0.84) TU and 1.35 (± 0.77) TU of tritiogenic 3He. From the derived tritiogenic 3He contents and the measured tritium concentrations of the drip water samples (2.02 ± 0.27 TU and 1.19 ± 0.40 TU), the 3H/3He ages (travel times) of the drip waters from the two caves were estimated as 12.9 ± 3.8 years and 13.5 ± 6.3 years. The latter age is probably overestimated because re-equilibration of He between drip water and cave air is expected to occur during sampling. For reliable derivation of 3H/3He age, it is advisable to take a drip water sample that is kept isolated from the cave air.

1. Introduction

[2] Speleothems that have formed in limestone caves preserve information that can be interpreted in terms of past hydrologic and climatic conditions. Therefore, they are becoming increasingly useful for paleoenvironmental studies [e.g., Gascoyne, 1992; White, 2004; McDermott, 2004; Fairchild et al., 2007], which can elucidate conditions on a time scale of recent years to hundreds of years or even tens of thousands of years in the past [Burns et al., 2002]. Drip water in speleothems can convey information about conditions that pertained at the time of rainfall. For that reason, it is important to consider the drip water travel time when using speleothems to interpret paleoenvironments of approximately hundreds of years correctly. Nevertheless, except for hydrological studies [Chapman et al., 1992; Rademacher et al., 2003; Fairchild et al., 2006], few studies have examined drip water travel time.

[3] We are now investigating this topic with the intention of reconstructing past climate variations in Asian equatorial regions using oxygen isotopes and other geochemical proxies recorded in Indonesian stalagmites [Watanabe et al., 2006]. In particular, we seek to examine the detection of the precipitation anomaly that reflects the El Niño Southern Oscillation (ENSO). In this study, we attempt comparison of proxy data from stalagmites using a meteorological dataset that includes local precipitation data of the past 50 years. For this attempt, it is important to obtain information related to the drip water travel time. Some methods have been used for estimating the groundwater age. Especially, numerous studies using tritium/helium 3 (3H/3He) method have been made of groundwater age [e.g., Schlosser et al., 1988, 1989; Solomon et al., 1992, 1993; Szabo et al., 1996; Katz et al., 2004; Zuber et al., 2004]. The present study is designed to develop a derivation method, using the 3H/3He method, to determine the drip water travel time in a limestone cave.

2. Geological Setting and Sampling, Analysis Methods

[4] We collected drip water samples from Cipicung and Ciawitali caves in the Sukabumi area of eastern Java, Indonesia (Figure 1). Both caves have almost level floors. The Cipicung cave has a passageway in which it is possible to stand and walk, with large chambers that are several meters high in some places. In contrast, the Ciawitali cave has very narrow passages that necessitate crawling. The respective thicknesses of the strata overlying the Cipicung and Ciawitali caves are about 50 m and 30 m.

Figure 1.

Locations of study areas: Dcip01-03 and Dciw01, and S01-03, respectively, show drip water (D) and spring water (S) sampling sites.

[5] Drip water samples used for measurement of rare gases, including 3He, were taken from the Cipicung and Ciawitali caves using different methods (Dcip03 and Dciw01). In the Cipicung cave, after we attached a rubber tube directly to a thin stalactite, drip water was pulled into a hypodermic syringe through a thin brass pipe while isolated from the cave air (Figure 2). Because the pulling rate of the piston rod of the syringe was somewhat lower than the free flow rate (drip rate) of drip water, more than several minutes were necessary to fill the syringe (about 50 mL) with the drip water sample. Gas bubbles were only slightly observed during and after sampling. For the sample storage, the drip water in the hypodermic syringe was transferred to a soft copper pipe (major diameter, 9.5 mm; inside diameter, 7.9 mm; length, 30 cm; capacity, about 40 mL) and was sealed using special clamps. In the Ciawitali cave, only a thicker stalactite dripped sufficient water; it was not possible to attach the rubber tube to it. For this reason, we abandoned air-isolated sampling and took the drip water sample as quickly as possible into the hypodermic syringe through a plastic funnel covering the stalactite.

Figure 2.

Schematic illustration of the sampling apparatus for drip water for rare gas analyses.

[6] After passage through a rubber tube connected to the stalactite or to the funnel covering the stalactite, the drip water samples used for tritium concentration measurements were collected in large plastic bottles set under the drip points. These sample waters yielded stable water isotope compositions (δD and δ18O). Furthermore, two drip water samples in the Cipicung cave (Dcip01 and Dcip02) and three cold spring water samples from different locations around the caves (S01, S02 and S03) were taken for reference purposes.

[7] Dissolved 3He, 4He, and Ne were analyzed using mass spectrometry (VG 5400; Waters Corp.) at the Geo-Science Laboratory, Japan after extracting dissolved gases from the drip water samples in the copper pipe. Tritium values were analyzed using the 3He in-growth method. Oxygen and hydrogen isotopic ratios (δD and δ18O) of the water samples were determined respectively using a mass spectrometer (Geo 20/20; PDZ Europa Ltd.) installed at the Stable Isotope Laboratory of IGNS, New Zealand, with CO2 equilibration method and the zinc reduction method. The obtained δD and δ18O data are shown using δ notation as the per-mil-deviation from the value of SMOW. Temperatures of the drip water and cave air were measured using a platinum thermometer (TFX 410; ebro Electronic GmbH and Co. KG). The drip rate was measured directly using a measuring cylinder and stopwatch. Analytical results are presented in Table 1 and Table 2.

Table 1. Results of Air and Water Temperature, Drip Rate, δD, and δ18O Measurementsa
SampleDateTa,°CTw, °CδD, ‰δ18O, ‰Drip Rate, mL/min
  • a

    Tw, water temperature; Ta, air temperature. The analytical precisions are ±0.1‰ for δ18O and ±1.0‰ for δD, respectively.

S01Jun. 03, 200624.523.6−7.3−44.5-
S02Jun. 03, 200624.623.2−7.3−44.7-
S03Jun. 04, 200628.022.8−7.3−46.2-
Dcip01Jun. 03, 200624.523.6−7.2−43.416
Dcip02Jun. 03, 200625.023.9−6.5−40.991
Dcip03Jun. 03, 200624.124.5−6.7−40.516
Dciw01Jun. 04, 200624.324.3−7.3−44.311
Table 2. Results of Rare Gas and Tritium Measurements
Sample4He ± σ, 10−8 ccSTP/g3He ± σ, 10−14 ccSTP/gNe ± σ, 10−7 ccSTP/g3H ± σ, TU
Dcip034.28 ± 0.046.32 ± 0.191.71 ± 0.012.02 ± 0.27
Dciw014.11 ± 0.045.92 ± 0.171.66 ± 0.011.19 ± 0.40

3. Results and Discussion

[8] The respective drip rates of the Cipicung and Ciawitali caves were 16 mL/min and 11 mL/min. The amounts of drip waters in the Cipicung and Ciawitali caves during one year, calculated by those results, are about 8 tons and 6 tons, respectively, assuming a constant drip rate throughout the year. Because the annual precipitation in the Sukabumi area is about 3000 mm/year (Y. Watanabe et al., Comparison between stable isotope time series of a stalagmite and meteorological data from west Java, Indonesia, submitted to Journal of Quaternary Science, 2008), the results of calculations described above indicate that the drip waters have only small catchment areas, perhaps several square meters. Furthermore, water isotope compositions of the drip waters and the spring waters indicate normal meteoric water (Table 1). Considering the information given above, we infer that the drip water is percolated rainwater. It traveled through the thick limestone strata overlying the cave and ultimately reached the cave ceiling.

[9] The respective Ne concentrations of the drip waters of the Cipicung and Ciawitali caves were 1.71 × 10−7 ccSTP/g and 1.66 × 10−7 ccSTP/g. Because it can be assumed that dissolved Ne in drip water is derived from the atmosphere, the Ne concentrations of air-equilibrated water were calculated as 1.63 × 10−7 ccSTP/g from the equations given by Kipfer et al. [2002] using salinity of the drip waters, altitudes and air temperatures of the sampling sites and water vapor pressure (quoted from meteorological data of Jakarta [National Astronomical Observatory of Japan, 1998]). The Ne concentrations of the drip waters in both caves are slightly higher than the calculated equilibrium concentration, which indicates that excess air was dissolving in the drip waters; the excess air contributions are estimated as 5% in Dcip03 and 2% in Dciw01. Here, the excess air contribution of the Dciw01 sample is somewhat lower than that of Dcip03. Moreover, the Ne concentration of Dciw01 (1.66 × 10−7 ccSTP/g) is close to the concentration of air-equilibrated water (1.63 × 10−7 ccSTP/g). As explained in description of the sampling method, the Dciw01 sample was exposed to cave air at the moment of sampling. Therefore, it was possible that it was affected somewhat by re-equilibration in the cave.

[10] The 4He concentration of air-equilibrated water is calculable as 4.01 × 10−8 ccSTP/g, as in the case of Ne using the described equations by Kipfer et al. [2002]. The other possible origin of 4He is the radiogenic 4He; the content is obtainable from analytical results of 4He of the drip water samples and the fore-going calculated 4He concentration of air-equilibrated water using the equation given by Schlosser et al. [1989]. It is estimated that 7.51 × 10−10 ccSTP/g and 4.90 × 10−10 ccSTP/g of radiogenic 4He are included respectively at the maximum in Dcip03 and Dciw01. These are less than 2% of the total 4He in the drip waters.

[11] Schlosser et al. [1989] demonstrated that tritiogenic 3He concentration of groundwater is given as

equation image

where 3Hetri is the tritiogenic 3He, 4Hetot is the measured 4He, 4Herad is the radiogenic 4He, 4Heeq is the 4He of air-equilibrated water, Rtot is the measured 3He/4He ratio, Ratm is the atmospheric 3He/4He ratio (1.38 × 10−6), Rrad is the mean radiogenic 3He/4He ratio of 2 × 10−6 [Mamyrin and Tolstikhin, 1984], and α is an air fractionation factor (0.984 at 25°C) [Benson and Krause, 1980]. Using the equation, the tritiogenic 3He in Dcip03 and Dciw01 are derived respectively as 5.37 ± 2.09 (× 10−15 ccSTP/g) and 3.36 ± 1.91 (× 10−15 ccSTP/g).

[12] Tritiogenic 3He content with tritium(3H) concentration enables calculation of the groundwater age [Schlosser et al., 1988]; the 3H/3He age of groundwater is calculable as

equation image

where τ is the 3H/3He age in years, T1/2 is the half-life of tritium (12.33 years [Lucas and Unterweger, 2000]), [3He] is the tritiogenic 3He content (ccSTP/g), [3H] is the tritium concentration in TU, and k is the coefficient used for converting 3He from units of ccSTP/g to TU (4.01 × 1014). Assigning the derived tritiogenic 3He contents and the measured tritium concentrations of the drip water samples to equation (2), the 3H/3He ages of Dcip03 and Dciw01 are obtained respectively as 12.9 ± 3.8 years and 13.5 ± 6.3 years. Here, the errors are given as 1σ. It is anticipated that the estimate of the value of the Dciw01 age is rather old because the 3H/3He age will be overestimated in cases where re-equilibration of He between the drip water and cave air is expected to occur. The reason is that re-equilibration of He might occur during the Dciw01 sampling in the same way as that of Ne might occur. For reliable derivation of the 3H/3He age, it is advisable to take drip water samples using the same method as that used for the Dcip03 sampling.

[13] The drip water travel times might be derived from the 3H/3He ages, but we were unable to verify these obtained travel times. Accordingly, we compared the initial tritium concentrations imprinted in the drip waters (12 to 13 years before 2006), which are calculable from the tritium concentrations at the time of sample collections, using the tritium-decay equation with past tritium concentrations of precipitations in Jakarta [International Atomic Energy Agency (IAEA), 2004]. The calculated initial tritium values fall between the past tritium concentrations of precipitation at Jakarta during 1993 to1994 (2.1 to 4.8 TU without one anomalous datum of 11.7 TU [IAEA, 2004]) (Figure 3), which suggests that the derived travel times are not greatly erroneous.

Figure 3.

Secular changes in tritium concentration of rain waters at Jakarta [IAEA, 2004] and plots of calculated initial tritium values of source rains of the drip waters (shaded asterisks). The initial tritium values were obtained from the present tritium concentrations and the 3H/3He ages of the drip waters using the tritium decay equation. The decay curves of the source tritiums are illustrated by the broken lines. The open asterisk on the Ciawitali decay curve shows an initial tritium value estimated on the assumption that 10% of the dissolved rare gas in the drip water would be lost by degassing during the sampling.

[14] We can also discuss the infiltration velocity of water before dripping on the cave ceilings. The mean infiltration velocities of waters above the drip sites of Dcip03 and Dciw01 were estimated using the derived travel times and thicknesses of strata overlying the caves, which mainly comprise limestone mass. They are, respectively, 1.31 × 10−7 m/s and 0.71 × 10−7 m/s for the Cipicung and Ciawitali caves. For this studied area, the permeability coefficient of the strata overlying the caves can be simplified as the foregoing mean infiltration velocities of waters because the hydraulic gradient of the water is almost unity, on the assumption of Darcy's Law. The expected permeability coefficients here fall within the reported permeability coefficients of limestone (10−9 to 10−5 m/s [Marsily, 1986]; 10−7 to 10−6 m/s [Kayane, 1980]). Furthermore, we made sketchy estimates based on precipitation, along with the porosity and thickness of limestone strata overlying the caves. Assuming that half of the precipitation, deducting the amount of evaporation (about 1500 mm/year [Kondoh, 1995]) from the precipitation (ca. 3000 mm/year (Watanabe et al., submitted manuscript, 2008)), is half recharged and that the limestone porosity is about 20% [Kodai, 1984], the linear vertical velocity is obtained as 3.75 m/year. Because the thicknesses of overlying strata of the caves are tens of meters (30 to 50 m), as described above, the recharged rain water can be expected to arrive on the cave ceiling 8 to 13 years after its precipitation. The derived drip water travel times from the 3H/3He dating method are close to the range of the expected arrival times. Results of these simple examinations provide further evidence that the derived travel times of the limestone cave drip waters using the 3H/3He dating method are reasonable.


[15] We are indebted to Tatsushi Ono and Wataru Ninomiya for their assistance in collecting samples. We thank workers of the Geo-Science Laboratory, Japan, and the Stable Isotope Laboratory of IGNS, New Zealand, for their technical assistance. This research was supported by the Kyoto University Active Geosphere Investigations for the 21st Century COE Program. Anonymous referees provided useful comments that led to considerable improvements in this manuscript.