An Estimate of Local Bomb-Produced 36 Cl Fallout Using the Depth Profile of Groundwater in the Tsukuba Upland , Central Japan

The depth profile of Cl/Cl ratio in groundwater was investigated in the Tsukuba Upland of central Japan. The obtained results clearly show the influence of bomb-produced Cl; the highest Cl/Cl ratio is about one order of magnitude greater than the natural background ratio (1 × 10). The vertical distribution of Cl is consistent with previous observations using H and Darcy’s law. From the profile, the total bomb-produced Cl fallout in the upland is 2.3 × 10 atoms/m after the correction for surface runoff (c.f. 2.4 × 10 atoms/m at the Dye-3 site, Greenland) and a scaling factor of 0.96 was obtained (c.f. 2.5 based on the simplified latitudinal fallout distribution model). We then reconstructed the local fallout history of 36 Cl based on the Dye-3 data (scaled with a factor of 0.96 for the Tsukuba Upland) and the mean Cl flux, produced in the atmosphere from cosmic rays and measured 30 atoms m s in the upland. The ratio of the maximum bomb-peak fallout to the average natural background flux of meteoric Cl is consistent with that of measured data in Nepal. The result implies that the simplified latitudinal distribution model for Cl deposition is not easily applicable for the prediction of the bomb-produced Cl fallout pattern.


INTRODUCTION
Knowledge of groundwater residence time is crucial for the development of groundwater resources and their continuous use.Various environmental tracers in groundwater (i.e.radionuclides, stable isotopes, anthropogenic substances, etc.) are effective tools in the investigation of groundwater flow and residence time.For modern or young groundwaters (a residence time <50 60 years), 3 H (tritium) has been extensively applied because of its large pulse induced during the atmospheric nuclear testing period.However, the 3 H bomb pulse becomes increasingly difficult to detect due to its short half-life (12.32 yr); therefore, it is being replaced by such tracers as tritiogenic 3 He, CFCs, SF6 and 85 Kr (e.g.Phillips and Castro, 2003).
The application of bomb-produced 36 Cl as a groundwater tracer was initially proposed by Bentley et al. (1982), which showed the 36 Cl depth profile in groundwater.Due to its very long half-life (3.01 × 10 5 yr) and conservative nature, bomb-produced 36 Cl acts as a stable environmental tracer in modern groundwater.With the exception of several studies of the unsaturated zone (e.g.Phillips et al., 1988), few studies have focused on the tracer properties of bomb-produced 36 Cl (Balderer et al., 2004;Corcho Alvarado et al., 2005;Tosaki et al., 2007).This may be partly attributed to the scarcity of information on the local fallout history of bomb-produced 36 Cl.
According to Green et al. (2004), the 36 Cl concentration profiles for ice cores from the Dye-3 site (Greenland) (Synal et al., 1990), the Inilchek Glacier (Kyrgyzstan) and the Guiliya Ice Cap (China) have almost identical shapes.From this result, one can assume that the fallout pattern of bomb-produced 36 Cl is essentially uniform in the northern hemisphere.Accordingly, the local fallout history of bomb-produced 36 Cl can be estimated by scaling the Dye-3 fallout values (Synal et al., 1990), which is currently the only available detailed data on the 36 Cl bomb pulse.
Therefore, the present study aims to reconstruct the local fallout history of bomb-produced 36 Cl.For this purpose, the 36 Cl depth profile of groundwater was investigated in an area of the Tsukuba Upland in central Japan.The total bomb-produced 36 Cl fallout was calculated and a tentative scaling factor was derived.The result will contribute to the future application of bombproduced 36 Cl to estimate groundwater residence time.

Climate, topography and geology
The study site is located in the central part of the Tsukuba Upland (approximately 60 km northeast of Tokyo) in the Kanto region, central Japan (Figure 1).The climate of the area is humid temperate, with an annual precipitation of 1235.6 mm and an annual mean temperature of 13.5°C (average values for 1971 2000; data from the Aerological Observatory at Nagamine near the study site).
The Tsukuba Upland is a Pleistocene upland surrounded by the Sakura River and Lake Kasumigaura on the east, the Tone River on the south, and the Kokai River on the west.At the northeast of the upland, the Tsukuba Mountains (including Mt.Tsukuba, 877 m a.s.l.) extend north and south.Although the upland surface is dissected by small rivers, the elevation typically falls in the range of 20 30 m for the most part (Unozawa et al., 1988).
Geologic information on the upland and its surrounding areas has been described in detail by Unozawa et al. (1988).Basement rocks in the area consist of granitic and metamorphic rocks, forming the Tsukuba Mountains.They are overlain by the Kazusa Group sediments of Pliocene to middle Pleistocene age.The overlying Shimosa Group sediments of middle to late Pleistocene age mainly constitute the upland.It consists of six formations (i.e. the Jizodo Formation, the Yabu Formation, the Kamiizumi Formation, the Kamiiwahashi Formation, the Kioroshi Formation and the Joso Formation in ascending stratigraphic order), with the upland surface covered by the Kanto Loam Formation, which is derived from volcanic ash.
The geologic columns for the study site were previously given in Unozawa et al. (1988) for 0 62.5 m, and in Taguchi (1981) for 0 300 m (Figure 2A).Underlying the surface soil are a loam layer (1.5 2.5 m; the Kanto loam), a clay layer (2.5 9 m; the Joso clay, which is the upper most part of the Joso Formation), a sand layer (9 30 m), a fine sand layer (30 47.5 m), and a sand & gravel layer (47.5 55 m) in descending order.

Hydrogeology
The water table is generally observed within the Kanto loam or the Joso clay.The lower part of the sand layer (ca.20 30 m) and the sand & gravel layer (ca.45 55 m) act as shallow confined aquifers; the latter belongs to dominant aquifers in the area (Unozawa et al., 1988).Several sand or gravel aquifers, respectively confined by clay layers, are additionally observed at greater depth down to ~300 m (ca.80 110 m, 130 145 m and 230 245 m) (Taguchi, 1981).
According to Yasuhara et al. (1991), the Joso clay exhibits remarkably low saturated hydraulic conductivity compared to the overlying Kanto loam (by 2 4 orders of magnitude) and the upper part of the underlying sands (by 3 5 orders of magnitude).The Joso clay, therefore, may have a great influence on the downward flow of groundwater.
Conversely, previous studies have revealed that the hydraulic head of the groundwater decreases with depth (Kayane and Li, 1983;Shimada et al., 1990), to ~55 m (Yasuhara et al., 1990).This observation suggests that the upland surface is essentially acting as a recharge area.Therefore, vertical groundwater movement especially in the central part of the upland can be expected at least for the upper 55 m.Since small rivers dissecting the upland surface are gaining streams for the most part (Unozawa et al., 1988), the shallow aquifers can be assumed to be recharged only from the surface.

Groundwater sampling
The observation wells of the Geological Survey of Japan (GSJ), the National Institute of Advanced Industrial Science and Technology (AIST) were used for depth-profile sampling of groundwater.The wells are located on the premises of the GSJ in the central part of the Tsukuba Upland (Figure 1).The six wells used in this study include three deep wells for continuous groundwater level monitoring (Taguchi, 1981).
Groundwater samples for 36 Cl analysis were obtained using a bailer sampler in February 2004.Additionally, samples for 3 H analysis were collected in 1998, 1999 and December 2003.Each sample mostly corresponds to the upper four layers below the Kanto loam (i.e. the Joso -10 -  clay, the sands, the fine sands, and the sands & gravels) and the two deep aquifers (Figure 2).Since the two deeper wells (Nos.5 and 6) have multiple screens, the corresponding samples were collected around the top screens.Due to the spatial variability of the layer's thickness, well No. 1 is actually screened in the uppermost part of the sands (just beneath the Joso clay).However, when the downward groundwater flow is dominant, it may be reasonable to assume that the sample taken from the well corresponds to the Joso clay.

Analyses of samples
The analyses for chloride (Cl ) concentration, 36 Cl/Cl ratio and 3 H concentration were carried out in the laboratory.The samples for Cl and 36 Cl analyses were filtered through a 0.20 m membrane filter (DISMIC-25cs, Advantec).The Cl concentrations of aliquots were measured by ion chromatography analysis (QIC Analyzer, Dionex).The 3 H concentrations of electrolytically enriched samples were determined by liquid scintillation counting (Tri-Carb 3100TR or 2250CA, Packard).
Water samples containing more than 1 mg of Cl were prepared for 36 Cl measurement by AMS (Accelerator Mass Spectrometry).The sample preparation scheme adopted in this study is essentially the same as that described in a previous paper (Tosaki et al., 2007).Each sample was heated on a hot plate for evaporative concentration to obtain 10 20 mL volume.The sample was then acidified with 1 mL of 13 M HNO3, and AgCl was directly precipitated by adding excess AgNO3.The AgCl precipitate was dissolved once in 3 M NH4OH, and saturated Ba(NO3)2 solution was added to the solution.This was allowed to stand overnight and sulfate (SO 2 4 ) was precipitated as BaSO4 .The sample was then filtered through a 0.20 m membrane filter, and was purified by repeating re-precipitation of AgCl with HNO3 and redissolution in NH4OH.After being washed with 0.01 M HNO3 three times and with 99.5% C2 H5 OH twice, the AgCl precipitate was dried in the oven at 130°C for 3 hours.
The 36 Cl/Cl ratios were measured by the Tsukuba AMS system at the Tandem Accelerator Complex, University of Tsukuba (Nagashima et al., 2000;Sasa et al., 2007).The 36 S interference was corrected by using a blank sample prepared from NaCl reagent; typical correction rate was 10% for samples with a 36 Cl/Cl ratio of ~1 × 10 12 .Relative ratio of 36 Cl/ 35 Cl (counts/ C) for a sample is then normalized to that for a diluted NIST 36 Cl standard (with a 36 Cl/Cl ratio of 4.47 × 10 11 ).The obtained 36 Cl/Cl ratio of the sample includes statistical error derived from uncertainties (1 ) of the sample, the standard and the blank.Details of the measurement and the 36 Cl/Cl calculations have been given in Nagashima et al. (2000Nagashima et al. ( , 2004)).The overall precision of the system is better than 5%, and the background level of 36 Cl/Cl measurement is below 2 × 10 14 .

Depth profile of 36 Cl in groundwater
Table I summarizes the measured 36 Cl and 3 H data. Figure 2B illustrates the depth profile of 36 Cl/Cl ratio in the groundwater.The obtained results clearly show the influence of bomb-produced 36 Cl in the upper ~50 m; the highest 36 Cl/Cl ratio is about one order of magnitude greater than the natural background level, mainly due to cosmogenic production in the atmosphere.This vertical distribution agrees with the depth profiles of 3 H concentrations (Figure 2C).It is also consistent with the previous observation in the central part of the upland by Kayane and Li (1983), who reported no detectable 3 H in the groundwater at depths greater than 36 m in 1982.In contrast, Yasuhara et al. (1990) calculated the downward groundwater flux through the Joso clay as 27 mm/yr based on Darcy's law; it leads to a mean residence time of 30 50 yr in the upper part of the underlying sand layer.The highest 36 Cl/Cl ratio observed in the sand layer is also consistent with their estimation.
As can be seen from Figure 2C, 3 H concentrations have gradually decreased from 1999 to 2003 due to both radioactive decay and groundwater flow, while 36 Cl/Cl ratio showed distinct difference from the natural background level (by one order of magnitude) in 2004 (Figure 2B).This suggests the potential usefulness of bomb-produced 36 Cl as a tracer in modern groundwater as described in previous studies (Bentley et al., 1982;Balderer et al., 2004;Tosaki et al., 2007).
It can be surmised that the deeper the sample, the greater the uncertainty about the corresponding recharge area.Thus, the deepest sample (No. 6; 233.2 244.6 m) is excluded from further discussion.The natural background 36 Cl/Cl ratio is then estimated as 1 × 10 13 from the result of sample No. 5 (Figure 2B).This is supported by the 36 Cl/Cl ratio for a sample from a shallow well screened in the Kanto loam (0.7 2.0 m): (1.19 ± 0.48) × 10 13 in October 2003; the well is located in the Terrestrial Environment Research Center, University of Tsukuba, which is about 6 km northwest from the sampling site at the GSJ.

Estimation of bomb-produced 36 Cl fallout
The total bomb-produced 36 Cl fallout at the Tsukuba Upland can be estimated by integrating bomb-produced 36 Cl observed in the profile (grayed area in Figure 2B).The following equations were used to calculate the total bomb-produced 36 Cl fallout: -11 - where D is the total bomb-produced 36 Cl fallout (atoms/ m 2 ), Abomb, i is the concentration of bomb-produced 36 Cl (atoms/L), hi is the thickness of the layer (m), ni is the porosity (%), Ri is the measured 36 Cl/Cl ratio, RBG is the natural background 36 Cl/Cl ratio, and [Cl]i is the Cl concentration (mg/L).
In the calculation, we assumed that obtained values are representative for the total thickness of the corresponding four layers below the Kanto loam, and that the porosity is identical to the effective porosity (e.g.Todd and Mays, 2005).The porosity used for the Joso clay was 62%, which is the arithmetic mean of the 19 measured values over the Tsukuba Upland by Yasuhara et al. (1991).For other layers, representative values were adopted from Morris and Johnson (1967): 39% for sand (as medium sand), 43% for fine sand, and 34% for sand & gravel (as fine gravel).The natural background 36 Cl/Cl ratio was set to 1 × 10 13 , as mentioned in the previous section.
Consequently, Equation (1) gives the total bombproduced 36 Cl fallout of 2.1 × 10 12 atoms/m 2 .Because the 36 Cl deposited onto the ground may be partly lost with surface waters, 2.1 × 10 12 atoms/m 2 would be the minimal estimate.General surface runoff rate in Pleistocene upland-alluvial lowland region is ~20% (e.g.Kotoda, 1968;Kondoh, 1985), whereas surface runoff rate measured in an experimental field at the central part of the Upland is ~5% (Itadera and Shimada, 1992).Here we assumed 10% of 36 Cl is lost by surface runoff.After incorporation of this loss, we obtained a total bomb-produced 36 Cl fallout of 2.3 × 10 12 atoms/m 2 .Dividing this value by 2.4 × 10 12 atoms/m 2 (fallout at the Dye-3 site; Synal et al., 1990), we obtained a scaling factor of 0.96 for the Tsukuba Upland (c.f.2.5 based on the simplified latitudinal fallout distribution model by Lal and Peters, 1967).It should be noted that the screen length of well No. 2 is wide (11 m) and the thickness of the corresponding sand layer is 21 m (Figure 2A).This indicates that the highest 36 Cl/Cl ratio in the profile is fairly representative of the layer.Therefore, the calculation described in this section will not lead to erratic results.

Reconstruction of the local fallout history of 36 Cl
The 36 Cl flux for 1560 1920 AD at the Dye-3 site is 20 ± 6 atoms m 2 s 1 (Synal et al., 1994), which is taken to be the natural background flux of meteoric 36 Cl.Subtracting this natural background flux from annual values for 1945 1985 (Synal et al., 1990), we obtained yearly values of bomb-produced 36 Cl fallout at the Dye-3 site.These values were then scaled using the factor of 0.96 for the Tsukuba Upland.
From the 36 Cl measurements of bulk precipitation samples collected monthly at the Natural Sciences building, University of Tsukuba, the average 36 Cl flux is ~30 atoms m 2 s 1 (from April 2004 to March 2006; un-published data).By adding this natural background flux, we obtained the local fallout history of 36 Cl (including both bomb-produced and cosmogenic components) for the Tsukuba Upland.
In order to allow comparison of the reconstructed fallout history with several measured data for ice cores at different locations, the ratios of the bomb-peak flux to the natural background flux of meteoric 36 Cl were calculated.It is worth noting that the value determined with the scaling factor of 0.96 reasonably agrees with data from Nepal (Table II).This may imply that the latitudinal fallout distribution (Lal and Peters, 1967) is not easily applicable for the prediction of the bombproduced 36 Cl fallout pattern.The obtained results also suggest that the local fallout history of 36 Cl can be reconstructed from groundwater in regions of near vertical recharge.Such studies, in turn, will enable the application of bomb-produced 36 Cl to date modern groundwaters.

CONCLUSIONS AND OUTLOOK
The present study investigated the 36 Cl depth profile in the groundwater of the Tsukuba Upland, central Japan.The total bomb-produced 36 Cl fallout in the upland was derived as 2.3 × 10 12 atoms/m 2 using the profile.The local fallout history of 36 Cl was estimated using the Dye-3 data with a tentative scaling factor of 0.96 (c.f.2.5 based on the latitudinal fallout distribution model).The ratio of the maximum bomb-peak fallout to the average natural background flux of meteoric 36 Cl is consistent with that of measured data in Nepal.The result implies that the simplified latitudinal distribution model for 36 Cl deposition is not easily applicable to predict the fallout pattern of bomb-produced 36 Cl.
Prior to the application of bomb-produced 36 Cl to estimate groundwater residence time, reconstruction of the local fallout history is necessary.Further study, that is, more data points in the depth profile, is needed to obtain a reliable scaling factor.However, the results of this study will contribute to the future application of bomb-produced 36 Cl to date modern groundwaters.(1990,1994)

Figure 1 .
Figure 1.Location of the sampling site in the Tsukuba Upland (after Unozawa et al., 1988).

Figure 2 .
Figure 2. (A) Geologic column for the sampling site in the Tsukuba Upland (after Taguchi, 1981; Unozawa et al., 1988).(B) Depth profile of 36 Cl/Cl ratio in the groundwater.Vertical error bars indicate widths of each screen.(C) Depth profiles of tritium concentration.

Table I .
Measured Cl concentrations, 36 Cl/Cl ratios, 36 Cl concentrations and 3 H concentrations in the groundwater 1 TU (tritium unit) means one 3 H atom in 10 18 1 H atoms. a Also screened in a deeper aquifer (136.5 140.3 m).b Also screened in deeper layers (255.6 259.4 m and 264.9 272.5 m).

Table II .
Ratios of the maximum 36 Cl flux to the average natural background flux for the36Cl bomb pulse at different locations