Introduction

Verification of the Comprehensive Nuclear-Test-Ban Treaty (CTBT) includes the prospect of an On-Site Inspection (OSI) subsequent to a suspected nuclear weapon test. The purpose of an OSI is to provide evidence for verifying whether an event was nuclear in nature, and one of the primary ways to make that determination is the analysis of radioactive noble gases. These noble gases, particularly the xenon isotopes 131mXe, 133mXe, 133Xe, and 135Xe are of particular interest as they may represent fission products resulting from a nuclear test. As they are chemically inert, short-lived, and low in natural background abundance, these noble gases are the most likely radionuclides to be observable at an underground testing location, even when the test is otherwise well contained [1].

In addition to production in nuclear explosions, radioxenon is also released from anthropogenic sources, including nuclear power plants, research reactors, and medical isotope production facilities [2–4]. Distinguishing between these various sources within atmospheric samples is a critical CTBT verification task. In order to distinguish a nuclear weapon signature from other anthropogenic sources a comparison of xenon isotopic ratios is often used [5, 6].

Measurements of radioxenon during an OSI are heavily dependent upon the underground transport mechanisms [1, 6–8]. One of the processes which govern subsurface gas transport is barometric pumping. In this process changes in barometric pressure act to impress or withdraw gas from the subsoil pores and fractures, with it generally acting to increase the rate of transport of a gas with a high subsurface concentration toward the surface [9, 10]. While primarily acting as an upward transport mechanism for signatures of underground nuclear explosions, barometric pumping is also predicted to have an opposite effect within the shallow subsurface, which is a potential sampling region during an OSI. During a period of increasing atmospheric pressure gas is pushed downward along fractures and into the pore spaces of the walls. Some of this atmospheric gas will then be retained within the pores [11]. Simulations have shown that during periods of increasing barometric pressure this mechanism should imprint atmospheric radioxenon into the shallow subsurface, potentially in detectable levels [7, 8, 10]. Simulations also show that over time this underground radioxenon can build up a measurable background concentration that is detectable irrespective of the presence of a plume. While the barometric pumping process will serve to remove some of this radioxenon back into the atmosphere much will remain until its eventual removal by radioactive decay. This poses the risk of contaminating the soil-gas environment, which could complicate the interpretation of radioxenon detections in an OSI.

In an effort to verify theoretical predictions of atmospheric radioxenon imprinting, a sampling campaign has been proposed in the vicinity of Chalk River Laboratories (CRL). At this location, on the Ottawa River approximately 160 km northwest of Ottawa, ON, Atomic Energy Canada Limited’s (AECL) National Research Universal (NRU) reactor irradiates highly enriched uranium (HEU) targets in order to produce 99Mo for medical applications of 99mTc, the daughter product of 99Mo. The 99Mo is also extracted on site through chemical dissolution of the HEU targets, thus releasing the radioactive xenon. This facility is the largest producer of 99Mo in the world and it has been estimated that the facility releases up to 1013 Bq/day of 133Xe [4]. It should be noted that this release quantity is below regulatory limits, but still great enough to complicate radioxenon monitoring for treaty verification.

In the vicinity of CRL, the river winds through a valley, with some significant hills to the east of the river (elevation changes of 200–300 m above the valley floor). To the west, the elevation changes more gradually, with the ground rising only approximately 50 m above the valley floor over a horizontal distance on the order of tens of kilometers.

Figure 1 shows an example estimate of imprinting for the Deep River site over the month of September 2013. This scenario was estimated using the underground gas transport code UTEX [6, 12] with geologic parameters chosen to match the generally sandy soil type around Deep River, Canada (high porosity and high permeability). Atmospheric xenon concentrations were obtained from surface level NaI detector measurements provided by Health Canada as part of the Fixed Point Surveillance Network [13, 14]. These preliminary calculations indicate that imprinting of 133Xe in the shallow subsurface may be on the order of tens of Bq/m3, and even at greater depths may still be within the detection limits of current systems. In addition, Fig. 1 indicates that the radioxenon may develop a background concentration within the soil gas that will linger for some time after a plume passes overhead.

Fig. 1
figure 1

Estimated imprinting of 133Xe in Deep River Canada. While the average soil concentration quickly falls away as the depth increases, the concentrations still fall within the detection limits of current radioxenon detection systems

Full size image

Site factors

In order to support this sampling campaign an analysis of the proposed sampling locations was performed in order to evaluate the site locations and to determine what conditions will be necessary for a detection of imprinted radioxenon. A variety of factors contribute to the decision about whether a site is suitable for use as a sampling location. The factors considered here are listed below:

  • Proximity—proximity to the source is a major factor when considering a site’s suitability. As the plume moves away from the site, the radioxenon becomes increasingly diluted and the probability of detection will likewise decrease.

  • Meteorological conditions—another major constraint on site location are the predicted weather conditions. Ideally, the site should be located such that the wind frequently blows from CRL to the sampling site. In addition, as barometric pressure has a direct effect on the imprinting mechanism; ideally the site would also be located such that favorable winds also correspond to periods of increasing pressure. It is worth noting, however, that Figs. 1 and 2 show that local meteorology may indicate that this correlation is not strictly required, especially with large releases.

    Fig. 2
    figure 2

    Frequency of increasing (a) and decreasing (b) 6-h pressure changes for cardinal wind directions at the Petawawa, ON airport over the month of September, 2011

    Full size image

  • Geology—site geology is another factor, which must be considered. Matrix porosity and fracture characteristics have been shown to have a great deal of influence upon subsurface gas transport [7, 8]. In addition, for ease of sampling, sites should have a geology conducive to the creation of a sampling hole and be above the water table.

  • Location—the geography of the region must be considered, both for its potential to affect atmospheric transport of the source plume, as well as its effect on site access. An additional factor that must be considered is the presence of water at or near the sampling site. There are a large number of lakes in the region and the Ottawa River also flows directly adjacent to CRL. Sampling sites must be far enough from any body of water and far enough above the water table that the risk of water intrusion into the sampling hole is minimized.

Experimental

The goal of this experiment will be to verify the presence of imprinted radioxenon and to correlate it with the local atmospheric concentrations and conditions. In order to accomplish this, two samples will be collected during each sampling period, an atmospheric sample and a subsurface sample. The subsurface sample will be obtained using currently accepted OSI procedures: a 2-cm hole will be drilled to a depth of 1 m or greater, tubing will be emplaced with a screen at the end, and the hole will be backfilled to 0.5-m with local soil while the remainder will be filled with a bentonite clay mixture. The soil gas will then be pumped into a sample collection bag. An atmospheric sample will be collected at the same time with the same sampling rate so that a correlation can be made. Local meteorological measurements will also be collected.

In order to better understand how the local meteorological conditions affect the source behavior, atmospheric transport modeling (ATM) was performed locally around the site. This was done using the AERMOD View software produced by Lakes Environmental Software, which itself uses the AERMOD model produced by the U.S. Environmental Protection Agency [15]. AERMOD is a steady state plume model used primarily for local (<35 km) transport modeling. The horizontal distribution of the plume is assumed to be Gaussian in both the stable boundary layer (SBL) and the convective boundary layer (CBL). In the SBL, the vertical distribution of the plume is also assumed to be Gaussian, but in the CBL the vertical distribution is described by a bi-Gaussian probability density function (pdf) [15]. ATM has been performed for CRL before, however those calculations have been on the regional and global scale rather than local [16, 17].

In order to better characterize potential site locations, the historical meteorological data was examined. Integrated Surface Hourly (ISH) Data from digital data set DSI-3505 which is archived at the National Climatic Data Center (NCDC), was acquired for Petawawa, ON, a location approximately 20 km southeast of Chalk River Laboratories [18]. This location was used due to the presence of an airport and corresponding weather station (AWS 71625). This data was used as the surface conditions input for AERMOD, while upper atmosphere data, which is also required by the model, was retrieved from the ESRL Radiosonde Database [19]. Ground surface characteristics such as the albedo, Bowen ratio, and surface roughness are also necessary for the operation of AERMOD, and were chosen to be most representative of the site conditions.

In addition to being used as inputs for the AERMOD code, the surface conditions were also examined to determine their potential impact on imprinting of atmospheric radioxenon into subsoil air. The imprinting process is primarily dependent on variations in atmospheric pressure, while the presence of radioxenon released from CRL is dependent on wind direction. Figure 2 shows both the wind direction and the corresponding atmospheric pressure change over 6 h periods in Petawawa, ON.

Results and discussion

Preliminary modeling indicates sufficient plume rise such that the smaller elevation changes have no noticeable effects, however some terrain effects are apparent in the cases where the plume interacts with the larger eastern hills. An overhead view of the area, with sites of interest labeled, can be seen in Fig. 3.

Fig. 3
figure 3

An overhead view of the proposed sampling locations. The white pins are proposed sampling locations, the red pin is the source location, and the blue pin is the location of the weather station used in this paper. Overlaid on top of the map is a wind rose showing the wind speed and direction distribution over the year 2013 (wind speed legend shown in Fig. 4). Note that the primary wind directions are along the river valley. (Color figure online)

Full size image

An examination of historical meteorological data indicates that the predominant wind direction in the region of Deep River follows the Ottawa River valley. Figure 4 shows the distribution of wind directions over a year’s measurement period, and when overlaid onto a site map, as shown in Fig. 3, the prevailing wind directions along the river are confirmed. In order to maximize the likelihood of detection of a plume released from CRL, the sampling location should be within the river valley.

Fig. 4
figure 4

A wind rose plot of wind speed and direction at the Petawawa, ON airport for the year 2013

Full size image

Each proposed sampling site was evaluated using the criteria previously listed. The first site, in close proximity to CRL, lies on a gently sloping region near the river. It is near an acid rain measurement location, and so will be referred to as the acid rain site. The local geology is a porous, sandy material that will be relatively easy to drill a hole into. This porous media also indicates that any barometric pumping will have a maximal effect on the sub-surface radioxenon levels. The site is elevated enough above the river that the likelihood of water interfering in the site is low.

The Deep River site is approximately 10 km NE of CRL, at the edge of the city of Deep River, ON. The location is somewhat inland from the river and easily accessible by road. The geology is similar to that of the acid rain site, and it lies along a similar wind trajectory as that site as well. However, due to its distance from CRL, the range of ideal winds is much smaller than the acid rain site. From ATM, wind directions between 100° and 140° has been shown to be ideal for measurement at Deep River. One concern is that this wind range most often corresponds to periods of decreasing pressure, the opposite of what is needed to cause imprinting. While this is a concern, the other elements of the location are satisfactory, and the preliminary calculations shown in Fig. 1 still indicate that substantial imprinting can be expected to occur.

A representative output from ATM calculations is shown in Fig. 5. This is the hourly maximum concentrations of 133Xe corresponding to a release of 1.38 × 108 Bq/s from CRL on 19 September, 2011. The wind direction at this point is 120 degrees, which falls right into the limits defined for the Deep River site. While the surface concentrations at Deep River can be verified experimentally [13], those at the closer CRL site must be determined using dispersion modeling. By comparing the concentration outputs from AERMOD, the dilution factor at the Deep River location is approximately ten times that of the CRL site when the meteorological conditions fall within the previously outlined parameters. Considering the large imprinting estimates shown in Fig. 1 for Deep River, it is to be expected that the concentration of radioxenon in subsoil gas will be significant at the site within close proximity of CRL.

Fig. 5
figure 5

A representative output from the AERMOD atmospheric transport code. This plot represents the 1-h maximum surface concentrations of 133Xe for a release of 1.38 × 108 Bq/s from CRL, indicated by the red circle, on 19 September, 2011. The dilution of 133Xe between the acid rain site, and the Deep River site (both indicated by white dots) is apparent, as is the plume dependence on wind direction. (Color figure online)

Full size image

Conclusions

Examination of the proposed sampling locations for an imprinting experiment near Chalk River Laboratories found that both proposed sites were acceptable for the needs of the experiment. The close proximity of the acid rain site to CRL makes it ideal for the intended sampling of imprinted radioxenon. While meteorological conditions were not ideal, the combination of site proximity, favorable geology and geography, and the high likelihood of a detection, even in non-ideal conditions, makes the proposed Deep River site also acceptable for use. This proposed sampling campaign will allow us to benchmark current underground xenon transport codes, a crucial step in understanding the potential backgrounds at potential OSI locations. In addition, this campaign will provide experience using OSI techniques and equipment and an opportunity to suggest future improvements.