Development of an autonomous station for measurements of artificial gamma activity in surface water bodies

Development of an autonomous station for measurements of artificial gamma activity in surface water bodies

Journal of Environmental Radioactivity 204 (2019) 42–48 Contents lists available at ScienceDirect Journal of Environmental Radioactivity journal hom...

1MB Sizes 0 Downloads 2 Views

Journal of Environmental Radioactivity 204 (2019) 42–48

Contents lists available at ScienceDirect

Journal of Environmental Radioactivity journal homepage: www.elsevier.com/locate/jenvrad

Development of an autonomous station for measurements of artificial gamma activity in surface water bodies

T

Michal Fejgl∗, Miroslav Hýža National Radiation Protection Institute, Bartoškova 28, Prague 4, 140 00, Czech Republic

ARTICLE INFO

ABSTRACT

Keywords: Gamma spectrometry Emergency monitoring 137 Cs Emergency Preparedness Fresh water monitoring

This paper reports on the structure of the autonomous station for monitoring artificial gamma activity in surface water bodies for the purposes of emergency preparedness of the Czech Republic. A simple design based on the NaI(Tl) submersible detector powered by a combined solar and wind source has been employed. Data transfer is provided by a satellite connection. The detection capabilities of the device have been tested for various unfavourable conditions, and the detection limits have been lowered by using the noise adjustment singular value decomposition (NASVD) method. The detection capabilities of the device fulfil the legal requirements for emergency monitoring, and are almost equal to the detection capabilities of other available devices with a more complicated and less versatile structure.

1. Introduction In the framework of emergency preparedness for the Czech Republic (CR), there is a need to introduce continuous gamma activity monitoring in surface water bodies. There are two reasons for this. In the event of a nuclear power plant (NPP) accident, continuous monitoring would be beneficial for radioactive plume deposition localization. In addition, it would improve the emergency monitoring system for surface water bodies. A survey of the literature on available water monitoring systems showed that none of the accessible systems complied well with the requirements for the emergency preparedness system of the CR. It was therefore decided to develop an original monitoring system that would perfectly suit the needs of emergency preparedness of the CR. 1.1. Motivation – radioactive plume localization The most important emergency preparedness concern of the governmental Radiation Monitoring Network (RMN) of the CR is an NPP accident with airborne primary inventory release. Mathematical models fitted by the source term of the accident and by the current meteorological situation are proposed as a leading tool for providing an assessment of the way in which the unintended reactor leak has been disseminated, and for rapidly determining the exposure of members of the public. Previous experience from the Chernobyl and Fukushima accidents



has demonstrated that complementary information describing the rate of ground deposition of contaminants would be very helpful for verifying the expectations of the models. Within the territory of the CR, some of this complementary information can be supplied by two monitoring networks, which are operated as a part of the RMN of the CR. The first monitoring network consists of 23 aerosol sampling stations placed across the territory of the CR. In the event of an NPP accident, aerosol filters would be changed at intervals of 6 h, and would be measured by high-resolution gamma ray spectrometry (SÚJB, 2016). This relatively long sampling period is a feature that reduces the benefit of the results. The second monitoring network consists of gamma detectors that continuously measure the dose rate (SÚRO, 2019). This tool can be very helpful, because it makes continuous measurements and there is a relatively dense network (180 detectors across the CR). However, experience of the Chernobyl and Fukushima accidents revealed that such monitoring networks are prone to become inoperative in the conditions of an NPP accident. In real accident situations, necessary complementary information has been obtained by making once-off field measurements, which are in principle sparse. The emergency preparedness system in some European countries is also able to exploit the results provided by devices that continuously or semi-continuously measure the gamma activity in river water. In the event of an NPP accident, results of this kind of monitoring can supply exceptionally valuable complementary information, due to the fact that the composition of river water reflects the contamination deposition

Corresponding author. E-mail addresses: [email protected] (M. Fejgl), [email protected] (M. Hýža).

https://doi.org/10.1016/j.jenvrad.2019.04.001 Received 20 November 2018; Received in revised form 29 March 2019; Accepted 1 April 2019 0265-931X/ © 2019 Elsevier Ltd. All rights reserved.

Journal of Environmental Radioactivity 204 (2019) 42–48

M. Fejgl and M. Hýža

within the whole drainage basin. This means that even a monitoring network consisting only of a limited number of water monitoring stations working in continuous regime can help to localize the contamination deposition area. A crucial condition is that the devices are not damaged as a consequence of the NPP accident, and that they continue in full operation during all phases of the accident. Implementing such a system into the RMN would reinforce the governmental emergency preparedness system for the CR.

1.4. Research intention – development of a monitoring system Project VI20172020083 of the Czech Ministry of the Interior is aimed at developing a system for monitoring the radioactive contamination of surface water. In the first stage of the project, between 2017 and 2019, the plan is to develop a Station for Artificial Gamma Activity Measurement (SAGMA), a stand-alone device working in an automatic regime. In the second stage of the project, between 2019 and 2020, a network will be built consisting of several SAGMAs, which will be placed in spots that are important from the point of view of water vulnerability from radioactive contamination. This network will be named System for Continuous Gamma Activity Monitoring (SCOMO). This system will be incorporated into the governmental Radiation Monitoring Network (RMN) of the CR. It was expected that SCOMO would be employed mainly in emergency monitoring. In the event of an NPP accident, the SAGMA detection capabilities should be sufficient to provide complementary information for evaluating the performance of the model and the contamination of river water sources. Sections 1.3. and 1.4. indicate that 137Cs is the radionuclide that is most frequently observed as a source of radioactive contamination of the water environment. It was therefore defined as an indicator of the SAGMA measuring capabilities. When searching for the most reasonable level of required detection limits for SAGMA, we selected Regulation 360/16 Monitoring of the radiation situation (SÚJB, 2016) as the best available indicator. The required Minimal Detectable Activity Concentration (MDAC) for 137Cs is 0.1 Bq L−1 for routine monitoring of various water samples within all monitoring networks in the CR. Exceptional requirements come into force during emergency monitoring, when MDCA of 5 Bq L−1 is required for local monitoring networks (in the proximity of a nuclear facility), or 10 Bq L−1 for border monitoring networks (for monitoring rivers leaving the territory of the CR), see Table 1. The decision was taken to strive for SAGMA detection capabilities that would enable the strictest MDAC level for 137Cs emergency monitoring to be achieved, i.e. 5 Bq L−1, under reasonable counting conditions.

1.2. Motivation - monitoring surface water radioactive contamination in the CR Surface water is an environment that is highly vulnerable to deposition of radioactive contamination from the atmospheric plume, and to other sources of radioactive contamination. As of 2016, surface water bodies covered 51.7% of drinking water sources in the CR (Fousová and Reidinger, 2017). Protection of surface water against radioactive contamination is therefore a task of high priority in the CR. In the CR, monitoring of the radioactivity in surface water is controlled by Regulation 360/16 Monitoring of the Radiation Situation (SÚJB, 2016). The prescribed monitoring of gamma activity in surface water in the CR is presented in Table 1. Different requirements are defined for emergency monitoring (the technical term for the monitoring regime for the fresh water environment in the CR in the event of an accident) and for routine monitoring (the technical term for the monitoring regime when no accident has occurred). Monitoring requirements are defined only for 137Cs and 131I, and only monitoring of grab samples is required. The current monitoring procedure is described in section 3. 1.3. Caesium 137 in a fresh water environment Cs-137 with a half-life of 30 years is most frequently used as a reference nuclide for radioactive pollution of the water environment. Its activity in surface water samples in Europe has fluctuated within six orders of magnitude in the course of the last half century. In 1963, as a consequence of nuclear weapon tests in the atmosphere, the ordinary activity concentration of 137Cs in rainwater was 2 Bq L−1 (Steinmann, 2015). After the Chernobyl accident, the activity concentration of 137Cs in surface waters near the site of the accident was strongly enhanced, e.g. the Pripiat River evinced 1600 Bq L−1, and even in fresh water samples from the western part of Europe the 137Cs activity concentration values were around 1 Bq L−1. At present, the 137Cs activity in surface waters in Europe is much lower. Nevertheless, a routine monitoring performed in the Czech Republic shows that residual activity of 137 Cs continues to be traceable in the surface water today. A usual 137Cs activity concentration value is about 1 mBq L−1 for most rivers and lakes in the CR. Only water samples from the river Odra, which drains the Jeseníky Mountains, is subject to a stronger impact of Chernobyl fallout. Samples from the Odra contain 137Cs in an activity concentration of about 4.5 mBq L−1 (SÚJB, 2018).

1.5. Gamma activity detection systems enabling continuous measurements in water Many sophisticated systems for determining the gamma activity in sea water have been described in the literature. The most common conception is a low-budget NaI(Tl) probe that makes continuous measurements in 4 π geometry of the gamma activity of ocean water, while the necessary electricity supply and data processing facilities are provided by a base located on board a raft or a boat floating on the surface of the water above the probe (Dulai et al., 2016; Osvath et al., 2005; Tsabaris et al., 2005; Sartini et al., 2010). An obvious difficulty that these systems have to cope with is the high background level caused by high activity of 40K in seawater. However, the background gamma activity levels are quite constant, and background subtraction is not complicated. The state-of-the-art dual HPGe–NaI(Tl) detection system for

Table 1 List of water samples required for137Cs and131I activity concentration monitoring, as required by regulation 360/16. nuclide

network

monitoring

water sample type

frequency

required MDAC

137

territorial, local local local local border border

routine routine routine emergency emergency emergency

surface, drinking liquid discharge underground surface, drinking surface, drinking surface, drinking

3 months 3 months 12 months 6h 1 week 1 week

0.1 Bq L−1 0.1 Bq L−1 0.1 Bq L−1 5 Bq L−1 10 Bq L−1 10 Bq L−1

Cs Cs 137 Cs 137 Cs 137 Cs 131 I 137

43

Journal of Environmental Radioactivity 204 (2019) 42–48

M. Fejgl and M. Hýža

seawater gamma spectrometry is described by Povinec et al. (2008). A 100 × 150 mm NaI(Tl) crystal was assembled with HPGe cooled to 120 K by a combined propane/He system. Both parts of the system are housed together with part of the spectrum acquisition and processing electronics, which are connected to the preamplifier of the HPGe detector and also to a shipboard computer and to a power supply through a 1200 m long coaxial cable, which is used for transmitting the data and the power. Data transfer is covered by a satellite line (Povinec et al., 2008). This device enables a Minimal Significant Activity Concentration (MSAC) of 0.019 Bq L−1 for 137Cs activity concentration to be achieved when 24 h integration time is used. Unlike the field of seawater gamma activity detection systems, the field of the freshwater gamma measurement systems is quite poorly covered. Continuous gamma activity monitoring systems were installed on the rivers in only a few European countries after the Chernobyl accident. Gamma activity measuring stations are operated within the French Hydrotéléray network monitoring system, operated by IRSN (IRSN, 2003). The monitoring stations employ NaI(Tl) detectors placed in 25 L canisters equipped with a lead shielding. A semi-continuous data acquisition cycle is started up when a new water sample is pumped into a canister every 2 h and each sample is measured until the next sample is pumped. These stations are placed about 20 km downstream from a nuclear power plant (NPP). Activity concentrations of 134Cs, 137Cs, 60Co and 131I are determined. A system that works in a similar way is used in Belgium within the TELERAD network, which works under FANC (the Federal Agency for Nuclear Control). NaI(Tl), LaBr3 detectors are also employed (FANC, 2005; Sonck et al., 2012; Gitzinger et al., 2013). Two monitoring stations with a similar structure but that make measurements in a continuous regime are in operation in Spain (Casanovas et al., 2013). The detection system is based on a 2 × 2 inch NaI(Tl) detector placed inside the throughflow vessel, which has a lead shielding. The Minimal Detectable Activity Concentration (MDAC) is 0.2 Bq L−1 for 137Cs, when a 60-min integration time is applied. The previously described river water monitoring systems employ a measuring canister and external shielding. This structure provides a lower detection limit, but implementation of the automated operation is hindered by the way in which the structure is conceived. The use of these kinds of detection systems is also affected by the limited portability of the devices. They can be installed only in a secured zone equipped with an electricity source. The SARA Water Measurement System working in automated regime is a really simple, inexpensive and failure-free gamma activity system operated in Switzerland by the Swiss Federal Office of Public Health. Since 2015, five automated monitoring stations have been in operation on particular sites on rivers, usually a few kilometres downstream from an NPP (Steinmann, 2015). Three-inch NaI(Tl) detectors measure the gamma activity in 4 π geometry, when the probe is covered by a water layer of at least 1-m from each direction. This water layer is an analysed sample and, at the same time, forms a shielding. The (unspecified) detection limit is about 1 Bq L−1 for 137Cs, when a 60-min integration time is used (Steinmann, 2015). Particular attention is dedicated to the subtraction of uneven background contributions of cosmic radiation as a consequence of variations in water surface level and radon progenies as a consequence of their wash-out from the air and soil after strong rains. Data is transferred using a GPRS modem, and the standard maintenance period is 6 months (Steinmann, 2015).

to the consequences of an NPP accident. An approach with a simple structure with a non-shielded submersible NaI(Tl) probe is well established in the field of ocean water monitoring (Povinec et al., 2008, Dulai et al., 2016; Osvath et al., 2005; Tsabaris et al., 2005; Sartini et al., 2010). Although the installation of this concept for river water monitoring involves some disadvantages related to fluctuations in the background level (Steinmann, 2015), this structure was chosen on the basis of its simplicity and robustness. A hybrid solar-wind power source was chosen as a power supply, and a satellite phone was chosen for data transfer. The energy and efficiency calibration of the test probe was performed in laboratory conditions. Subsequently, the detection capabilities were examined in semi-field conditions, and the spectral data were used for developing a detection algorithm that takes the natural background fluctuations into account. 2.1. Hardware The entire system consists of two parts, an underwater probe and an on-land external control unit. The submersible unit itself contains a 3 × 3 inch NaI(Tl) detector in a waterproof plastic case (SBG.D3 type, by NuviaTech Instruments). The detector's photomultiplier is connected to a compact pulse height analyser with the conversion gain adjustable from 256 to 2048 channels (NuNA MCB3 by NuviaTech Instruments). It is capable of processing up to 50 000 CPS maintaining the dead time correction error below 5%. The energy range is set to 50–1900 keV, which covers the ROI of the vast majority of important radionuclides. The energy resolution at 661 keV 137Cs is 7.5%. The total weight of the unit is approximately 10 kg. Data transmission from the multichannel analyser to the external control unit is provided via an Ethernet connection. The control unit consists of the embedded PC and a communication module, which sends data to the central server via GPRS or a satellite connection. The power supply of the entire system is modular. The user can choose between a standard connection to the power grid and/or to a solar panel and a wind turbine accompanied by a 900 Ah lithium cell that ensures 14 days of battery run. This transforms the device into a fully autonomous system. 2.2. Energy calibration A slight spectrum gain shift due to changes in water temperature is observable. This problem is solved by the automatic spectrum stabilization managed by Gamwin commercial software (GAMWIN, 2019), using the 1462 keV peak of natural 40K. Energy calibration was performed using a set of calibration standards and other natural background peaks. The quadratic equation E [keV ] = A + B × CHNL + C × CHNL2 was chosen, as a linear model is not sufficient due to the slight response nonlinearity of the probe. The resolution was calibrated by fitting the measured data to a square root model. FWHM [keV ] = A + B × E [keV ] . 2.3. Detector efficiency calibration For efficiency calibration purposes, 241Am, 131I, 134Cs, 137Cs, 60Co solutions were prepared. The measurements were made in a cubeshaped plastic tank 1 m3 in volume. Each solution was chemically stabilized (by a stable carrier and a mineral acid) to minimize the wall deposition effect. The activity of the solutions was at a level of 10 kBq m−3, which provided favourable counting statistics even for a measurement lasting 1 h. All measurements were run in sequencing mode in order to observe potential time-dependent behaviour of the detector response. This would potentially reveal deteriorated homogeneity of the solution. In addition, an independent calibration solution check was performed by regular sampling and measurements, using an HPGe detector. Homogeneity analysis proved that, after the initial mixing phase,

2. SAGMA structures A non-shielded submersible structure with a NaI(Tl) probe was chosen to be used for constructing SAGMA. The selection of the construction concept was supported by literature analyses (section 1.5.). Shielded throughflow vessel concepts (Casanovas et al., 2013; IRSN, 2003; FANC, 2005; Sonck et al., 2012; Gitzinger et al., 2013) evince low detection limits, but this complicated structure seems to be vulnerable 44

Journal of Environmental Radioactivity 204 (2019) 42–48

M. Fejgl and M. Hýža

Fig. 1. Cs - 137 calibration solution stability measured by 3 × 3 inch NaI(Tl) in a calibration tank. A linear fit does not yield a statistically significant slope coefficient. Homogeneity after the initial mixing phase can therefore be assumed.

the solutions remained stable during calibration for all utilized radionuclides. An illustration of the mixing process is depicted in Fig. 1 for the case of 137Cs. In addition to the experimental data, a Monte Carlo model (MCNP code) of the probe was developed and validated by experimental measurements. By combining experimental data and simulation, a calibration curve was created for energies covering the entire measuring range (See Fig. 2). Since the calibration tank is only an approximate model of the real situation in the river, the Monte Carlo model was also used for a calculation of the finite volume correction (the detection efficiency ratio between tank and river geometry) to obtain the correct detector efficiency in the river. For high energy gamma rays (60Co), the correction is approximately 0.85. This means that the calibration tank can be assumed to be a reasonable approximation of the river geometry. In addition to the efficiency calibration, the tank measurements provided a detector response matrix for the measured radionuclides. This was later used for the spectrum analysis.

River flows continuously through the canal. The probe was installed in a stable position 0.8 m from the concrete bottom and 1.0 m from the water surface. The water surface level is stabilized by a spillway. Long-term gamma spectrometric data collection covering all seasons was performed in 2017 and 2018. These data were processed as a gamma activity reference background at this particular location. The time course of the gamma spectra shows quite a stable concentration of gamma emitters in the river water. The activity of the gamma emitters is considerably enhanced only after rains. After strong storms and torrential rains, the total count rate in the river water spectrum was even doubled. The gamma activity enhancement is caused by the wash-out of radon and its progenies from the air and from the riverside surface (see Fig. 3). These rain events can significantly raise the detection limits and the probability of a false positive alarm due to spectral interferences between the monitored radionuclides and radon decay products. During the test period, approximately 40 thousand spectra (1024 channels) with 10 min integration time were collected and uploaded in the ANSI N42 format to the central server. This dataset was subjected to quality control, consisting of a check of the spectrometric and technological parameters associated with each spectrum. Then, a detection algorithm based on this cleaned dataset was developed. Our approach is based on principle component regression (PCR), where information from the whole spectrum is taken into account, instead of a local analysis such as a peak search using a spectrum derivative. This method

2.4. Data acquisition and processing The preliminary tests were carried out in semi-field conditions in a calibration canal of the T. G. Masaryk Water Research Institute in Prague. The calibration canal is 1.8 m deep, and water from the Vltava

Fig. 2. The efficiency calibration curve, corrected to the finite volume of the calibration tank. The labelled points indicate the radionuclide used for calibration. 45

Journal of Environmental Radioactivity 204 (2019) 42–48

M. Fejgl and M. Hýža

Fig. 3. Elevated natural background during rainfall in comparison with a non-rain event spectrum.

A comparison was made of the 137Cs and 131I MDAC values obtained by a standard background subtraction method and by the NASVD background subtraction method. The results are displayed in Tables 2 and 3. Gamwin software, based on the Currie (1968) approach, was employed as a standard procedure, where the MDAC (at the 95% confidence level) is defined by the equations:

belongs to the deconvolution algorithm family, which treats spectrum analysis as a regression procedure. In the PCR method, the analysed spectrum is fitted by a set of background and artificial spectral components. The background components are obtained as a low rank approximation of matrix M, the columns of which represent the measured clean background spectra. This can be done easily by singular value decomposition (NASVD) of M. The artificial components can be obtained by measurements, or by an Monte Carlo simulation of the particular radionuclide of interest. This approach is widely used in the field of airborne gamma ray spectrometry (Minty and Hovgaard, 2002; Schubert, 2015), but it can also easily be implemented in other fields of application. In this work, we have adopted the same procedure as was initially developed for continuous aerosol measurements using NaI(Tl), described in Hýža and Rulík (2017). The detection limits of the NASVD algorithm were determined by a simulation in which the artificial impulses were injected into background spectra. These modified spectra were then analysed by a detection algorithm, and the result (a regression coefficient) was plotted against the virtually added activity. As there is always background noise in the algorithm response (a nonzero regression coefficient for non-contaminated spectra), it is necessary to choose a cut-off value above which the response will be considered as significant. The response noise is also significant for the contaminated spectra. A simple comparison of the response and the cut-off level is not possible, and it is therefore necessary to construct a prediction band in order to limit the false negatives. In our setup, the MDACs were determined in such a way that the false negatives and false positives were less than 5% (Fig. 4.). Estimates were made on the basis of an analysis of approx. 5 × 104 simulated spectra with randomly added activity ranging from roughly 0 Bq to (2 x MDAC) Bq. In order to quantify the detection limits for longer integration times, we applied this procedure for datasets created by summing the original 10 min spectra (running sum). As a result, five datasets were created, containing spectra for 10 min, 1 h, 4 h, 12 h and 24 h integration time. As MDACs are affected by the natural background fluctuations, we divided the whole dataset into two groups: 1.) spectra collected during and after rain events, called “rain event spectra”, and 2.) an “all spectra” group, containing both rain influenced spectra and uninfluenced spectra. This grouping allowed the rain effect to be quantified, and thus provided an estimate of the MDACs for the worst-case scenario.

MDAC = 2.71 + 4.65 ×

B

and

B = BKG + S 2 +

LTS LTS × AB + LTB LTB

2

× SB 2

where. BKG – background under peak, S – background uncertainty, LTs – measurement time of the sample spectrum, LTB – measurement time of the background spectrum, AB – net area of the interfering peak, SB– uncertainty of the interfering net peak area in the background spectrum. 3. Results and discussion A simple structure for a non-shielded and autonomous detection system was built for determining gamma radioactivity in fresh water. The device was calibrated in laboratory conditions and in long-term testing in semi-field conditions. Minimal Detectable Activity Concentration (MDAC) values for 137Cs and 131I were calculated on the basis of the results. Two methods were used: 1.) standard analysis with Gamwin software, and 2.) the NASVD method (as described in section 2.4.). The NASVD method provides greater sensitivity, see Fig. 5, which shows that lower injected activity is sufficient to exceed the MDAC level. The estimated MDAC levels for our experimental setup are presented in Table 2. In addition, Table 3 contains MDAC estimates for “rain event” spectra only, and illustrates the detection capability of SAGMAs in challenging measuring conditions. Tables 2 and 3 show that the difference between MDAC levels obtained by the standard method and by the NASVD method is minimal when a longer integration time is applied and when the “all spectra” background is subtracted. When 10-min integration time is applied for the “all spectra” background, the MDAC level obtained by the NASVD method is lowered by a factor of ̴ 3 in comparison with the MDAC level obtained by the standard method. Background subtraction by NASVD instead of the standard method yielded a reduction of MDAC values by 46

Journal of Environmental Radioactivity 204 (2019) 42–48

M. Fejgl and M. Hýža

Fig. 4. Method for determining the MDAC in the NASVD algorithm approach. An example of the 137Cs response to the 5x104 simulated spectra with added activity between 0 and 600 Bq m−3 was calculated. The background cut-off was chosen as the 95% tolerance interval of the background response, and the width of the prediction band (one sided) was also chosen to be 95%. The intersection of these two lines defines the MDAC level.

factors of 3–5 for all integration times, when the “rain event spectra” background dataset was analysed. For an evaluation of the SAGMAs as a part of the SCOMO contribution for the governmental RMN, it is necessary to compare the detection capabilities of SAGMA with the legal requirements covered by Regulation 360/16 Monitoring of the radiation situation (SÚJB, 2016) and the current monitoring procedure is also taken into consideration. Regulation 360/16 specifies the required MDAC levels for 137Cs monitoring and also for 131I monitoring, separately for routine monitoring and for emergency monitoring (Table 1). The current procedure for water monitoring involves carrying out grab sampling of water once in the prescribed period (Table 1). The activity in water is measured using high-resolution gamma ray spectrometry in the Marinelli beaker geometry. This procedure is performed within routine monitoring with required MDAC levels of 0.1 Bq L−1 for both 131I and 137Cs. The same mode of 137Cs and 131I analyses is expected when emergency monitoring is to be employed, but the MDAC level requirements will be raised to 5 or 10 Bq L−1 (Table 1.). SAGMA working with the NASVD background subtraction method

Table 2 MDAC values for the “all spectra” background. MDAC [Bq L−1] All spectra

Integration time

RN I

10 min 3.6 1.29 4.2 1.41

131 137

Cs

Method Standard NASVD Standard NASVD

1h 1.48 0.52 1.7 0.57

4h 0.74 0.26 0.84 0.28

12 h 0.25 0.16 0.28 0.16

24 h 0.13 0.13 0.14 0.12

Table 3 MDAC values for the “rain event spectra” background. MDAC [Bq L−1] Rain event spectra

Integration time

RN I

10 min 5.3 1.42 8.4 2.26

131 137

Cs

Method Standard NASVD Standard NASVD

1h 2.9 0.69 3.1 0.83

4h 1.8 0.36 1.7 0.37

Fig. 5. A comparison of artificially contaminated spectra at the MDAC limits of the standard analysis approach (MDAC = 0.84 Bq L−1) and the NASVD analysis approach (MDAC = 0.24 Bq L−1) and 4 h integration time.

47

Journal of Environmental Radioactivity 204 (2019) 42–48

M. Fejgl and M. Hýža

fulfils the lower required MDAC levels of 5 Bq L−1 for 137Cs and 10 Bq L−1 for 131I for emergency monitoring, even in the most unfavourable conditions (rainy event background and 10-min integration time). This shows that, in occasion of an accident, incorporating SAGMA into the governmental Radiation Monitoring Network (RMN) would lead to the acquisition of sufficient information about the activity concentrations of these two crucial radionuclides in water samples, without the need to take grab samples. It would therefore ease the load on the RMN. SAGMA was developed as a tool for emergency monitoring aiming to be a strong tool for reinforcing the governmental system of emergency preparedness in the Czech Republic (CR). The final MDAC levels achieved by this tool are surprisingly low, because of NASVD method. It could therefore also be helpful for monitoring the radiation situation within routine monitoring. When the required 137Cs MDAC level is prescribed to be 0.1 Bq L−1 for routine monitoring, it can be achieved using the NASVD method by measurements with an integration time only slightly exceeding one day. A comparison of SAGMA and the SARA Water System, a monitoring station with a similar structure (described in section 1.5.) shows clearly the efficiency of the NASVD method. SAGMA achieves MDAC levels lower by a factor of ̴ 2. The potential of the NASVD tool is even clearer when the MDAC values obtained by the detection system are compared with the MDAC values obtained by a throughflow detection system with lead shielding measuring in a continuous regime, as described by Casanovas et al. (2013). For 24 h integration time and “all spectra”, the MDAC values of our device exceed MDAC as reported by Casanovas et al. (2013) ̴ 15fold. For 10-min integration time, the difference factor is reduced to less than 1.4. However, the portability of the device and its ability to make measurements in any location, even without an electric power source and during a blackout of normally available services, are also major benefits.

because of the NASVD method for background subtraction. It fulfils the requirements of Regulation 360/16 on monitoring the radiation situation (SÚJB, 2016) for emergency monitoring, when 10-min integration time is used. Simultaneously, when the integration time only slightly exceeds one day, the requirements for routine monitoring are fulfilled. This makes it possible to provide continuous monitoring of 137Cs and 131 I activity in surface water, and to replace the laborious grab sampling method. Acknowledgement This study was supported by institutional funding from the Ministry of the Interior of the Czech Republic (project VI20172020083). References Casanovas, R., Morant, J.J., Salvadó, M., 2013. Implementation of gamma-ray spectrometry in two real-time water monitors using NaI(Tl) scintillation detectors. Appl. Radiat. Isot. 80, 49–55. Currie, L.A., 1968. Limits for qualitative detection and quantitative determination. Appl. Radiochem. Anal. Chem. 40 (3), 586–593. Dulai, H., Kamenik, J., Waters, C.A., Kennedy, J., Babinec, J., Jolly, J., Williamson, M., 2016. Autonomous long-term gamma-spectrometric monitoring of submarine groundwater discharge trends in Hawaii. J. Radioanal. Nucl. Chem. 307, 1865–1870. FANC, 2005. Radiological Monitoring in Belgium. FANC report, Brussels. Fousová, E., Reidinger, J., 2017. Zpráva O Stavu Vodního Hospodářství ČR V Roce 2016. Ministry of Agriculture of the Czech Republic, pp. 132. GAMWIN, 2019. Třebíč, Czech Republic. Available: https://nuvia.cz/en/produkty/244gamma-spectroscopy-software-gamwin, Accessed date: 12 February 2019. Gitzinger, C., Henrich E, E., Turai, I., 2013. Doel Nuclear Power Station and the National Network of Environmental Radiological Monitoring Belgium. European Commission Document. Hýža, M., Rulík, P., 2017. Low-level atmospheric radioactivity measurement using a NaI (Tl) spectrometer during aerosol sampling. Appl. Radiat. Isot. 126, 225–227. IRSN, 2003. IRSN Activity Report 2002. Accessible on the Web. https://www.irsn.fr/ EN/publications/corporate/Documents/irsn_annual_report_2002.pdf Access verified on 27th of October 2018. Minty, B., Hovgaard, J., 2002. Reducing noise in gamma-ray spectrometry using spectral component analysis. Explor. Geophys. 33 (3), 172–176. Osvath, I., Povinec, P.P., Livingston, H.D., Ryan, T.P., Mulsow, S., Commanducci, J.F., 2005. Monitoring of radioactivity in NW Irish Sea water using a stationary underwater gamma-ray spectrometer with satellite data transmission. J. Radioanal. Nucl. Chem. 263 (2), 437–440. Povinec, P.P., Osvath, I., Comanducci, J.F., 2008. Underwater gamma-ray spectrometry. Radioact. Environ. 11, 449–477. Sartini, L., Simeone, F., Pani, P., Lo Bue, N., Marinaro, G., Grubich, A., Lobko, A., Etiope, G., Capone, A., Favali, P., Gasparoni, F., Bruni, F., 2010. GEMS: underwater spectrometer for long-term radioactivity measurements. Nucl. Instrum. Methods Phys. Res. 626–627, 145–147. Schubert, G. (Ed.), 2015. Treatise on Geophysics, Volume 11: Resources in the NearSurface Earth, 11.14. Tools and Techniques: Radiometric Methods, second ed. Elsevier9780444538024, . Sonck, M., Desmedt, M., Claes, J., Sombré, L., 2012. TELERAD: the Radiological Surveillance Network and Early Warning System in Belgium, Proceedings of IRPA12: 12. Congress of the International Radiation Protection Association: Strengthening Radiation Protection Worldwide - Highlights, Global Perspective and Future Trends. Steinmann, P., 2015. Radioactivity in river water below nuclear plants, the new network monitors the Aare and the Rhine on an ongoing basis. Aqua Gas 10. SÚJB, 2016. Regulation Number 360/2016 Sb.: Monitoring of the Radiation Situation, vol. 143. pp. 5642–5689. SÚJB, 2018. Annual Report 2017. Accessible on the Web. https://www.sujb.cz/ dokumenty-a-publikace/vyrocni-zpravy/vyrocni-zpravy-sujb/ Access verified on 27th of October 2018. SÚRO, 2019. Radiation Monitoring Network. Accessible on the Web. https://www.suro. cz/en/rms Access verified on 13th of March 2019. Tsabaris, C., Thanos, I., Dakladas, T., 2005. The development and application of an underwater γ-spectrometer in the marine environment. Radioprotection 40 (S1), 677–683.

4. Conclusion The system for continuous gamma activity monitoring (SCOMO), assembled from several Stations for Artificial Gamma Radioactivity Measurement (SAGMA) can potentially become a strong tool to reinforce the Czech governmental Radiation Monitoring Network (RMN). In an emergency situation, SCOMO can provide continuous determination of surface water radioactive contamination, and can simultaneously supply complementary information to verify mathematical models. The emergency preparedness systems of some European countries are equipped with similar monitoring stations. However, these stations are not in fact protected against the consequences of a severe NPP accident. In such an event, they would probably be out of service due to a general electric power blackout. SAGMA, which should finally be launched into routine operation within SCOMO as a part of the governmental RMN in 2020, aims to become a real “stand-alone” monitoring station. SAGMA is not vulnerable to the consequences of an NPP accident, and field measurements will be provided even in the event of a general outage of public services. SAGMA's autonomous structure allows it to be used as a portable station that can be relocated to a site of interest during an accident. However, a problem when SAGMA is relocated is the limited efficiency of the NASVD tool when the background dataset does not cover the background variations. The detection capabilities of SAGMA do not lose much in a comparison with other on-site stations with a more complicated structure,

48