Evaluation of Hungarian monitoring results and source localization of the 106Ru release in the fall of 2017

Anthropogenic 106Ru has been detected in the environment from late September to early October 2017 by several European environmental radiological monitoring networks. The paper presents the comprehensive evaluation of Hungarian monitoring results related to the occurrence of 106Ru in various environmental compartments (airborne particulates, deposition, plants, and terrestrial indicators), which was implemented to determine the temporal and spatial variation of the contaminant on a national scale and also to verify the findings based on the data arising from environmental monitoring at a local scale in Budapest. Difficulties in direct comparison of the diverse reported data were also considered; results arising from varied sampling periods were corrected with account taken of the relation between the sampling duration and 4-day-long plume residence (estimation based on the daily monitoring of air and backward trajectory analysis). Integrated analysis of air and deposition measurements and meteorological data was also performed; the deposition processes were investigated by establishing the correlations of activity concentrations measured in the atmosphere and in the deposition samples. In order to study the temporal distribution and spatial localization of the 106Ru contamination and to interpret the measurements at ground level, backward trajectory analysis was performed with HYSPLIT model. The backward trajectory simulations suggested that the release had probably occurred during the last week of September 2017 from the geographical area between Volga and the Urals. In addition, assessment of the doses due to the 106Ru release was implemented considering external exposure from cloudshine and groundshine and internal exposure via inhalation.


Introduction
In a short period from late September to early October of 2017, several European networks involved in environmental radiological monitoring reported the detection of 106 Ru isotope in environmental samples. Correspondingly to the European observations, 106 Ru was also detectable by the stations of the Hungarian environmental radiological monitoring networks. Following the first registered detections, the European monitoring networks were formally requested by the Incident and Emergency Centre (IEC) of the International Atomic Energy Agency (IAEA) (IAEA 2017a) to provide measurement results related to the appearance of 106 Ru in the environment in order to collect and analyze the measurement data and-if necessary-to recommend the implementation of public protective actions. The reported 106 Ru activity concentration values higher than the minimum detectable activity (MDA) concentrations in the atmosphere over Europe varied in the range of 0.8 μBq·m −3 -145 mBq·m −3 (IAEA 2017b; IRSN 2018); however, it must be noted that the air sampling duration at the data supply monitoring stations widely varied.
Large-scale emission of gaseous ruthenium oxides from the nuclear fuel can occur in high temperature and oxidizing atmosphere through volatilization, in which ruthenium oxides (RuO, RuO 2 , RuO 3 , and RuO 4 ) are generated from the metallic state of ruthenium. However, during the examined event, 106 Ru was detected solely in the environment, not as a minor component in a mixture of radionuclides of artificial origin, like following the Chernobyl nuclear power plant accident in 1986 (Steinhauser et al. 2014). The absence of other fission products excluded the assumption of a potential release from a nuclear reactor, which would have resulted the occurrence of other artificially produced radionuclides. The release of volatile ruthenium oxides (RuO 4 is the most volatile of them) can also occur through certain technical phases and processes of nuclear fuel reprocessing (dissolution process of spent fuel in nitric acid, denitrification, evaporation, and vitrification of highly active waste concentrate) (Kleykamp 1988). 106 Ru is commonly utilized for medical application, in particular for ocular brachytherapy for small-to medium-sized uveal melanoma (Stöckel et al. 2018) and retinoblastoma treatments (Schueler et al. 2006). The reported air concentrations were unlikely caused by medical treatment due to the small activity level (typical activity is up to 50 MBq (IAEA 2004)) of the ocular 106 Ru brachytherapy sources.
Since detection of 106 Ru in various environmental compartments happens to be a scarce phenomenon, studies on the direct measurements on the appearance of 106 Ru contamination in the environment, the dispersion, and the radiological consequences of the release attain significance.
The objective of the evaluation of the available Hungarian environmental monitoring results was to analyze the concentration levels of 106 Ru activity on a national scale and assess the dose consequences to contribute to the information related to national radiation conditions and radioactive contamination. The temporal and spatial behavior of the contaminated plume was also studied with backward trajectory analysis. Simulations were used to describe the atmospheric transport and dispersion of air parcels to interpret measurements on ground level, in order to locate the potential release region of the contaminated air masses as well as to estimate the travel time from the release at the source location and the residence time of the contaminated plume over the territory of Hungary. Additionally, the occurred incident causing unplanned discharge of 106 Ru to the environment provided an opportunity to assess the data supply procedure related to the notification obligations.

Materials and methods
Evaluation of the monitoring of 106 Ru in environmental media In this study, occurrence of 106 Ru in various environmental media was investigated, based on total beta counting and gamma spectrometric analysis of environmental samples (airborne particulates, wet and dry deposition, grass, other plants) collected at different locations in Hungary by the National Environmental Radiological Monitoring System (NERMS). According to the Government Decree 489/2015 (XII. 30.) (Hungarian Government 2015), NERMS has the obligation to acquire, analyze, register, and evaluate results related to environmental radiation measurements within the territory of Hungary and has to contribute to the fulfillment of the national and international notification and information obligations related to national radiation conditions and radioactive contamination. Following the first reported 106 Ru detections in national and European level, NERMS had the obligation to collect the measurement data from the environmental radiological monitoring stations and report to the competent authorities of the IAEA (IAEA 2017a).
According to the regulation, NERMS members as public administration organizations (authorities), special facilities (nuclear power plant, research reactors, spent fuel interim storage facility, radioactive waste repository, radioactive waste treatment and disposal facility, research centers) and other institutions are obliged to participate in the activity of the NERMS. The distribution of the monitoring network locations throughout the country provides representative geographical coverage of the Hungarian territory. The monitoring programs of NERMS members correspond to the obligatory measured samples and quantities specified in the pertinent governmental decree (Hungarian Government 2015), which requirements are in compliance with the European Commission recommendations determined in 2000/473/Euratom (Commission of the European Communities 2000). According to the regulatory requirements, measurement of airborne particulates, ambient gamma dose rates, surface water, drinking water, milk, and mixed foodstuff is obligatory on a routine basis.
Key technical considerations regarding the measurement of low-activity, pure β-emitter 106 Ru had been discussed by Jakab et al. (2018) and by Hult and Lutter (2017). In the following, general review will be given on the Hungarian monitoring of environmental compartments in relation to the 106 Ru detections.
The activity concentration of aerosol-bond 106 Ru in ground-level air was determined based on the continuous sampling of aerosol particulates. The air samplings were typically performed on glass-fiber filters at a constant air flow rate for a definite time period. The air flow rate at the measurement stations varied between the range of 3 and 250 m 3 ·h −1 depending on the flow rate of the operated air pumps. Sampling time varied from 1day-to 2-week-long intervals at different measurement locations depending on the actual decision between routine or increased sampling frequency. Extended sampling periods were generally required at monitoring stations equipped with low-flow-rate air samplers to ensure sufficient detection limits. The air filters were generally changed in the mornings on a routine basis. The continuously collected air samples were subjected to routine methods of radionuclide analysis, typically total beta counting with proportional counters and gamma spectrometry. Although the total beta counting method is not able to result in specific activities for individual β-emitters present in aerosol air filters, after all, it provided a rapid screening to indicate the elevated range of radiation levels in sampled ground-level air due to the 106 Ru contamination mostly by means of the highenergy beta decay of its short-lived descendant 106 Rh. Gamma spectrometry provided a tool to confirm the 106 Ru presence in environmental samples and nuclidespecific determination of 106 Ru also through 106 Rh, without the necessity of chemical separation or any specific pretreatment of the measured air filters. In the majority of cases, high-purity germanium (HPGe) detectors with relative efficiencies of 20-60% were used for the identification and quantification of 106 Ru/ 106 Rh. MDA of 106 Ru in air samples varied on a wide range between 0.01 and 1.2 mBq·m −3 due to the different volumes of sampled air, the diverse relative detection efficiencies of the used detectors, and variant counting times.
The majority of the reported Hungarian air measurement results were based on aerosol air filters sampled over a 7-day-long period. However, as it was noted, the duration of air sampling varied in time at certain measurement stations. Because of the varied sampling durations, the direct comparative analysis of the measured values was difficult. At those measurement points, where the sampling period exceeded the duration of the residence time estimated on the available daily monitoring results, the integration over the whole sampling interval would have resulted in an underestimation of the expected air concentration. In order to make the results comparable and to provide reliable estimation, the relation of the length of sampling interval to the residence time of the contaminant had to be taken into consideration. Measurement results therefore were corrected by weighting with the ratio of the sampling duration to the estimated average residence time of 106 Ru in the territory of Hungary at a representative measurement point. This definition implies the assumption of an identical residence time for each sampling location, thus slightly increasing the overall uncertainty of the final result but providing input values for an overall dose assessment for the country.
Deposited activity of 106 Ru radionuclide was also monitored additionally to the aerosol measurements. Primarily combined sampling of wet and dry deposition (fallout) was carried out with weekly to monthly sampling frequencies. The continuous sampling was complemented with subsequent radionuclide analysis (gamma spectrometry), following the pretreatment of the sample (in particular concentrating with evaporation). The measured activities were related to the collecting surface of the sampling vessels, which varied from 0.15 to 1.0 m 2 . Because of its lower sampling frequency, measurement of deposition samples was unable to detect the contamination rapidly and to monitor short-term variation of radiation levels. Nevertheless, the monitoring of the deposited activity enabled to measure the accumulated coverage of radionuclides on the ground and to follow the dispersion of discharged radionuclides in the environment due to the meteorological conditions. Plant and terrestrial indicator (in particular grass) measurements were typically implemented promptly, so the routinely used drying or ashing processing steps were skipped and the sample pretreatment was restricted to compaction. It was all the more feasible as the vast majority of 106 Ru present in these samples is assumed to occur due to surface deposition and not due to root uptake. At several stations, freshly harvested plant and grass samples were analyzed, so the measured activities were related to the fresh weight of the samples, which limited the sensitivity (i.e., MDA) of plant measurements.

Dose estimation calculations
Three possible pathways of exposure were considered for dose estimation: (a) cloudshine dose due to gamma radiation from the passing radioactive plume, (b) groundshine dose due to radioactivity deposited on ground surface, and (c) inhalation dose due to particulates. The effective external and inhalation doses for reference population subgroups of 3-month-old infants; 1-, 5-, 10-, and 15-year-old children; and adults were derived on the basis of environmental monitoring data in compliance with the recommendations specified in the pertinent ICRP documents.
Dose assessment was performed on the basis of national average air and deposition monitoring data considering conservative assumptions (continuous exposure during the radioactive plume residence, whole fraction time spent in the radiation field in absence of shielding). Effective dose coefficients for the radioactive plume and for the deposition on the ground surface were obtained from the Federal Guidance Report No. 12 (Eckerman and Ryman 1993). The age-dependent effective dose coefficients for inhalation of 106 Ru radionuclide and the inhalation rates were applied from the ICRP Publication 119 (Eckerman et al. 2012). Inhalation rates were determined from the default daily air intakes averaged over a 24-h period for each population subgroups. The daily inhalation was taken to be 2. 86, 5.16, 8.72, 15.3, 20.1, and 22.2 m 3 for 3-month-old infants; 1-, 5-, 10-, and 15-year-old children; and adults, respectively. The used procedure provides a rather conservative dose assessment as it assumes that the air concentration does not depend on whether the persons stay indoors or outdoors.

Backward trajectory simulations
The temporal and spatial behavior of the contaminated plume was studied with backward trajectory simulations. Inverse atmospheric dispersion modeling could be used in a variety of simulations describing the transport and dispersion of air parcels to interpret the 106 Ru measurements and to locate the possible origin region of the contaminated air masses, as well as to estimate the travel time from the release at the source location and residence time of the contaminated plume over a certain area. The simulations of the trajectories were performed with the Hybrid Single-Particle Lagrangian Integrated Trajectory model (HYSPLIT) developed by the US National Oceanic and Atmospheric Administration (NOAA), which is a hybrid model combining the Lagrangian approach and the Eulerian methodology (see detailed description in Stein et al. 2015). The required meteorological input data were obtained by the Global Data Assimilation System (GDAS). The publicly accessible meteorological dataset was obtained from the 3-, 6-, and 9-h forecasts as model output based on the 3hourly archive data came from NCEP's GDAS.
Since the 106 Ru activity concentration dispersion pattern in Hungary showed low temporal and spatial variation, the backward trajectories were calculated from one fixed representative receptor point located in Central Hungary (at 47.151900 latitude and 18.867300 longitude coordinates) to cover the territory of the Hungarian region. A total of 12 simulations were performed backward-in-time from 12 UTC 4 October to 00 UTC 27 September. The total run time, which corresponds to the trajectory duration, was 120 h, which provided 5day-long data of the atmospheric transport of air masses. At each simulation, three vertical levels of trajectory were calculated at 50, 75, and 100 m height above the ground-level receptor with respect to the meteorological data in each pressure level.

Results and discussion
Measurement results of Hungarian monitoring of 106 Ru Measurement results from 22 monitoring stations were evaluated based on the available data provided by the monitoring data supply centers of NERMS. 106 Ru was detectable in environmentally monitored constituents, as air (aerosol air filters), deposition, plants, and terrestrial indicators (in particular grass).

Results of total beta measurements of aerosol air filters
The daily monitoring of aerosol air filters to detect airborne particulate 106 Ru enabled the determination of short-term variation of 106 Ru activity concentration in ground-level air as it was described by Jakab et al. (2018). Evaluation of the national daily monitoring results enabled the estimation of the date of 106 Ru occurrence and residence time of the radioactive plume in the atmosphere over Hungary. The comparative analysis of the available national daily monitoring results also allowed the verification of the findings based on the daily measurements at the KFKI Campus in Budapest.
The available daily total beta activity concentrations (see Fig. 1) provided a rapid indication of the contamination. The first detections of increased level of total beta activity due to the appearance of 106 Ru in the ground-level air were observed in the air filters sampled between 30 September and 1 October. The arrival date of the radioactive plume agrees with the first detections in Stockholm, Sweden (Ramebäck et al. 2018), and corresponds to the reported observations in Austria and the Czech Republic, based on the results published by IRSN (2018). The presence and contribution of 106 Ru to the elevated total beta activity concentration was proven by subsequent nuclide-specific gamma spectrometry analysis of the filters. The maximal reported atmospheric total (gross) beta activity value (35.0 ± 2.0 mBq·m −3 ) was observed in Budapest by the Environmental Protection Service of MTA EK, between the mornings of 1 October and 2 October. This sample was evaluated according to the routine procedure of the service; thus, contribution of short-lived radon progeny was eliminated. The results obtained from daily sampling later than 4 October showed a steady decrease and the elevated level of total beta activity decreased to the range of the average background radiation level originated from the presence of natural beta-emitter components (1.4 ± 0.8 and 1.8 ± 0.4 mBq·m −3 in the territory of MTA EK and the measurement station of OKI KI SSFO, respectively) as the subsequent measurements generally did not determine increased quantity of total beta activity (as it can be seen in Fig. 1). It can be observed that the 106 Ru activity concentration in ground-level air decreased gradually because of the removal from the plume due to deposition mechanisms in addition to the passing of the plume. The daily variation of total beta activity levels indicated a maximum of 4-day-long residence time correspondingly to the daily monitoring results observed in Budapest by the Environmental Protection Service of MTA EK (Jakab et al. 2018). According to the national monitoring data, it was verified that the 106 Ru contamination was present over the territory of Hungary from 30 September until the morning of 4 October.
Results of gamma spectrometric measurements of aerosol air filters Table 1 gives a general overview of the measured 106 Ru activity concentrations in the atmosphere over Hungary during the given sampling intervals. Figure 2a shows the 106 Ru activity concentration in aerosol air filters obtained from the gamma spectrometry analysis, regarding the time period of the sampling from 25 September to 16 October 2017. The figure also visualizes the geographical locations of the measurement points. The histogram columns on the diagrams illustrate well the variant sampling durations, and the differences between initial and endpoints of the samplings.
Average value of 106 Ru activity concentration in the ground-level atmosphere over Hungary was 10.6 ± 1.3 mBq·m −3 integrated over the whole 2-week-long period (25 September-9 October 2017), covering all cases of 106 Ru detections. To ensure the comparability of the measurement results, correction of the values was executed by weighting with the ratio of the sampling duration to the estimated residence time of 106 Ru over the sampling interval at each measurement point. The corrected values (see in Fig. 2b) gave an average activity concentration of 25.0 ± 2.0 mBq·m −3 over the 4-daylong estimated exposure time (from the morning of 30 September to the morning of 4 October) of 106 Ru in the atmosphere over Hungary. Based on the corrected results, it can be concluded that the measurement results obtained by different monitoring data supply centers are in a good agreement. 106 Ru activity concentration dispersion pattern in Hungary showed low temporal and spatial variation, resulting relatively uniform levels of 106 Ru contamination in the country. From the data reports of a few monitoring data supply centers, measurement results of simultaneously sampled air filters (i.e., samples collected over the same interval) were available from the same locations. Parallel measurement results enabled the validation of the sampling and measurement procedures as well as the refinement of the outcomes of the calculations. Based on the comparison of simultaneous samplings, it can be determined that air monitoring results observed in the same area over a given sampling interval show a good agreement. For the comparative analysis of parallel air sampling, in the following section, 106 Ru activity concentrations observed in simultaneously sampled aerosol air filters in Paks will be taken as an example. In the vicinity of the Paks Nuclear Power Plant (NPP), an environmental radiological monitoring system is operated, in which aerosol particulates are sampled at 9 stations-located within a 2-km radius around Paks NPP-routinely on a weekly basis. Simultaneous sampling was performed in Paks at 9 stations between the interval of 25 September and 2 October (arithmetic mean, 13.5 mBq·m −3 ; standard deviation, 1.0 mBq·m −3 ), whereas parallel sampling was carried out from 2 October to 6 October at 3 stations (arithmetic Fig. 1 a Temporal variation of total beta activity concentration of aerosol air filters from 25 September to 9 October 2017, reported by MTA EK. Sampling was performed with 100-150 m 3 ·day −1 flow rate air samplers. At stations 2 and 6 daily sampling was executed, except the weekends, where averaged values integrated over 3 days (from Friday morning to Monday morning) were presented. At station 5, sampling was performed continuously on a daily basis, including the weekends. b Temporal variation of total beta activity concentration of aerosol air filters from 25 September to 9 October 2017, reported by the monitoring data supply center ERMAH (OKI KI SSFO). At the measurement site daily sampling was performed, except the weekends, from which measurement results were not available mean, 6.6 mBq·m −3 ; standard deviation, 0.7 mBq·m −3 ) and simultaneous measurement results are available from the remaining 6 measurement stations from the period between 2 October to 9 October (arithmetic mean, 4.0 mBq·m −3 ; standard deviation, 0.6 mBq·m −3 ). The simultaneous results from the 9 measurement points and parallel results from the 3 and the remaining 6 measurement stations varied with 7.7%, 10.3%, and 14.4% relative standard deviation, respectively. These relative standard deviations (i.e., standard deviation related to the average value) in the air measurements can be interpreted as an acceptable variation considering the typical uncertainty level of environmental measurements.
In the case of the air filters sampled in Paks, the 7day-long sampling period exceeded the duration of the 106 Ru contamination presence, which could cause an underestimation of the air concentration, as the measured value would be integrated over the total sampling interval (i.e., the sample activity divided by the total Ru activity concentration of simultaneous samplings over the same sampling period from the same area. The number of the averaged values, which were used to calculate arithmetic mean, is indicated in parenthesis b Daily collected air filters obtained in one-week intervals were combined and measured collectively providing weekly average values over the sampling period c106 Ru activity concentration is given in < MDA format, when the measured value was below MDA for the given sample d Uncertainty for the individual sample has not been determined; approximate uncertainty was calculated based on the average uncertainty typical for the gamma analysis of aerosol air filters sampled air volume of that week). The measurement results therefore were corrected similarly as the national values (by weighting with the ratio of the sampling interval length to the estimated residence time of the plume). The corrected values gave an average activity concentration of 49.4 ± 3.8 mBq·m −3 for the period between 30 September and 2 October, which value was referred to as the maximal 106 Ru activity concentration in the atmosphere over Hungary by some reports (IAEA 2017b; IRSN 2018). Table 2 shows the detected 106 Ru activity concentrations in deposition samples. Based on the available meteorological datasets, a dry period lasted in the interval between 24 September and 2 October in Hungary. During this period, naturally occurring 7 Be was detectable in deposition; this cosmogenic radionuclide is suitable for quality control as the average and the range of specific activities of 7 Be are known for each measurement location. Under dry conditions, 106 Ru from the passing contaminated air masses deposited on the surface via dry deposition mechanism (resulted mainly from the downward vertical fall of atmospheric aerosols onto the surface) in the first days of the plume residence. 106 Ru activity concentrations in deposition sampled before 2 October were below the detection limits (typically 2.1 Bq·m −2 ) at each measurement point, which indicates low deposition rates and high remaining activity in the plume. According to the dry deposition model, the dry deposition velocity (v d ) of 106 Ru aerosols, defined as the ratio of the 106 Ru activity concentration deposited on the ground surface (dry deposition rate, D < 2.1 Bq·m −2 ) to the time integral of the average corrected concentration of 106 Ru activity in air from the morning of 30 September to the morning of 2 October (C air = 1.8 Bq·h·m −3 ), would be lower than 3.2 × 10 −4 m·s −1 . Because of the particle size dependency of the dry deposition velocity, the low value of the dry deposition velocity compared to the default values used by operational atmospheric dispersion models (reference values generally lie in the range of 10 −3 m· s −1 ) follows from the presence of particles with small aerodynamic diameter (≤ 1 μm).

Results of the deposition measurements
Integrated analysis of deposition measurement results and available meteorological data (see Fig. 3) indicated that deposition to the ground surface of 106 Ru was strongly affected by the occurrence of precipitation. On a national scale, the first contiguous rainfall event with intensities exceeding the measurement resolution occurred on 3 October with about 1.2 mm·h −1 national average precipitation intensity related to the precipitation durations. Apart from some rare locally prevailing precipitation events, the spatial and temporal distribution of the precipitation in the territory of Hungary showed low variation during the plume residence. The evaluation of the Hungarian deposition values verifies the conclusion made on the basis of the deposition measurements by the Environmental Protection Service of MTA EK, namely that during the plume residence, wet deposition (due to the removal of radionuclides effected by precipitation mechanisms) was the dominant contributor to the deposition onto the surface as 106 Ru was only detectable in deposition samples after 2 October (Jakab et al. 2018). This conclusion corresponds to the findings of Ramebäck et al. (2018) as well, where wet deposition as a consequence of the rainfall on early October was also identified as the major removal process. The Hungarian average 106 Ru activity concentration in deposition was 8.4 ± 1.0 Bq·m −2 .
Based on the weekly deposition sampling, trends of deposition in time could be determined. As it was formerly stated, 106 Ru was only detectable between 2 and 9 October, in weekly collected deposition samples. 106 Ru activity in deposition sampled after 9 October remained below the detection limit in spite of the subsequent heavy rainfalls in the period of 22-23 October and on 29 October. The undetectable 106 Ru activity concentrations in deposition confirmed that the cloud had already passed by that period. Parallel deposition measurement results showed a slightly higher variation than the simultaneous air measurements. The results of simultaneous monthly deposition measurements in the 4 measurement stations of MTA EK and parallel results from 2 measurement stations in Paks varied with 17.7% and 15.2% relative standard deviation, respectively. The differences between the deposition results observed in the same area over a given sampling interval can arise from the sampling uncertainty due to the natural variability of deposition, and the formation of local turbulent wakes generated by the ambient air flow in the area of the sampling equipment. Despite the variances between the parallel deposition measurements, the measured activities in deposition samplescorrespondingly to the meteorological datasufficiently indicated the temporal variation of the deposited 106 Ru concentration and provided input data that could be used in atmospheric deposition models and dose assessments. Table 3 shows the detected 106 Ru activity concentrations in plants and terrestrial indicators.

Results of plant and terrestrial indicator measurements
The vegetation sampling locations covered the territory of the Hungarian region sufficiently (see Fig. 4). The contamination of external plant and grass surfaces was also caused predominantly due to the wet deposition resulted from the rainfall event on 3 October. Grass samples were taken after 2 October yielded a maximum of 10.7 ± 0.5 Bq·kg −1 106 Ru activity concentration related to the fresh weight.
The objective of the monitoring of plants could be the determination of the doses to members of the public from the ingestion of deposited radionuclides in foodstuffs. Plant surfaces are primarily contaminated by direct deposition of aerosol-bond radionuclides or by indirect contamination via resuspended radionuclides. Grass could be a direct pathway of the deposited radionuclides to animals by their rapid uptake of radionuclides. 103 Ru detection in the atmosphere over Hungary Predominantly, 106 Ru have been detected solely in the atmosphere over Europe; nevertheless, simultaneous detection of 103 Ru and 106 Ru was reported by a few European environmental radiological monitoring networks. According to IRSN (2018), the 10.6 ± 1.5 mBq·m −3 average level of 106 Ru was accompanied by 3.1 ± 0.5 μBq·m −3 average 103 Ru activity concentration in air filters sampled in Sweden, Austria and the Czech Republic in the period between 25 September and 4 October. The ratio of average 106 Ru and 103 Ru activity concentrations corresponded to 3930. The cumulative fission product yield of shorter-lived 103 Ru (T 1/2 = 39.3 days; Bé et al. 2016) for 235 U thermal and  Koning et al. (2006). According to the fission yields, assuming that 106 Ru and 103 Ru were generated together in a reactor fuel of an appropriate light water reactor, the Ru species detected in the atmosphere over Europe were about 2 years old. In aerosol filters sampled at the Hungarian monitoring stations 103 Ru was not detected, even at the monitoring stations equipped with the highest-flow-rate (240-250 m 3 ·h −1 ) air samplers. MDA of 103 Ru with gamma spectrometry measurements was in the range of some tenths of microbecquerels per cubic meter, which exceeded the highest 103 Ru activity concentration detected in Europe with one order of magnitude. Fig. 3 a, b Relation between the daily summation of precipitation amounts and deposited 106 Ru activity concentrations. Panel a shows the weekly (visualized data for one station) and monthly measurements (visualized data for the weighted mean of four simultaneously collected samples taken between 2 October and 6 November) of MTA EK. Panel b shows the monthly measurements of Paks (visualized data for the weighted mean of two simultaneously collected samples for which reliable 106 Ru data was available) Estimation of public exposure due to the 106 Ru contamination Based on the observed 106 Ru activity levels in air, it can be concluded that 25.0 mBq·m −3 106 Ru activity concentration in ground-level air would lead to 9.4 × 10 −4 nSv·h −1 external gamma radiation dose rate, whereas 106 Ru deposition of 8.4 Bq·m −2 would result in 6.4 × 10 −3 nSv·h −1 ground radiation dose rate (calculated with the dose coefficients for air submersion and for exposure to contaminated ground surface given by Eckerman and Ryman (1993)). As the ambient gamma dose rate monitors have detection limits typically in the order of nanosievert per hour, it can be stated that 106 Ru occurrence in the environment was not sufficient to cause a detectable increase of the ambient gamma dose rate.
Considering the inhalation dose of the examined population subgroups, the results varied with 21% relative standard deviation due to the age dependence of the dose coefficients and inhalation rates. The most exposed group of the population for internal exposure due to inhalation was the adult reference subgroup. The findings of the deposition model calculations based on the deposition measurements indicated the presence of particles with small aerodynamic diameter. Therefore, effective dose coefficients for inhalation of particulate aerosols with default activity median aerodynamic diameter (AMAD) of 1 μm were used. Internal dose from inhalation of airborne 106 Ru radionuclide to the adult population group was estimated as 146 nSv, assuming constant inhalation of 25.0 mBq·m −3 activity concentration with exposure duration corresponding to the 4-daylong estimated residence time of 106 Ru in Hungary. As a simplifying assumption, the resuspension mechanisms moved by the action of wind or by disturbances of the soil or external plant surfaces were not taken into account in the calculation of inhalation. External dose from cloudshine to the adult population group was 0.09 nSv, assuming a continuous exposure over the 4day-long estimated residence time in 9.4 × 10 −4 nSv·h −1 external gamma radiation dose rate as stated above. External dose from groundshine to the most exposed group in the first week following the initial deposition was 1.1 nSv with assumption that the radioactivity was removed from the ground surface only via radioactive decay. It must be noted that radioactive progeny was considered in external dose coefficients, the parent 106 Ru and progeny 106 Rh remain in secular equilibrium in the radioactive plume and in the subsequent deposition on the ground surface. External doses received by members of the examined subgroups were less variant because of the slight dependence of the dose coefficients on age.
The dose assessment showed that internal exposure due to the inhalation of contaminated air had paramount contribution to public exposure, whereas external pathways via exposure from cloudshine and groundshine had negligible contribution to the total effective dose which itself was also negligible. The total effective dose to the adult reference population subgroup in the Ru activity concentration is given in < MDA format, when the measured value was below MDA for the given sample estimated residence time of 106 Ru was 147 nSv, which is 0.006% of the 2.4 × 10 6 nSv worldwide average annual effective dose from natural sources of ionizing radiation (UNSCEAR 2000). This practically means that the received dose due to 106 Ru exposure equals with the received dose from natural background radiation by spending about 2 h outdoors (as the populationweighted average absorbed dose rate in outdoors is 57 nGy·h −1 reported by UNSCEAR 2000).
Trajectory analysis of the 106 Ru contamination on European scale Figure 5 represents the results of the trajectory analyses backward-in-time from 12 UTC 4 October to 00 UTC 27 September performed with HYSPLIT atmospheric transport and dispersion modeling system. The geographical location of the potential release zone was estimated with the assumption that the detected 106 Ru was discharged to the atmosphere from a ground-level release point. Based on the executed simulations, it can be concluded that the most probable region of the release lies in the geographical area between the Volga and the Urals. The findings of the spatial localization are consistent with the outcomes of the simulations described by IRSN (2018) and Sørensen (2018). The backward trajectory analysis result confirms the findings based on the assessment of the aerosol air filters that the radioactive cloud reached the territory of Hungary at the end of September 2017 from the determined release direction. As it can be seen on the trajectories inversely modeled from 4 October 2017, the wind direction changed in this time period, which eliminated further transport of potentially contaminated air masses from the region of origin resulting in consequent reduction of the environmental radiation levels of the atmosphere. The release, with regard to the backward trajectory simulations, could occur during the last week of September 2017.
Atmospheric dispersion model calculations are associated with uncertainties; these uncertain components (e.g., uncertainties of meteorological variables) have effect on the outcomes of the atmospheric dispersion modeling. In consequence of the above ascertainment,  latitude and 18.867300 longitude coordinates) was indicated with a black star on each map the overall uncertainty of the inverse trajectory modeling was relatively high; therefore, the more precise determination of the contaminant's spatial localization was not feasible from the available data.

Conclusions
Results of Hungarian radiation monitoring and environmental sampling stations were evaluated in terms of the occurrence of 106 Ru in the environment over the period from late September to early October 2017. 106 Ru activity concentrations in different environmental compartments (air, deposition, and plants as terrestrial indicators) were analyzed. Comparative analyses of the reported datasets, provided by the Hungarian monitoring data supply centers of NERMS, were implemented to determine the temporal and spatial variation of the radiation levels on a national scale and also to supplement and verify the outcomes of the evaluation based on the local data arising from the environmental monitoring in Budapest. The dispersion patterns of the 106 Ru activity concentration in Hungary showed low temporal and spatial variation, which resulted in relatively uniform radiation levels at a national scale. The hypothesis on the maximum of 4-day-long residence time was verified based on the additional Hungarian daily monitoring data and the backward trajectory analysis. According to the investigation of the daily variation of Hungarian activity levels in ground-level air, the 106 Ru contamination circulated over the territory of Hungary from 30 September until the morning of 4 October, correspondingly to the daily monitoring of air in Budapest. To enable direct comparability, the diverse available data were corrected with account taken of the relation between the sampling duration and the estimated 4-day-long residence time of the radioactive plume. With the correction of the raw national air monitoring data, average 106 Ru activity concentration of 25.0 ± 2.0 mBq·m −3 has been calculated over the estimated residence time of 106 Ru in groundlevel air. Integrated analysis of deposition measurement results and available meteorological data verified that the deposition of 106 Ru to the ground surface was dominantly influenced by the occurrence of rainfall episodes. Results indicated that wet deposition mechanism was the major contributor of the removal process from the plume to ground, which led to an average of 8.4 ± 1.0 Bq·m −2 deposition on the ground surface prior to the plume passage.
Backward trajectory analysis was used for the temporal distribution and spatial localization of the 106 Ru contamination. The simulations were performed with HYSPLIT model on the basis of monitoring results and publicly accessible meteorological data. Based on the simulations, the geographical area between the Volga and the Urals was determined as the possible origin region. According to the backward trajectory simulation, the release occurred probably during the last week of September 2017.
Dose assessment was performed considering external exposure from cloudshine and groundshine and internal exposure via inhalation. The doses received by members of infant; 1-, 5-, 10-, and 15-year-old children; and adult reference population subgroups were estimated with conservative assumptions, on the basis of national average air and deposition monitoring data. According to the dose assessment, the estimated exposure was negligible; however, inhalation has been determined as the main exposure pathway, which contributed predominantly to the total effective dose (147 nSv on average) in short term. External pathways via exposure from cloudshine and groundshine had negligible contribution to the total effective dose which itself was also negligible.