137Cs distribution on the territory of Romania 30 years after Chernobyl accident

To estimate the contribution of Chernobyl 137Cs contamination, in 1993 and especially 2016, its total inventory was determined by gamma-ray high-resolution spectroscopy in 62 and respectively 747 soil samples covering the entire Romanian territory. This permitted to estimate the 137Cs inventory as varying between 0.4 and 187 as well as between 0.2 and 94.2 kBq/m2 for years 1993 and 2016, respectively. By representing the spatial distribution of 137Cs inventory in Voronoi polygons, it was possible to evidence a decrease of the total 137Cs inventory over entire Romanian territory with a factor of about 3 from about 3.6 TBq to less than 1.2 TBq, exceeding in this way the natural decay which suggests that a certain amount of 137Cs was washed out by precipitation and, at a lower extent, was incorporated into plants. At the same time, by evaluating the maximum contribution of 137Cs to the population exposure, in 1993 as well as in 2016, the supplementary annual effective dose did not exceed, in the majority of sampling points the value of 0.2 mSv/year.


Introduction
The Chernobyl accident, which took place at the dawn of 26 May 1986, is considered, together with Fukushima accident, the worst nuclear events in the history. On the seven levels of AIEA International Nuclear and Radiological Event Scale, both Chernobyl and Fukushima are classified as major accidents occupying the highest position, the 7th on this scale (IAEA, 2008).
The fission product 137 Cs has a half-life time of 30.05 ± 0.08 years and a special affinity for clay minerals, especially illite, which binds it (Cornell, 1993;Pennington et al., 1973). Despite this finding, small amount of 137 Cs deposited on pastures or other agricultural lands can be absorbed by plants roots, transferred to edible parts of plants, and then to final human consumers which could induce internal irradiation (Burger & Lichtscheidl, 2017). For this reason, the 137 Cs presence in soils should be periodically checked. In the first year after accident, the presence of short-living radioisotopes significantly overshadowed 137 Cs contribution, but as this category of fission products decayed, radiocesium relative contribution increased, making it the most important gamma ray emitting radionuclide for both internal and external exposure. At present, 36 years after the accident, the natural decontamination processes have reached a stationary phase, the natural decay being one of these processes which makes 137 Cs presence to gradually decrease with a factor of about 2 10 every 300 years.
The first systematic investigation of the 137 Cs distribution on the Romanian territory, corrected for the radioactive decay at 10 may 1986, was presented in the Atlas of Caesium Deposits on Europe (De Cort et al.,. 1998). At the same time, due to the complex meteorological conditions in spring, 1986, the radioactive contamination following Chernobyl accident presented an irregular character, confirmed by more post-accident investigations. Judging on the date presented on the Romania Map (De Cort et al., 1998), the 137 Cs contamination reached maximum values up to 100 kBq m −2 , mainly in the southwestern regions and in some places on the northern border with Ukraine, while on the rest of territory, the contamination was lower or even almost absent as the case of north-eastern Danube Delta (Duliu et al., 1996). In this regard, some studies performed after May 1986 evidenced a continuous presence of 137 Cs in Romanian soil (Margineanu et al, 1998(Margineanu et al, , 2018Paunescu et al., 1992).
Once deposited, 137 Cs diffusion in soil is very slow so that, according to (Krštic et al., 2004), even 50 years after the Chernobyl accident, its largest amount will be still in the first 10 cm layer of soil, where it can be easily washed out by precipitation. Under these circumstances, it is of a scientific interest to compare 137 Cs actual distribution with previous determinations, as the experimental conditions in Romania permit now a better analysis, mainly due to the presence of the ultra-low background laboratory (ULBL) located in the Slanic-Prahova Unirea salt mine (Margineanu et al., 2008(Margineanu et al., , 2018. Therefore, to get a more detailed picture of the late post-Chernobyl environmental contamination, between years 2016 and 2018, about 750 samples of arable and at a lesser extent of pasture land were collected and gamma spectroscopy was analyzed to determine the specific activity of remaining 137 Cs. It should be pointed out that the arable land was preferred for this kind of investigation as repeated deep plowing has homogenized the 137 Cs containing soil horizon, which makes more reliable data concerning its activity concentration, and subsequently the 137 Cs inventory. Moreover, arable land represents an important reservoir of radiocesium for plants and subsequently for human consumption, and, thus, a relevant source of internal contamination.
To evidence the time evolution of 137 Cs environmental contamination over the entire Romanian Page 3 of 10 848 Vol.: (0123456789) territory, we have reconstructed the spatial distribution of 137 Cs inventory based also on the existing data collected in 1993 (C. Dovlete, personal communication) and those existing in the Atlas of Caesium Deposits on Europe (De Cort et al.,. 1998). It is also worth mentioning that before Chernobyl accident, the atmospheric nuclear tests have contributed 1.5 to 3 kBq/m 2 to 137 Cs land contamination (De Cort et al., 1998).
The results of this study, which covers entire Romanian territory excepting the Danube Delta ( Fig. 1), will be further presented and discussed.

Sampling
For this study, 747 samples of arable and, at a lesser extent, pasture soil were manually collected following the quality assurance sampling procedures according to ISO 17025 of the Horia Hulubei National Institute for R&D in Physics and Nuclear Engineering (IFIN-HH) [Internal procedures] in more campaign between 2016 and 2018. Accordingly, at each sampling point, a minimum three soil columns covering an area of about 10 m 2 were collected using a 5-cm (78.5 cm 2 ) metallic corer. This procedure would compensate for the local variability of 135 Cs, by diminishing the inherent errors which could be about 50 % in the case of single sampling (Khomutinin et al., 2020). Finally, for each case, the resulting material was mixed together and thoroughly homogenized.
As sampling points covered the entire territory of Romania, the investigated soils were mainly Cernisols, Luvisols, loess Cambisols, and only in very few cases, at alpine plains, Spodosols (Stanila & Dumitru, 2016). Due to this diversity, the soil organic matter (SOM) varied between 3 and 5%, being mainly concentrated in the first upper soil horizons. To avoid cross-contamination, after each sampling, the corer was thoroughly washed and soaked with disposable paper napkins. Samples consisting of 5 cm soil columns, in the case of 1993 sampling (Dovlete, personal communication) and 20 to 25 cm in the case of 2016-2018 campaign, were labeled in the field with collecting date and geographical coordinates, the last one being determined using a commercial GPS. It is worth mentioning that in the case of 1993 samples, as majority of sampling points were placed on pastures or uncultivated areas, the soil sampling was restrained to the superficial soil layer of 5 cm, this particularity being considered when the 1993 137 Cs inventory was calculated. Taking into account the slow penetration of 137 Cs in soil (Krštic et al., 2004), we can suppose that for the 1993 samplings, the first 5 cm of topsoil retained the great majority of Chernobyl radiocesium.
In both cases, after being collected, soil samples were cleaned of vegetable matter such as fragments of grass and grass roots, leaves, and wood pieces, as well as of pebbles. Finally, amounts of about of 0.5 kg of cleaned soil was dried at room temperature, homogenized again, and sieved using an 18 mesh (1 mm) sieve. After that, 100 g of each sample of processed soil was sealed in cylindrical plastic boxes for further radiometric determinations. As we were interested only in 137 Cs activity concentration determination, the time interval between sample preparation as presented before and gamma spectrometric measurements was irrelevant.

Radiometric determination and statistical data analysis
All radiometric measurements were performed in the Slanic-Prahova low backgrounder laboratory (Margineanu et al., 2008). Due to the high number of samples collected during 2016-2018 campaigns, we have used two Canberra HPGe detectors with a relative efficiency of 22.3 % and 130 %, respectively, located in the Slanic Prahova ULBL. Both detectors were calibrated using the same certified reference sources (CRS). To supplementary reduce the background, the first detector was shielded by three layers consisting of 2 cm of OFHC Cu, covered by 5 cm of old lead of with a 210 Pb activity concentration less than 25 Bq/kg, all shielded by an external recent Pb layer of 10 cm thick. In the case of the second detector, the shielding was the same excepting the inner OFHC Cu of which thickness was increased to 5 cm. In this way, the radiation shielding factor was about 3600-4000 for both instruments (Margineanu et al., 2008, Tudor et al., 2019, which permitted us to reach, for each of them, a minimum detectable activity (MDA) of 0.5 Bq/kg for a 24-h measuring time.
The two high-resolution gamma spectrometers have been calibrated in efficiency using IAEA 241 Am, 137 Cs, and 60 Co and by participating in proficiency tests organized by the IAEA (2014, 2015 TM 2000 Geometry Composer routine. A special attention was paid so that both detectors be reciprocal compatible by periodically measuring the same soil samples alternatively using each of them. Moreover, the total time of gamma spectra acquisition for a single soil sample varied between 22 and 24 h. Under these conditions, the total experimental uncertainty concerning the 137 Cs activity concentrations was no greater than 5 %. In the case of 2016-2018 campaign, all 137 Cs activity concentration values were recalculated such that to correspond to May 2016. The PAST 4.1 software (Hammer et al., ) was used for the statistical analysis of the data.

Mapping
For a better comparison with the 1993 activities, the sample activities collected between 2016 and 2018 were recalculated to May 2016, i.e., 30 years after Chernobyl accident. Both set of data, i.e., the 1993 and 2016 ones, have generated two geospatial databases, further processed by ArcMap 10.4 Spatial Analyst specific tools. This permitted obtaining spatial distribution maps of 137 Cs inventory over the entire Romanian territory (Fig. 2a, b). As the sampling points were not perfectly aligned to form a regular network, we have decomposed the investigated area in Thiessen-Voronoi polygons (Burrough et al., ) (Fig. 2a, b-insets), followed by an inverse distance weighting (IDW) interpolation with a cell size of 100 to get the spatial distribution of 137 Cs inventory (Longley et al., 2015) (Fig. 2a, b). Both maps were represented using the Lambert azimuthal projection. The same data were used to calculate the total 1993 and 2016 137 Cs inventory corresponding to entire Romanian territory. This was done by means of the ArcMap 10.4 Surface Volume Tool.

Results and discussion
By analyzing the data provided in Fig. 2 and Fig. 3, it can be remarked that the local values of 137 Cs inventory vary between 0.4 and 187 kBq/m 2 and between 0.2 and 94.2 kBq/m 2 for the 1993 and 2016 data, respectively. The Shapiro-Wilk, Anderson-Darling, and Lilliefors and Jarque-Bera tests showed that for both sampling periods, the data concerning 137 Cs are neither normal nor log-normal distributed, as histograms reproduced in Fig. 3 show. Here it should be taken into consideration the presence of pre-Chernobyl, atmospheric nuclear weapon tests, of about 2 kBq/m 2 (De Cort et al., 1998), but this value was estimated and not experimentally determined.
In interpreting cesium data, it should be considered the fact that all measurements were done at certain interval from May 1986 so that the recorded patterns reflect the inventory distribution seven and respectively 30 years after the Chernobyl accident. In the case of 1993 determination, the map reproduced in Fig. 2a presents some well evidenced maxima on a line from Toaca (Ceahlau Mountain) to Parang Mountain and further till Danube River as well as a local maximum not far away from Bucegi Mountains in the vicinity of Campina city, South-Eastern Romania. These data confirm some officially unreported data according to which few days after Chernobyl accident, on the Bucegi Mountains Plateau, the 131 I contamination was extremely high, suggesting that, at its turn, the 137 Cs activity concentration at the soil level should follow the same tendency. On the other hand, it should be considered that 1993 results reflect the status quo 7 years after Chernobyl accident.
As for the 2016 determinations, the map reproduced in Fig. 2b illustrates the presence of four maxima of which position, excepting somehow those of Campina city maximum, is quite different from the 1993 ones. But, even in these conditions, the maximum recorded values are, as mentioned before, significantly lower than the 1993 higher values. Moreover, at careful analysis of the same data, it can be remarked that only in 7 out of all 747 samples the 137 Cs inventory exceeds 37 kBq/m 3 , i.e., about 1 % of the total locations.
By comparing the 137 Cs distribution maps corresponding to 1993 (Fig. 2a) and to 2016 (Fig. 2b), it can be remarked that the 1993 maxima from Toaca (Ceahlau Mountain) to Parang Mountain almost disappeared excepting small pockets around Alba Iulia and Drobeta Turnu Severin cities. In this regard, it should be considered the fact that some sampling points of 1993 and 2016 campaigns did not coincide, and, as mentioned above, in 1993 campaign, the majority of analyzed soil comes from pastures, while in 2016 campaign, soil samples were mainly collected on ploughed areas.
In spite of these facts, the total 137 Cs inventory calculated for the entire territory of Romania and corrected by the pre-Chernobyl radiocesium was of 3.6 TBq for 1993 and of 1.2 TBq for 2016 determinations (Table 1). By comparing the 2016 data with the 1993 ones, it results that during this 23 years interval, the total 137 Cs inventory corresponding to the entire Romanian territory has dropped by factor of about 3, which, at a first glance, is higher than the value of 1.7 which results for this interval of time by considering for 137 Cs a half-life period of 30 years.
These findings suggest that beside decay, other natural factors such as local levigation under the action of precipitation have a significant contribution to the diminution of 137 Cs inventory, as it was reported in the case of radioactive contamination following Fukushima accident (Taniguchi et al., 2019). Most probably, a significant amount of 137 Cs was transported throughout rivers until the Black Sea, its presence in the uppermost sediment layers being reported in the past decades since Chernobyl accident (Delfanti et al., 2014;Florea et al., 2011;Ilie et al., 2022). Also, this hypothesis is sustained by the fact that in spite of a large number of soil samples collected in the proximal vicinity of Danube River and its main tributaries (Fig. 1), the specific activity of 137 Cs was almost as in the case of other sampling points far away from river beds or floodplains (Fig. 2b).
On the other hand, given the great variability of 137 Cs inventory evidenced in both cases, this kind of determinations should be periodically repeated for the same areas to elucidate this phenomenon.
Another aspect that should be taken into account concerns the human exposure to 137 Cs gamma rays. According to Bellamy et al. (2019), for 137 Cs/ 137m Ba ground surface contamination, the reference person effective dose rates are of 5.0110 −16 and 3.0910 −16 Sv/Bq m 2 /s for newborn and adults, respectively. Given the fact that, for average rural population of which the outdoor average activity does not exceed 8 h/day, with very, very few exceptions which correspond to 1993 situation, the exposure would be no greater than 0.08 and 0.03 mSv/year for years 1993 and 2016, respectively (Table 1). This finding suggests that, on average, on the Romanian territory, the supplementary average exposure due to 137 Cs, even immediately after Chernobyl accident, does not exceed 0.15-0.2 mSv/year, significantly lower than the value of 1 mSv/year which is considered the maximum acceptable value for nonprofessional exposure. Data availability All data generated or analyzed are included in this article.
Code availability Not applicable for this section.

Declarations
Ethics approval Not applicable for this section Consent to participate Field study was carried out on agreement of local authorities. The study did not involve endangered or protected zone.

Consent for publication
All authors have consented to publish this study.