INTRODUCTION

The Antarctic continent has been considered one of the most pristine and unspoiled regions on the planet, in addition to being geographically isolated from other continents (^{Cary et al., 2010}). Despite being an area far from the major sources of radioactive elements such as the northern hemisphere and the western Pacific (^{Friedlander et al., 2005}), the atmosphere and the marine environment of the polar region have already been affected by human activities (^{Focardi et al., 1993}; ^{Legrand and Mayewski, 1997}; ^{Bargagli, 2000}). Additionally, due to the increase in human activities on the continents and the impact of climate change, the environmental conditions in this area are rapidly changing, with anthropic footprints in Antarctica being small compared to the total size of the continent, however, the impacts are not uniformly distributed (^{Bhardwaj et al., 2018}).

Throughout the world, great efforts have been made for the characterization and monitoring of radioactive elements (^{Friedlander et al., 2005}), due to their impact on ecosystems and human health, since these can be transferred through the trophic chain, resulting in a generally higher concentration within the higher levels organisms of the chain, which is known as Biomagnification (^{Connell, 1989}). Radionuclides can be of natural origins such as ²²⁶Ra, ²¹⁴Pb, ²⁰⁸Tl, ²¹⁴Bi, ²²⁸Ac, and ⁴⁰K, but they can also come exclusively from anthropogenic processes, mainly derived from nuclear explosions, such as ¹³⁷Cs (^{Ferreira et al., 2013}). Although the cessation of atmospheric nuclear tests in the 1980s, and the decay associated with its half-life and dilution, have produced a reduction in ¹³⁷Cs concentrations in recent decades (Friedlander et al., 2005), the signal of this radionuclide still exist in the marine environment. Additionally, nuclear accidents like those of Chernobyl and Fukushima, have again released large amounts of radionuclides into the environment (^{Marzano and Triulzi, 1994}; ^{Matisoff et al., 2011}).

For the Antarctic region, in the South Shetland Islands and Admiralty Bay, there are several studies of radionuclides, which have focused on a variety of matrices such as terrestrial and marine biological material, surface soils, and marine sediments (Ross et al., 1994; ^{Godoy et al., 1998}; ^{Mietelski et al., 2008}; ^{Sanders et al., 2010}; ^{Ferreira et al., 2013}). Most of which have focused on the concentration of radioactive elements and their spatial and temporal changes, but few works have focused on analyzing radiological risk, except for Godoy et al. (1998).

Although it seems to be clear that the main source of transport of radionuclides of anthropic origin, such as ¹³⁷Cs, is through atmospheric circulation (^{Godoy et al., 1998}), little is known about their transport to the deep areas of the ocean. Likewise, there is less knowledge of the sources and main transport routes of radionuclides of natural origin. Taking into account that marine sediments record materials from adjacent continental areas, ocean basins, as well as elements transported by currents or deposited on the sea surface and subsequently carried to the ocean floor, these can be useful in the study of the concentration of radionuclides.

This study aimed to characterize the concentration of radionuclides of natural and anthropogenic origin in marine sediments from localities with contrasting influences, on King George Island and Bransfield Strait (Antarctica), to establish a baseline to define the level of radiological risk for the environment and human health in the sampling areas. In the same way, from the distribution of the concentrations, the possible sources and transport routes of the radioactive elements studied were defined.

STUDY AREA

The study area corresponds to the northwestern part of the Antarctic continent, where 3 locations were visited, the Orca Submarine Mount (Bransfield Strait), Admiralty Bay, and MacKellar Inlet (South Shetland Islands) (figure 1). Mount Submarine Orca, located approximately between Latitude 62.20 to 62.30° S and Longitude 58.14 to 58.42° W, which corresponds to a volcanic site directly associated with the tectonic depression of the rift zone in a post-arc environment, the result of subduction processes between the Fenix ​​and Antarctic plates (^{Lawver et al., 1996}). It is still unknown if it is an area with hydrothermal activity, but in the Bransfield Strait área, active hydrothermalism has been identified, associated with the presence of volcanic centers, the most important being the one on Deception Island. The depths in Orca range between 600 m at the top of the mountain walls and 1600 m at the external bottoms adjacent to the slope of the mountain.

Figure 1 Geographic location of the study area on the Antarctic continent; a. showing the location of the Orca Seamount, Admiralty Bay, and Mackellar Cove; b. and details of the marine sediment collection stations in each of the locations; c. and d.. Images were taken from Google Earth. 

Admiralty Bay is located on King George Island, being the largest bay in the South Shetland Islands (Ferreira et al., 2012). The Bay in nature is similar to a fjord, with three branches: Inlet Ezcurra to the southwest, Mackellar to the north, and Martel to the northeast (Rakusa-Suszczewski, 1980; 1995). It has a surface that covers 122.08 km², of which the main body is 52.3 %, and an average depth of 198.6 m, but the deepest area can reach 535 m (Rakusa-Suszczewski, 1995).

For its part, the Mackellar Cove, as mentioned above, is located within the Admiralty Bay, to the north, and the Machu Picchu Antarctic Scientific Station, which belongs to Peru, is located there. Mackellar and the other inlets of the bay are shallower than the main body (^{Rakusa-Suszczewski, 1995}) and are more influenced by the contributions of glaciers, which have shaped their morphology. The hydrological and also hydrochemical conditions of the Inlet are very different from those of the main body of the bay, as well as the suspended material, which has higher concentrations in this area (Rakusa-Suszczewski, 1995), possibly contributed by icebergs.

MATERIALS AND METHODS

Collection and treatment of samples

The collection of marine sediment samples was carried out during the second cruise of the Twenty-sixth Scientific Campaign of Peru to Antarctica - ANTAR XXVI, carried out between January 24 and February 19, 2019, within the framework of the project “Characterization radiological analysis of the sediments extracted from the depth of the sea in the Bransfield Strait and environmental samples in the ECAMP”. The collected samples corresponded as follows: 7 sediment samples in the Orca seamount, 1 sediment sample in Admiralty Bay, and 7 samples in the surroundings of Mackellar Inlet (Figure 1). Additionally, data from radiometric measurements were obtained from marine sediment samples collected from Mackellar Inlet between 1997 and 2001 and from pits around the Peruvian station of Macchu Pichu located in Punta Crepin. The samples were collected by Van Veen dredge at bottom depths between 15 m and 30 m in Mackellar Inlet, 423 m in Admiralty Bay, and between 1080 m and 1636 m in Seamount Orca. After their collection, the samples were placed in sealed polyethylene bags and refrigerated at 4 °C for two months, and then dried at 105 °C overnight, sieved to a diameter of less than 0.5 mm, and taken to the laboratory for its conditioning and measurement.

Table 1 shows the details of the sediment sampling points extracted around the Orca seamount. 08 points were planned of which point 2 (MOS2_sed) was not collected due to 03 unsuccessful dredge sets. Approximately 2 kg of mass was collected for each sample and the dose rate was measured, in some cases being less than the natural background (0.01 µSv/h).

Table 1 Data from sediment sampling points collected around Orca Seamount. 

The details of the sampling points of the sediments extracted in the Admiralty Bay and the surroundings of the Mackellar Inlet are described in Table 2. For each sample between 0.5 kg and 1 Kg of mass approximately was collected and the dose rate measured, being in all cases slightly higher than the natural background (0.01µSv/h).

Table 2 Data from sediment sampling points collected in Admiralty Bay and interior of Mackellar Cove (Station 2-17). 

Sample Analysis

For the measurement of the radionuclides of interest:¹³⁷Cs, ²²⁶Ra, ²¹²Pb, ²¹⁴Pb, ²⁰⁸Tl, ²¹⁴Bi, ²²⁸Ac y ⁴⁰K It was conditioned between 180 g to 380 g of dry sample, in polypropylene containers of 300 cm³ volume (diameter = 8.2 cm and height = 5 cm), which were hermetically sealed, weighed in a 1 mg precision balance and stored by a month to achieve a secular balance between ²²⁶Ra y ²²²Rn (^{Ibrahiem et al., 1993}).

All measurements were performed using a p-HPGe detector with a relative efficiency of 70 %, and a resolution of 1.9 keV for the 1332.5 keV peak of the ⁶⁰Co and a peak/Compton ratio of 70. The detector is coupled to an automatic sample exchanger system, Tema Sinergie model SC 100/20, with a 100 mm thick Pb shield lined with a sheet of electrolytic Cu and another of Cd of 1 mm each. For the acquisition and spectral treatment, the Genie 2K version 2.0 program (Canberra) was used, correcting both the contribution of the natural background and the spectral interference of ²¹⁴Bi at the 661.65 keV peak of the ¹³⁷Cs

The counting time ranged from 65000 s to 120000 s with a dead time < 1 %, previously the measurement system was calibrated in energy using a standard source of ¹⁵²Eu, and in absolute efficiency using a known mass quantity of multiple standard solutions of Eckert & Ziegler, Code UA591, with calibration date April 1, 2012, containing ¹³³Ba, ⁵⁷Co, ¹³⁹Ce, ⁸⁵Sr, ¹³⁷Cs, ⁶⁵Zn, ⁸⁸Y, ⁵⁴Mn deposited in a properly dried and sieved soil sample. Considering the decay time elapsed up to the date of measurement of the samples, only the gamma peaks of ¹³³Ba (276.38 keV, 302.85 keV, 356.013 keV, 383.848 keV) and ¹³⁷Cs (661.65 keV) were feasible to be used.

For optimal counting statistics, a distance of 18 mm between the sample and the detector was established as the measurement geometry, considering a possible effect of coincidence in the gamma peaks of interest (Talavera et al., 2001; Haluk et al., 2010). Given the difference in density observed in the samples, from 0.6 g cm^-3 a 1.3 g cm^-3, the gamma attenuation effect was corrected applying the transmittance method (Cutshall, 1981), using a calibrated source of ¹⁵²Eu.

The activity concentration of the radionuclides found in the marine sediment samples was calculated using equation 1, according to Reza et al., 2012.

Where C _A is expressed as Bq kg^-1, ε is the absolute efficiency for gamma peaks of specific energy, T _V is the counting live time in seconds, W is the amount of mass expressed in Kg, B _r is the fraction of particles that disintegrate by a defined disintegration mode for each energy, according to LNHB databases (Laboratorie National Henri Becquerel) and A_neta is the net area of the evaluated gamma peaks of 186,21 keV (²²⁶Ra), 238.63 keV (²¹²Pb), 661.6 keV (¹³⁷Cs),), 351.93 keV(²¹⁴Pb), 583.18 keV (²⁰⁸Tl), 609.316 keV (²¹⁴Bi) and 1460.82 keV (⁴⁰K), calculating the uncertainty of C_A by propagating errors.All results reports, both samples and references are given in dry weight.

Determination of radiation hazard indices

The radiological risk of the radionuclides present in the sediment samples was estimated using the following indices based on the determination of activity concentration of ⁴⁰K, ²²⁶Ra (by ²¹⁴Bi), and ²³²Th (by ²²⁸Ac): the radium equivalent activity concentration index ( Ra _eq ) (equation 2), the gamma radiation hazard index ( Iγ ), the externally absorbed dose rate at 1 m above ground level ( D ) and the corresponding annual effective radiation dose (AED)

Where C _Ra , C _Th , and C _K are the activity concentrations of ²²⁶Ra, ²²⁸Ac, and ⁴⁰K in Bq kg^-1 respectively.

The value of the radiation hazard index Iγ was estimated by equation 3

The absorbed dose rate of external gamma radiation in the air at 1 m ground level was calculated according to equation 4 (UNSCEAR, 1988)

Where D is the dose rate in nGy h^-1 and C _Ra , C _Th and C _K have the same meaning as equation 1.

For the evaluation of the health effects of the absorbed doses of gamma radiation, the annual effective dose was calculated using equation 5, using the conversion of the absorbed dose to the effective dose through the conversion coefficient (0.7 Sv Gy^-1) and external load factor (0.2) (UNSCEAR, 1988),

Quality Assurance and Statistical Evaluation

To ensure the quality of the measurements obtained in this investigation, the reference samples were analyzed: CRM IAEA Soil 6 and CRM IAEA Moss Soil 447 provided by the International Atomic Energy Agency. Table 3 shows the values obtained expressed as Bq kg^-1 and their combined uncertainty calculated by the propagation of errors with coverage factor k = 1.

Table 3 Activity concentration expressed in Bq kg^-1 in dry mass of radionuclides in reference samples, uncertainty combined with coverage factor k = 1. 

IC = confidence interval at 95 %

The reported values indicated that the analytical methodology was under statistical control. The discrepancy regarding the activity concentration of ²²⁶Ra in CRM IAEA Moss Soil-447 is explained by the presence of ²³⁵U (^{Shakhashiro et al., 2012}), whose peak of 185.72 keV and 57 % gamma emission interferes in the analysis of ²²⁶Ra (186.21 keV peak) by gamma spectrometry. Such overlapping of gamma peaks cannot be resolved by the measurement system, resulting in an additional 50 % contribution in the value of ²²⁶Ra.

The minimal detectable activity (MDA) was determined according to Curie (1968). The calculated values are shown in table 4, varying by counting time, sample mass, detector efficiency, and the evaluated gamma peak background.

Table 4 MDA values in Bq kg^-1 for Mackellar Cove, Orca Seamount, and reference samples used. 

The variability between the Mackellar Inlet and Seamount Orca samples was studied to establish possible transport pathways for radionuclides in both areas. To do this, the analytical results of ²²⁶Ra, ²¹²Pb, ²¹⁴Pb, ²⁰⁸Tl, ²¹⁴Bi, ²²⁸Ac, and ⁴⁰K were processed utilizing multivariate main component extraction (PCA) and hierarchical classification methods using the SPSS statistical program.

To evaluate whether it was feasible to apply factor analysis (PCA) to the data set obtained, the KMO test (Kaiser-Meyer-Olkin) and Bartlett’s sphericity test were performed, obtaining a value of 0.718 > 0.5 and a significance = 0, respectively, showing that the data matrix (15x7) is adequate to carry out the PCA. Similarly, a simple correlation between Depth and radioisotope concentration was carried out, to assess whether the route of arrival to marine sediments was from the continent or the ocean floor.

RESULTS

The activity concentration results of the radionuclides of interest calculated for the samples from Mackellar Inlet, Admiralty Bay, and Orca Seamount, showed the presence of ¹³⁷Cs in the Mackellar Inlet area at activity levels higher than the MDA value, except for the samples from stations BA1, E02, E03 and E08 (Table 5). On the contrary, in the sediment samples from the Orca seamount area, ¹³⁷Cs were not detected, except for the sample from Station MOS4, which presented a high measurement uncertainty.

Table 5 Results of radiometric measurements in marine sediment extracted from Mackellar Inlet (E), Admiralty Bay (BA), and Orca Seamount (MOS). 

M* = Identification of sample in dendrogram

The historical results (different collection years) for sediment samples extracted from the same point in Mackellar Inlet (table 6), which were measured in the present study, were in general at levels above the MDA for ¹³⁷Cs, except for those corresponding to 1999, with a gradual decrease in the activity concentration value.

Table 6 Results of radiometric measurements in marine sediment extracted from the same sampling point in Mackellar Cove, Coordinates: 62º05’29 “S, 58º28’10” W, in different years. 

The correlation analysis from a factor analysis (PCA), for the samples from Mackellar Cove (N = 7), Admiralty Bay (N = 1), and Orca Seamount (N = 7), showed a high ratio of ²²⁶Ra, ²¹⁴Pb, and ²¹⁴Bi belonging to the ²³⁸U radioactive series. Likewise, a high linear correlation between ²¹²Pb, ²⁰⁸Tl, and ²²⁸Ac belonging to the radioactive series of ²³²Th and a minimum relation with ⁴⁰K. However, an inverse relationship was observed between ⁴⁰K and radionuclides of the ²³⁸U radioactive series. The hierarchical classification using the average link method between groups is shown in figure 2, where a grouping of stations 4 to 8 of Mackellar Inlet is observed and a second group with the rest of the stations.

Figure 2 Cluster analysis of matrix data (15x7). 

According to the PCA, the total variance in the evaluated data matrix can be explained by the first component constituted by the radionuclides of the radioactive series of ²³⁸U (²²⁶Ra, ²¹⁴Pb, and ²¹⁴Bi) with a 69.19 % variance. Meanwhile, a second component corresponding to ⁴⁰K explains 23.27 % of the variance, making together a total explained variance of 92.46 %. The analysis of the relationship between Depth and the concentration of activity of the elements (figure 3), showed a relationship directly between the variation of ²³⁸U and ²²⁶Ra as a function of Depth, while for ⁴⁰K the relationship is inverse, although with a low level of correlation. For ¹³⁷Cs it was not possible to establish a numerical relationship, but a greater number of events with a concentration greater than MDA was observed in Mackellar Inlet (table 5).

Figure 3 Relationship of Depth with a concentration of 40K. 226Ra, and 238U in the sediment samples from the different collection stations of the study. 

In the values of the radiation hazard parameters in Admiralty Bay, Mackellar Inlet, and Orca Seamount (Table 7), the lowest level was observed for Station in Admiralty Bay (BA1) and the highest value for Station 4 in Orca Seamount (MOS-4). For the radium equivalent activity concentration index (Raeq), a variation from 56.9 to 88.8 Bq kg^-1 was observed. For the radiation hazard index (Iγ), the levels varied between 0.18 and 0.27. Likewise, the values of the dose rate (D) were obtained, which varied between 31.89 and 48.35 nGy h^-1, and the values of the Annual Effective Dose (AED), which varied between 0.04 and 0.05 mSv a^-1.

Table 7 Radiation hazard parameter values at Admiralty Bay, Mackellar Inlet, and Orca Seamount. 

Maximum Allowable Limits (LMP): Ra_eq =370 Bq kg^-1; I_γ < 1;D = 55 nGy h^-1; DEA = 0.067 mSv a^-1

DISCUSSION

The concentration of radionuclide activity in the marine sediment samples indicated a significantly lower level of radioactivity for ¹³⁷Cs, compared to other areas worldwide (Table 8), which can be explained by the distance between the Antarctic continent and the source areas of this element (^{Friedlander et al., 2005}). While ⁴⁰K and ²²⁶Ra showed to be within the ranges observed in marine sediments from various locations around the world, with the exception of those reported for ⁴⁰K by Lambrechts et al. (1992) and Kurnaza et al. (2007), who showed ranges whose upper limit is above 1000 Bq kg^-1, likewise ^{Abdel-Halim and Saleh (2016)} report concentrations of ²²⁶Ra up to 499.2 Bq kg^-1 for the Alexandria coasts.

The concentrations of ¹³⁷Cs that are reported for some stations in Mackellar Inlet correspond to the levels of activity historically observed according to Table 6, considering the natural decay factor. A probable explanation for the presence of higher concentrations of ¹³⁷Cs activity in shallow marine sediments (Mackellar Inlet) with respect to the deeper zone (Orca seamount) of the study area, is based on the fact that the transport of this radioisotope is mainly by atmospheric route (^{Godoy et al., 1998}), subsequently accumulating in snow, sea ice, soils and terrestrial vegetation, which has been corroborated with the results of ^{Mietelski et al. (2008)}, who found lower levels of contamination in samples from the marine environment than those from terrestrial ecosystems. Subsequently, this material is contributed to marine sediments through land runoff and ice drift into the open sea.

Table 8 Comparison of radioactivity (expressed as Bq kg^-1) in marine sediments in the study area with other areas worldwide. 

The variability of naturally occurring radionuclides both in Mackellar Inlet and Orca Seamount is defined mainly by radionuclides of the radioactive series of ²³⁸U (²²⁶Ra, ²¹⁴Pb, ²¹⁴Bi) and by ⁴⁰K, meanwhile, those of the radioactive series of ²³²Th (²¹²Pb, ²⁰⁸Tl, ²²⁸Ac) show general homogeneity in both zones. According to the cluster analysis, the MOS4, MOS5 and MOS6 stations are conveniently grouped considering the Location of such stations in that locality (figure 1d).

The activity concentration of ²³⁸U and ²²⁶Ra found, suggests an origin associated with volcanic activity in the study area, the values ​​of ²²⁶Ra coincide with that reported by ^{Iyengar (1990)}, for different types of rocks with this origin. The direct relationship of ²²⁶Ra concentrations with Depth confirming the geological evolution of the area proposed by several authors (^{Schreider et al., 2015}; ^{Kozlenko and Kozlenko, 2019}) and the temporal changes from the source of material to the mantle, which alters the geochemistry of volcanic rocks from different areas of the South Shetland Islands (^{Fretzdorff et al., 2004}; ^{Lee et al., 2008}). The highest ²²⁶Ra values ​​(29.6-38.8 Bq kg^-1) were found in the deepest and most recently active zone, Orca seamount, in which alkaline magmatic or volcanic igneous rocks are found (Fretzdorff et al., 2004; Lee et al., 2008), which have a concentration of ²²⁶Ra between 11.1 and 48.1 Bq kg^-1 according to Iyengar (1990), and a lower concentration (18.5-24.8 Bq kg^-1) in the shallower zone (Admiralty Bay and Mackellar Cove ) in Where ^{Birkenmajer (1980)} identified andesitic and rhyolitic lavas considered intermediate volcanic rocks, for which Iyengar (1990) reported ²²⁶Ra values of 18.5 Bq kg^-1.

For their part, the ⁴⁰K activity concentrations found in this study in Admiralty Bay and Mackellar Inlet, coincide with those reported by ^{Godoy et al. (1998)}, who registered values between 402 and 607 Bq kg^-1. The ⁴⁰K as the ²³²Th and the ²³⁸U is available in magma and thus enters the crystalline structure of minerals, because during its transport process after the disintegration of the rock it does not undergo significant enrichment or depletion, it provides a species radioactive label of the source rock (^{Anjos et al., 2007}). The inverse relationship between ⁴⁰K and Depth suggests the continental zone as a source of this element, possibly associated with the type of volcanic rocks that make up King George Island, which are different from the Orca Seamount, as stated above. Taking into account that the Orca seamount does not show current volcanic and hydrothermal activity, based on the explorations carried out in the area on the ANTAR XXVI and XXVII cruises, and the records of ^{Klinkhammer et al. (1996)} and Klinkhammer et al. (2001), it is necessary to explore in future works the concentration of activity of naturally occurring radionuclides in other areas of the Strait with active hydrothermalism, to define whether there are spatial changes derived from this condition.

Regarding the indices used to evaluate radiation hazard, it was found that all concentrations of radionuclides were below the maximum permissible limits (MPL) suggested by UNSCEAR (1988), which indicates that although there is the presence of these radionuclides, they are not yet present, represents a health hazard. For the radium equivalent activity concentration index (Raeq), the values were below the limit set at 370 Bq kg^-1, for the radiation risk index (Iγ) the values were less than unity. Likewise, the dose rate values (D) were below 55 nGy h^-1 and the Annual Effective Dose (AED) values did not exceed the maximum value of 0.067 mSv a^-1 as suggested by UNSCEAR (1993).

Taking into account the extension of the Antarctic continent and the accumulation of ¹³⁷Cs in the ice, as well as its subsequent contribution to the adjacent ocean and sediments, it is possible to think that Antarctica can be considered a large reservoir of radioactive elements of anthropic origin. ^{Pourchet et al. (2003)} set the activity of ¹³⁷Cs deposited throughout Antarctica at 760 TBq (terabecquerels), this is equivalent to around 20,000 Ci, although it is a very large amount of activity, it only represents 0.08 % of the deposited ¹³⁷Cs worldwide according to the same reference. Derived from the above, it can be said that the melting and release of these elements due to climate change may represent a danger for the organisms of this region and humanity, therefore measurements in sediment cores can be useful in evaluating the contribution of these radionuclides over time and their relationship with climate changes in the region, to predictive models.