Delayed inputs of hot 137Cs and 241Am particles from Chernobyl to sediments from three Finnish lakes: implications for sediment dating

Anomalous peaks in the 137Cs and 241Am records from three lake sediment cores from southern Finland appear to be due to the presence of micron-size hot radioactive particles, almost certainly originating in fallout from the 1986 Chernobyl accident. Since the imbedding sediments all post-date 1986 by several decades, it appears that they were initially deposited on the catchments of these lakes and transported to the lakes some years later. The activities of the particles were determined using a sequential splitting process. Two of the particles were found to contain 137Cs with activities of 64 ± 4 mBq and 266 ± 15 mBq respectively. The third contained 241Am with an activity of 17 ± 2 mBq, but no evidence of significant amounts of 137Cs. The delayed input of such particles into the sedimentary records highlights the need for care in using 137Cs or 241Am as chronostratigraphic markers in areas subject to significant levels of contamination from Chernobyl fallout.


Introduction
Artificial radionuclides such as 137 Cs and 241 Am play an important role in dating lake sediments, and the wide range of environmental records they contain. Matches between well dated features in the 137 Cs (or 241 Am) fallout record and clearly identifiable features in the activity versus depth record are used to date specific sediment layers (Foucher et al. 2021). Although of limited value on their own, these chronostratigraphic dates play an important role in validating the more detailed chronologies obtained from 210 Pb records, or correcting 210 Pb dates in the event of any discrepancies (Appleby 2001). Since the early 1970s, concentration peaks identified as recording the period of maximum fallout from the atmospheric testing of thermonuclear weapons have been used to identify the 1963 sediment depth (Pennington 1981;Appleby et al. 1991). In many parts of north-west Europe, including Finland (Arvela et al. 1990), high levels of fallout from the Chernobyl reactor fire have resulted in the presence of a second peak that can be Abstract Anomalous peaks in the 137 Cs and 241 Am records from three lake sediment cores from southern Finland appear to be due to the presence of micronsize hot radioactive particles, almost certainly originating in fallout from the 1986 Chernobyl accident. Since the imbedding sediments all post-date 1986 by several decades, it appears that they were initially deposited on the catchments of these lakes and transported to the lakes some years later. The activities of the particles were determined using a sequential splitting process. Two of the particles were found to contain 137 Cs with activities of 64 ± 4 mBq and 266 ± 15 mBq respectively. The third contained 241 Am with an activity of 17 ± 2 mBq, but no evidence of significant amounts of 137 Cs. The delayed input of such particles into the sedimentary records highlights the need for care in using 137 Cs or 241 Am used to identify the 1986 depth (Erlinger et al. 2008). Fallout from this source took place during a relatively short period of time, from 26th April 1986 through to the middle of May (Persson et al. 1987). Radioactive debris deposited directly onto the surface of a lake will in most cases have a relatively short residence time in the water column. Although some will exit the lake via its outflow, a significant fraction will be delivered to the bed of the lake and incorporated into the sediment record for that year. The record in subsequent years will contain further inputs from fallout deposited on the catchment and delivered to the lake by various catchment/lake transport processes (Appleby et al. 2019). Inputs via this indirect pathway will generally be much lower than direct inputs at the time of the accident, and sediments with highest concentrations of Chernobyl radionuclides can for the most part confidently be dated to 1986. Although concentrations in post-1986 sediments frequently vary in response to factors such as different levels of catchment runoff, or post-depositional migration coupled with preferential absorption onto fine-grained clays, small features in this part of the record are for the most part easily distinguished from the 1986 peak itself. In a recent exception to this, the 137 Cs record in a core from a Norwegian fjord was found to contain a very distinct peak in a sediment sample deposited several decades after 1986 (Appleby et al. 2022). Detailed analyses showed that this feature was due to the presence of a single micron-size hot 137 Cs particle. Although its activity was just 15 mBq, the 137 Cs concentration in this near surface sample (15.6 Bq kg −1 , 61% of which was due to the particle) was twice as high as in the slice containing the concentration peak attributed to direct fallout from the 1986 event (7.4 Bq kg −1 ). The most likely origin of the particle was fallout onto the landscape during the passage of the Chernobyl cloud and transport to the fjord some decades later, though in this case a second potential source was discharge into the marine environment from nuclear installations such as those at Sellafield or Dounreay in the UK.
The study by Arvela et al. (1990) showed that although there were significant amounts of Chernobyl fallout in many parts of Finland, highest levels were mainly recorded in the south of the country (Fig. 1). In a recent palaeolimnogical study of lakes in southern Finland, three sites were found to have anomalous features in their Chernobyl fallout records. At two sites, Koirusjärvi and Nuasjärvi, the 137 Cs record had two distinct peaks. At a third site, Kiantajärvi, a single sample had an unusually high 241 Am activity.
Much of the Chernobyl fallout over Finland was carried on particles in the sub-micron range (Kauppinen et al. 1986), though there were reports that it did include "hot particles" of up to 10 µm in size (Anttila et al. 1987;Raunemaa et al. 1987;Saari et al. 1989). The term "hot particle" is not well defined (Sandalls et al. 1993), but in this context is normally used to describe highly radioactive particles up to around 10 µm in diameter released into the environment from a nuclear accident or device. Although activity levels in particles released by the Chernobyl accident were initially dominated by short-lived radionuclides such as 141 Ce, 144 Ce, 95 Zr, 95 Nb, 103 Ru and 106 Ru, in the longer term they are mainly due to 137 Cs, and to a lesser extent, 241 Am. The objectives of  Arvela et al. (1990). Also shown are the locations of the study sites, and a comparison lake, Karipääjärvi, from a region with a low level of Chernobyl fallout. Results from Karipääjärvi have been reported earlier in Ruppel et al. (2013Ruppel et al. ( , 2015 the present paper were to investigate the causes of the observed anomalies in the sediment records, the possibility that they might be due to delayed inputs of hot particles from the catchment, and any possible implications for the use of 137 Cs and 241 Am as chronological markers in dating lake sediments.

Site description
The study lakes (Nuasjärvi, Kiantärvi and Koirusjärvi) are all located in the southern boreal to middle boreal vegetation zones in central Finland (Fig. 1). Mean annual precipitation (1991Mean annual precipitation ( -2020 for the area ranges from 585 to 678 mm. Average temperatures  lie between − 8.2 and − 9.5 °C during January, and 16.2-17.1 °C during July (Jokinen et al. 2021). The lakes have ice cover from November to May. All have relatively large catchment areas, though the exceptionally large value for Koirusjärvi reflects the fact that it is hydrologically connected to a much larger lake, Kallavesi. For more detailed physiogeographical and limnological data, see Table 1.

Sample collection and preparation
Sediment cores were retrieved from the sedimentation basins of the lakes in October 2020 using a HTHkayak gravity corer (Renberg and Hansson 2008). The cores were subsampled on site at 0.5-cm intervals between 0 and 5 cm depth, and at 1-cm intervals between 5 and 20 cm depth. All samples were stored in plastic ziplock bags in a dark cold room at +4 °C within eight hours of retrieval. The samples were freeze-dried prior to radiometric analysis. Dry bulk densities were determined from the water content data.

Radiometric analysis
Subsamples from each core were analysed by direct gamma assay using Ortec HPGe GWL series welltype coaxial low background intrinsic germanium detectors (Appleby et al. 1986) in the Liverpool University Environmental Radioactivity Laboratory, preparatory to dating by 210 Pb and 137 Cs. 210 Pb was determined via its gamma emissions at 46.5 keV, and 226 Ra by the 295 keV and 352 keV γ-rays emitted by its daughter radionuclide 214 Pb following 25 days storage in sealed containers to allow radioactive equilibration. 137 Cs and 241 Am were measured by their emissions at 662 keV and 59.5 keV respectively. Detection limits were ~ 8 mBq for 210 Pb, ~ 5 mBq for 226 Ra and ~ 2 mBq for 137 Cs and 241 Am. The absolute efficiencies of the detectors were determined using calibrated sources and sediment samples of known activity. Corrections were made for the effect of self-absorption of low energy γ-rays within the sample (Appleby et al. 1992). Supported 210 Pb in each sample was assumed equal to the 226 Ra activity, and unsupported (fallout) 210 Pb calculated by subtracting this from the total 210 Pb activity.

Results
Results of the radiometric analyses carried out on each core are shown in Figs. 2, 3, and 4. Table 2 lists some of the basic radiometric parameters for each site, including the maximum unsupported 210 Pb activity, unsupported 210 Pb inventory, mean 210 Pb supply rate, and 137 Cs inventory. The mean annual atmospheric 210 Pb flux is estimated to be significantly less than 100 Bq m −2 y −1 (Paatero et al. 2015). Significantly higher 210 Pb supply rates at all three core sites can be attributed to factors such as sediment focussing, and allochthonous inputs of fallout deposited on the catchment. All three sites have relatively large catchment areas compared to the size of the lake.
The unsupported 210 Pb activity versus depth records all deviated significantly from a simple  (Appleby and Oldfield 1978). This has generally proved to be the most reliable means for dating non-exponential records. Because of an apparent hiatus in the Kiantajärvi record, for this core the model was applied in a piecewise way using the methods outlined in Appleby (2001). A table of dates for each site, including their uncertainties, is given in Viksted et al. (2022, p. 68). The 137 Cs inventories have been corrected for decay since 1986, the year of fallout from the Chernobyl nuclear accident. The high 137 Cs inventories, most notably at Koirusjärvi, confirm that all three sites were significantly impacted by fallout from the 1986 Chernobyl accident. This is further supported by the fact that the values of the inventories are consistent with the fallout data shown in Fig. 1. Data from Finnish sites less affected by Chernobyl fallout suggest that contributions from nuclear weapons test fallout are unlikely to be more than around 1000 Bq m −2 .
The 137 Cs record in the Koirusjärvi core (Fig. 2a), the most heavily impacted site, has two well-defined peaks, in the 6-7 cm and 11-12 cm sections. Although these features would normally be assumed to record two distinct events, fallout from the 1986 Chernobyl accident and the 1963 fallout maximum from the atmospheric testing of nuclear weapons, the high concentrations (> 1500 Bq kg −1 ) suggest that both peaks contain 137 Cs originating in fallout from the Chernobyl accident, with the deeper peak in all probability dating from that time. This is confirmed by the 210 Pb dates, which place 1986 within the 11-12 cm section. Sediments in the 6-7 cm section containing the more recent 137 Cs peak were dated to 2009 ± 2 years, more than two decades after the Chernobyl fallout event. The year of maximum fallout from the atmospheric testing of nuclear weapons, 1963, is placed within the 14-15 cm section. The absence of a 137 Cs peak at this depth can be attributed to downwards migration of the very high levels of Chernobyl 137 Cs (Klaminder et al. 2012). To determine whether the original 6-7 cm subsample was representative of the slice as a whole, a second subsample from the same slice was analysed and found to have a similar 210 Pb activity but an 11% lower 137 Cs activity. The possibility that the discrepancy was due to the presence of a hot particle was examined using a sequential splitting method (e.g. Falk et al. 1988;Pöllänen et al. 1999). The original higher activity subsample was divided into two equal parts that were then reanalysed. They again had similar 210 Pb activities, but this time there was a 40% discrepancy in the 137 Cs concentrations, 2049 ± 20 Bq kg −1 , compared to 1477 ± 20 Bq kg −1 . Repeated splitting of the high activity subsamples followed a similar pattern. By the fifth split the 137 Cs concentration in the high activity half had increased to 5994 ± 109 Bq kg −1 , compared to 1334 ± 64 Bq kg −1 in the low activity half. Since the low activity halves all had similar 137 Cs concentrations, the most likely explanation of these results is that the 137 Cs peak in the 6-7 cm sample was mainly due to the presence of a single hot 137 Cs particle. Excluding this particle, 137 Cs activity appeared to be uniformly distributed throughout the sediment matrix. The mean concentration in the matrix based on a weighted average of measured activities in the low activity splits was calculated to be 1441 ± 81 Bq kg −1 . This is very similar to the mean of the 137 Cs activities in the adjacent 5-6 cm and 7-8 cm samples, 1270 ± 14 Bq kg −1 and 1442 ± 16 Bq kg −1 , respectively.
For each high activity split the total 137 Cs activity can be divided into two components, one due to the sediment matrix and the other due to the presumed hot particle. The activity due to the matrix is calculated by multiplying the mass of the split by the 137 Cs concentration in the matrix. Excluding the value from the first split where the uncertainty was necessarily very large, the results, shown in Table 3, yield a consistent value for the activity of the hot particle of between 256 and 281 mBq with a weighted mean of 266 ± 15 mBq. Higher weights were given to smaller splits where the uncertainties were necessarily smaller. A similar analysis of concentrations in the 11-12 cm section showed that in this case the a b Am concentration was recorded in the 6-7 cm sample. The markers show the 241 Am concentrations in the first three splits containing the hot particle   (Fig. 3a) also has two distinct peaks, a broad feature between 7 and 10 cm, and more narrowly defined feature within the 4-4.5 cm sample. There is also a small peak in 241 Am concentrations slightly deeper in the core, between 9 and 12 cm. High concentrations again suggest that both 137 Cs peaks originate in fallout from the 1986 Chernobyl accident, with the deeper and more substantial feature being a direct record of that event. This is confirmed by the 210 Pb results, which place 1986 at a depth of around 8 cm. Since the 210 Pb dates also place 1963 at a depth of around 10 cm, the 241 Am peak most probably records the early 1960s peak in fallout of its parent radionuclide 241 Pu from the atmospheric testing of nuclear weapons (Appleby et al. 1991). The relatively small gap between the 1963 and 1986 dates is attributed to a much lower sedimentation rate at that time. The CRS model calculations suggest a value of around 0.1 cm y −1 , compared to 0.24 cm y −1 for the post-1986 period. The absence of a 1963 137 Cs peak can again be attributed to downwards migration of Chernobyl 137 Cs, as can the presence of significant amounts of 137 Cs at depths predating the 1953 onset of global fallout from the atmospheric testing of nuclear weapons.
Sediments in the 4-4.5 cm section containing the more recent 137 Cs peak were dated to 2009 ± 2 years. An analysis carried out on a second subsample from this section suggested that the anomaly was again due to the presence of a single hot particle, and this was confirmed by carrying out a sequential splitting analysis as described above. During this process 137 Cs concentrations in the hot splits increased from an initial value of 593 ± 8 Bq kg −1 to 1952 ± 58 Bq kg −1 in the 5th split (Table 4). The results again suggested that, excluding the hot particle, 137 Cs was relatively uniformly distributed throughout the section with a mean concentration of 487 ± 19 Bq kg −1 . Using this value, calculations of the 137 Cs activity in each hot split due to the presumed hot particle yielded a consistent value of between 57 and 64 mBq, with a weighted mean of 64 ± 4 mBq. A similar analysis of sediments in the 8-9 cm section showed that in this case the 137 Cs activity was uniformly distributed throughout the sample and that the high concentrations were due to elevated levels of fallout. Successive splits had relatively similar 210 Pb and 137 Cs concentrations.
Results from the Kiantajärvi core suggest that there is a hiatus in the sediment record at a depth of around 12 cm. At this depth there is an abrupt jump in unsupported 210 Pb concentrations from 57 ± 17 Bq kg −1 in the 12-13 cm sample to 337 ± 22 Bq kg −1 in the 11-12 cm sample, and a similar jump in 137 Cs concentrations from 76 ± 4 Bq kg −1 below to 337 ± 8 Bq kg −1 above. Although there is no clear feature identifying the 1986 Chernobyl event, high 137 Cs concentrations in all samples above 12 cm suggest that they all post-date 1986. Near constant values in the unsupported 210 Pb concentrations suggest that sediments above the hiatus span no more than a few years. One of the samples within this part of the core, at 6-7 cm, had an unusually high 241 Am activity of 25 ± 1 Bq kg −1 . At all other depths 241 Am concentrations were below detection limits. Since fallout from the Chernobyl accident was known to include hot  (Falk et al. 1988), a sequential splitting analysis was carried out to determine whether that could account for the high activity in this sample. The results, shown in Table 5, were consistent with the presence of a single hot particle. At each stage of the analysis one of the splits was shown to contain all of the 241 Am activity, and the other negligible 241 Am activity. Calculated values of the activity in the hot split had a consistent value of between 15 and 21 mBq. The weighted mean was 17 ± 2 mBq. There was no evidence of significant 137 Cs activity on the hot particle, possibly indicating that it was a Group 1 particle as defined by Falk et al. (1988). All of the splits had essentially the same 137 Cs concentration.
Although there are large uncertainties over the dating of this core, it is almost certain that sediments containing the hot particle were deposited at some time during the past decade.

Discussion and conclusions
Although their precise origin is to some extent immaterial, the hot particles reported here almost certainly originated in fallout from the Chernobyl cloud during its passage over Finland in late April and early May 1986. All three cores are from areas known to have been heavily impacted by fallout from Chernobyl. Further, measurements carried out in the immediate aftermath of the accident showed that hot particles from the cloud were deposited in Sweden (Persson et al. 1987) and Finland . Releases from the Ignalina nuclear power plant in Lithuania (Marčiulionienè et al. 2015) appear to have been relatively localised. Atmospheric releases from other nuclear installations in the Former Soviet Union (Suokko and Reicher 1993) were almost certainly either too localised or too remote. Although the Kyshtym accident in 1957 did release a large amount radioactive material into the atmosphere, the prevailing winds carried it mainly in a north-north east direction from that site (Norwegian Radiation Protection Agency 2007).
A study of hot particles collected one year after the Chernobyl accident (Falk et al. 1988) showed that they included fragments from the damaged reactor as well as highly radioactive inclusions formed within the fuel elements. Although initially the radioactivity would have been mainly associated with short-lived fission products, only longer-lived products such as 137 Cs and some transuranic elements including 241 Am are presently above detection limits. One such particle from Sweden (Falk et al. 1988) had a 137 Cs activity of 400 mBq (179 mBq decay corrected to 2021) but negligible 241 Am. Another from the Gotland Deep in the Baltic Sea (Pöllänen et al. 1999) contained both 137 Cs (59 mBq decay corrected to 2021) and 241 Am (100 mBq). Particles deposited in Finland appear to have been of a similar size, up to around 10-µm in diameter . Assuming a similar composition to those collected in Sweden, the Nuasjärvi and Koirusjärvi 137 Cs particles are estimated to have diameters of between 6-11 µm and 10-18 µm respectively, and the Kiantajärvi 241 Am particle a diameter of around 6 µm.
Hot particles from the study sites account for only a small fraction of the 137 Cs inventory in the sediments. The sediment matrix accounted for 97% of the total 137 Cs activity in the core slices from Nuasjärvi checked for homogeneity. In the Koirusjärvi core the corresponding figure was 96%. There was no evidence of the presence of hot particles in the remaining slices, in all probability they are likely to account for less than 1% of the total 137 Cs inventory. Since particles deposited directly onto the surface of the lake would have been either incorporated in the sediment record for 1986 or lost from the lake via the outflow, those detected in the present study were almost certainly deposited on the catchment and only transported to the lake some years later. Studies of catchment/lake transport (Appleby et al. 2019) have shown that significant quantities of 137 Cs deposited on the catchment may continue to be transported to the lake decades after the last fallout event. Regardless of contribution or pathway, these findings do, however, show that at sites significantly impacted by fallout from the Chernobyl accident, delayed inputs of hot particles originating in fallout from the 1986 Chernobyl accident can have a disproportionate impact on 137 Cs and 241 Am records and hence on their reliability as chronostratigraphic markers. This has relevance not just for the 137 Cs/ 241 Am dates themselves but perhaps most importantly for their use in validating 210 Pb chronologies. It is widely recognized that neither of the standard 210 Pb models can be relied on absolutely and independent validation is essential, even where the 210 Pb dates appear to be unequivocal. The importance of reliable chronostratigraphic dates for validating 210 Pb dates is highlighted by data from Karipääjärvi, a small lake around 500 km further north in Finnish Lapland (Fig. 1). Figure 5 shows radiometric records from a core collected from this site in 2010. The results have been previously reported in Ruppel et al. (2013Ruppel et al. ( , 2015. The 137 Cs record has two distinct peaks, within the 1-1.25 cm and 3-3.25 cm sections, respectively. The presence of a small 241 Am peak within the 3-3.25 cm section showed that the earlier peak in this case recorded the 1963 weapons test fallout maximum. Confidence in the reliability of the 137 Cs date allowed it to be used as a reference point (Appleby 2001) to correct a small but significant discrepancy with the raw 210 Pb dates. a b Fig. 5 Fallout radionuclides in the Karipääjärvi sediment core showing a 137 Cs and 241 Am and b unsupported 210 Pb concentrations versus depth. Substantially lower levels of Chernobyl fallout are reflected in the greatly reduced 137 Cs concentrations compared to the study sites. Although the 137 Cs record is superficially similar to that at Koirusjärvi (Fig. 2a), association of the deeper peak at 3-3.25 cm with a similar but smaller 241 Am peak suggests that this feature most probably records the 1963 fallout maximum from the atmospheric testing of nuclear weapons. The 210 Pb calculations, which take account of dilution due e.g. to episodes of more rapid sedimentation, suggest that the later 137 Cs peak almost certainly records fallout from the 1986 Chernobyl accident The superficial similarity of the Karipääjärvi records to those from Koirusjärvi shows that such confidence is not always justified and that 137 Cs dates can be misleading. Where there is any doubt, samples containing potentially anomalous features should be investigated more thoroughly for possible causes other than direct fallout.