Introduction

The Krasnoyarsk Mining and Chemical Combine (MCC) is situated on Yenisei River 60 km downstream from Krasnoyarsk city and started operation in 1950. It includes three reactors: AD (operating period 1958–1992), ADE-1 (1961–1992), and ADE-2 (1964–2010). All three are thermal-neutron uranium-graphite channel-type reactors. The AD and ADE-1 reactors produced exclusively weapon-grade plutonium and worked in straight-through mode, discharging cooling waters directly into the river. The ADE-2 reactor had a closed-type cooling system, and aside from plutonium production, it served as a source of steam for electricity production and water heating for heating stations and the electrical grid of Zheleznogorsk city. In 1964, a radiochemical plant began operation as a part of the MCC. Its main task consisted of plutonium extraction from irradiated uranium. The plant also included a spent fuel regeneration factory. The ‘Severny’ emplacement area was designed to accept and dispose of liquid low-level and intermediate-level radioactive waste and entered into service in 1967. In 2015, a new commercial mixed oxide (MOX) fuel fabrication facility (MFFF) was launched. Currently, dismantling at AD and ADE-1 is almost completed, and decommissioning activities by radiation-proof on-site disposal methods are in progress. The ADE-2 reactor currently operates in shutdown mode.

The impact of the Krasnoyarsk Mining and Chemical Combine on the Yenisei River floodplain environment has been studied for more than 40 years after 137Cs anomalies of 15 kBq m−2 (four to eight times the global fallout) were found in the bottom sediments of the Kara Sea in 1971 (Vakulovsky et al., 1995). Sampling floodplain sediments between 6 and 250 km downstream of MCC in 1973 revealed high contamination with artificial radionuclides, at 460 kBq m−2 137Cs to 90 kBq m−2 (Nosov et al., 1993; Vakulovskii et al., 2008; Vakulovsky et al., 1995). The calculated mean annual 137Cs and 90Sr fluxes into the system in 1985–1995 were, respectively, 6.0 and 1.6 TBq year−1 in the proximal zone within 15 km away from the MCC and 1.8 and 1.5 TBq year−1 in the distal zone at 1360 km downstream (Vakulovsky et al., 1995).

Systematic studies of the MCC-derived radioactive pollution (mainly gamma-emitting 137Cs) of floodplain components began in 1991, shortly before the straight-through reactors were stopped. Then the bottom sediments near the MCC contained 0.3–1.5 kBq kg−1 137Cs (Nosov, 1996). After the reactors had been stopped in 1994, 137Cs decreased slightly to 0.2–1.0 kBq kg−1 and 370 kBq m−2 (Nosov, 1996). Although the long-lived isotopes in the floodplain decreased to n\(\cdot\)10−2 Bq L−1, up to 80% of the 137Cs transported to the river system was adsorbed onto particulate matter, while the total contamination depended on secondary release from polluted bottom and floodplain sediment (Nosov, 1996). In general, the total contamination of the floodplain soil and bottom sediments varied little, but for a 5–15% decrease by short-lived isotopes decay (Bolsunovsky et al., 2002; Nosov, 1996).

The floodplain outside the flooded areas has been free from the MCC-related contamination, with the soil 137Cs level about 5–12 Bq kg−1 global fallout (Bolsunovsky et al., 1999; Nosov, 1996), except for local anomalies produced by winter effects, such as rafting of mud-bearing ice (Nosov, 1996). The contamination of bottom sediments decreases away from the MCC, and the specific activity of radiocaesium depends on the sediment composition and depth (Bondareva, 2012; Semizhon et al., 2010). Here, 137Cs is higher in the lower strata (Standring et al., 2009) and is the least mobile in shallow sediments (5–15% of mobile forms) among other long-lived isotopes (Bondareva, 2012; Bondareva & Bolsunovskii, 2008).

The Yenisei floodplain sediments contain hot particles (60–70 per km2), which affect the mobile forms of 137Cs (Bondareva & Bolsunovskii, 2008). The 137Cs specific activity in some hot particles can reach 30 MBq per particle (Bolsunovsky & Melgunov, 2014; Bolsunovsky & Tcherkezian, 2001; Bolsunovsky et al., 2017; Gritchenko et al., 2001; Sukhorukov et al., 2000, 2004, 2009). In natural conditions, hot particles dissolve gradually upon interaction with percolating water or pore moisture, while the 137Cs they carry become involved in secondary migration (~ 0.5% per flood cycle) (Chuguevskii et al., 2010; Chuguevsky, 2019).

Pollution has also been studied in water biota. Contamination with long-lived gamma-emitting isotopes was found in fish caught upstream and far downstream of the effluent discharge (Bolsunovsky & Dement’ev, 2010; Nosov et al., 1993; Vakulovsky et al., 1995; Zotina et al., 2011). In aquatic plants growing near the MCC, the highest 137Cs activity concentration reached 660 Bq kg−1 in 1997–1998, 350 Bq kg−1 in 1999–2009, and as low as 70 Bq kg−1 in 2010 (Bolsunovsky, 2004; Bolsunovsky et al., 2002, 2011). Cytogenetic and other toxic effects in aquatic plant cells showed a positive correlation with 137Cs in the root layer of bottom sediments (Medvedeva et al., 2014), while 137Cs specific activity correlated inversely with shoot length (Zotina et al., 2014). Aerial plant parts and fungi contained up to 300–520 Bq kg−1 of 137Cs in 1993–1994 (Nosov, 1996). Concentration of 137Cs did not decrease significantly in water biota of the Yenisei after the shutdown of the last reactor plant because the discharges of this radionuclide to the Yenisei continued at the same level. On a longer-term scale (since 1973 and since 1991), concentration of 137Cs in fish muscle had significantly decreased, following the decrease in annual discharges of this radionuclide to the Yenisei, and the ecological half-life of 137Cs was estimated as 6.5–12.8 years (Zotina et al., 2019).

Since the beginning of radioactive pollution research, the 137Cs inventory in the Yenisei floodplain sediments has been a subject of considerable interest. It was estimated to be 23 GBq in the Atamanovsky Spit, 5 km downstream of MCC (Sukhorukov et al., 2000), 145 GBq in Beryozovy Island at 19 km, 148 GBq in Mikhin Island at 180 km, and 16 GBq in Cheryomukhovy Island at 250 km (Linnik et al., 2006). The total measured between the MCC and Dudinka (located 1900 km away) in 1972 was 5300 GBq (Tertyshnik, 2007).

Most studies have focused on gamma-emitting radionuclides, meaning data reports on radiostrontium in different floodplain components are limited to fewer publications. Monitoring during 1997–2002 (Sukhorukov et al., 2004) showed an uneven distribution of 90Sr along a soil profile near the MCC. The 90Sr specific activity varied from 8 to 6000 Bq kg−1 as a function of depth at sites in close proximity to the MCC and decreased markedly away from the combine to 40–350 Bq kg−1, and down to 10–20 Bq kg−1 in soil samples from islands and the river’s east bank. The activity was noted to be higher (50 Bq kg−1) in a bay of the Balchug Channel. Relatively high values were measured locally in the near impact zone (620 and 210 Bq kg−1) and in non-flooded islands (110 Bq kg−1), possibly due to atmospheric fallout (Sukhorukov et al., 2004). A similar 90Sr pattern was observed in bottom sediments, uneven spatial distribution and a general decrease away from the MCC, from 100 to 180 Bq kg−1, or locally up to 250 Bq kg−1, to within 10 Bq kg−1 (Sukhorukov et al., 2004).

According to modelling, 90Sr influxes to the Kara Sea during 1992–2000 were controlled by run-off from the drainage basin rather than by the operation of the MCC, which was responsible for no more than 12% of total 90Sr flux (Platovskikh et al., 2003). A large portion of 90Sr (20–30%), more than 137Cs, was released from water-logged floodplain soil, i.e. Sr transport into the Yenisei was mostly with intrasoil run-off (Legin et al., 2008). Radiostrontium has high migration mobility in shallow bottom sediments, which carry up to 60–70% of mobile forms (Bolsunovsky & Melgunov, 2014).

Previous studies have revealed an intricate pattern of radionuclide contamination of the Yenisei floodplain. The environment generally recovers as the contaminated floodplain deposits become buried under more recent clean sediments. However, no systematic studies of terrestrial vegetation in the near impact zone of the MCC have been undertaken before, especially in relation to the soil and rhizosphere. Living plants can liberate fixed elements and play an important role, either positively or negatively, in the self-cleaning action of a floodplain ecosystem. In this respect, it is important to constrain the redeposition of radionuclides in the flood plain soil, their partitioning between the soil and rhizosphere, the bioavailability of mobile and fixed forms, and the controls of their distribution among the root layer in the soil, the rhizosphere, and the aerial parts of plants.

Sampling and methods

Sampling sites

During August–September 2004, 2011–2016, we sampled different components of the floodplain ecosystem at five sites on islands and on the banks of the Yenisei in the near-impact zone of the Krasnoyarsk MCC (Fig. 1).

Fig. 1
figure 1

Study area and sampling sites. A Shivera Village area, reference site, 5 km upstream of effluent; B: Atamanovskaya Spit, 6 km downstream of effluent; C: Atamanovsky Island, 7 km downstream of effluent; D: Beryozovy Island, 15 km downstream from disposal site; E: Balchug Channel, 18 km downstream of effluent (Google Earth images)

Full size image

Site A (reference area): Potekhin Island and the Yenisei west bank, situated 200 m away from Shivera Village and 8 km upstream of the discharge. Both sampling points are located on the beach, which is covered with thick grass and becomes flooded systematically with short heavy floods.

Site B: upstream (upper) and downstream (lower) ends of Atamanovskaya Spit (an island), near the right bank of the river, located 6 km downstream of the effluent discharge from the reactor cooling systems. Yearly floods have produced a thick layer of alluvium covered with thick grass and shrubs. The island is located in the impact zone of the MCC wastewaters that the Yenisei current diverges towards the right bank. During the floods, which cover the whole spit area, the flow regime is short and intense.

Site C: upper and lower ends of Atamanovsky Island, 7 km downstream of the discharge. Sampling was close to the river, which becomes flooded at minor rises in water level. The island comprises two zones, grown, respectively, with grass and shrubs and with trees (including pine plantations); the former is flooded yearly or periodically, with short and heavy floods, and the latter never floods, even at the high water levels.

Site D: the upper end of Beryozovy Island, towards the west, 15 km downstream of the discharge. Most of the island is covered with mixed forest and eludes yearly flooding, but the area near the water, with grass vegetation, is flooded more or less systematically, with short heavy floods.

Site E: swampy stagnant bay on a meander of the Balchug Channel, 18 km downstream of the discharge. The channel separates Beryozovy Island from the east bank of the Yenisei. The riverside becomes swampy and accommodates particulate matter during low floods or acts as an additional channel for the stream during short high floods.

Sampling procedure and preconditioning of samples

The sampled plants mainly belonged to the genus of sedge (Carex L.): mostly Carex riparia Curt. and a minor portion of Carex versicaria L. It was simple to select multiple areas with similar conditions as sedge grows profusely in the Yenisei floodplain, on the islands, and riverside. Sedge can tolerate long floods, and as flooding recedes, it can restore its above-ground biomass quickly. Furthermore, sedge has a developed root system that regenerates rapidly after root mass has been withdrawn, and thus, it can be sampled each year at very nearly the same place.

Areas of maximum radioactivity at each site were chosen with an SRP-68–01 scintillation radiometer to ensure that the plants contained enough radionuclides for analysis. All aerial mass of plants was cut over an area of 2–5 m2, washed in the river water, and rinsed in distilled water on the site. Then two 20 cm × 20 cm to 40 cm × 40 cm cubes of substrate, spaced at 60–70 cm, were cut out from the ground and divided manually into soil and roots with rhizosphere. The rhizosphere samples, defined as the soil that clings to roots after being gently shaken (Séguin et al., 2004), were separated from the roots after being dried together. Thus, soil and rhizosphere were collected as bulk samples. Their mineralogy was similar at all sampling sites. The phase composition of soil and rhizosphere consisted of silica and feldspars (plagioclase and K-feldspar). Pelitic minerals, Mg-Fe-chlorite and illite, constituted a sizeable share; there was also a limited amount of amphibole and traces of smectite and siderite.

Since 2011, plants at sites B and C (Atamanovskaya Spit and Atamanovsky Island) were sampled from both the upper and lower ends of the Spit and the Island in order to monitor rapid radioactivity changes. The substrate was sampled in this way from 2015 onwards.

Methods

The specific activity of 137Cs was measured by semiconductor gamma spectrometry, with lead and tungsten shields, on several instruments: a EURISYS MEASURES EGPC 192-P21/SHF 00-30A-CLF-F coaxial well HPGe detector with a low-background cryostat (France); an AMETEC (ORTEC) GWL-220–15-XLB-AWT well HPGe detector (USA); an IFTP DGDK-100 V coaxial Ge(Li) detector (Dubna, Russia); BDEG-OGK-3 K coaxial HPGe detector (Dubna, Russia). The data were processed using the AnGamma software SPC ‘Aspect’, (Dubna, Russia) (SPC “Aspect”, 2000).

The activity of 137Cs was detected and estimated according to the gamma-ray line Eγ = 661.66 keV (quantum efficiency 0.852), using 250 cm2 Marinelli beakers for better performance. Measurements, applied to 160–360 g of aliquots homogenised by stirring, lasted from 1 to 24 h depending on 137Cs activity level, to provide a photo peak area accuracy to within 5–10%. The detection limit of 137Cs at this geometry was from 2 to 5 Bq kg−1. Samples of plant ash and low-activity soil and rhizosphere samples (≤ 10 cm3) were measured on well HPGe detectors in a low-background mode providing 0.01 Bq sensitivity (that corresponded to the measurement 5 g sample at 84,000 s).

The 90Sr specific activity was determined indirectly from the measured radiation of its 90Y progeny in secular equilibrium (VIMS, 2003). The sizes of aliquots were 250–400 g for plants and 50–100 g for their substrate. The substrate samples were annealed at 600 °C in a muffle furnace, while the plant samples were ignited by incremental heating to 480 °C at 50 °C temperature steps, with 30 min exposure at each step. 90Y was obtained from samples with stable strontium dissolved in 6 M HCl and stable yttrium added to the solutions cleaned from all beta emitters that could interfere with 90Sr measurements. The two isotopes attained equilibrium in 14 days corresponding to five 2.67-day half-life periods, to an accuracy no worse than 3%. The chemical output of the Sr carrier (60–95%) was determined by atomic absorption after Sr separation. The radiochemical grade of the 90Y progeny was checked according to repeated measurements after a five-half-life cycle. The 90Y specific activity was measured on a RUB-01P beta radiometer with a BDGB-06P low-background detector in a room with low cosmic and natural background, with additional lead and tungsten shielding. The detection limit of 90Sr was 0.1 Bq kg−1, at a total uncertainty within 20%.

Results and discussion

Soil and rhizosphere were not sampled during 2011, 2013, and 2014 as the plant samples collected in 2011 and 2013 showed decreasing specific activities of 137Cs and 90Sr, despite the effects of the Fukushima nuclear accident felt both globally and in the Yenisei floodplain (Bolsunovsky & Dementyev, 2011; Steinhauser et al., 2014). Soil and rhizosphere monitoring resumed in 2015 and 2016 after markedly higher concentrations of radionuclides were measured in the plants’ samples of 2014 (no soil and rhizosphere samples were taken that year due to logistical difficulties).

137Cs and 90Sr in soil and rhizosphere

Soil and rhizosphere (substrate) 137Cs and 90Sr showed time-dependent variations, albeit with long time gaps (Table 1). The substrate samples contained relatively high 137Cs at sites most proximal to the MCC. The content decreased downstream away from the combine and with time but increased abruptly in the Balchug Channel due to favourable conditions for silt deposition (Korobova et al., 2014; Linnik et al., 2004a, 2004b, 2014) and a large amount of organic matter in the soil.

Table 1. 137Cs and 90Sr specific activities in bulk soil and rhizosphere (Bq kg−1 dry weight, with measurement error)
Full size table

The partitioning of 137Cs between soil and rhizosphere varied both in space and in time. However, the respective pattern in the Balchug Channel was different, possibly, due to high percentages of organic carbon at the site: samples of 2004 from the Atamanovskaya Spit contained 0.82 wt. % Corg in soil and 0.66 wt. % in the rhizosphere, while the respective values for the Balchug Channel were 0.59 and 2.99 wt. % (Corg was analysed at the Institute of Soil Science, Novosibirsk). As mentioned above, the organic matter of the rhizosphere accumulates most of the radiocaesium and creates favourable conditions for the formation of its mobile and/or bioavailable species. In 2004, at the beginning of the monitoring project, soil 137Cs at the Balchug site was twice the rhizosphere value (1270 ± 130 Bq kg−1 against 620 ± 60 Bq kg−1). The difference may result from plant uptake of bioavailable 137Cs from the rhizosphere or, more likely, secondary loss and redistribution of mobile species in soil. In the subsequent years, substrate 137Cs at the site became 2–6 times lower. The difference between soil and rhizosphere persisted but reduced to 410 ± 40 against 300 ± 30 Bq kg−1, respectively, in 2015 and became even reverse in 2016: 200 ± 20 Bq kg−1 in soil against 300 ± 30 in the rhizosphere. The reversal may be due to changes in the way of 137Cs transport to the ecosystem or in the water level that affected the delayed soil run-off. The background 137Cs contents at the reference site likewise showed a descending trend, especially in the rhizosphere where they decreased from 13 ± 1 Bq kg−1 in 2004 to 1.0 ± 0.1 Bq kg−1 in 2016.

Substrate 90Sr had quite complex space and time patterns. According to 2004 data, its specific activity likewise decreased away from the MCC, but no increase was observed in the Balchug Channel. The reason is that radiostrontium mainly migrates in the dissolved form (Sysoeva et al., 2005), unlike 137Cs, which is adsorbed on suspended particles (Nosov et al., 1993) and stays at accumulations of fine particulate matter. The complex behaviour of mobile 90Sr appears in changes of its rhizosphere–soil difference: 2–4 times higher specific activity in rhizosphere near the MCC and almost equal values far from the combine. In 2015, the specific activity of substrate 90Sr increased both in the reference sample and in samples from the Atamanavoskaya Spit sites and Beryozovy Island. The relatively high 90Sr specific activity at the reference site indicates atmospheric fallout influents. The greater 90Sr specific activity in soil relative to the rhizosphere value at the reference site can be explained as an effect of the Fukushima event. However, to remain in the soil, atmospheric fallout should be fixed tightly on airborne particles. This idea is inconsistent with a lower percentage of aerosol 90Sr than that of insoluble 137Cs: 18% against 20–50% (Moiseev & Ramzaev, 1975), nor with the selective 90Sr increase at other sites, and with the lack of evidence for the atmospheric fallout of 137Cs. Thus, the fallout episodes, if any, were most likely local and singular and were instead caused by the MCC operation where new facilities were installed in 2015.

137Cs and 90Sr in plants

Time-dependent 137Cs and 90Sr variations in the plants’ aerial parts were monitored during 2004, 2011, and 2013–2016, which provided a more complete idea of radioactive pollution in the area. In 2004, the 137Cs in plants from the flooded zones of islands and the Yenisei west bank revealed the same trend decreasing away from the MCC as for substrate (Table 2). However, no 137Cs anomaly was observed at the Balchug site, likely because mobile caesium moved from the rhizosphere, but no plant uptake occurred. This pattern persisted for the whole period of observation.

Table 2. 137Cs and 90Sr specific activities in aerial parts of sedge (Bq kg−1 dry weight, with measurement error)
Full size table

The situation generally remained the same at almost all sites when monitoring resumed in 2011. The 137Cs activity decreased at sites near the MCC in 2013 but changed little at remote sites. In 2014, 137Cs returned to nearly the previous values at sites proximal to the MCC, but the decreasing trend continued far from the combine facility. Further, 137Cs decrease was observed in 2015 at proximal sites but was especially prominent at distal sites and in the Balchug Channel, where the specific activity became comparable with the background (2.4 ± 0.2 and 1.2 ± 0.1 Bq kg−1, respectively). However, in 2016, all sites showed an increase while the background 137Cs remained stable.

The 90Sr variations over the monitoring period share some similarity with the 137Cs patterns but also have some features of difference. In the same way that 137Cs, the specific activity of 90Sr generally decreased away from the MCC, but the time-dependent variations were different; the 90Sr values were lower during 2011 and 2013, were higher during 2014 and 2015, and fell to the background almost everywhere in 2016. Note that the 90Sr background increased in 2013 to 19 ± 4 Bq kg−1. In the absence of substrate 90Sr data for that year, any explanation for the increase can only be hypothetical. One cause might lie with the global fallout from the Fukushima nuclear accident, but the 90Sr specific activity in plant samples decreased at all sites instead of increasing. Note that the 137Cs background remained at the same level (0.5 Bq kg−1) in 2013, though much of the isotope was released during the Fukushima disaster (UNSCEAR, 2015). The difference may be due to the different sorption capacities of aerosols and substrates with respect to the two radionuclides: up to 95% of 137Cs is adsorbed on soil (Moiseev & Ramzaev, 1975) while 70–90% of 90Sr exists in the exchangeable form (Sysoeva et al., 2005). On the other hand, the bioavailability of the radionuclides (Burger & Lichtscheidl, 2019; Sysoeva et al., 2005) is different as the release of 90Sr from soil is two or three orders of magnitude greater than for 137Cs (Legin et al., 2008).

The 90Sr increase of 2015, when the monitoring of substrate 137Cs and 90Sr resumed, is worth special consideration. The specific activity of 90Sr in plants was higher near the MCC, including at the reference site with a background of 7.0 ± 1.4 Bq kg−1 (though lower than in 2013). On the other hand, the 90Sr specific activity in plants sampled far from the MCC was below or at the background in 2015: 3.6 ± 0.7 Bq kg−1 on Beryozovy Island and 7.1 ± 1.4 Bq kg−1 in the Balchug Channel. Note that soil and rhizosphere 90Sr values differed more strongly on Beryozovy Island than at the Balchug site in 2015: 12 ± 2 and 5.0 ± 1.0 Bq kg−1 against 4.8 ± 1.0 and 4.7 ± 0.9 Bq kg−1, respectively. This provides some evidence of the local spread of the pollution: the amount of dust aerosol that can carry 90Sr was insufficient at remote sites to allow radiostrontium to get into the soil and pass to rhizosphere or plants (case of Beryozovy), or the isotope failed to reach the substrate and remained only on the plant surface (case of Balchug).

Relative 137Cs and 90Sr in the system ‘soil – rhizosphere – plants’

The data of 2004, 2015, and 2016 were used to plot specific activity diagrams of the two radionuclides for the ‘soil – rhizosphere – plant’ system (Fig. 2). The soil/rhizosphere 137Cs specific activity ratios demonstrated positive correlation at the Pearson coefficient r = 0.95–0.99 over the whole period of observation, whereas the respective ratio of plant/soil varied markedly from year to year (R2 = 0.0043–0.6496), with negative, or near zero, correlation that reached r = 0.81 in 2016 only. The plant/rhizosphere 137Cs ratios likewise vary with time though not as markedly as in sedge and soil (R2 = 0.0920–0.5255); the correlation was almost absent in 2004 (r = 0.13) but increased to r = 0.562015 in 2015 and to r = 0.72 in 2016. The soil and rhizosphere 90Sr specific activities correlate quite well (r = 0.73–0.98), the variance being the greatest in 2004 (R2 = 0.5353) and the smallest in 2015 and 2016 (R2 = 0.9566 and 0.9650, respectively). On the other hand, 90Sr in sedge and soil correlate better than 137Cs, with moderate variance: r = 0.87–0.97 and R2 = 0.7588–0.9420; the sedge-rhizosphere correlation is at r = 0.76–0.87 and R2 = 0.5751–0.7503, respectively.

Fig. 2.
figure 2

137Cs vs. 90Sr correlation in different components of system ‘soil-rhizosphere-plants’

Full size image

The behaviour of the two radionuclides in the system changes upon absorption by plants owing to differences in uptake mechanisms. The uptake of 137Cs occurs in the same way as for K, its chemical analogue: transport into the root xylem and stems and leaves, mainly through symplasts via potassium-selective and non-selective ion channels. Conversely, 90Sr, an analogue of Ca, enters mainly via apoplasts and non-selective channels (Burger & Lichtscheidl, 2018, 2019; Ehlken et al., 2002; Roca & Vallejo, 1995; Sokolik et al., 1996). Other dissimilar features may be due to the metabolism of rhizosphere microorganisms which can either immobilise or release the bioavailable forms (Burger & Lichtscheidl, 2018, 2019).

The correlation between 137Cs and 90Sr in the aerial parts of plants was studied (Fig. 3). It was somewhat positive (r = 0.77–0.95) during 2004 and 2014–2016, though at significant variance, especially in later years (R2 = 0.5855–0.6929). The correlation changes most likely indicate a change in relative influxes of bioavailable 137Cs and 90Sr into the system. In 2011 and 2013, the variance increased (R2 = 0.0024 and 0.1991, respectively) while the correlation became absent or slightly negative (r = 0.05 and − 0.45, respectively). The 137Cs and 90Sr changes had some other cause from previous years, though the relative specific activities in 2014 approached those of 2004. In general, the revealed patterns do not contradict the hypothesis of local 90Sr influxes.

Fig. 3.
figure 3

137Cs vs. 90Sr correlation in aerial parts of plants in different monitoring campaigns

Full size image

137Cs and 90Sr in ecosystem components from different parts of islands

The specific activities of 137Cs and 90Sr in ecosystem components sampled in the upstream and downstream ends of the Atamanovskaya Spit and Atamanovsky Island differ markedly (Tables 1 and 2). As determined previously (Korobova et al., 2014, 2016; Linnik et al., 2004a, 2004b, 2005, 2006, 2014), the radionuclides are localised at accumulations of particulate matter carried with water, mainly in the lower ends of islands and in stagnant bays or in places composed of very fine or fine sand, silt, peat, or light-textured loam. This matches, for instance, the downstream end of the Atamanovskaya Spit located 75–110 m away from the east bank of the Yenisei (108 m away at the upper end and 85 m away at the lower end). Data from 2016, when substrate 137Cs began to increase, show that the amount of 137Cs transported mainly on particulate matter (Nosov, 1996) was much higher in the lower end of the spit than at its head, by factors of 8.8 in soil and 5.5 in the rhizosphere. The respective difference for Atamanovsky Island was smaller (2.1 times for soil and 1.6 times for the rhizosphere). The 137Cs specific activity in ecosystem components in the upper and lower ends of the islands shows the lack of correlation, with r =  − 0.17 and 0.10 for the substrate and plants, respectively (Fig. 4a), which corresponds to the case of transport on suspended particles. Unlike radiocaesium, 90Sr migrates mainly in the dissolved form and is localised preferably at different ends of the spit and the island. On the spit, it is higher in the lower end, whereas in the island, higher values are measured at the head. The difference is by factors of 1.8 for soil and 1.4 for rhizosphere in the spit, and 1.5–2.0 in the island. Note, however, that the setting was reversed in 2015 (though the data are incomplete): 90Sr in rhizosphere sampled in the downstream end of the island was 4.1 times greater than the respective value for the upstream end. Substrate 90Sr values in the upper and lower ends of the islands showed a high correlation at r = 0.93, which confirms migration mainly in the dissolved form, but the correlation for plant samples was as low as r = 0.33 (Fig. 4b).

Fig. 4
figure 4

Correlations of 137Cs and 90Sr in system ‘soil-rhizosphere-plants’ between upstream and downstream ends of islands

Full size image

Data from the lower end of Beryozovy Island were not considered for several reasons, including the location on the border of the near impact zone of the MCC, the size of the island being much larger than those upstream, complex island geometry and topography due to an abrupt eastward river turn, and complex current patterns around the island (Linnik et al., 2006, 2014).

Excess over background

Excess element content over the background (N) is often used to estimate pollution for different ecosystem components. In our case, the excess was calculated as a ratio of specific activities of the two radionuclides (A) to the respective background value at the reference site (Abg):

$$N=\frac{A\left({\mathrm{Bq} \ \mathrm{kg}}^{-1}\text{dry weight}\right)}{{A}_{bg}\left({\mathrm{Bq} \ \mathrm{kg}}^{-1}\text{dry weight}\right)}$$
(1)

The specific activity of 137Cs exceeded the background in all components of the ‘soil – rhizosphere – plants’ system. The excess for the aerial plant organs was the greatest (by 336–433 times) in 2011 and 2016 in the lower end of the Atamanovskaya Spit and the upper end of Atamanovsky Island and remained high in the other years of observations (Table 3). The 137Cs specific activities in soil and rhizosphere (Table 3) were from 14 to 200–300 times greater than the background in different years and at different sites and peaked in 2016 in the lower end of Atamanovskaya Spit (900 and 600 times, respectively).

Table 3. 137Cs excess above background
Full size table

The excess of 90Sr above the background was much smaller and even negative in some cases. For plants, it was the greatest in 2011 in the upper (74 times) and lower (30 times) ends of Atamanovskaya Spit, reduced to almost zero in 2013, and was within ten times in the subsequent years (Table 4). For soil, a minor excess was observed in 2016 in the lower spit end, while for the rhizosphere, it was the highest in 2015 and relatively high in 2016 (Table 4).

Table 4. 90Sr excess above background
Full size table

Transfer factor

Plant uptake of radionuclides from the soil is commonly expressed as soil-to-plant transfer factor (TF), normally expressed as a ratio of specific activities in plant tissue and specific activities in soil

$$TF=\frac{{A}_{plant}(\mathrm{Bq}\ {\mathrm{kg}}^{-1}\text{dry weight})}{{A}_{soil}(\mathrm{Bq}\ {\mathrm{kg}}^{-1}\text{dry weight})}$$
(2)

The transfer factor is a key parameter used to predict the effect of radionuclides released in a nuclear fuel cycle (NFC) on the ecosystem and to calculate further the radiation dose the humans receive through the food chain.

The time-dependent TF patterns in the MCC proximal zone differed for 137Cs and 90Sr, which showed increasing and decreasing trends, respectively. The 90Sr TF decreased to 0.036–0.136 at all sites sampled in 2016 (when the background was 0.262) relative to 1.905–2.024 (0.714 background) in 2004 (Table 5). Typically, TF can decrease for several reasons: the absorption volumes change when the nuclides become immobilised in the substrate (ageing effect), isotopes may migrate down the soil profiles to outside the reach of roots (IAEA, 2010), or floods can carry the mobile forms away from the soil which has low sorption capacity (Sysoeva et al., 2005). However, these processes are inconsistent with the fact that the specific activity of substrate 90Sr returned to the level of 2004 in recent years. The TF decrease may indicate that 90Sr became less available for plants because its greater portion was supplied to the system as adsorbed on suspended particles rather than in the dissolved form (Bolsunovsky et al., 1999).

Table 5. Soil-to-plant transfer factor (TF)
Full size table

The TF of 137Cs in 2004 and 2015 was notably lower than that of 90Sr and increased from 0.027 to 0.757, at background values of 0.254 and 0.595, respectively. In 2016, however, it rose above the 90Sr TF at all sites and reached 1.427 (background 0.909). The data of 2016 show the highest 137Cs TF in the upstream parts of the islands, markedly exceeding the values from the lower ends. As noted above, the cause of this difference is that particulate matter, with adsorbed 20–50 % of 137Cs influxes to the river, becomes deposited in the downstream ends of the islands (Nosov et al., 1993; Vakulovsky et al., 1995; Linnik et al., 2006) and is less available for plant uptake (Carver et al., 2007; Ivashkevich & Bondar, 2008). Meanwhile, dissolved 137Cs is more available at the island head.

Furthermore, the TF values were consistently lower in the Balchug Channel than elsewhere, whereas the specific activities of the radionuclides are high in soil but low in plants, possibly, because the soil at the site was rich in highly sorptive organic carbon (see ‘137Cs and 90Sr in soil and rhizosphere’). However, the 90Sr TF increased to 1.479 in 2015 and exceeded both the background (0.444) and the values from all other sites (0.241–1.276), while the 90Sr specific activities in both substrate and plants were around the background. Data from a water station in Atamanovo Village showed a high water level between late March and late May in 2014 and 2015 and approached the decade maximum in 2014 (Information and Analytical Center of the Register and Cadastre, 2014). The water level in the river was also high in 2016, but rather in January–February and middle-late May, in the periods beyond the sedge spring vegetation season. The high 90Sr TF may result from the water regime providing high soil moisture while 90Sr loss from water-logged soil generally exceeded that of 137Cs, and most of 90Sr, including its bioavailable forms, was carried away with intra-soil run-off (Ovsyannikova et al., 2000; Legin et al., 2008).

Conclusions

The radioactive contamination in the Yenisei floodplain ecosystem in the near impact zone of the Krasnoyarsk MCC has a complex pattern. The specific activity of 137Cs in soil and rhizosphere decreased throughout observations, while 90Sr increased. The river water level fluctuations at some sites (Balchug Channel), especially the time and duration of high stand, may cause a strong influence on the partitioning of isotopes between soil, rhizosphere, and plants (aerial parts). Variations of the soil-to-plant transfer factor (TF) indicate continuing radionuclide fluxes into the system, as well as changes in the shares of bioavailable and fixed species, especially in the case of 90Sr. These changes, including local atmospheric fallouts, apparently controlled the pattern of 90Sr in the ‘soil – rhizosphere – plant’ system observed during 2014–2016. The specific activities of 137Cs and 90Sr in different ecosystem components exceeded the background values in all years of monitoring (more strongly for 137Cs), though the specific activity of 90Sr fell below the background in a few cases.