Sources of plutonium in the atmosphere and stratosphere-troposphere mixing

Plutonium isotopes have primarily been injected to the stratosphere by the atmospheric nuclear weapon tests and the burn-up of the SNAP-9A satellite. Here we show by using published data that the stratospheric plutonium exponentially decreased with apparent residence time of 1.5 ± 0.5 years, and that the temporal variations of plutonium in surface air followed the stratospheric trends until the early 1980s. In the 2000s, plutonium and its isotope ratios in the atmosphere varied dynamically, and sporadic high concentrations of 239,240Pu reported for the lower stratospheric and upper tropospheric aerosols may be due to environmental events such as the global dust outbreaks and biomass burning.

Scientific RepoRts | 5:15707 | DOi: 10.1038/srep15707 residence time of gaseous chemical components such as CO 2 and SF 6 , were calculated. The obtained residence time of gaseous chemicals in the mid-latitude stratosphere was in the range of 1.1 to 2.1 y (ref. 15).
Recently Corcho Alvarado et al. 16 reported interesting results of investigations of plutonium isotopes and 137 Cs in stratospheric and tropospheric aerosols, which included new data observed during the period of 2007 to 2011. They found higher 239,240 Pu and 137 Cs concentrations, and higher 238 Pu/ 239,240 Pu activity ratios in the lower stratosphere and lower troposphere than expected. The observed levels of 239,240 Pu in stratospheric aerosols were from two to four orders of magnitude higher than that in the ground-level air. They also suggested that the stratospheric mean residence time of plutonium and 137 Cs should be 2.5-5 y, arguing that radionuclides attached to fine aerosol particles (< 0.02 μ m in diameter) could have a longer stay in the stratosphere, and therefore radionuclides injected there mainly during the early 1960s have still been present during the 2000s in the stratosphere. However, studying long-term variations  of plutonium isotopes in the stratosphere and surface air of the Northern Hemisphere we have found that the dominant processes affecting plutonium concentrations in the upper troposphere should be global dust events and biomass burning, and that its apparent residence time in the atmosphere did not change from 1.5 ± 0.5 years.
Nuclear power plant accidents such as Chernobyl and Fukushima were also sources of plutonium in the environment, although they were much lower scale events. Plutonium isotopes released from these accidents were characterized by higher 238 Pu/ 239,240 Pu, 241 Pu/ 239,240 Pu activity ratios and higher 240 Pu/ 239 Pu atom ratios than those ratios derived from nuclear tests 6 . The 238 Pu/ 239,240 Pu, 241 Pu/ 239,240 Pu activity ratios and 240 Pu/ 239 Pu atom ratio for the Chernobyl accident were 0.5, 85, and 0.41, respectively 17,30,31 , while for the Fukushima accident these ratios were 1.2, 108, and 0.30-33, respectively 32,33 . The plutonium isotopic signature may help therefore to better understand sources of anthropogenic radionuclides, and their behavior in the upper and lower atmosphere.
For better understanding of plutonium levels in the stratosphere it is therefore important to elucidate sources of the bomb-derived plutonium and other radionuclides observed in stratospheric aerosols in the 2000s. Possible radionuclide sources are the large-scale nuclear tests carried out in 1961-62, the Chinese nuclear tests (especially those conducted in 1976 and 1980), resuspension of plutonium from deserts, and biomass burning. To achieve this aim, we collected data of plutonium isotope concentrations in stratospheric air and in surface air in the Northern Hemisphere. Unfortunately, a continuous data set of both stratospheric and ground-level plutonium levels during the period of 1960-2010 has not been possible to construct. We describe here long-term variations (1964-2010) of plutonium isotopes in the stratosphere and surface air of the Northern Hemisphere using available data.
Temporal variations of 239,240 Pu in the stratosphere and surface air. Large data sets on plutonium isotopes in stratospheric and ground-level aerosols is available from the Environmental Measurements Laboratory (EML, USA) 34 , which conducted high-altitude aerosol monitoring programs from the early 1960s to the early 1980s (ref. 35). We used 238 Pu and 239,240 Pu activity concentrations in the Northern Hemisphere stratospheric air (20-40 km altitude), in which unreliable results with high measuring uncertainties were removed (Fig. 1). The data obtained by Corcho Alvarado et al. 16 from 1973 to 2009 for the lower stratosphere (10.1-14.2 km altitude) were included in Fig. 1 as well. Further, plutonium isotopes results obtained for surface air at mid-latitude region of the Northern Hemisphere were also included in Fig. 1 38 , Vilnius (Lithuania, 54° 42′ N, 25° 30′ E) 6 and Milford Haven (UK, 51° 43′ N, 5° 02′ W) 39 . All plutonium isotope concentrations in surface air were determined at monthly or quarterly basis.
A pronounced peak of 239,240 Pu concentration in surface air of New York occurred at delay of about 17 months after the 1976 Chinese thermonuclear test (Fig. 1). The surface 239,240 Pu concentration decreased then with the apparent residence time of 1.3 ± 0.3 y (similar to that observed for stratospheric aerosols) until the end of 1980. The surface 239,240 Pu concentrations showed seasonal variations -a late spring maximum and a winter minimum, in contrast of the stratospheric 239,240 Pu levels 40 . After the 1980 Chinese nuclear test (total yield of 0.6 Mt in 16 October 1980), a small peak of 239,240 Pu occurred in the upper stratosphere, which is consistent with the result that most of plutonium from the 1980 Chinese nuclear test was injected into the lower stratosphere and the AME layer just above tropopause 14 . Irrespective of location of surface monitoring sites (New York, Beaverton, Tsukuba, Milford Haven), the surface 239,240 Pu levels showed marked increase in spring 1981. After 1982, a decrease with the apparent stratospheric residence time of about 1.3 ± 0.3 y was observed until 1984, which means that the surface plutonium until the early 1980s was controlled by the stratospheric inputs. These findings confirm therefore that significant amounts of 239,240 Pu and fission products in the upper stratospheric air in the 1970s and in the early 1980s were derived from the series of the Chinese nuclear weapons tests. The 239,240 Pu levels observed in the lower stratosphere during 2007-2008 were by about two orders of magnitude larger than those observed in surface air 6,16,31 . The 239,240 Pu concentrations are expressed per standard cubic meter (15 °C, 101.325 kPa), however, the pressure at 10 km of altitude is about one order of magnitude lower than that in ground-level air. The thermodynamics indicates that sampling volumes in high altitudes are greater than the SCM, therefore it is difficult exactly compare radionuclide concentrations measured at high altitudes with those measured at ground-level air. Another point is a difference of sampling periods between surface and high altitudes measurements. High altitude sampling was usually carried out only for hours, whereas sampling periods of surface air were one to three months. Therefore short-term sporadic events occurring at high altitudes need not be visible in monthly or three months mean values observed at ground-level air.
There are no data available on plutonium concentrations in stratospheric aerosols during the period from June 1986 to October 2004 (Fig. 1). On the other hand, the plutonium isotope levels in surface air were measured in several monitoring sites, most of which were located in Europe. The plutonium concentrations in surface air were as follows: < 2.5 to 9.5 nBq m −3 during the period from 1986 to 1989 (as quarterly means) at Milford Haven 39   The observed temporal variations of the 238 Pu concentrations in stratospheric air in the period from 1965 to 1986 showed an exponential decrease (Fig. 2), with an apparent stratospheric residence time of 1.7± 0.4 y, consistent with previous results 43 . The result reveals that stratospheric 238 Pu concentrations showed altitude-depended distribution: higher 238 Pu concentrations occurred in the upper stratosphere (20-40 km altitude), whereas lower 238 Pu levels were observed in the lower stratosphere (10.1-14.2 km altitude), which suggests that temporal variations of atmospheric 238 Pu were controlled by the upper stratospheric 238 Pu derived from the SNAP-9A burn-up, and partly due to the Chinese nuclear explosions. The 238 Pu concentrations measured in the stratospheric air in the 2000s were, however, of the same order of magnitude as that during the 1970s and 1980s. It has been suggested 16 that the high 238 Pu levels observed during the 2000s were still due to the SNAP-9A burn-up. It is difficult to consider, however, that the supply of the SNAP-9A-derived 238 Pu from the upper stratosphere in the 2000s was the same level as that during the 1970s, i.e. two orders of magnitude higher than levels observed in 1986, and when between 1986 and 2007 typical global fallout 238 Pu levels were observed in the lower stratosphere 16  . This finding suggests that there are more than two sources of radionuclides in the upper atmosphere, because it is likely that radionuclide composition in the stratospheric aerosols is homogenized throughout a long-time mixing.

Discussion
Redistribution of plutonium in the atmosphere. Corcho Alvarado et al. 16 revealed that significant amounts of plutonium isotopes and 137 Cs existed in the lower stratosphere during the 2000s, which could be transported to the lower troposphere. The recent observations of 239,240 Pu, 14 C and 137 Cs concentrations in surface air 8,38 did not show, however, typical spring maxima, which were observed until the late 1980s when stratosphere-troposphere radionuclide transport was dominant (see also Fig. 1 for 239,240 Pu). These maxima occurred due to the transport of lower stratospheric air masses containing high radionuclide levels to the upper troposphere. Similar variations have been observed for cosmogenic 7 Be produced by interactions of cosmic ray particles with nitrogen and oxygen atoms mainly in the lower stratosphere and the upper troposphere. Higher 137 Cs concentrations in winter months and lower concentrations in summer months are apparent for the last decade, which are more similar to variations of terrigenic 210 Pb (a decay product of 222 Rn) than for the 7 Be variations of the stratospheric origin 8,[47][48][49][50] . Clearly, there is no more stratospheric influence on the 137 Cs concentration in surface air, otherwise the 137 Cs maxima would be observed in the late spring, similarly as the 7 Be spring maxima, which are still observed in the groundlevel air. This change in the 137 Cs record in the atmosphere is connected with the fact that resuspended 137 Cs has become important source of tropospheric radioactivity. Similarly 14 C variations observed in the troposphere after 1990s have been due to 14 C decreases during winter caused by the Suess effect 8,51 . Therefore the absence of spring maxima in 239,240 Pu, 137 Cs and 14 C records during the 2000s suggests that the stratospheric transport does not play anymore a dominant role, but a resuspension from soil or other processes may be responsible for observed radionuclide variations in the atmosphere.
Global desert dust events. Desert dust events, well known as Saharan dust and Asian dust (Kosa), have been loading large amounts of soil particles into the atmosphere. Saharan dust transport has been known as the biggest global event redistributing aerosols in the atmosphere 52 . The Asian dust clouds, and similarly the Saharan dust clouds were transported in the mid-latitude region around the globe 53 . Since soil particles contain anthropogenic radionuclides primarily derived from global fallout and the Chernobyl accident, Saharan and Asian dusts could cause sporadically increasing anthropogenic radionuclide levels in the atmosphere. Asian dusts were characterized by temporal variations of 239,240 Pu and 137 Cs deposition 4,46,54,55  Biomass burning. Other strong candidates of anthropogenic radionuclide variations in the troposphere are forest and grassland fires when huge amounts of particles of submicrometer size are released to the air 64 . The radionuclides derived from the Chernobyl accident as did global fallout had been contaminated in wide areas of Eurasia 9 , which could be lifted back to the atmosphere. The wildfire events could be enhanced by specific meteorological conditions, such as temperature inversions and/or rain events at remote places, causing secondary deposition of 137 Cs (ref. 65,66). Smoke plumes from biomass fires, could reach several kilometers height, and they can travel distances as long as several thousand of kilometers. They could even penetrate the tropopause, and reach the lower stratosphere 67,68 . The biomass burning events could be under specific conditions combined with Saharan dust events causing thus global aerosol impact on the atmosphere. Under specific meteorological conditions they could stay in the atmosphere for several weeks. As around 50 Mha of forest and roughly ten times more grassland are burnt annually, the biomass burning represents important way of radionuclide transport in the environment 69 . The biomass burning plumes originating in Eurasia may redistribute global fallout and Chernobyl deposited sources of anthropogenic radionuclides (e.g. 137 Cs and 239,240 Pu) in the atmosphere, which could change concentrations, as well as isotope ratios of these radionuclides in the atmosphere 6 16 .
A much bigger volcano eruption than the Eyjafjallajökull one was the Mt. Pinatubo (15°08′ N, 120°21′ E) eruption, which occurred on June 12-16, 1991, and was one of the 20 th century′ s greatest volcanic eruptions 71 . As a result of this powerful eruption, 15-20 megatons of SO 2 were injected into the stratosphere, as the eruption columns reached 40 km in altitude. Sulfuric acid and/or sulfate aerosols transformed from SO 2 can effectively attach radionuclide-bearing particles, and remove them from the stratosphere by a residence time of about 13 months for the Pinatubo aerosol cloud 72 . Unfortunately, there are no stratospheric/tropospheric data available during the 1990s to confirm/discard the volcano hypothesis. As shown in Fig. 1, no enhanced 239,240 Pu concentration after the Pinatubo eruption, as did 239,240 Pu deposition 4,46 was observed in surface air. It is likely therefore that remnants of radionuclides derived from the nuclear tests from 1950 to 1980 and from the satellite burn-up in 1964 were already removed from the stratosphere before injection of the Pinatubo aerosol cloud. This has been supported by observations of 7 Be and 137 Cs concentrations in summer of 1991 (ref. 47). Therefore there is no obvious evidence of enhanced removal of stratospheric anthropogenic radionuclides due to stratospheric injection of sulfate and ash by large-scale volcanic eruptions.
Sea-spray effects. It has been pointed out that sea spray may be another potential source of plutonium in the atmosphere 73,74 . However, the contribution of plutonium from sea salt to atmospheric plutonium deposition is much lower than that from soil (< 0.3%) 4 . It was estimated that a contribution of plutonium from major constituent in sea salt (chloride) in surface air of Tsukuba was below 1 μ Bq m −3 . The plutonium/chloride ratio in seawater was 0.5 μ Bq of Pu per gram of Cl (assuming that plutonium in seawater is homogeneously attached on sea-salt particles) 75 . The calculated maximum contribution of sea-spray plutonium in the atmosphere is less than 0.006 nBq m −3 , which is by 2-3 orders of magnitude lower than from other potential plutonium sources.
We may conclude that our studies of long-term variations (1964-2010) of plutonium isotopes in the stratosphere and troposphere of the Northern Hemisphere suggest that plutonium levels in ground-level air followed the stratospheric trends until the early 1980s. In the 2000s, plutonium and its isotope ratios in the atmosphere varied dynamically, and sporadic high concentrations of 239,240 Pu reported for the lower stratospheric and upper tropospheric aerosols may be due to environmental events such as the global dust outbreaks and biomass burning. Long-term measurements of plutonium isotopes in the stratosphere and troposphere revealed that the plutonium concentrations in the stratosphere and the troposphere decreased with apparent residence time of 1.5 ± 0.5 y. The plutonium concentrations in surface air, irrespective of sampling sites in the mid-latitude regions, decreased following the changes in the stratospheric plutonium concentrations.
Anthropogenic radionuclides in the troposphere and the lower stratosphere have been useful tools for better understanding of dynamical processes of aerosols in the atmosphere. Knowledge of their temporal variations has also been important pre-requisites for climate change studies, and assessment of radioecological impacts of nuclear facilities accidents on the atmospheric and terrestrial environments 76 .