Introduction

While the majority of atmospheric radioactivity can be accounted for by cosmogenic production and decay products of the 238U and 232Th decay series, some originates from nuclear power reactors, nuclear accidents, and past nuclear-weapon testing. Airborne radioactivity from these natural and man-made sources can enter the human body through inhalation and through settling on foodstuffs. For example, an isotope of beryllium, 7Be, produced by cosmic-ray bombardment of nitrogen and oxygen atoms in the troposphere, is commonly observed in ground-level air and on vegetation due to its lengthy half-life (53 days). Likewise, deposited radiostrontium (90Sr) and 137Cs from past atomic detonations has been observed in milk throughout New York State (NYS) for several decades [1]. In addition, studies have documented the release and deposition of alpha- and beta-emitting isotopes from the 238U and 232Th decay series during coal combustion.

To detect airborne radioactivity, the NYS Department of Health initiated a sampling and analysis program for scheduled surveillance around reactor and nonreactor sites over 25 years ago. This continuous surveillance provides a method for the determination of normal radio-activity levels and for monitoring of reactor emissions, and provides a basis for post-incident dose reconstruction should an emission event or accident occur. Airborne activities of gross-beta particles, tritium, and gamma-ray emitters, determined as part of the surveillance program, are often near or below the analytical detection limits. A summary of results of the surveillance program are presented elsewhere [2]. In this paper, activities of gamma-ray emitters that were observed in NYS following the atmospheric release of radionuclides from the Fukushima Daiichi complex of nuclear power reactors in Japan are provided.

As a result of an earthquake and subsequent tsunami affecting the east coast of Japan on March 11, 2011, fission-product radionuclides were released from the nuclear power reactors to the atmosphere. Although the radioactive plume traveled across the Pacific Ocean, primarily in the direction of the Arctic, and underwent dispersion and deposition, measurable amounts of gamma-ray emitters were detected at several locations in NYS. Our data are provided for comparison to concentrations of atmospheric radioactivity typically observed at the sampling locations, as well as provide an estimate of the radiation dose to the citizens of NYS as a result of the release from the Japanese reactors.

Experimental

Locations of the six air-monitoring sites in six counties in NYS are shown in Fig. 1. The sites at Albany (A) are not affected by the operation of reactor facilities that are known to release radionuclides to the atmosphere. Samples analyzed from these locations exhibit normal concentrations of naturally occurring radionuclides, plus the usual influence from past nuclear detonations and global use of nuclear energy. The six nuclear power reactors operating in NYS are located near three sites (B, E, F). Additional air-monitoring sites are located near Brookhaven National Laboratory (G) and a small training reactor operated by the United States (US) Department of Energy (D).

Fig. 1
figure 1

Map of New York State (NYS), showing the locations of the sampling sites

Full size image

At the air-monitoring surveillance sites, airborne particulates were collected for ~7 days each, by drawing air through a cellulose-fiber filter (5-cm dia; Grade 54; Ahlstrom Corp., Mt. Holly Springs, PA) at a typical flow rate of ~60 L/min (Table 1). The air samplers are enclosed in galvanized steel housings ~1.5 m above the ground. Staff from county health departments changed the filters approximately weekly and delivered the samples to the laboratory within days of collection. Gas-phase 131I was collected by placement of a 5.8-cm dia canister (model BG-300; Radeco, Plainfield, CT) containing activated charcoal in series after the particulate-collecting filter. Charcoal canisters were packaged separately, but delivered with the air-particulate filters.

Table 1 Characteristics of air sampling conducted at various locations in New York State
Full size table

For the period of March to May 2011, additional air sampling was conducted at site C (~15 km from site A) using identical equipment as that deployed at the surveillance air-monitoring sites. Samples were changed twice weekly, thus the air volume was only half that collected at the surveillance sites. In contrast, additional sampling at a location ~1 km from site A (denoted as A*) utilized a high volume (~1.2 m3/min) pump, a 10-μm particle cutoff head, and large (20 × 25 cm2) polycarbonate membrane (1 μm; Poretics Corp., Livermore CA) or quartz (Whatman QMA, Clifton, NJ) prefilters. A large charcoal-impregnated filter (20 × 25 cm2; Schleicher & Schuell Inc., Keene, NH) was placed in series after the prefilter to collect gaseous airborne components. In addition to airborne samples, fallout samples were collected in open buckets exposed to atmospheric precipitation and dust at the two ALB sites (A and C). Buckets containing no liquid precipitation were washed with distilled water to obtain the dry fallout sample.

Gamma-ray spectrometry measurements of the filters, charcoals, and fallout were conducted using high-purity germanium (HPGe) detectors in low-background lead shields. Detector efficiencies ranged from 20 to 133 %, relative to a 3″ × 3″ NaI(Tl) detector, and resolutions were ~2.0 keV at 1,333 keV. During standardization of the HPGe detectors, efficiencies for 134Cs were corrected for coincidence summing. The filter and charcoal samples were counted directly on the detector surface for 400–1,000 min, while fallout samples were counted for 300–1,000 min. Data were acquired with a multiplexed system coupled to an Ethernet network. Spectra were collected and analyzed using the Genie 2000 gamma spectroscopy system (Canberra Industries Inc., Meriden, CT). Results, errors, and detection limits for gamma-ray spectrometry data are reported at the 95 % confidence level (~2 sigma). Concentrations of 131I and 7Be were corrected for decay from the midpoint of the collection period to the midpoint of the counting period. For all analyses, quality control was ensured by the use of NIST-traceable standards and participation in external proficiency-testing studies.

Results and discussion

Although a wide variety of radionuclides were emitted from the damaged reactors at Fukushima, in this study we report on the detection of 131I (t1/2 = 8.03 days), 134Cs (t1/2 = 2.07 years) and 137Cs (t1/2 = 30.1 years), as well as naturally occurring 7Be (t1/2 = 53.2 days), in the atmospheric samples. Detection of additional reactor-produced radionuclides in the collected air samples has been reported elsewhere [3]. Figure 2a illustrates that only the annihilation peak (511 keV) existed in the spectra of a typical detector background, while photopeaks for the isotopes of interest were prevalent on a typical 5-cm prefilter (Fig. 2b) collected during the sampling period. 131I, 7Be, and 137Cs each emit a single primary gamma-ray photopeak (364, 477, and 662 keV, respectively), while 134Cs has major photopeaks at 605 and 796 keV.

Fig. 2
figure 2

Comparison of gamma-ray spectra for typical detector background (a) and air-particulate filter (b)

Full size image

Fukushima-related radionuclides were first detected in NYS on filters collected more than a week after the reactor accident, and activities continued to rise for several weeks before diminishing. As shown in Fig. 3, similar concentration patterns were observed for 134Cs at the six surveillance sites that employed identical sampling apparatus. Slight differences are likely due to local rainout by precipitation. While it is evident that 134Cs was not detected on filters collected immediately after the accident and prior to arrival of the plume, the high detection limits reported for some filters (e.g., OSW 4/22) are a result of reduced counting times (e.g., 400 min) and/or use of a lower-efficiency detector. Considering the low levels, activities calculated based on the two gamma-rays of 134Cs were somewhat correlated (r 2 = 0.67). For the sampling periods given in Table 1 in which activities were above detection limits, the site-averaged 134Cs varied from 47 to 94 μBq/m3 at the six sites and was 69 μBq/m3 overall. In comparison of results above detection limits, the overall average for NYS is somewhat less than the average (95 μBq/m3) reported for Washington State [4], equivalent to the average (72 μBq/m3) observed in Europe [5], and greater than the average activity (39 μBq/m3) measured in Greece [6]. These differences in activity are likely the result of variations in weather (rainout and dispersion) and wind patterns impacting the sites, and proximity of the sites to Japan.

Fig. 3
figure 3

Activities of airborne 134Cs determined at six surveillance sites in NYS

Full size image

The activities and distribution for 137Cs on the filters (Fig. 4) were similar to that observed for 134Cs. For the sampling periods given in Table 1 in which activities were above detection limits, the site-averaged 137Cs varied from 40 to 111 μBq/m3 at the six sites and was 68 μBq/m3 overall. The latter is 24 times less than the average observed on the east coast of the US following the Chernobyl accident [7]. As with 134Cs, the overall average of results above detection limits determined in this study is about half of the average (130 μBq/m3) reported for Washington State [4], equivalent to the average (76 μBq/m3) reported for northern Europe [5], and greater than the average activity (44 μBq/m3) measured in Greece [6]. As with 134Cs, differences in reported values are likely the result of variations in weather and wind patterns, and proximity of the sites to Japan. The similarity of 134Cs and 137Cs activities in these studies implies that the ratios of 137Cs/134Cs are near unity. For filters from the six sites, in which both primary gamma-rays of 134Cs were detected, the ratio of 137Cs/134Cs averaged 1.13 ± 0.25. The two isotopes were well correlated (r 2 = 0.82). In this study no decay correction was applied for 134Cs (2 % over three weeks) since this adjustment was well within the measurement errors.

Fig. 4
figure 4

Activities of airborne 137Cs determined at six surveillance sites in NYS

Full size image

Of the radioisotopes reported here, only 131I exists in both particulate (131Ip) and gaseous (131Ig) phases in the atmosphere. Historically, no 131Ip has been observed above the detection limit from samplers surrounding the NYS reactor sites. For the current study, 131Ip was measurable at all routine sampling sites (Fig. 5), with an overall average of 460 μBq/m3 for the sampling periods listed in Table 1 in which results exceeded detection limits. Site averaged 131Ip activities ranged from 320 to 550 μBq/m3, with maximum measured activities provided in Table 2. At each sampling site, the 131Ip activity peaked near the beginning of April. As reported elsewhere [5], and apparent by comparing Figs. 4 and 6, both particulate and gaseous 131I were detected before 137Cs at every surveillance air-monitoring site (except SUF). This observation is likely due to the greater airborne concentrations and lower detection limit for 131I relative to the radiocesium isotopes. The overall average of the 131Ip activities were about half of the values (1,000 μBq/m3) reported for Washington State [4], comparable to results measured in most European countries [5], and about double the average activity (240 μBq/m3) that was measured in Greece [6] during the same period. Both peak and average 131Ip levels determined in this study were about one-third of values determined in the eastern US following the Chernobyl accident [7].

Fig. 5
figure 5

Activities of airborne particulate 131I determined at six surveillance sites in NYS

Full size image
Table 2 Maximum airborne activities (μBq/m3) from sampling sites in New York State
Full size table
Fig. 6
figure 6

Activities of airborne gaseous 131I determined at five surveillance sites in NYS

Full size image

The gaseous component (131Ig) was collected on activated charcoal canisters at the six surveillance sites. Except for the 40K inherent to the charcoal and the short-lived progeny of collected radon, no other radionuclides were observed on the charcoal. As with 131Ip, no gaseous 131I has been detected from near the NYS reactor sites during several decades of sampling. However, a decade ago, 131Ig activity was detected on the charcoal canisters at the Albany (A) background site [2], presumably due to emissions from incineration of dried sewage sludge containing excrement from radioiodine therapy treatments. During an 8-month period when the sampler was adjacent to the incinerator, the airborne 131Ig concentrations averaged 1.1 ± 0.9 mBq/m3 (131Ip was not measured). At the current air-sampler location (~1.5 km from the incinerator), airborne 131Ig concentrations are consistently below the detection limit (roughly 0.04 mBq/m3). For the current study, 131Ig was collected at five of the six sampling sites (Fig. 6), with a range of 0.2–3.4 mBq/m3, and an overall average of 1.2 mBq/m3. Few Fukushima-related studies have reported 131Ig levels, thus comparisons are sparse. Though no individual results were provided, Masson et al. [5] reported peak 131Ig levels in European countries that exceeded those measured in NYS, possibly due to transport mechanisms and/or particle-to-gas conversion during transport and collection. The average 131Ig level determined in the current study was about half of that measured in the eastern US following the Chernobyl accident [7]. A comparison of Figs. 5 and 6 indicates that 131Ip and 131Ig levels peaked at the same time. Activities of 131Ig and 131Ip in the air samples were well correlated (r 2 = 0.80).

Over a three-year sampling period in the eastern US, gaseous iodine (stable) comprised nearly 90 % of the total atmospheric iodine [8], while sampling at the location following the Chernobyl accident found 131Ig comprised ~65 % of the total 131I [7]. For samples from the current study in which both phases were measured above the detection limit, 131Ig comprised, on average, 75 % of the total 131I (range = 60–88 %). This average is identical to the 77 % (±14 %) measured in some European countries [5]. In contrast to ratios of 137Cs/134Cs near unity, ratios of 131Ig/131Ip (Fig. 7) at the six sampling sites averaged 3.4 and varied greatly (1.5–7.5). In comparison, 131Ig/131Ip ratios in Washington state [4] varied between 2 and 20, with an average of ~5. The range of radioiodine ratios is likely a result of variations in weather (deposition, rainout, and dispersion) incurred by the emissions while transported from Japan. Of the radioisotopes determined in this study, poor correlations existed between 131Ig and 137Cs (r 2 = 0.18) and between 131Ip and 137Cs (r 2 = 0.14). Similarly poor correlations of these isotopes in the emissions from Japan have been reported elsewhere [5, 6].

Fig. 7
figure 7

Activity ratios of gaseous 131I to particulate 131I determined at five surveillance sites in NYS

Full size image

Naturally occurring 7Be was above the detection limit on all of the particulate filters, with values ranging from 1.4 to 5.2 mBq/m3 and an average of 3.2 ± 0.9 mBq/m3. As summarized elsewhere [2], 7Be activities are historically similar among the sites over a 13-year period, with an average activity that was nearly identical to that measured during the current study. A seasonal pattern of elevated 7Be levels occurs during the summer months due to the increased solar flux that occurs during this period, with concomitant increased production of 7Be. Due to the different sources, one would not expect 7Be levels to correlate with the released fission products, and correlations with the measured activities (131Ip r 2 = 0.36; 137Cs r 2 = 0.23) supports this conclusion.

In addition to the air-filter and charcoal samples collected from the six surveillance sites, supplemental sampling was implemented at two sites (A* and C) from March to May. Since these samples did not mimic the sampling schedule of the routine monitoring sites, the results are discussed separately. Samples from site C were collected for ~3 days each, resulting in reduced air volumes and fewer activities above detection limits. Ratios of 137Cs/134Cs at site C averaged 1.13, a value identical to that determined at the reactor surveillance sites. 131Ip was detected on 70 % of the filters, with an average (0.58 ± 0.46 mBq/m3; range = 0.04–1.46 mBq/m3) which is identical to that (0.55 mBq/m3) determined at site A. 131Ig was detected on 85 % of the charcoal canisters, with an average (1.3 ± 1.1 mBq/m3; range = 0.1–3.4 mBq/m3) which is nearly identical to the average (1.4 mBq/m3) determined at site A. These similarities are expected based on the close proximity (~15 km) of the two samplers. The ratios of 131Ig/131Ip averaged 2.6 (range = 1.2–4.9). On average, 131Ig comprised 65 % of the total 131I. 7Be was detected on all of the particulate filters, with values from 1.9 to 4.7 mBq/m3 and an average (3.4 ± 0.8 mBq/m3) that was comparable to activities determined at the reactor surveillance sites.

About 1 km from site A, high-volume sampling was conducted at ~3-day intervals with an average volume of ~5,500 m3. While the large volume allowed the detection of several radioisotopes [3], only a few are discussed here. Activity concentrations of 137Cs, which averaged 13 ± 9 μBq/m3 and ranged from 1 to 29 μBq/m3, were above detection limits on 90 % of the filters (Fig. 8). These concentrations were significantly less than those observed on the open-faced filters that were used at the six surveillance sites, possibly due to the 10-μm particle cutoff inlet used at site A* and/or inclusion of additional sampling periods containing low activities. Concentrations of 134Cs averaged 14 ± 10 μBq/m3 (range = 1–29 μBq/m3) and correlated well with 137Cs (r 2 = 0.97). For filters on which both isotopes were detectable, the ratios of 137Cs/134Cs averaged 1.05. During the first month of sampling at the site, 131Ip was measured on every filter (Fig. 9), with an average of 125 μBq/m3 (range = 8–360 μBq/m3), but 131Ip was not detectable (<1 μBq/m3) after April 22. This 30-day time frame of exposure to the Fukushima-related radioisotopes was also observed at the other sampling locations in NYS. Concentrations of 131Ig were detected on 84 % of the charcoal filters, and for a time period that exceeded that for 131Ip. The collected 131Ig activity averaged 155 μBq/m3 (range = 1–510 μBq/m3). The ratios of 131Ig/131Ip averaged 1.5 ± 0.4, with 131Ig comprising, on average, 58 % of the total 131I. Lastly, naturally-occurring 7Be was detected on all of the particulate filters, with activities from 250 to 1,270 μBq/m3 and an average of 750 μBq/m3. A comparison of average activities determined at site A* with those from nearby site A shows the latter’s concentrations were 4.0–4.8 times greater for 134Cs, 137Cs, 131Ip, and 7Be, and nearly 9 times greater for 131Ig. The greater discrepancy for 131Ig may be an artifact of the high flow rates (low contact time) through the charcoal filter paper used at site A* relative to lower flow through the thick charcoal canisters used at site A.

Fig. 8
figure 8

Airborne activities of 134Cs and 137Cs determined from high-volume sampling at site A*

Full size image
Fig. 9
figure 9

Airborne activities of particulate and gaseous 131I determined from high-volume sampling at site A*

Full size image

Gamma-ray emitting radioisotopes originating from Japan were also measured in fallout (wet and dry deposition) collected at sites A and C. Four of the six weekly samples collected at site A from March 19 to April 22 had 131I concentrations ranging from 220 to 3,800 mBq/m2-day. Considering the volume of rain (0.2–1.9 Lw) collected with the fallout samples, the liquid concentrations of 131I ranged from 110 to 1,470 mBq/Lw. In comparison, 131I activities in two deposition samples in Greece [6] were near 100 and 700 mBq/Lw. Deposition of 137Cs (n = 2) averaged 160 mBq/m2-day (46 mBq/Lw), while 134Cs (n = 1) deposition was 220 mBq/m2-day (400 mBq/Lw). Naturally occurring 7Be was detected in all six deposition samples from site A, with an average of 4.5 Bq/m2-day (3.1 Bq/Lw). Collection of fallout (n = 8) at site C, at intervals of 3–4 day each, had 131I concentrations above detection limits on half of the samples, with activities ranging from 350 to 5,100 mBq/m2-day (70–1,800 mBq/Lw). Deposition of 137Cs (n = 2) at site C averaged 180 mBq/m2-day (63 mBq/Lw), while 134Cs (n = 1) deposition was 140 mBq/m2-day (49 mBq/Lw). 7Be was detected in seven deposition samples at an average of 12 Bq/m2-day (4 Bq/Lw).

Dose assessment

Human exposure to the measured radioisotopes occurred through three pathways: inhalation, ingestion, and external irradiation. Only the radiological dose associated with inhalation of the observed radionuclides is provided, as similar calculations for ingested and external (submersion) radiation showed these two exposure pathways to be much less in magnitude than the inhalation dose. For a given time period, the committed effective dose equivalent (E(τ)) from inhaled gamma-ray emitters can be estimated, using Eq. 1, by factoring the measured concentrations (Ci) of the airborne radioisotopes during the period, an inhalation (I) rate (1.5 m3/h) [9], and adult dose coefficients (Di) of 20, 20, and 39 nSv/Bq for 131I, 134Cs, and 137Cs [10], respectively.

$$ {\text{E}}_{(\tau )} = {\text{S}}\left( {{\text{D}}_{\text{i}} *{\text{ C}}_{\text{i}} *{\text{ I}}} \right) .$$
(1)

For the sampling periods in which the isotopes were detected at the six air-monitoring surveillance sites in NYS, the resulting committed effective dose values averaged 0.33, 0.05, and 0.10 nSv/day to the total body from the inhaled 131Ip, 134Cs, and 137Cs, respectively. Similarly, the committed effective dose values from 131Ig averaged 0.87 nSv/day, with a range of 0.13–2.46 nSv/day at the surveillance sites. The total committed effective dose values for the 6-week sampling period ranged from 4.9 nSv at WES to 7.9 nSv at WAY. These two sites are located ~400 km apart. Comparison of the total derived committed effective dose from inhalation of the airborne radioisotopes to that (~7 μSv/day) received from natural sources (primarily radon and its decay products), demonstrates that the activities of the Fukushima-related radiation provided an insignificant additional dose.

The absorbed dose to the thyroid gland from 131I is dependent on age, with infants being the most affected. For this sensitive population (1 year olds), the inhalation Di is 3.2 μSv/Bq [10] for I2 uptake to the thyroid. When 131Ip was detected at the six air-monitoring surveillance sites in NYS, the committed dose equivalent to the thyroid from 131Ip averaged 49 nSv/day, with a range of 6–136 nSv/day. Using an identical approach, the committed dose equivalent from 131Ig averaged 130 nSv/day, with a range of 20–370 nSv/day at the sites. The total absorbed dose to the thyroid gland, determined by combining the contribution from 131Ip and 131Ig at each site for the 6 week sampling period, ranged from 690 nSv at WES to 1,040 nSv at ALB. This dose is trivial when compared to the US average daily radiation dose of ~17 μSv (6,200 μSv/year) from all sources.

Conclusions

Results of airborne samples collected following the Fukushima reactor accident and analyzed as part of an environmental surveillance network in NYS have been presented. Weekly filters of collected airborne-particulate matter had detectable 134Cs, 137Cs, 131I, and 7Be activity at all sites. Airborne activities of gaseous 131I, collected on activated charcoal, surpassed those collected on particulate matter. Airborne activity levels in NYS were considerably less than those measured in the eastern US following the Chernobyl accident. During the six-week sampling period, the health impact of Fukushima-related radionuclides in NYS was negligible, as demonstrated by comparing the increase in committed effective dose (<0.2 nSv/day) to that (~7 μSv/day) received from natural sources during the period, and posed no concern to public health.