Radioactive contamination of southeast Abai oblast, Kazakhstan, from the Chinese nuclear weapons testing program at Lop Nor: an analytical review

Abstract

Between 1949 and 1962 the Soviet Union performed atmospheric tests of nuclear weapons at the Semipalatinsk nuclear test site (SNTS) in Kazakhstan, resulting in widespread contamination of the surrounding region with radioactive fallout. Settlements in the southeast Abai oblast of Kazakhstan, close to the border with China, are not thought to have received significant fallout from the SNTS. There is, however, evidence that the study area, including Makanchi, Urdzhar and Taskesken villages, was contaminated by atmospheric nuclear tests performed by China at the Lop Nor NTS between 1964 and 1980. We identified the most reliable data indicating contamination from the Lop Nor tests from archive documents. Prompt sampling of soil performed in the area revealed elevated levels of total beta activity in the days and weeks following the Lop Nor tests. The highest activities were recorded following the thermonuclear tests in June 1967 and June 1973. Tooth enamel dosimetry using electron paramagnetic resonance methods suggests residents of the study area have been exposed to excess doses of 50–60 mGy but provides no information on the source and timing of exposure. Currently, evidence of contamination of the study area from nuclear weapons testing at Lop Nor is based on limited radiation measurement data. Therefore, work will continue on the search for archival data on radiological and meteorological monitoring carried out in the study area at the time of the Lop Nor testing campaign.

INTRODUCTION

Atmospheric nuclear weapons testing performed at the Semipalatinsk nuclear test site (SNTS) in Kazakhstan in 1949–1962 resulted in radioactive contamination of populated areas adjacent to the test site. The radioactive contamination of the area surrounding SNTS has received considerable attention [1–3]. Settlements east of SNTS, in the present day Abai oblast (oblasts are administrative division units, similar to provinces or regions), roughly equivalent of the historic Semipalatinsk oblast, were the most heavily contaminated [2]. Settlements southeast of Abai oblast, including the villages of Makanchi, Urdzhar and Taskesken, received only limited contamination from the SNTS tests [1]. This area is situated 900–1100 km away from the Lop Nor NTS (LNNTS) in China where 22 atmospheric nuclear weapons tests were performed between 1964 and 1980 [4].

The key institute involved in monitoring of radioactivity contamination of the environment and following-up the population affected by SNTS tests is the Scientific Research Institute of Radiation Medicine and Ecology (SRIRME), located in the city of Semey (formerly Semipalatinsk), Kazakhstan. SRIRME stores historical reports and archival data on radioecological measurements performed after nuclear weapons tests at SNTS, and to lesser extent after the tests at LNNTS, enabling retrospective exposure assessment [1, 5].

From the mid 1980s through mid 1990s reports were published based on official medical statistics on elevated cancer mortality rates in the population of southeast Abai oblast, compared to previous years [6], suggesting that contamination from LNNTS could be responsible. However, there was no systematic effort reviewing the existing evidence of contamination to assess exposure levels and potential health risks in the residents of southeast Abai oblast settlements.

Here, we review the available evidence of this contamination, focusing on (i) soil sampling campaigns performed in Kazakhstan at the time of the Lop Nor tests, (ii) tooth enamel dosimetry using electron paramagnetic resonance (EPR) methods, and (iii) contribution from Semipalatinsk NTS to contamination of the area.

ANALYSIS

Characteristics of southeast Abai oblast study area

In this paper, the study area corresponds to the present day Makanchi and Urdzhar raions (raions are oblast subdivision units, similar to districts) in southern Abai oblast, 370–650 km southeast from SNTS and 900–1100 km northwest from the LNNTS (Fig. 1). The study considered the three most populated villages, Makanchi, Urdzhar and Taskesken, in Makanchi and Urdzhar raions in the southeast of Abai oblast, Kazakhstan. The territory of the Makanchi and Urdzhar raions is agricultural, with an emphasis on livestock and crop production. The relief of the study area is mostly flat steppe with mountains stretching along the northern and southern borders. The predominant soil types are gray soil, mountain chestnut and mountain chernozem. There are no urban-type settlements. All settlements are rural-type villages with predominantly single-family houses built from adobe. According to the All-Union Census, the population size of the study area in 1970 (during the period of LNNTS activity) was 93 000 people, predominantly (around 90%) of whom were ethnic Kazakhs [7].

Fig. 1.

Regional map, showing the locations of test sites and study region. Note: Semey was known as Semipalatinsk City before 2007.

Characteristics of nuclear tests

The Lop Nor (or Lop Nur, Lob Nor) nuclear testing site is located in the Xinjiang autonomous region of China, around 1500 km southeast of the Semipalatinsk NTS and around 1000 km from the border between China and Kazakhstan (Fig. 1). Between 1964 and 1980, 22 atmospheric nuclear weapons tests were conducted at the LNNTS in western China. The key parameters of the tests are given in Table 1 [4, 6].

Table 1.

Atmospheric nuclear tests performed at Lop Nor by China

Test date	Yield (kt)a	Type	Release	Height above groundb
16 October 1964	20	Fission	Tower	30 m
14 May 1965	40	Fission	Aircraft	150–350 m
09 May 1966	300	Thermonuclear	Aircraft	1.0–1.5 km
27 October 1966	20	Fission	Missile	150–350 m
28 December 1966	300	Thermonuclear	Tower	100–150 m
17 June 1967	3000	Thermonuclear	Tower	100–150 m
24 December 1967	20	Fission	Tower	30 m
28 December 1968	3000	Thermonuclear	Aircraft	6–8 km
29 September 1969	3000	Thermonuclear	Aircraft	6–8 km
14 October 1970	3000	Thermonuclear	Aircraft	6–8 km
18 November 1971	20	Fission	Tower	150–350 m
07 January 1972	20	Fission	Tower	100–150 m
18 March 1972	100	Thermonuclear	Aircraft	1.0–1.5 km
27 June 1973	2500	Thermonuclear	Aircraft	1.0–1.5 km
17 June 1974	600	Thermonuclear	Aircraft	1.0–1.5 km
23 January 1976	20	Fission	Ejection to atmosphere	Underground
26 September 1976	100	Fission	Tower	30 m
17 November 1976	4000	Thermonuclear	Aircraft	1–1.5 km
17 September 1977	20	Fission	Tower	30 m
15 March 1978	20	Fission	Tower	30 m
14 December 1978	20	Fission	Tower	30 m
16 October 1980	600	Thermonuclear	Missile	8.0 km

^aUNSCEAR [4].

^bGusev et al. [6].

More than 100 fission products and their decay products as well as products of activation of the fuel, bomb construction and other surrounding materials with relatively short half-lives are produced in nuclear explosions and contribute to radioactive fallout. Measurements of elevated levels of radioactivity in the environment performed shortly after tests allow determination of whether the given area was contaminated by fallout. The most informative data available for the study were archival records of measured radioactivity in soil samples, described below.

Evidence from soil sampling

The results of measurements of total beta activity in soil samples are given in the English language report by Gusev et al. [6] published by the US Department of Defense. Unfortunately, most of the hard copy documents from the archives of SRIRME containing data relevant to the tests at LNNTS are not thought to have survived the disruption caused by the breakup of the Soviet Union in 1991.

According to Gusev et al. [6], soil samples were collected during the period 1967–1981 in Makanchi, Urdzhar and Taskesken. The sampling site in each village was selected to represent undisturbed soil surfaces as far as possible from buildings or trees that might have influenced local deposition. Large and sharp stones were removed from the site where the soil sample was taken. At each village, soil samples were collected using a sampling shovel with dimensions of 0.1 × 0.1 m at depth of 0.01 m.

The total beta activity of the samples was measured using B-2 beta radiometer with MST-17 or SBT-10 Geiger-Mueller counters. Between 1967 and 1973, beta radiometers were calibrated using an equilibrium ⁹⁰Sr + ⁹⁰Y calibration source [6]. The average beta emission energy of radioactive debris in the soil samples was not considered, and this energy differed from the average energy of beta emission of ⁹⁰Sr + ⁹⁰Y calibration source. Later, beta radiometers were calibrated using a set of calibration sources (¹⁴C, ²⁰⁴Tl, ⁴⁰K, ⁹⁰Sr + ⁹⁰Y, ⁹⁰Y) with average energy of beta emission ranging from 0.05 to 0.933 MeV, meaning measurement results accounted for the average energy of beta emission in the soil samples. Soil samples were measured in an aluminum dish with a diameter of 3.2 cm and a height of 0.6 cm.

The report by Gusev et al. [6] contains activity values for individual radionuclides in soil samples in Bq kg⁻¹. These figures were probably derived from the measured total beta activity, since the activity of long-lived ¹³⁷Cs and ⁹⁰Sr appear to fluctuate between high (≈5 × 10³ Bq kg-1) and relatively low levels (<50 Bq kg⁻¹) in a manner that cannot be explained by radioactive decay, soil migration or weathering, given the timescale. The report by Gusev et al. [6] also includes values of exposure rates in air (in mR h⁻¹) at 24 hours post detonation, ground deposition density (in Bq m⁻²) and activity concentration in milk (in Bq L⁻¹) and vegetation (in Bq kg⁻¹) for ¹³¹I, ¹³³I, ¹³⁵I, ⁹⁰Sr and ¹³⁷Cs. These figures were previously presented in the archival reports of SRIRME [8, 9]. However, these figures were estimated from test parameters (yield, height of explosion, height of cloud, wind speed, distance to settlement, etc.) using equations given in archival reports, rather than being directly measured. For this reason, we consider total beta activity figures for soil to be the most reliable data indicating contamination from the Lop Nor tests. Table 2 summarizes the data on the measured total beta activity in soil reported by Gusev et al. [6]. The delay between the test and sampling ranged from 4 to 94 days and was less than two weeks for 6 tests. The majority of contamination followed the tests performed on 17 June 1967 and 27 June 1973 (both thermonuclear), in which a total beta activity of around 10⁶ Bq kg⁻¹ shortly (8–13 days) after detonation was measured. Soil samples were also collected in 1963, before the first Lop Nor nuclear test in 1964. Average total beta activities for these samples were 1120, 994 and 1080 Bq kg⁻¹ for Makanchi, Urdzhar and Taskesken, respectively [6]. These activities are lower than those recorded following the 1967 and 1973 tests by three orders of magnitude.

Table 2.

Total beta activity in soil sample (Bq kg⁻¹) for sites representing Makanchi, Urdzhar and Taskesken following Lop Nor nuclear tests [6]

Test date	Sample date	Delay (days)	Total beta activitya (Bq kg−1)
			Makanchi	Urdzhar	Taskesken
28 December 1966	01 April 1967	94	1.95 × 104	1.74 × 104	1.47 × 104
17 June 1967	25 June 1967	8	1.38 × 106	1.09 × 106	1.02 × 106
18 March 1972	10 April 1972	23	1.05 × 103	1.03 × 103	8.26 × 102
27 June 1973	10 July 1973	13	8.52 × 105	8.17 × 105	6.94 × 105
26 September 1976	30 September 1976	4	8.04 × 103	6.96 × 103	5.65 × 103
17 September 1977	30 September 1977	13	1.71 × 103	1.43 × 103	1.21 × 103
15 March 1978	20 March 1978	5	8.00 × 103	6.98 × 103	5.79 × 103
14 December 1978	15 March 1979	91	2.29 × 102	1.88 × 102	1.61 × 102

^aArithmetic mean of five samples.

Evidence from EPR of tooth enamel measurements

EPR, also known as electron spin resonance (ESR), can be used as a retrospective dosimetry technique, allowing dose estimates to be obtained many years after exposure [10]. Zhumadilov et al. [11] analysed tooth enamel samples from 22 adults who resided in Makanchi, Urdzhar and Taskesken between 1966 and 1981. In each case, the year of enamel formation was before 1950 (range: 1941–1948). Estimated excess doses were calculated by subtracting 0.8 mGy per year of tooth enamel age. Samples were also taken from five individuals from the town of Kokpekty, which the authors considered to have not been contaminated by any nuclear tests [11]. In this case, the year of enamel formation was between 1964 and 1975. A summary of the results is shown in Table 3. Note that according to Zhumadilov et al. [11], ‘the negative values obtained arise because the measurements were performed near the threshold of sensitivity of the method’.

Table 3.

Summary of EPR dosimetry results of Zhumadilov et al. [11]

Residence	Sample size	Excess dose (mGy)
		Mean ± SE	Range ± SE
Makanchi	6	62 ± 28	5 ± 26–123 ± 32
Urdzhar	6	64 ± 30	23 ± 26–118 ± 32
Taskesken	10	49 ± 27	−6 ± 26–107 ± 31
Kokpektya	5	−19 ± 36	−66 ± 39–24 ± 39

SE: standard error.

^aControl group considered as unexposed.

Contribution of Semipalatinsk NTS to contamination of the study area

Gordeev et al. [2] identified eleven atmospheric SNTS nuclear tests between 1949 and 1962 resulting in effective doses from external irradiation exceeding 5 mSv for adult residents around SNTS. Of these, only two resulted in radioactive clouds traveling southeast of SNTS, in the direction of the study area (southeast Abai oblast), i.e. those performed on 24 September 1951 (total yield of 38 kt) and 12 August 1953 (total yield of 400 kt). Figure 2 shows a map of isodose lines for unshielded ground level external doses in Roentgen (in air, 1 R = 8.77 mGy) for these two tests [12]. Radioactive cloud trajectories for the other nine tests are also shown, based on Gordeev et al. [2].

Fig. 2.

Map showing isodose lines for unshielded external doses in Roentgen for the atmospheric nuclear weapons tests conducted in SNTS on 24 September 1951 (test #2) and 12 September 1953 (test #4) which might affect the study area [12]. Also shown are the radioactive cloud trajectories for nine SNTS tests that did not affect the study area, based on [1,2]: 29 August 1949 (test #1), 5 May 1954 (test #13), 30 October 1954 (test #18), 29 July 1955 (test #19), 02 August 1955 (test #20), 16 March 1956 (test #26), 24 August 1956 (test #28), 22 August 1957 (test #41), and 07 August 1962 (test #148). Point measurements for Karaul, Barshatas and Ayagoz represent exposure rate at 12 hours, post-detonation.

The radioactive cloud from the 12 August 1953 test swept round to the east before passing to the north of the study area (Fig. 2). The time of arrival (TOA) of the edge of the radioactive cloud at the village Ayagoz, located about 100 km north of Taskesken, was 6 hours post detonation (H + 6 h). An exposure rate at H + 12 h, , of 1.0 mR h⁻¹ was reported for this location by Gordeev [13]. The arithmetic mean of individual doses from external irradiation from this test for sixteen residents of Ayagoz included in the study by Land et al. [14] was 0.15 mGy. Therefore, doses from external irradiation to residents of the study area as a result of this test were significantly lower than 1 mGy.

The radioactive cloud resulting from the 24 September 1951 test could potentially have reached the study area. The nearest village with available radiation data is Barshatas with a TOA of 10 h and reported value of 50 mR h⁻¹ [13]. The arithmetic mean of external doses for eight residents of Barshatas included in the study by Land et al. [14] from this test was 5 mGy. Considering the distance between Barshatas and the settlements of the study area i.e. Taskesken (200 km), Urdzhar (250 km), and Makanchi (300 km) as well as the average wind speed of 26.4 km h⁻¹ [13], doses from external irradiation from fallout from this test to residents of the study area did not exceed 3 mGy. In addition, exposure rate of 4 mR h⁻¹ measured at 20 h after detonation in Taskesken was reported by Logachev et al. [15]. Considering a TOA of 17 h after detonation, this resulted in doses from external irradiation to residents of Taskesken around 1 mGy.

Underground tests performed at SNTS after 1963 are also unlikely to be the source of contamination of the study area. Three underground SNTS tests were performed in the 60 days preceding the 17 June 1967 LNNTS test, one on 20 May 1967 (yield of 20–150 kt), and two on 28 May 1967 (<20 kt) [16]. No SNTS tests were performed in the 60 days preceding the 27 June 1973 LNNTS test, though three SNTS tests were performed on the same day as the sampling, all of which were underground and below 20 kt in yield [16].

DISCUSSION

In this paper, we have gathered and critically reviewed available archival and published data providing evidence of contamination of southeast Kazakhstan from the Chinese nuclear weapons tests performed at the Lop Nor test site. Results of total beta activity measurements in soil sampled after the tests suggest radioactive contamination of Makanchi and Urdzhar raions in the southeast Kazakhstan occurred in the days and weeks following the atmospheric nuclear weapons tests conducted in LNNTS. High total beta activity was detected in the surface layers of soil following two thermonuclear tests at Lop Nor in particular, performed on 17 June 1967 and 27 June 1973 [6]. Tests performed at the SNTS around the same time were of much lower yield and detonated underground, meaning they were unlikely to be the source of detected activities. France, which was not a signatory of the Partial Test Ban Treaty, performed two tests in the south Pacific in June 1967. However, due to weak exchange between air basins of the southern and northern hemispheres, it is also unlikely that these tests caused contamination of Kazakhstan.

The study area was not the only territory of the former USSR that received contamination from the Lop Nor tests. Elevated activities during the period of Chinese nuclear tests were recorded by a network of monitoring stations throughout the Soviet Union, including the cities of Saint-Petersburg (formerly Leningrad) and Petropavlovsk-Kamchatsky and territories of Uzbekistan and Kyrgyzstan [17]. Outside the USSR, Burchfield et al. [18] attributed raised activity levels of ⁹⁰Sr in Arkansas, United States, between 1973 and 1981 to the Lop Nor nuclear tests. Simpson et al. [19] also reported elevated levels of fission products activity (¹³¹I, ¹⁴⁰Ba + La, and ¹³⁷Cs) in cow’s milk in Rhode Island and New Hampshire (USA) following testing at Lop Nor on 26 September 1976.

China operated an extensive system of environmental monitoring stations, beginning in the early 1960s [20], although publicly available data are limited. Total beta activity levels reached a peak in 1962, before the first Chinese tests, with smaller peaks observed throughout the late 1960s and 1970s, corresponding to the Lop Nor nuclear testing program. Deposition of ¹³¹I reached 10 kBq m⁻² in Lanzhou (east of Lop Nor) following the 27 June 1973 test [20]. There are no available data on radiation measurements at the time of testing campaign at LNNTS for sites located between Lop Nor and the border with Kazakhstan.

Contemporary sampling studies provide little evidence for major contamination of the study area from Lop Nor nuclear tests. An extensive program of soil sampling was performed in the late 1990s and early 2000s by Yamamoto et al. [21–26] But this largely focused on sites in the vicinity of SNTS. This contemporary sampling campaign determined levels of ¹³⁷Cs and isotopes of plutonium in soil and their associated ratios. Ground contamination by ¹³⁷Cs varies in the study area from 1.5 to 5.0 kBq m⁻² at the time of measurements in the early 2000s. These activities were slightly higher than global average levels but lower than for sampling sites closer to SNTS (12.1–21.2 kBq m⁻²). The mean ratio of ²⁴⁰Pu/²³⁹Pu is soil within the study area was ≈0.10 (based on three samples) suggesting contamination in addition to global fallout, though this could be either from SNTS or LNNTS.

Previous research has suggested EPR dosimetry can detect excess radiation exposure from nuclear weapons tests [3]. The EPR results performed by Zhumadilov et al. [11] provide evidence that the residents of Makanchi, Urdzhar and Taskesken have received relatively high radiation doses, although it is not possible to distinguish the exposure source or timing. Enamel formation was before 1950 for all subjects, so exposures could also be due to atmospheric nuclear weapons tests conducted at SNTS. However, thorough analysis and mapping of radioactive cloud trajectories after atmospheric tests performed at SNTS shows that population exposure in the study area from these tests was in the order of few mGy only.

To confirm the extent and levels of radioactive contamination of the study area from LNNTS, a search for other radiation measurements is required. This includes exposure rates in air, total beta activity in milk from dairy animals along with meteorological information, such as precipitation, wind speed and direction. Work will continue on the search for data that could possibly be available in various organizations, including meteorological stations of the State Committee for Hydrometeorology, Sanitary and Hygiene Centres of the Ministry of Health of the Republic of Kazakhstan, veterinary laboratories of the State Agro-Industrial Committee of the former USSR, Production Association ‘Typhoon’ (former Institute of Experimental Meteorology). However, this task requires significant effort and resources. A further source of information could be retrospective luminescence dosimetry of quartz-containing bricks, from buildings that were constructed before the period of testing. This method has been previously used to investigate doses from the SNTS tests in nearby settlements [27, 28]. However, given the EPR tooth enamel dose measurements, and assuming a reduction factor due to being indoors of ≈0.4–0.6, doses to external materials are unlikely to exceed ≈150 mGy, thus may be below the minimum detectable dose by this method.

The long-term goal of the study is to assess the exposure levels and potential radiation-associated health risks in the residents of Makanchi, Urdzhar and Taskesken raions contaminated by radioactive fallout from atmospheric nuclear weapon tests conducted at Lop Nor NTS.

In conclusion, evidence of contamination of Kazakhstan by the Chinese nuclear weapons testing program at LNNTS is mostly limited to the results of sampling performed close to the time of the Lop Nor tests. These samples revealed highly elevated levels of total beta activity in the superficial soil layer in the days and weeks following the tests. In addition, EPR results provide evidence that the residents of the study area received relatively high radiation doses, although it is not possible to distinguish the exposure source or timing. Contemporary evidence is limited as most of the major contaminating radionuclides have decayed away. Current levels of ¹³⁷Cs and ratios of plutonium isotopes neither confirm nor contradict contamination of Kazakhstan from Chinese nuclear tests. Work will continue on the search for the results of radiological and meteorological monitoring carried out in the study area at the time of the Lop Nor testing campaign. If the results became available, it would substantially contribute to the body of evidence.

CONFLICT OF INTEREST

All authors declare that they have no competing interests.

FUNDING

This work was supported by the agreement on grant financing of scientific start-up projects of scientific and pedagogical personnel of the NCJSC Semey Medical University for 2022–2025 (grant number 378, dated 12 September 2022) and is based at the Research Institute of Radiation Medicine and Ecology of the NCJSC Semey Medical University.

DATA AVAILABILITY

All data presented in this paper are publicly available.

DISCLAIMER

Where authors are identified as personnel of the International Agency for Research on Cancer/World Health Organization, the authors alone are responsible for the views expressed in this article and they do not necessarily represent the decisions, policy or views of the International Agency for Research on Cancer/World Health Organization.