Radioactive Fallout and Potential Fatalities from Nuclear Attacks on China’s New Missile Silo Fields

Abstract

China is constructing three new nuclear ballistic missile silo fields near the cities of Yumen, Hami, and Ordos as part of a significant buildup of its nuclear weapon arsenal. Once operational, these missile silos will likely be included as targets in U.S. strategic counterforce war plans. Such plans call for using one or two nuclear warheads to strike each silo. Such attacks can generate large amounts of radioactive debris that are then transported by local winds and can deliver lethal doses of radiation to population hundreds of kilometers away. Here, we compute radioactive fallout from counterforce attacks on the three new alleged missile silo fields using the combination of a nuclear war simulator and modern atmospheric particle transport software and recent archived weather data. We find that the construction of these new silos will put tens of millions of Chinese civilians at risk of lethal fallout including in East China. In particular, the relatively short distance between the Ordos missile field and Beijing and the local winds patterns for the region, suggest that about half of the 21 million inhabitants of the Chinese capital could die following a counterforce strike, even if given advanced warning to shelter in place.

Acknowledgements

The authors gratefully acknowledge the NOAA Air Resources Laboratory (ARL) for the provision of the HYSPLIT transport and dispersion model used in this publication. Figures 1, 3, and 4 use map data from Mapbox and OpenStreetMap and their data sources. Figure 2 use data from C. Beccario (https://earth.nullschool.net/). We thank Frank von Hippel and Zia Mian for useful discussions and feedback on the manuscript.

Disclosure statement

Ivan Stepanov is the lead developer of the Nuclear War Simulator, a commercial software published by Slithering Software UK Ltd used in this study. Sébastien Philippe declares no conflicts of interest.

Code availability statement

The source code for the HYSPLIT transport and dispersion model is available upon request from the NOAA Air Resources Laboratory (https://www.ready.noaa.gov/HYSPLIT_linux.php).

Data availability statement

All data generated or analyzed during this study are either included or cited in the published article and are available from the corresponding author upon reasonable request.

Notes

1 Warrick, Joby, “China is Building more than 100 New Missile Silos in its Western Desert, Analysts Say,” Washington Post, 30 June 2021; Korda, Matt and Hans Kristensen, “China is Building a Second Nuclear Missile Silo Field,” The Strategic Security Blog, Federation of American Scientists, 26 July 2021, https://fas.org/blogs/security/2021/07/china-is-building-a-second-nuclear-missile-silo-field/; Broad, William J. and David E. Sanger, “A 2nd New Nuclear Missile Base for China, and Many Questions about Strategy,” New York Times (26 July 2021); Lee, Rod, “PLA Likely Begins Construction of an Intercontinental Ballistic Missile Silo Site near Hanggin Banner,” China Aerospace Studies Institute, 12 August 2021, https://www.airuniversity.af.edu/CASI/Display/Article/2729781/pla-likely-begins-construction-of-an-intercontinental-ballistic-missile-silo-si/. In addition to the three identified silo sites, China has also built an important ICBM training site at Jilantai, which may achieve some operational capabilities. See: Hans Kristensen, “China’s Expanding Missile Training Area: More Silos, Tunnels, and Support Facilities,” The Strategic Security Blog, Federation of American Scientists, 24 February 2021, https://fas.org/blogs/security/2021/02/plarf-jilantai-expansion/.

2 Silo coordinates were obtained from the Federation of American Scientists, Data as of 15 January 2022. See: https://public.tableau.com/app/profile/kate.kohn/viz/YumenMissileSiloField/Dashboard1; https://public.tableau.com/app/profile/kate.kohn/viz/HamiMissileSiloField/Dashboard1; https://public.tableau.com/app/profile/kate.kohn/viz/OrdosMissileSiloField/Dashboard1.

3 U.S. Central Command recently disclosed the presence of a U.S. ballistic missile submarine (SSBN) in the Arabian sea. From such a station, a U.S. SSBN could target China’s new missile silo fields without flying missiles above Russia. See: Moriyasu, Ken. “Stealthiest U.S. submarine makes rare appearance in Arabian Sea,” Nikkei Asia, 20 October 2022.

4 Drell, Sidney D., and Frank Von Hippel. “Limited Nuclear War,” Scientific American 235, no. 5 (1976): 27–37; Von Hippel, Frank N., Barbara G. Levi, Theodore A. Postol, and William H. Daugherty, “Civilian Casualties from Counterforce Attacks,” Scientific American 259 (1988): 36–43; McKinzie, Matthew, Thomas B. Cochran, Robert S. Norris, and William M. Arkin, The U.S. Nuclear War Plan: A Time for Change (Washington, DC: Natural Resources Defense Council, 2001), 1–198; Helfand, Ira, Lachlan Forrow, Michael McCally, and Robert K. Musil, “Projected US Casualties and Destruction of US Medical Services from Attacks by Russian Nuclear Forces,” Medicine & Global Survival 7 (2002): 68–76.

5 Rolph, G. D., F. Ngan, and R. R. Draxler, “Modeling the Fallout from Stabilized Nuclear Clouds using the HYSPLIT Atmospheric Dispersion Model,” Journal of environmental radioactivity 136 (2014): 41–55; Philippe, Sébastien, Sonya Schoenberger, and Nabil Ahmed, “Radiation Exposures and Compensation of Victims of French Atmospheric Nuclear Tests in Polynesia,” Science and Global Security 30 (2022): 1–33, DOI: 10.1080/08929882.2022.2111757.

6 Stein, A.F., Draxler, R.R, Rolph, G.D., Stunder, B.J.B., Cohen, M.D., and Ngan, F., “NOAA’s HYSPLIT Atmospheric Transport And Dispersion Modeling System,” Bulletin of the American Meteorological Society 96 (2015): 2059–2077, http://dx.doi.org/10.1175/BAMS-D-14-00110.1

7 Ivan Stepanov, Nuclear War Simulator, build 394, www.nuclearwarsimulator.com

8 This yield is typical of the U.S. arsenal. The silo-based LMG-30G Minuteman III intercontinental ballistic missiles carry 300-kiloton warheads and Sea-launched ballistic missile carry warheads in the 90 to 455-kiloton range, see: Kristensen, Hans M., and Matt Korda, “United States Nuclear Weapons, 2021,” Bulletin of the Atomic Scientists 77 (2021): 43–63.

9 The size, surface and volume of fallout particles are all log-normally distributed. Each representing the 1st, 2nd, and 3rd moments of the distribution, respectively. We use the 2.5 moment of the distribution to model to a first order the effect of fractionation between refractive and volatile nuclide who are typically located within the particle volume and on the particle surface, respectively. See Bigelow Jr, Winfield S. Far field fallout prediction techniques. Air Force Institute of Technology, PhD Thesis, 1983, 1–172, https://apps.dtic.mil/sti/pdfs/ADA151871.pdf. For our particle source term, we use the DELFIC mean diameter and standard deviation for ground bursts (obtained from particle measured at the U.S. Nevada test site) dividing the size-distribution into 100 size bins such that the integrated activity within each bin is equal. The inverse cumulative distribution function is used to determine the minimum and maximum particle size of each bin. The representative particle size of each bin is then calculated as the geometrical mean of the edges of the bin. See Norment, Hillyer G., “DELFIC: Department of Defense Fallout Prediction System. Volume I-Fundamentals,” Atmospheric Science Associates, Bedford MA, 1979, 1–101, available at https://apps.dtic.mil/sti/pdfs/ADA088367.pdf.

10 The spatial distribution of cloud particle is obtained from the model developed by Conners, Stephen P., “Aircrew Dose and Engine Dust Ingestion from Nuclear Cloud Penetration,” Air Force Institute of Technology, PhD Thesis, 1985, 82–83, https://apps.dtic.mil/sti/pdfs/ADA159246.pdf

11 For a yield of 300 kT, the standard deviation is about 2 km. See Hanifen, Dan W. Documentation and Analysis of the WSEG-10 Fallout Prediction Model. Air Force Institute of Technology, 1980, 7 and Edward Geist, GLASSTONE-nuclear weapons effects modelling in Python, 15 August 2019, https://github.com/GOFAI/glasstone

12 The GDAS data is also available on a 0.5 by 0.5 and 0.25 by 0.25 degree-grids. We use the 1-degree grid to optimize storage space as we run HYSPLIT locally. See, NOAA Air Resources Laboratory (ARL), “Global Data Assimilation System (GDAS1) Archive Information,” (2004). https://www.ncei.noaa.gov/products/weather-climate-models/global-data-assimilation

13 See Hanifen, Dan W., “Documentation and Analysis of the WSEG-10 Fallout Prediction Model,” Air Force Institute of Technology, 1980, 1–105 and Edward Geist, GLASSTONE-nuclear weapons effects modelling in Python, 15 August 2019, https://github.com/GOFAI/glasstone

14 In HYSPLIT, wet removal is defined as β_wet = 8 × 10^–5 P^0.79 with P the precipitation rate in mm/h. We used the HYSPLIT default scavenging coefficient (8 × 10^-5), applied both to below and within-cloud scavenging, which apply to a wide range of particle sizes. See Sportisse, Bruno, “A Review of Parameterizations for Modelling Dry Deposition and Scavenging of Radionuclides,” Atmospheric Environment 41 (2007): 2683–2698.

15 This method has been used for example in earlier iteration of the U.S. DOE DELFIC fallout prediction code. See: Norment, Hillyer G., “DELFIC: Department of Defense Fallout Prediction System,” Volume I-Fundamentals, op.cit.

16 Standard values for K are available from the 1979 DELFIC source code. Typical values range from 6.11 10⁷ to 7.2 10⁷ (Gy/h)/(kT/m²) depending on the nuclide and the fission energy. See: Norment, Hillyer G. DELFIC: Department of Defense Fallout Prediction System. Volume II. User’s Manual. Atmospheric Science Associates, Bedford MA, 1979, 161, and Bigelow Jr, Winfield S., “Far Field Fallout Prediction Techniques,” III-4.

17 This is sometimes referred as a 50% fission fraction in the literature. See, McKinzie, Matthew et al., “The U.S. Nuclear War Plan: A Time For Change,” Op. cit.

18 Florczyk, Aneta J., Christina Corbane, Daniele Ehrlich, Sergio Freire, Thomas Kemper, Luca Maffenini, Michele Melchiorri et al., “GHSL Data Package 2019,” Luxembourg, EUR 29788, no. 10.2760 (2019): 290498.978-92-76-13186-1, doi:10.2760/290498, JRC 117104

19 McClellan, Gene, David Crary, and Darren Oldson, Approximating the Probability of Mortality Due to Protracted Radiation Exposures. DTRA-TR-16-054, Applied Research Associates, Inc. Arlington United States, 2016.

20 Buddemeier, B. R., and M. B. Dillon, “Key Response Planning Factors for the Aftermath of Nuclear Terrorism,” No. LLNL-TR-410067. Lawrence Livermore National Lab (LLNL), Livermore, CA (United States), 2009.

21 Yang, Chen, and Shuqing Zhao, “A Building Height Dataset Across China in 2017 Estimated by the Spatially-Informed Approach,” Scientific Data 9 (2022): 1–11.

22 We ran 365 counterforce attacks twice for each day of the year (once with all silo fields and once with Ordos only) as well as 10 attacks on each silo of the Ordos field in the span of 24H on 29 April 2021 (820 simulations in total).

23 Kristensen, Hans M., Matthew McKinzie, and Theodore A. Postol, “How US Nuclear Force Modernization is Undermining Strategic Stability: The Burst-Height Compensating Super-Fuze,” Bulletin of the Atomic Scientists, March 1, 2017, https://thebulletin.org/2017/03/how-us-nuclear-force-modernization-is-undermining-strategic-stability-the-burst-height-compensating-super-fuze/

24 We use a semi-empirical model based on data from U.S. atmospherics test fallout in Nevada with f_d = 0.45345^λ\65 where λ = (h/Y^1/3) is the scaled height of burst in feet per kT^1/3. See, Norment, “DELFIC: Department of Defense Fallout Prediction System,” Volume I-Fundamentals, 53.

25 For the dimensions of the 10,000 psi kill volume, see Glasstone, Samuel and Philip J. Dolan, The effects of nuclear weapons, U.S. Department of Defense, 1977, 111.

26 We extracted data from the GDAS1 weather archive using the ARLreader python package, see Martin Radenz, Yin Zhenping, ARL reader, 29 April 2020, https://github.com/martin-rdz/ARLreader

27 McClellan, Gene, David Crary, and Darren Oldson. Approximating the Probability of Mortality Due to Protracted Radiation Exposures. DTRA-TR-16-054, Applied Research Associates, Inc. Arlington United States, 2016.