Supporting Defense and Safety using Satellite Imagery

Supporting Defense and Safety using Satellite Imagery


Danish Ahmad
Software System Architect
Cres-Tech Pvt. Ltd.
[email protected]

In the current era, proliferation of nuclear weapons poses a great threat to international peace and security. Defense is the most sensitive, critical and important issue for any regime. In these circumstances the Nuclear powers, International Atomic Energy Agency (IAEA), United Nations have an important role to stop this massive development. I present various case studies for observation and remote monitoring of such nuclear facilities.

To start with, any conscientious person will go for “Identification of Enrichment Plants” first. There are two reliable identification and monitoring approaches, Feature Oriented Identification of Enrichment plants and Thermal identification to ensure working of plants.

Feature Oriented Identification
With the advancement in Remote Sensing, high resolution imagery (of 1m) from Commercial satellites (like IKONOS, LANDSAT etc.) is easily available. One can easily identify the huge typical structures of a nuclear constellation like Reprocessing Plant, Gaseous Diffusion Plant and Reactors. Let’s look at there respective observable features [2]:

Nuclear Facility Observable Features
Reactors Have cooling towers or natural water body (with intake and discharge ports), a high narrow stack (or its shadow); a reactor building, security perimeter; rail roads; an isolated site etc.
Reprocessing plants A very high stack (or its shadow); a long “canyon like” building; some holding ponds or reservoirs for waste or sludge; security perimeter; rail roads; an isolated site etc.
Gaseous Diffusion Plants Large area (roof) process buildings (roof of most buildings have ventilation shaft) ; cooling towers or a nearby river or lakes; a nearby fossil fuel power plant; large electric switchyard (substation); waste management and disposal facilities; security perimeter; rail roads; an isolated site etc.

Let’s view few satellite images of important nuclear facilities of Russia, India and Pakistan. One can easily pin-point in 1m IKONOS images almost all the features mentioned above:

Fig 1. 1m resolution IKONOS satellite image (captured July 2000) of site of plutonium production reactors at Tomsk-7, Russia.
[Credit: Reference 2 and]

Fig 2. 1m resolution IKONOS satellite image (captured Feb 2000) of reprocessing plant at Trombay, India.
[Credit: Reference 2 and]

Fig 3. 1m resolution IKONOS satellite image (captured Feb 2000) of heavy water production facilities near Khushab reactor site, Pakistan.
[Credit: Reference 2 and]

Facing this new challenge of widely available high resolution satellite imagery, states in future will opt. deceptions and anti-satellite imaging counter measures to make their dedicated nuclear facilities hidden. For example Concealment can be done by building sites that are difficult to detect from satellite imagery, but this could increase safety concerns as no state up-till this time has tried it, for instance if one tries to go underground the cost would be extremely high. Such program will require many personnel, instrument and such activities will be easily observable in these high resolution images.

Consider one smart solution by Russia at Krasnoyarsk-26, where three plutonium production reactors, a reprocessing plant, and associated storage and processing facilities were built entirely inside a granite mountain on the side of Yenisey river. From declassified Corona satellite images one can clearly identify security fence around site, the railroads and the entrances to the underground site. Cooling water reservoirs were visible near the river which could be used to provide cooling water piped in from the river. The discharge point of the cooling water could be detected by the visible photos through the absence of ice on river in the winter. Also ventilation shafts and a high stack used to release gaseous fission products from reactor or reprocessing plants. Also it can detect the underground reprocessing waste injection wells to the northwest of the underground complex. [2]

Thermal Identification
The above discussion was more features oriented for identification of sites and a bit less scientific, but we have set a reasonable ground of site identification from above example. Adding to the discussion, it will be very important to monitor the status of military reprocessing plants that were shutdown under Fissile Material Cut-off Treaty (FMCT). For such scenarios more reliable result are required, that can be achieved using thermal infrared sensors operating in infrared wavelength region (8-14 micrometers in atmospheric window) to reveal heat signatures (as absence of ice in previous example).

Fig 4. Atmospheric Transmission at various Wavelengths

If the reprocessing plant is closed the activity level must be very low but a very small scale reprocessing plant again requires TIR imaging. So removal of heat from reactor core is essential for shutdown that is no vaporization is occurring (no water-vapor plume visible as in UK’s site). As already discussed if reactor has cooling pond TIR can still indicate the change in temperature. If the reactor is using water of river through ports, the temperature variance of water from input stream to discharge stream is a clear indicator.

Fig 5. Four Calderhall Magnox reactors at Sellafield, United Kingdom. The four reactors made a significant contribution to the UK’s plutonium for weapons.
[Source: Reference 4 and ]

Another important indication of working facility is the roof temperature of nuclear sites and facilities. Most of the heat from a gaseous diffusion plant (GDP) is discharged from its cooling towers. These temperature changes are highly emphasized in Hui Zhang’s work. Let’s discuss his case study on Portsmouth enrichment plant.[3]

Fig 6. Landsat-5 TIR Image of Portsmouth GDP taken March 1994, The hot roofs of buildings X-333, X-330 and X326 are clearly visible.
[Source: Reference 4]

After error removal in his studies on Portsmouth the three reactors (X333, X330, X326) showed a temperature difference from surrounding to be +21.7, +19.7 and +17.7 respectively in March 1994 from a low resolution LANDSAT5 satellite TIR imagery. Given that the process buildings are clearly visible in the thermal bands of images, the analysis shows that thermal imagery from commercial satellites can reveal that a GDP appears “hot” relative to surroundings, and by implication that the shut-down status of a plant could be identified.

Night time or early morning thermal imagery, might be used to minimize the effects of solar heating and confirm the source of the effect. But even without such imagery, the ~5 – 20 °C increases in roof temperatures induced by internal activity alone can be clearly seen with existing and future satellite thermal images.

Conclusion and Recommendations
I conclude that features like availability of water, difference in water temperature, water vapor plumes, building structures (especially GDP) and activities are features and high spatial resolution Very Near Infrared Imagery (VNIR in TIR) provide evidence and status of a reprocessing plant. Landsat-7 and ASTER imagery can be typically utilized for this purpose. From FMCT to a global, verified ban on the production of fissile materials for nuclear explosives would, therefore, be a key building block in a comprehensive strategy to contain and eliminate nuclear weapons.


  • Zhang, H., “Strengthening IAEA Safeguards using high resolution commercial satellite imagery”, International Atomic Energy Agency (IAEA-SM-367/16/01), 29 Oct-Nov 2001.
  • Zhang, H., “Strengthening IAEA Safeguards using high resolution commercial satellite imagery”, Symposium on International Safeguards: Verification and Nuclear Material Security, Vienna Austria (Oct-Nov 2001).
  • Zhang, H., and Hippel, F., “Monitoring Large Enrichment Plants Using Thermal Imagery from Commercial Satellites: A Case Study” , Science and Global Security, Vol. 9 (2001), pp. 143 – 163.
  • Zhang, H., “Uses of commercial satellite imagery in FMCT verification”, (2000).