Independent Commission on the Verifiability of the CTBT: Bhupendra Jasani: Contribution of Remote Sensing Satellite to CTBT Verifiability

Contribution of Remote Sensing Satellite to CTBT Verifiability

Bhupendra Jasani

Introduction

42 years since the first international proposal for a complete ban on testing nuclear weapons was made in April 1954 by India, the Comprehensive Nuclear Test Ban Treaty (CTBT) was finally opened for signature on 24 September 1996. In its Preamble, it is recognised "that the cessation of all nuclear weapons test explosions and all other nuclear explosions,…constitutes an effective measure of nuclear disarmament and non-proliferation in all its aspects,…". Thus, the State Parties committed themselves not to carry out nuclear tests. It was also accepted that "the most effective way to achieve an end to nuclear testing is through the conclusion of a universal and internationally and effectively verifiable comprehensive nuclear test-ban treaty,…". With this in mind, several verification methods have been recognised in the Treaty, all of which depend on the detection of a nuclear explosion after it occurs. However, from the point of view of non-proliferation of nuclear weapons, this would be too late and would not fulfil the above aim of "non-proliferation in all its aspects…". Ideally, in order to achieve truly the non-proliferation goals of the CTBT, it would be useful to have a method that could detect a potential nuclear test so that the State involved could be persuaded not to carry it out. It is suggested here that, to some extent, the use of commercial remote sensing satellites can satisfy such a requirement.

Historical precedents

Photographic reconnaissance satellites belonging to the former Soviet Union detected the preparations by South Africa of its planned nuclear test in 1977.¹ However, South Africa was persuaded not to carry out its test. In 1981, it was reported that India had begun preparations for a test in the Rajasthan desert.² These observations were carried out until 1984. On 15 December 1995, news was leaked that US military observation satellites had detected considerable activities at India's nuclear test site.³ It was assumed to be related to continued nuclear test preparations. After considerable diplomatic flurries supported by satellite imagery⁴ , India was apparently persuaded not to go through with its plans.

It has been argued that satellites failed to observe the Indian nuclear test preparations in 1998. It is hard to believe that satellites observed preparations in 1981 through 1984 and again in 1995 but suddenly stopped looking at the Rajasthan Desert in 1998. In fact, it was reported that satellite imagery indicated that a test was imminent in May 1998. The information, however, came too late to the decision makers.⁵ It should be realised that it is always difficult to predict the exact time of such an event unless communications are also monitored closely. If a State wishes to hide a nuclear test, it will either encrypt all communications or remain silent before a test. The above indicates that optical reconnaissance satellites have been used for monitoring preparations of nuclear tests. In fact satellites are an important element of the national technical means (NTM) of verification. The NTM consists of methods of collecting information using technical equipment not dependent on any co-operation by other countries.

Vast amount of energy are released from a nuclear explosion. This energy is emitted in the form of thermal and light radiation, blast and shock waves and nuclear radiation consisting of gamma rays, X-rays, neutrons and charged particles as well as fission and fusion products. Beside optical cameras, various types of nuclear radiation detectors such as for example gamma ray, X-ray and neutron detectors, and optical instruments are deployed on board spacecraft.⁶ In the US, satellites carrying such devices were called Vela satellites. Subsequently, such sensors have also been carried on board the US Defense Support Program (DSP)⁷ satellites and on global positioning system (GPS)⁸ satellites. However, all of these types of satellites were primarily developed and deployed for defence purposes, and as such, data from them are not generally available to the international community, particularly those generated from the Vela, DSP and GPS satellites. While the data generated by such satellites are not commercially available, thay are shared only with a few very close allies. Even these may not get all the information they require. Thus, for multilateral treaties more open verification methods need to be explored. Commercial remote sensing satellites now have this potential.

Why commercial satellites?

The CTBT does not exclude the possibility of using satellites in its verification procedures. While this technique is not one of the several verification methods listed in the Treaty, States Parties to the Treaty are urged to look at this technique. Article IV.11 of the CTBT for example states that

"Each State Party undertakes to cooperate with the Organization and with other States Parties in the improvement of the verification regime, and in the examination of the verification potential of additional monitoring technologies such as…satellite monitoring…".

Satellites offer the possibility of monitoring a large area of the Earth quickly and repeatedly. Not only this, but they could provide an improved factor of at least 7 in terms of area coverage compared with that obtained from aerial surveillance by aircraft. For example, a modern aircraft, such as the US SR71, flying at an altitude of some 25km at a speed of 1km/sec, is capable of filming slightly more than 250,000km² of the earth's surface in an hour.⁹ A satellite, such as the French satellite SPOT (resolution¹⁰ 10m), or the Indian IRS-1C or -1D (resolution 5.8m), travelling at some 7km/sec at an altitude of 800km, could observe about 1,750,000km2 of the earth's surface in an hour. A satellite carrying a sensor with a resolution of 1m, such as the US Ikonos-2, could cover about 277,000 km² in an hour, nearly the same area as that covered by an aircraft. Unlike for over-flights by aircraft, no permission would be required from States over which satellites pass. Furthermore, since a satellite orbits at an altitude of at least 150km, well beyond national airspace, and since it is unmanned, humans are not exposed to retaliation from an adversary, unlike reconnaissance aircraft pilots. Moreover, the quality of data from commercial observation satellites has improved some 100 fold since 1972 (see Table 1) when the first such spacecraft was launched by the US. Finally, data from commercial observation satellites could be purchased by anyone. Considerations like these must give much impetus to the development of multilateral technical means of verification (MTM).

Electromagnetic (EM) radiation reaching a sensor on board a satellite can be recorded on film or electronically in digital form. The latter, recorded over specific spectral regions of the EM radiation, are assigned brightness values. Thus, such data are not in colour. In the case of the Landsat-5 satellite, there are seven bands (see Figure 1). Colour images are then obtained when selected bands are channelled through red, green and blue colour guns in a computer display monitor. If bands 1, 2 and 3 are assigned colours blue, green and red respectively, the resulting colour image will be very close to an image formed by human eyes. Such an image is known as a true colour composite. Spatial resolutions of such multi-spectral sensors range from 4m to 120m. While a panchromatic band spans over a wide range of wavelengths (see Figure 1), the spatial resolution is much better (see Table 1). For example, the latest US commercial satellite Ikonos-2, launched in 1999, has a resolution between 0.8 and 1m. Thus, in a multi-spectral satellite image (for example, a combination of bands 2, 3 and 4 of a Landsat), after a nuclear explosion, a localised spectral changes can be detected owing to the change in surface structure. Surface fracturing or a crater can be detected by high-resolution panchromatic image.

Therefore, with the development of commercial remote sensing satellites, even the participation by the international community in the verification process of a treaty such as the CTBT is now possible. Satellites are non-intrusive and information acquired by them is openly available. Moreover, a number of States are launching and operating their own commercial remote sensing satellites with high-resolution sensors on board. Thus, authentication of data becomes possible. There is a considerable potential for detecting changes in a scene owing to nuclear tests both by eye and with the use of mathematical techniques using computers. The latter are most useful for detecting spectral changes in a scene. It has often been argued that optical sensors are very limited because clouds frequently cover the earth's surface. Civil radar satellites that have day and night and all weather capabilities now overcome this obstacle and can be used to detect changes, by interferometric methods, before and after a test.

Figure 1 This shows spectral sensitivity of the French, the Indian and the US satellites. Number of bands in each case is also indicated.

Satellites cannot always detect the nuclear test preparations. For example, India conducted its first nuclear test in 1974 at a site at Pokharan test range in the Rajasthan desert. This test came as a surprise since apparently satellites did not detect the preparations.¹¹However, subsequently in the same region satellites did detect test-related activities in 1981.¹² It should also be emphasised that satellites are not the only method used for verification of treaties. Information derived from many sources is usually required and used. Data from satellites could act as an additional very important source of information. This data can also be used to trigger on-site inspections.

Cost of satellite derived data

It is often argued that the cost of satellite imagery will be so high that their use for verification becomes prohibitive. For monitoring a CTBT, generally some specific known sites are to be monitored so that large area scanning of the earth's surface is not necessary. At present there are six known nuclear test sites. These are: (1) the US Nevada site (the Yucca Flats and Frenchman's Flats); (2) the Russian Novaya Zemlya; (3) the French site in the Pacific at Moruroa and Fangataufa; (4) the Chinese site near Lop Nor; (5) the Indian Pokharan site in the Rajasthan desert; and (6) the Pakistani site at Ras Koh in the Chagai Hills region. In addition, only a few States listed in the CTBT with significant nuclear activities are likely to develop nuclear weapon programme. Thus, the area to be monitored may not be so large. Also, once a site has been identified and recorded initially, it does not have to be monitored continuously. Only the test locations need to be monitored periodically. Thus, images of smaller sizes could be acquired. This will have a considerable impact on the cost of imagery.

The cost of images is not always simple to estimate because the cost of an individual scene can be very different when bought singly or as one of a larger order. Moreover, scenes that are archived and are older than a certain date may be cheaper by as much as 40 percent. On the other hand, if a satellite is specifically targeted to acquire a scene, then the image will cost considerably more. The retail prices of data from various remote-sensing satellites are shown in Table 2. The cost can be reduced if extracts of full scenes are purchased. However, in this case the exact location of the site needs to be known so that only a small scene needs to be purchased. Initially the sites could be identified using, for example, the SPOT or the Indian IRS-1C satellites since they cover larger areas. Once the site of interest is identified, then a high-resolution image could be acquired. Table 2 gives some estimates of the cost of various types of imagery products. It should be remembered that as more and more countries launch their own satellites and enter the market, the cost is bound to decrease.

Some conclusions

As a result of considerable improvement in the capabilities of commercial remote sensing satellites, their use could significantly enhance the verification of the CTBT. Not only can the location of a test be determined accurately but its preparations can also be detected possibly in time to avert the test. This is important as the ideals of non-proliferation are then truly fulfilled. Moreover, it would be difficult to hide from satellite observations a nuclear test in a seismic event because, on a multi-spectral image, an explosion would record very localised spectral and surface structural changes that would not be the case in an earthquake. Perhaps there are two most important aspects of monitoring from space; it is non-intrusive and it could be used by anyone because satellite imageries can be acquired commercially.

It is reasonable to assume that countries without any significant nuclear research and/or nuclear power programme and any national security concerns are less likely to embark upon nuclear weapons testing. This would reduce the number of countries to be monitored. Furthermore, as more and more countries launch and operate their own satellites, not only the cost of imageries will be reduced but also it would be possible to authenticate information from various sources.

It is, therefore, suggested that such satellites should be a part of the CTBT verification regime. The Treaty, while at present does not include verification by satellites, it does suggest that the use of satellites could be investigated at some future date. The CTBTO together with States Parties should begin to look at satellites now and if it concludes that the technique can "enhance the efficient and cost-effective verification of this Treaty…" it "be incorporated in existing provisions in this Treaty, the Protocol or as additional sections of the Protocol, in accordance with Article VII, or, if appropriate, be reflected in the operational manuals in accordance with Article II, paragraph 44." It is suggested here that this technique should now be explored in collaboration with the CTBTO.

Table 1 Current and some of the future commercial remote sensing satellites belonging to various countries are summarised below.

Country Satellite	Date of launch of first satellite	Resolution in pixel size (m)
Country Satellite	Date of launch of first satellite	Panchromatic	Multi-spectral	Thermal/IR
OPTICAL
*Brazil/China* Zi Yuan CBERS I & II	1999-2001	20
CBERS III & IV	-	5	10	40-80
France SPOT-4	1998	10	10
SPOT-5	2002	2.5
India IRS-1C,-1D	1995, 1997	5.8	25
IRS-P5	1999-2000	2.5
*Israel* Eros-A	1999	1
Japan ALOS	2003	2.5	10
*Russia* Resurs-F series	1989-98	2
*USA* KH-1 to 4 KH-4A KH-4B	June1959-Dec 63 Aug1963-Oct 69 Sept1967-May 72	7.6 2.7 1.8
Landsat-5	1984		30 (bands 1-5,7)	120 (band 6)
Landsat-7	15 April 99	15	30 (bands 1-5, 7)	60 (band 6)
Ikonos-1 did not achieve orbit, -2	1999	0.8-1	3-5
Quickbird-1,-2	1999	0.8	4
Orbview-3 Orbview-4	2000	1-2 1-2	4 4 & hyperspectral 8
Earlybird-1 ASTER	1997 18 December 99	3	15 15 (bands1-3), 30m (bands 4-9)	90 (bands 10-14)

RADAR
*ESA* ERS-1, & -2	1991 & 1995	25
*Japan* JERS-1	1992	18
*Canada* Radarsat	1995	9-100
*Russia* Almaz-1B	1998	4-15
*USA* SIR-C	1994	8-30

Table 2 Cost of data from some of satellites.

Satellite	Area (km²)	Data format	Cost US$ (cost/km²)
*Photographic*
Russian KVR-1000	1,600	Film positive Film negative	3,300/(2.10) 3,520/(2.20)
Digital
Russian KVR-1000	16 160 1,000 1,600	Panchromatic	1.100/(68.75) 1,760/(11.00) 2,750/(2.75) 4,400/(2.75)
French SPOT	3,6000	Multispectral level 1B (3 bands) Panchromatic level 1B Multispectral level 1A/1B (3 bands) Panchromatic level 1A/1B Multispectral level 2A (3 bands) Panchromatic level 2A	~ 1,200/(0.33) ~ 1,600/(0.44) ~ 2,000/(0.56) ~ 2,800/(0.78) ~ 2,800/(0.78) ~ 3,400/(0.94)
Indian IRS-1C	4,900	Panchromatic Multispectral (4 bands)	2,500/(0.51) 2,800/(0.57)
US Ikonos		Panchromatic Multispectral (4 bands)	Min. 1,000 or (12-17) - US; 2,000 or (29-44) - Int.
Landsat	22,500	Multispectral (7 bands)	~ 4,400/(0.18)
*Radar (digital)*
European ERS-1, -2	6,400	Synthetic aperture radar (SAR)	~ 1,200/(0.19)
Canadian Radarsat	2,500	SAR	~ 2,500/(1.00)

References and Footnotes

¹ Jasani, B., "Military Satellites," World Armaments and Disarmament, SIPRI Yearbook 1978, (Almqvist & Wiksell and Stockholm International Peace Research Institute, Stockholm, 1978), pp.69-103.

²Senator Alan Cranston, "Nuclear Arms Race in South Asia Endangers US Security Interests," Congressional Record, US Senate, April 27, 1981, pp.7375-7377.

³Weiner, "India Suspected of Preparing for A-bomb Test", The New York Times, 15 December 1995, p.A6.

⁴ Richelson, J., "Examining US intelligence failure", Jane's Intelligence Review, vol. 12, no., 7, September 2000, pp.41-44.

⁵ Ibid, Richelson, p.43.

⁶"Satellite Instruments for Monitoring the Limited Test Ban Treaty", Verification Technology, Sandia Technology, Sandia National Laboratories, Albuquerque, New Mexico, November 1984, pp.8-11.

⁷ Whelm, C. R., Guide to Military Space Programs, Arlington, VA: Pasha Publications Inc., 1994, p.81.

⁸"GPS to Test Nuclear Detection Sensor", Aviation Week & Space Technology, vol. 111, no. 9, 27 August 1979, p.51.

⁹ Leopold, G., "Verification technology grows in tandem with pressure to obey SALT II", Defense News, vol. 1, no. 50, 1986, p.11.

¹⁰ Resolution may be defined as the smallest distance between two identical objects at which they can be resolved by a sensor as two objects. Such a definition is applied to a photographic image. In the case of a modern digital device, this resolution may be defined as the area on the ground represented by each sensor cell or a pixel. Thus, smaller the pixel, smaller the area and thus, finer the resolution.

¹¹ Jasani, B., "Reconnaissance satellites", World Armaments and Disarmament, SIPRI Yearbook 1975, (Almqvist & Wiksell and Stockholm International Peace Research Institute, Stockholm, 1975), pp381-384

¹² Ibid., Cranston, p.