'We have an adequate scientific database for designing ... a credible nuclear deterrent'

Print edition : January 02, 1999
Interview with Dr. R. Chidambaram.

According to Dr. R. Chidambaram, Chairman, Atomic Energy Commission (AEC), and Secretary, Department of Atomic Energy (DAE), "a dozen new ideas and systems" were tried out in the five nuclear tests carried out at Pokhran in Rajasthan on May 11 and 13. "And all of them worked perfectly well." With the data yielded by the tests, he said, "we have now built an adequate scientific database for designing the types of devices that we need for a credible nuclear deterrent. So from a scientific point of view we advised that we could now announce a moratorium on testing because no more tests were considered necessary by us."

Dr. Chidambaram asserted that "India carried out the tests based on today's knowledge of physics, engineering, materials science and electronics, (and) there is a kind of leapfrogging here, and each one of the tests should be considered equivalent to several tests carried out by other nuclear weapon states over decades."

Dr. Chidambaram, whose term as AEC Chairman was extended by two years from December 1, 1998, was in Chennai on December 9. In a 70-minute interview he gave T.S. Subramanian, he answered a range of questions. Excerpts:

There is a controversy about the total yield of the five nuclear tests conducted at Pokhran in May 1998. Roger Clarke, a British seismologist, has agreed with the assessment of the Department of Atomic Energy that the total yield of the three tests conducted on May 11 was around 60 kilotonnes. But another group of seismologists disputes this. For example, University of Arizona geophysicist Terry C. Wallace wrote in the journal Seismological Research Papers that the yield was 10 to 15 kilotonnes on May 11, and 100 to 150 tonnes on May 13. Bhabha Atomic Research Centre (BARC) scientists S.K. Sikka, Falguni Roy and G.J. Nair wrote in the September 10 issue of Current Science (published from Bangalore) that the interference between the seismic waves from the two main explosions on May 11 would have led to a lowered estimate of the seismic signal strength at stations situated in the eastern and western directions. They have argued that only the data from the stations situated in the northern and southern directions should be taken into account. Can you put the record straight?


It is always difficult to correlate the seismic magnitudes with yields unless it is a well-calibrated testing site like Nevada in the United States or eastern Kazakhstan in the erstwhile Soviet Union. It is also susceptible to deliberate manipulation, as happened between the Soviet Union and the United States. In our case, for the tests on May 11, there is the further complication caused by separated but simultaneous explosions, when the seismic signals interfere, as you mentioned, and their unfamiliarity with the Pokhran geology. The latter is important because the strength of the seismic signal is determined by the way the explosive energy couples into the geological medium, and there are strong regional differences. In fact, each seismic station has to be calibrated, and this is obvious from the range of seismic magnitudes reported by various global seismic stations. A small difference in body wave magnitude of a little over 0.2 corresponds to a halving of the yield estimate. And for any underground nuclear explosion, seismic body wave magnitudes are known to range over 1.0 or even more, which indicates the pitfalls in yield estimates from seismic signals, unless they are done carefully and correctly. This has been done by BARC scientists, using four different methods, and the details have been published in the November 1998 issue of BARC Newsletter.

The first method is to look at the body wave magnitude, or mb. Here is where the asymmetry from the seismic records of the various stations in the world comes into the picture. Because the two (main explosives on May 11) were located in shafts oriented east-west with a separation of one km, the seismic signals produced from them superpose with a phase lag depending upon the direction. However, if one looks at the signal which has been recorded in the northern direction, for which the phase lag is zero, one can clearly see the difference. In fact, if we plot mb versus the orientation, say zero degree for the north, 90 degrees for the east and -90 degrees for the west, you get a bell-shaped curve; in other words, if you allow for this orientational effect, the body wave magnitude for the two tests is 5.4, which corresponds to about 60 kilotonnes, which we had announced immediately after the experiments. This is also consistent with our yield calculations, that is, based on our computer calculations. The design values and the announced experimental yields soon after the tests were 15 kilotonnes for the fission device and 45 kilotonnes for the thermonuclear device (popularly known as the hydrogen bomb); and there was also the small sub-kilotonne device, with a yield of 0.2 kilotonne.

The second method is more straightforward. As I mentioned earlier, the one problem is that you must know the geological medium in which the device has been emplaced before you venture into yield estimation because that decides how much energy couples into the (geological) medium from the device. Then you must calculate the absorption along the path from the point of detonation to the seismic station. Since there is no global or universal model for the earth, these absorptions along the various paths could be different. Unless the site has been calibrated well, you can make serious mistakes in estimating the yields from seismic magnitudes. On the other hand, if the site and the seismic station have been properly calibrated together, these effects can be eliminated.

There was only one experiment to calibrate Pokhran, that is the PNE (peaceful nuclear explosion) experiment carried out by us in May 1974. So in the second method used by the BARC scientists, for eight stations around the world for which we have the data, they have looked at the difference in body wave magnitudes for the PNE experiment done in May 1974 and the tests done on May 11, 1998. There is a small correction coming from the orientational effect I mentioned above. This correction is 0.1 for two stations, 0.2 for one station and zero for the other five. The average body wave magnitude difference is 0.5, which translates into a yield ratio of 4.5. That is, the total yield of the tests done on May 11, 1998 is 4.5 times the yield of the test done in May 1974.

The International Data Centre (IDC), Arlington, U.S., gives the yield of our May 1974 test as between 10 and 15 kilotonnes. We have evaluated it more accurately as between 12 and 13 kilotonnes. This is accepted by leading seismologists in the world. If you multiply these numbers by 4.5, you again get a yield of about 50 to 60 kilotonnes.

The third method used by the BARC scientists was to look at surface waves, which are less susceptible to geological variations. By looking at the surface wave magnitude - Ms, as it is called - from four stations of the United States Geological Survey and three of our own stations, the average comes to 3.62. That is, Ms is equal to 3.62. Then, using the standard formula which relates Ms to the yield, the yield works out to 58 kilotonnes.

All the above methods of measurement are based on internationally available seismic data, which - with all its defects - is the only way anyone can look at other countries' underground nuclear explosions. In fact, seismic monitoring is one of the methods for the international monitoring system of the Comprehensive Test Ban Treaty (CTBT).

The fourth method that the BARC scientists used was to look at the close-in acceleration measurements and compare them with the U.S. data. The U.S. data are available for the dry hard rock of the Nevada test site and the sedimentary formations in Colorado, where the U.S. has carried out two peaceful nuclear explosion experiments - Rulison and Rio-Blanco. When they scaled the yields, our data fit with the Colorado data and not with the Nevada test site data. This is what we expected because the Pokhran medium is closer to the sedimentary formations in Colorado, where these two experiments were carried out, than with the rock in the Nevada test site. So here again, the acceleration measurements are consistent with the yield of about 60 kilotonnes. If we accept the Nevada site parameters, the Pokhran-II yield will go much higher, to 100 kilotonnes.

So we have plenty of data now. From all these data, it is very clear that the yields of the devices we tested on May 11 are exactly what we announced immediately after the tests.

As you correctly said, a number of international seismologists have agreed with our yields within the limits of experimental error. But this one group of seismologists in the U.S. seems to have ignored some of the factors that I mentioned before, and in particular the fact that the U.S. geostations are crowded round an angle westerly to the northern direction. As I explained in the beginning, they are likely to get lower body wave magnitude signal, and this seems to be the main reason for the lower value given by them. On the other hand, it also raises serious questions regarding software and analytical inadequacies in the global seismic network for monitoring simultaneous but separated underground explosions around the world.

It is not surprising that they did not detect the sub-kilotonne explosions of May 13 because the threshold for the CTBT in terms of sensitivity of detection is around one kilotonne when the Global Seismic Network is fully operational, which is not likely to be before 2002. The only station from which they could monitor our Pokhran tests was at Nilore, in Pakistan, which, I think, is at a distance of about 740 km from Pokhran.

About a couple of hours back you said in your commemorative address at the Indian Overseas Bank foundation day that seismic data could be "manipulated". You also said that the International Data Centre (IDC) had misdiagnosed the tests as "an earthquake on the India-Pakistan border." Can you elaborate on that?

(Laughs.) Even though the Prime Minister had announced on May 11 itself that we had carried out the tests, the Internet message from the International Data Centre for several days after this was that there had been an earthquake on the India-Pakistan border at the time of the tests! That gives you an idea of how well seismic monitoring works!

It is also that during and after the signing of the Threshold Test Ban Treaty, which set the limit for testing at 150 kilotonnes, the Americans routinely overestimated the yields of the Soviet tests in order to accuse them of violating the Treaty. In particular, I remember one instance when the Americans accused the Soviet Union of testing a device at 300 to 500 kilotonnes, well above the 150-kilotonne limit. But later, when the Americans went and examined the geology of the site where the Soviet Union had carried out the test, they agreed that the yield was indeed below 150 kilotonnes. This emphasised the importance of knowing the geology of the site well or the need to calibrate it properly before one can make any statements about the yield of the device tested, from seismic measurements.

The shaft and damaged equipment at the site at Pokhran where Test 1, of a thermonuclear device, was carried out on May 11, 1998.-BY SPECIAL ARRANGEMENT

Another interesting example was that after the CTBT was signed, the Americans accused Russia of carrying out a sub-kilotonne explosion in the Arctic region, while seismologists all over the world knew that it was an under-sea earthquake, 100 km away from the Arctic test site!

So my feeling is that unless one has competent seismologists and good analytical software, one can make honest mistakes. But there is always the possibility of manipulating data, as in these two examples I cited.

How important is seismology in measuring the yields of nuclear explosions?

In terms of verification of compliance to the CTBT, the most important component of the international monitoring system is indeed seismology. There are more than a hundred primary and secondary stations distributed around the world to look at the signals. But the problem is that the experience so far of most people has been on experiments done at recognised or identified test sites. How well the seismologists analyse the data from the global seismic network of a new (unidentified) test site has to be debated.

There is another serious question that has to be taken into consideration. Since testing has ceased in all the nuclear weapon states, I am told that the best seismic analysts who were working on such problems in the Western countries have left, and quite possibly the job is now being done by less competent seismologists. The situation may worsen in the years to come.

I must mention in this context that the analysis of the Indian data that I referred to in the beginning has been done by seismolgists at BARC, who have a very high international reputation and 34 years' experience in seismology.

What are the basic characteristics of a neutron bomb, a hydrogen bomb, a fission device, and a plutonium bomb? When I met you last time you said there was no difference between a neutron bomb and a hydrogen bomb.

Basically, there are two types of devices. One is the pure fission device and that can use either plutonium or highly enriched uranium. Here, you start with the configuration which is sub-critical and then you increase the density of the nuclear material by implosion or you assemble pieces of material together or both so that you reach a super-critical configuration. And if you start a chain-reaction at the proper time, from that time onwards until the system becomes sub-critical again through disassembly or a little later, there will be energy released. This is a typical fission device. You can put into this device fusion material which essentially acts as a source of additional neutrons once the fission device operates and thereby increases the yield of the device. You can call it a fusion-boosted fission device. Well, you can decrease the energy of the fission device so that it comes to the sub-kilotonne range. This has to be done very carefully because when you are trying for marginal super-criticality, there is always a chance that if you make a mistake, it may not reach criticality and it may become a fizzle.

The very satisfying thing from India's point of view about the three sub-kilotonne tests was that our computer calculations were so exact that we could predict very accurately the yields of the three devices that we tested.

In the thermonuclear device, or hydrogen bomb as it is called, there are at least two stages. The first stage is a fission device and one uses the radiation energy from this stage to detonate the second stage, which contains the fusion material. In advanced hydrogen bomb designs, the primary stage is a fusion-boosted fission device in order to get increased radiation energy density.

In the case of the thermonuclear device that we tested, the primary stage was a fusion-boosted fission device. For a nuclear scientific community of India's capability, there was really no need to test separately a fusion-boosted fission device. That is why we decided to use it as the primary in the thermonuclear device. In fact, it is very interesting that The Times, London, said that India had carried out "sophisticated tests of modern nuclear weapons". As I have said before, the five tests that we carried out are based on advanced designs of 1998 vintage.

In the five tests that we carried out, we tried out a dozen new ideas and systems. And all of them worked perfectly. We have now built an adequate scientific database from the tests for designing the types of devices that we need for a credible nuclear deterrent. So from a scientific point of view we advised that we could now announce a moratorium on testing because no more tests were considered necessary by us.

Histogram of the average azimuthal variation of mb values in respect of the May 1998 nuclear tests at Pokhran, as reported by the International Date Centre in the U.S., the Kyrgyz network (KNET) and the U.S. Geological Survey monitoring stations, in 20-degree intervals in the northern hemisphere.-

Negotiations on the CTBT are, of course, a political question, and I have no comments on that.

What is the yield range of fission and thermonuclear devices?

I know that fission devices have been designed up to a yield of at least 100 kiltonnes. This was the work of Theodore Taylor. Thermonuclear devices can be made in the range of megatonnes. (One megatonne is equal to 1,000 kiltonnes.) But they have been made as small as a couple of kiltonnes also. I remember one U.S. peaceful nuclear explosion test called Cabriolet, where they tested a small thermonuclear device for excavation purposes whose yield was as low as 2.3 kilotonnes, of which the fission trigger was probably only a hundred tonnes. Even though they called it a PNE device, this is also the prescription for a neutron bomb! So thermonuclear devices can go from a couple of kilotonnes or even less to megatonnes. The consequent conclusion is that from the yield of a device, you cannot really make out what kind of device has been tested.

Will the five nuclear explosions themselves constitute a minimum nuclear deterrent as is being claimed?

The deterrent weapons have to be built from the three types of devices which we successfully tested. You should not just count the number of tests. First, the number of tests has to be related to the number of types of devices that you want to develop. The Americans, for example, carried out more than a thousand tests. But they developed 70 to 80 types of devices. The Chinese have carried out 45 tests.

The second point to remember is that suppose you were testing three decades ago, you would carry out your design on the basis of the then existing knowledge of physics, materials science, engineering and electronics. Suppose your knowledge improves, suppose you had a better computer available to you, suppose your electronics improved as it indeed did over the decades, you would redesign and test again. But basically, it could be the same device. This will get repeated again and again over the next few decades.

So if in 1998 India had carried out the tests based on today's knowledge of physics, engineering, material science and electronics, there is a kind of leapfrogging here, and each one of the tests should be considered equivalent to several tests carried out by other nuclear weapon states.

Some countries brought in the entry-into-force clause, which says India should accede to the CTBT by September 24, 1999. Otherwise, the CTBT would collapse. Don't you think this is an assault on India's sovereignty ?

It is a very unusual clause because so far as I am aware, there is no other treaty in which there is an enforced requirement of signature of a list of states before the entry-into-force.

Yes, 44 states which have access to fissile material. Article XIV of the CTBT states so.


Why were the five tests carried out in such a rapid sequence? Was it out of fear that if they were not carried out now, international pressure would mount on India later to accede to the CTBT?

Comparison of seismic waveforms from the nuclear tests at Pokhran in May 1974 and in May 1998. The date are from the same channel of Gauribidanur seismic monitoring station in Karnataka.-

It was a very carefully planned series of tests, and the design and planning was carried out over a number of years. As time passed, the designs were continuously improved. That is why we were prepared to carry out tests at short notice. Of course, if we had tested a few years back, the designs would not have been as advanced as they were when we tested them in May 1998. Yes, indeed it was a rapid sequence because I do not know of any other instance where five tests were carried out within a span of 48 hours when the devices were of three different basic designs.

But there have been instances. The Soviet Union once exploded eight nuclear devices simultaneously.

I don't think anybody has exploded within 48 hours such a variety of devices as we did.

If India had not conducted the five tests, what would have been its stand?

I will put it this way. The one test that we carried out in May 1974 was surely inadequate. But the five tests we have done now, after such careful planning, have given us enormous and very valuable database.

Does this database give us the capability to do sub-critical experiments in the future, which are not barred by the CTBT?

There is a multiplicity factor, called K, which is the ratio of the neutron population in a given generation to that in a previous generation. K equal to one is the 'Lakshman rekha' for the CTBT. You are not supposed to cross this line if the CTBT comes into force and you are a signatory to this Treaty. This means that if you plan experiments which take K close to 1, your calculations must be accurate enough to ensure that crossing beyond K equal to 1 does not happen. The fact that we carried out three very precise sub-kilotonne tests, which take the system marginally above K equal to 1, means that our capability to do such critical tests, if found necessary in the future, is now established.

I must also mention that from whatever reported data I have seen about such sub-critical tests by the United States, they have been far from criticality. I must add that at the moment, we are not planning any sub-critical tests.

What about India's capability to go in for computer simulation?

This again has to be related to the fact that there was a perfect match between the calculated and the measured yields in all our tests. In the early days, when they carried out a computer design calculation for a device, the actual yield would be far away from the calculated yield, often being off by a factor of two or more. And then they will adjust some parameter in the calculation to get a match between the calculated and the measured yields. If they did several tests, they would have several parameters to play with and thus have a computer simulation package. The errors were even more in the case of a thermonuclear device.

But today, since we have got experts in practically every branch of physics which goes into the design of the various types of nuclear devices, we did not have to do any empirical parameter fitting but could go ahead with first principle calculation. That is why I am confident that we have the so-called computer simulation capability in this field.

In what ways were the nuclear tests of May 1998 an improvement over the nuclear explosion of May 1974? The 1974 test was earlier called a PNE experiment but now it has been admitted that it was indeed a nuclear bomb.

The 1974 test was important since it gave us the confidence to carry out the 1998 tests. The design of the fission device has been vastly improved, and the others are, of course, new designs. From physics design-wise, there is no difference between a nuclear weapon and a PNE explosive. Obviously, the same physics has to go into both. What is different between the two are the packaging and the detonation methodologies. The experiment that we carried out in 1974 was a PNE phenomenology experiment.

Were the nuclear tests of May 1998 inescapable? Could India not have got the data through computer simulation?

This is like saying that only theoretical physics is enough and that we don't need experiments.

What was the yield of Pakistan's two tests? Pakistan claims it conducted six nuclear tests - five on May 28 and one on May 30.

The signal at Gauribidanur (station in Karnataka) which we recorded for the May 28 test (of Pakistan) was between five and ten kilotonnes. The Americans estimated it at six kilotonnes. The yield from the test carried out on May 30 (by Pakistan) was about half as much. However, as these estimates are based on seismic body wave magnitude again, these can be off by a factor of two.

Pakistan claims that it carried out five tests in the same horizontal shaft (on May 28), which is unusual because, after all, tests are intended for gathering data. And separating data from such closely spaced devices would be very difficult. So it is very unusual - their claim that they conducted five tests in a (single) shaft. It is much easier to dig a separate shaft for each device like India did and like everyone else usually does, except in very special situations. For example, in the Rio-Blanco tests, which the Americans did for gas stimulation, they emplaced three 30-kilotonne devices in the same vertical shaft. But these were not tests as such but actually used for a PNE application in order to fracture a large rock mass.

Pakistan claimed that it tested a hydrogen bomb.

From whatever I understand, they only tested fission devices.

The site of Test 4, May 13, 1998. The yield from this test was 0.2 kilotonne, and many monitoring stations failed to detect the two sub-kilotonne explosions on that day.-BY SPECIAL ARRANGEMENT

What is the status of the Koodankulam Power Project? Has work started on preparing the detailed project report (DPR)?

Actually, the supplement to the Inter-Governmental Agreement was signed by me and Dr.Yevgeny Adamov, Minister for Atomic Energy of the Russian Federation, when he came here in June 1998. Later the Chairman and Managing Director of the Nuclear Power Corporation of India Limited (Y.S.R. Prasad) went to Moscow to sign the agreement on the preparation for the detailed project report. This will take about two years to prepare.

I know that a very good seminar on the Koodankulam Atomic Power Project was held in Chennai, which you probably attended.

I did attend it.

What made me very happy was that the abstracts of the papers, which were presented by the Russians and the Indians, were translated into Tamil and distributed to village leaders from the Koodankulam area, who had come to attend the seminar. From whatever I hear about that seminar, the panchayat leaders as well as leaders of the fishing community were very happy about the project and said that they fully supported the project. This they did by not only attending the seminar but after a detailed visit to Kalpakkam (atomic power station) where they visited the reactors and also talked to the villagers in the area. I am convinced that the Koodankulam project will be very valuable to the people in the area.

We are funding a project at the M.S. Swaminathan Research Foundation (MSSRF), Chennai, on nuclear and biotechnological tools in coastal system research which will also involve field-level demonstration projects at a few places along the coastal areas. The DAE has also established a Homi Bhabha Chair for Nuclear Science and Rural Society at the MSSRF.

Thank you.
This article is closed for comments.
Please Email the Editor