'Other Countries, Other Scenarios: a Brief Overview of Nuclear Weapons Programs Around the Globe, and Possible Crisis Scenarios
Stephen Young, British American Security Information Council
I will be talking today on a few topics - other countries, and other scenarios, is the intentionally vague title I've given it - so I can fill in what needs to be said, and add relevant points here and there.
I will begin with a brief summary of Russian Y2K program, minimal as it is, and then look briefly at the arsenals of other nuclear weapon states - China, France, and the UK - and what we know about their Y2K programs. Then I will sketch a brief scenario, based on a few seemingly unconnected events, and see where that leads us. Finally, I will make some concluding policy recommendations.
Russia, as we all know, is a mess. Bruce has described the status of their nuclear arsenals. I do not have the time today to go into all the problems with Russian early warning systems, except to show you one slide, which I stole from Ted Postol of MIT. It also appeared in the Washington Post not too long ago, so many of you saw it. It shows that Russia has some huge gaps -blind alleys - where its radar could never detect incoming missiles. Those gaps mean that Russia is constantly vulnerable to attack, particularly from sea-based systems, which of course is where most of our arsenal is.
On Y2K and Russian nuclear weapons, what we had for a long time was a series of denials. The Russians systems use completely different technologies, we were told, so there is no danger of a millennium bug problem. To some extent, this is true - Russia relies much more on mechanical and electro-mechical equipment than the US, and the web of interconnected systems is probably less wide and deep.
But there are a few things that are known problems, or would seem to be. Most ironically, the Russian military may have some real problems with software and hardware that they stole from the US back in the 1970s and early 80s. This technology, according to a few experts, was reverse engineered, and used by the Russians. It is likely that these stolen systems do have Y2K bugs in them, but exactly what was stolen and how it is used is unknown, at least to us.
More recently, Russia has become more open about some of the Y2K problems it has. In February, the Russians stated that they would need $3 billion to fix the bug. How much of this is defense-related and how much otherwise was not specified. Last week, Maj. General Vladimir Dvorkin stated that Russia had less than $4 million to spend on its Y2K efforts on its nuclear arsenal and early warning systems. According to Dvorkin, there is no danger of Y2K problems with the nuclear weapons themselves, as they are not controlled electronically. But Dvorkin confirmed that Russia's early warning systems and command and communications do have problems.
For example, in a separate report, Dvorkin said that 74 control centers of Russia's Strategic Nuclear Forces were judged in "critical" condition, vulnerable to Y2K problems.
According to other officials, repairs should be done by June, to be followed by testing. As Dvorkin put it at one point, "I have no grounds for saying we will not solve this problem." Now if that doesn't instill confidence in all of us, I don't know what will.
Other than Russia
Outside of the US, the most is known about the UK and France. The UK only has one nuclear weapon system remaining in its arsenal, the Trident submarine-launched ballistic missile. The UK has three nuclear subs carrying these missiles, and a fourth is under close to coming "on-line" as they say. The UK has stated that at most 200 warheads will be actively deployed on these subs - making is the smallest arsenal of the five official nuclear-weapon states.
Most interestingly, the UK is the only country to declare that its forces are, day to day, not on high alert status. Last year, as part of a comprehensive defense review, the British stated that their nuclear forces are, as they put it, "routinely at a "notice to fire" measured in days rather than the few minutes' quick reaction alert sustained throughout the Cold War. However, the UK has also made very clear that they do not consider this a de-alerting step, different than the proposals bruce just outlined. As the "notice to fire" status is unverifiable, it may reduce the likelihood of accident, it does little to reassure other countries.
For its early warning data, the UK depends primarily on the United States. The UK operates a US-owned ground-based radar facility in the UK, and it gets data directly from that source. It is also provided with filtered data - just how filtered is a secret - from US sources around the globe and on satellites. So any problems that the US faces in its early warning, will also affect the UK.
The UK's military Y2K program, according to their Ministry of Defence, is the biggest government millennium bug effort. There are roughly 30,000 military systems to examine and out of those, roughly 1000 systems have been deemed "mission critical" for crucial air operations, logistical support, joint command capabilities, and the nuclear deterrent.
If you look at the British MOD reports on Y2K progress, they appear to be about six months behind the US in getting systems past the repair stage and into the testing process. This is not surprising, as the UK's deadline for compliance is this fall, while in the US its this month. According to the MOD, about half of the mission-critical systems were fixed and tested as of December 1998, and it plans to have 98% completed by September, 100% by December. At least, that's the plan.
However, the November '98 MOD Y2K report - the most recent available - indicates that, as they put it, "Shortage of skilled in-house staff has been identified as a potential risk to the programme across MOD" In other words, they don't have the people to fix the problems, and are having to try to hire consultants from outside. How successful they will be is an open question.
The second question is the timeliness of the British effort, and its reliance on the US. As most of you know, the UK buys its Trident missiles from the US, and equips them with its own nuclear warheads. Thus, it is relying on the US for Y2K compliance of the missile. As the US Navy is the service the furthest behind in its Y2K fixes, one would think that the British would be concerned. In fact, in at least one Trident-related system, the US plans to have a "work-around" or a substitute system, in place by December 15, 1999. That does not leave much room for error.
On to the French
The French nuclear arsenal is larger than the British, and includes two weapons systems. First, like the UK it has four ballistic missile submarines, although it currently only has one of its most modern submarines on line - one more will be introduced this year and two more over the next six years. Of the approximately 450 warheads in the French arsenal, the vast majority are on submarine-launched ballistic missiles on these subs. The French are building a new ballistic missile for the subs, and it was primarily to develop a new warhead for this missile that the French conducted the series of nuclear tests in 1996 that drew so much international criticism. The other weapon in the French arsenal is the ASMP - Aire-Sol Moyenne Portee - or air-to-ground missiles, with a range of up to 180 miles. One carrier has planes capable carrying the ASMP, the rest are on Mirage 2000 aircraft, with a total of 60 warheads or so.
The French Ministry of Defense also has a sizable Y2K program. It has a five-part effort involving a first survey of what systems needed to be looked at. This stage divided up French military systems into three categories: T-1 - essential or strategic systems, similar to the mission-critical category used by the US, that must function fully for operations - T-2 - important systems, that must either function, or have work-arounds or safety measures introduced, and T-3 - less important systems.
Second, was an evaluation of what needed to be done on each system. That was supposed to be completed by September of last year, but was not finished until December. It's never a good sign when you miss one of your first deadlines.
The third stage is correction - fixing the problems that have been identified. By October 98, the French claim that 30 percent of their systems had finished this stage, 50 percent were in this stage, and 20 percent were still being analyzed.
The fourth stage is validation - ensuring that the corrections fixed the problems. You may have heard mentioned this morning that during a test of Y2K fixes on a nuclear power plant in the US, all the computer screens went blank, and it took seven hours to figure out what had gone wrong with the test. So validation is not only important, its extremely tricky. French validation is only expected to be completed by October of this year - which obviously leaves little room for error if things don't go as planned.
And finally, you have the fifth stage - putting systems on line again. The deadline for this is also October, so the step from validation to on-line is apparently a short one. For those systems that are not completed by the first of October, there is a back-up deadline of November 1st for having plans to deal with the failure.
And a few words on China
China has roughly 400 nuclear warheads in its arsenal, on a variety of platforms. Ten or so are on ballistic missiles that have intercontinental range. Experts believe that these missiles are liquid-fueled, and normally they are kept empty, in other words, at a minimum hours from launch. Chinese technology is generally considered to be anywhere from 10 to 30 years behind the West, and I have seen nothing worth mentioning in terms of a Y2K program for China.
And even less about India and Pakistan, the newest nuclear states.
Both countries are trying to weaponize their small nuclear arsenals. India might have enough materials for 200 warheads, Pakistan might have enough for 20. Neither yet has long-range missiles fully developed, both are developing missiles capable of hitting each other and, in India's case, China. Their Y2K programs are also unknown.
Let me paint a scenario for you.
Dec 31 comes and goes. Scattered failures in power systems and other industries, but no drastic failures. One nuclear power plant in Ukraine has serious problems when backup power supplies fail and the manual shutdowns have problems. It's not as bad as Chernobyl, but a small region is evacuated for safety's sake. But, hip hip hooray, we all survive the millennium.
Mid-January comes and goes, and the Cheyenne-based Center for Y2K Strategic Stability shuts down, having done its job, in theory.
February 26 - North Korea launches a second missile across Japanese territory - similar to the one it launched a few months back. This time the missile fails even more dramatically, and debris actually lands on Japanese territory. Japan's military goes on high alert. February 28, midnight. As you know, leap years happen every four years. The exception to this rule in on the century mark - every hundred years - when leap years do not happen. The exception to the exception is, every four hundred years, the leap year DOES happen. It's the vagaries of the yearly cycle - a year isn't really a 365 days, its 365 plus. Anyway, 2000 is one of those years. But some computer systems didn't get that far, and unexpected failures occur (of course, failures are often expected - they're only unexpected when they are worse … than expected.)
As a result, a wide number of Russia's early warning and command and communications systems fail, because some think its February 29, and some think its March 1, so they can't talk to each other. US systems, which tested for this date, do much better. But the Russian system - which already has significant holes - becomes swiss cheese.
And, on February 29, in this leap year, with its economy still sinking, Primakov dies. Yeltsin, of course, is already more or less dead, and Russian leadership is confused.
And then, well, it really doesn't matter what happens next - a US satellite launch goes astray, the sun reflects off some clouds, an isolated computer chip fails, India and Pakistan start an actual war, a Russian nuclear submarine collides with a US hunter-killer submarine - all of which have happened - and we get a crisis scenario, where the chances of global nuclear catastrophe escalate dramatically. Now the fact is, its extremely unlikely the US will make a serious mistake. Its also unlikely that Russia will. The problem is, we don't know how unlikely it is. In 1960, when the US built a early warning system in Greenland, in its first week it declared a full-scale Russian attack underway. Why? Because the moon rose, and programmers hadn't accounted for that.
The worst thing is, none of these things have to happen - no Korean missile, no Y2K failures, no Primakov dying, no chip failure, and we still have serious problems. How many people have heard of the 1995 missile launch off Norway? (Describe if Bruce doesn't, with thanks to Ted.) US-Russian relations were pretty good at that point, and we were still less than five minutes away from a Russian decision to launch a counter-attack. What happens if something goes wrong in the future? Bruce has described the dangers of the high alert status of US and Russian forces - the Y2K problem is only one, very serious, instance where that danger increases.
This danger was highlighted last week by Gen. Dvorkin. He was quoted in an AP wire story saying, "The risk of making the wrong decision is higher when international tensions escalate. The risk of such mistakes, including those caused by the unresolved Y2K problem, would be eliminated if international tensions eased, especially in conflict regions such as Iraq and Yugoslavia." Now clearly, there is some self-interest working here - Russia does not like US policy in Iraq or Yugoslavia, so this type of statement is, at least in part, an attempt to change that policy. But it is also a fact - in times of tension, stress increases, the likelihood of mistakes increases, and the whole situation becomes more dangerous.
So what do we do?
Bruce also described some very reasonable options for reducing alert status, and those should be pursued. But there are more steps.
Now this is an interesting point. In the UK, the government's Y2K program has threatened to shut down companies that fall behind in their Y2K compliance programs. The headline from the wire report reads "Britain Threatens Y2k Bug Laggards with Closure". Now, unfortunately, this statement came out of the UK's Action 2000 program, and is directed at health and safety inspections of private British companies. One wonders whether they are taking the same approach with their nuclear arsenal. Somehow I doubt it, but I, for one, wish they would.
Where do we stand?
A few words about what we are doing. In September, Clinton and Yeltsin agreed to share early warning data on nuclear launches, and to an enhanced pre-launch notification system, so that incidents like the 1995 rocket off Norway would be more unlikely.
Sharing early warning data, however, is no easy task. Originally, Yeltsin requested, and the US agreed, that they should build a joint facility in Russia to share this data, gathered from US and Russia satellites and radar systems. However, the agreement was vague, and it was not clear if, for example, data about US or Russian launches would be included, or just data about other countries launches. The reason for that former data being an issue - you would think it would be the most important element of the agreement - is concern about secrecy and spying capabilities. If you tell me what you can see, I can more easily figure out ways to avoid you seeing it. So at some levels in the Russian government, there is real concern that the only reason for the Clinton-Yeltsin agreement was US attempts to gather intelligence on Russian capabilities.
As a result of those difficulties, plus that fact that sharing early warning data is actually a complicated thing to do, has led to delays in the implementation of the early warning data sharing. Originally, the US hoped to have the Joint Warning Center established before 2000, so that it could help address concerns about Russia going blind at the millennium. But because of these delays which, at least according to US officials, are primarily due to Russian slowness, the Center will not be ready in time.
So instead, as Mike briefly described, last month, the US proposed a Center for Year 2000 Strategic Stability, aimed directly at addressing the millennium bug issue. Under the US proposal, the Center would be established outside Cheyenne Mountain, a US facility, and provide US data to Russian observers. Links would also be made to Russia facilities, although what would go over those links is not yet clear. This center would be established sometime this fall - perhaps in time for the 9/9/99 date - and would shut down in mid-January, once the millennium has passed.
In essence, what the US is offering is to provide filtered data to the Russians on US-detected launches globally. The problem is, it is very difficult to create a system where the Russians are absolutely confident that they are not being bamboozled, that the data they are seeing has not be filtered in such a way that US launches are being hidden. Now, of course, we are not going to do that. And it is very difficult to figure out how we could give the Russians unfiltered data. But the problem is, how do the Russians know we are not lying to them?
So I am offering today a fairly simple proposal that would add one more layer of confidence. The US should place cameras that can watch each US missile silo cover, and provide the Russians with a constant data stream so that they can tell their have been no launches. You could even put those camera images up on the web, so everyone could watch. You would have to build-in some system checks, so the Russians could tell the video wasn't being altered, but it is technically feasible. Unfortunately, it does not address the issue of the majority of the US force, missiles on Trident submarines. But it's a start.
There are also other steps that we should be taking.
One easy thing to do would be to not make the current situation worse. We are planning to put what are called D-5 warheads on all of our Trident missiles. The D-5 would replace the C-4 warhead, which is currently deployed on our Pacific-based submarines. (D-5s are already deployed on our Atlantic-based subs.) The D-5 have a greater yield, are more accurate, and better able to attack silos, than the C-4. In military terminology, the D-5 has hard target kill capability. If we simply decide not to put D-5s on those missiles, we will both reduce the threat to Russia's arsenal, and save money. Reducing the threat to Russia's arsenal has the benefit of lessening the pressure on Russia to launch its missiles early. If we attack with D-5s, the Russians have no assurance that any of their ICBMs will survive, so they have to launch 'em or lose 'em. If we attack with C-4s, we increase Russian confidence that at least a few of their missiles were survive. Thus, a second step, would be to remove the D-5 warheads already deployed on Atlantic-based Tridents, and replace them with C-4s.
A second option, and for this I am extremely grateful to Ted Postol of MIT, would be for the US to provide financial aid to allow Russia to launch new-generation early warning satellites. Dr Postol estimates that it would cost $50-$70 million per year to launch and operate six new satellites that would greatly improve Russia's own early warning capabilities.
Even further, Dr. Postol recommends that the US begin cooperative research efforts with Russia to develop early warning science and technology, which could then be used by both countries. An additional step would be for the US to sell or provide certain types of sensor technologies to Russia on a satellite by satellite basis. This technology - once put in space - would be locked away forever. The Russians would benefit from it, but would be unable to steal it, so concerns about technology transfer would be limited.
In a related area, you should also be aware that the Congressional Budget Office is, at the request of Senator Daschle, undertaking a review ofwhat are called "nontraditional cooperative arms control measures" to improve US-Russian nuclear security. A preliminary section of the study, on improving Russia's early warning, has already been released, and I commend it to you. It describes five steps that could be taken by the US to support Russian early warning capabilities. These would include paying the salaries of Russian scientists, funding Russian research in early warning technology, and providing assistance for Russian purchases of US sensor technology, and area where the Russians are weak. These recommendations would obviously ring warning bells in some minds, particularly in providing Western technology that has traditionally been very closely guarded.
And the reason for this is very simple. At present, the greatest threat to the US survival is an accidental or mistaken Russian launch. If the Russian government takes a sharp turn for the worse, more dangers appear. Anything that the US can do to reduce those dangers, to increase Russian confidence in the survivability of their arsenal and in the capability of their early warning systems, is in our interest more than anyone else's.
One point the CBO study makes, almost in passing, is well worth mentioning, and I'll conclude with this. All the early warning in the world is not as useful as reducing the size of nuclear arsenals. As long as both the US and Russia continue to maintain huge nuclear arsenals on high alert, the world, and the US in particular, faces annihilation. It would come, most likely, through mistake, through miscommunication, through bad data, just the kinds of errors that will almost be inevitable due to the Y2K bug. It is not too late to act, and in fact we must act.
noradiation.org |
