GCHQ

by Richard Aldrich

  The biggest limitation on Britain’s local ground-based sigint capability was range of intercept. Warsaw Pact command and control used VHF radio, and this could only be heard at a maximum distance of forty or fifty kilometres. This reflected both transmitter power and also the need for line-of-sight interception. Accordingly, it was possible to provide adequate comint to divisional commanders, but problems would occur at the level of corps and above, which needed to see further behind the front line. In short, sigint on Warsaw Pact reserve divisions would have arrived too late for the commander of British forces in Germany to react to it. Limitations of range also required sigint collectors to be based as far forward as possible, exposing them to risk and forcing them to move frequently, which was not ‘conducive to the best COMINT collection’.21 General pessimism was reinforced by the knowledge that at the level of divisional headquarters and above, Soviet communications security was excellent. In 1969, the monitoring of large-scale Soviet exercises showed that their ability to ‘successfully use communications security measures results in lean intelligence collected by the SIGINT organisation’.22

  It was only during the mid-1970s that new technology began to provide good alternative sources. Unattended ground sensors, small electronic devices which detected the movement of vehicles, began to become available in Europe. Initially trialled in Vietnam under the ‘Igloo White’ programme, by the early 1980s these had become a formidable intelligence instrument. Many of the early models – nicknamed ‘bump-counters’ – were tested during real Warsaw Pact exercises in Eastern Europe, having been put in position by Brixmis personnel.23 Had war broken out prior to 1970, British commanders in Germany would also have turned to traditional air reconnaissance using aircraft with cameras.24 However, this source of intelligence would have diminished quickly, because a horrendous aircraft casualty rate of 60 per cent was expected in the first week of any war with Russia.25 Both captured enemy documents and PoW interrogation were unlikely to be helpful in a fast-moving war.26

  Did all this planning for a future war matter? After all, deterrence was supposed to ensure that the north German plains would remain the Cold War’s frozen front, with little likelihood of real conflict. In fact, by the early 1960s a number of crises had created a climate of growing anxiety. Confrontations over Berlin and Cuba, together with the escalating conflict in Vietnam, made war seem somewhat closer. Alarmingly, American sigint had failed to give much warning about the emerging Cuban Missile Crisis in 1962. The Six-Day War in 1967 and the invasion of Czechoslovakia in 1968 were also unsettling for commanders. More generally, throughout the 1960s there was a growing awareness that NATO’s conventional inferiority in numbers, especially in northern Germany, might call for the early use of nuclear weapons to stem the tide of a Warsaw Pact attack. A better intelligence flow was required, not only for warning, but also for decision-making in any nuclear crisis.27

  Indeed, in 1968, even before the invasion of Czechoslovakia, London and Washington began thinking the unthinkable. Up until recently, sharing anything other than the lowest level of sigint with the European allies had been anathema. However, they now considered upgrading the flow of sigint to the new NATO political headquarters that had just moved to the outskirts of Brussels. Bill Millward, Director of Requirements at GCHQ, had already been out to Brussels for talks with British officials there. However, these plans were plagued by repeated espionage flaps at NATO. In the autumn of 1968 the CIA secured a major counter-espionage coup when it recruited Colonel Ion Iacobescu, the deputy head of the Paris station of the Romanian foreign intelligence service (known as DGIE). Iacobescu revealed that his boss in Paris, Mihai Caraman, was running two high-grade spies inside NATO’s Brussels headquarters. One of them was a Turkish staff officer, Colonel Nahit Imre, who had been recruited by DGIE in Ankara, and had been controlled from Paris since his appointment as Deputy Financial Controller of NATO. The other had been recruited by Caraman himself. This was François Roussilhe, who worked as Head of Translation in the NATO registry. The KGB had been so impressed by Caraman’s enterprise that it sent one of its own top officers, Vladmir Arhipov, to work alongside him in Paris. Once Caraman’s deputy defected to the West, everything began to unravel.28

  In late 1968 Nahit Imre was publicly revealed to be an agent of the Romanian intelligence service.29 Once he had been interrogated, his disposal was left to the Turkish authorities. His departure was monitored by the British security team at NATO, who reported that ‘when he went on board the Turkish aircraft he was closely followed by two Turkish “gorillas”’. He was tried secretly in Turkey, and executed by firing squad the following year.30 In London, the Imre case caused real consternation. Ronnie Burroughs of the Foreign Office immediately saw the connection between this case and the recent decision to improve sigint for NATO, and wrote to the senior UK representative on the North Atlantic Council asking ‘whether we are right to proceed with our plans to set up an all-NATO GCHQ cell at Evere [near Brussels]’.31 London eventually decided to press ahead, but was clearly unsettled.32

  In early 1969, NATO turned to deal with the Roussilhe case. François Roussilhe was a French employee of the international staff who had worked for NATO since 1952, and had gradually risen to be Chief Clerk of the Document Translation Centre. Since 1963 he had provided the Romanian intelligence service with at least five thousand NATO documents, including top-secret material. His motives had apparently been purely materialistic: he had been paid in cash – largely in gold – and up to date had received around a million French francs.33 The UK Chiefs of Staff took a ‘very grave view’ of these penetrations, and were determined that they should not be swept under the carpet. There were many voices in London now arguing that Britain was putting its own forces at risk by passing too much information to NATO.34

  The experienced MI5 representative on NATO’s Special Committee, Dick Thistlethwaite, had wanted to run Roussilhe as a deception agent ‘to confuse or reduce Soviet assessment’ of the information it had received. MI5 went so far as to propose a ‘NATO disinformation cell’ that would actually turn weak NATO security into an advantage. The British sensed an opportunity to repeat the fabulous achievements of the wartime Double-Cross Committee and the famous D-Day deception. However, this was blocked by the CIA, which disliked the sheer organisational complexities of doing such a thing in Brussels. The British complained that the CIA was timid, and that ‘Roussilhe was a lost opportunity’.35

  The greatest damage caused by these agents was the compromise of operating procedures for nuclear command and control mechanisms. At the same time the British accepted that where more general political policy on nuclear weapons was concerned, there was an avant-garde school of thought which held that ‘we should ensure sufficient leakage in order to induce the Soviets to enter the dialogue at this point in our discussions’. Over a hundred copies of Nuclear Planning Group documents were routinely distributed in English and French, and many regarded these as effectively open documents that would soon reach Moscow.36

  Remarkably, notwithstanding the repeated security breaches in the late 1960s, the GCHQ cell for NATO at Evere went ahead. Without a flow of sigint, NATO’s Military Committee was increasingly an obstacle to rapid mobilisation in time of crisis. This was illustrated by a strategic alert in early August 1969, triggered almost entirely by sigint. Monitoring of Soviet units by GCHQ detected an unusual lull in air activity, combined with what seemed to be pre-orders for large-scale troop movements. There had also been a slight rise in signal activity on the Soviet General Staff network. The intelligence suggested ‘preparations of a highly unusual character’ in Russia and other Eastern Bloc countries. The JIC thought the most likely explanation was an impending Soviet move inside Eastern Europe, against Romania or possibly Yugoslavia. There was even an outside chance of an attack on China.37 The signals traffic between Moscow and the Strategic Air Command, the Navy and the Strategic Rocket Force remained normal, so all intelligence agencies were convinced that the Soviets were not
contemplating hostile action against NATO. Nevertheless, the British Chiefs of Staff found this pattern of Soviet activity ‘unusual and disturbing’. They suggested that at the very least there should be some response from NATO, ‘so that the Russians would be made aware that we knew of their activities’. They asked General Andrew Goodpaster, the Supreme Allied Commander Europe (SACEUR), who was in charge of all NATO forces in Europe, about the possibility of sending such a signal to the Soviets, ideally by covert means.38

  On 10 August 1969 Britain instituted a covert mobilisation in collaboration with SACEUR, making war preparations, but only insofar as they would not be publicly visible or cause alarm. Border patrols were increased, and all RAF combat aircraft in Germany were dispersed. Battle-flight practice missions were flown, and all personnel were brought to two hours’ notice of readiness. However, efforts to stimulate NATO’s Military Committee into action were again frustrated by the difficulties over the release of sigint.39 This was explained to Denis Healey, the Defence Secretary, by one of his senior officials. The NATO Military Committee had suffered ‘paralysis’ on this occasion ‘because some of its members were not allowed to see the relevant intelligence material’, since much of it was sigint and so was CAN/UK/US Eyes Only. He added: ‘As early warning will very often be of the COMINT sort we could obviously meet similar difficulties again, and perhaps in more serious circumstances.’ By 12 August, further sigint collected from Hong Kong suggested an easing of the immediate situation. However, the issue of NATO sigint had now been identified as an ‘unfortunate state of affairs’ that simply had to be addressed.40

  The GCHQ cell at NATO also went ahead because of the recommendations of the Working Group on Sigint set up by NATO’s Special Committee. This suggested that the cell might focus only on intelligence that might influence war warning and ‘a timely decision to release nuclear weapons’. Sigint for NATO was specifically about intelligence during the critical period between the beginning of a crisis and a request for the release or use of nuclear weapons. In other words, there was never any intention to supply NATO with a broad range of sigint under normal peacetime conditions. Instead, the new arrangements for sigint only sought to regularise the position that had occurred during the Middle East War of 1967, when GCHQ and NSA had provided NATO with most of its intelligence, requiring the laborious production of specially tailored documents that provided detail on the situation, while hiding its source.41 In 1971, continued anxiety about a possible Soviet invasion of Romania or Yugoslavia prompted the American decision to ask the British to be allowed to run regular SR-71 Blackbird spy-plane missions from RAF Mildenhall, providing the ability to collect imagery and sigint at short notice.42

  Offsetting some of these anxieties, the late 1960s saw the beginning of arms control agreements and confidence-building measures that were only possible because of intelligence. Improved intelligence from satellites and sigint gave both sides more faith in these agreements. In this way, espionage activities of all kinds seem to have made the international system more stable by the late 1960s. Arguably, the emergence of a regime of reassuring measures, with its emphasis on the avoidance of crisis, also reflected a growing recognition that the command systems on both sides were fundamentally ill-suited to rapid decision-making under pressure. Some British officials had even begun to recognise that, taken together, the intelligence efforts of both sides contributed to a collective calming of nerves. Indeed, during the 1960s the penetration of the NATO registries by Eastern Bloc spies was so complete that the Warsaw Pact had no choice but to conclude that the intentions of Western countries were genuinely defensive and benign.43 Yet no amount of espionage ever provided complete reassurance. Accordingly, there was little respite for the intelligence agencies, and by the late 1960s ever greater resources were poured into the increasingly high-tech world of sigint and satellites.44

  14

  Staying Ahead – Sigint Ships and Spy Planes

  …no other European NATO countries will have purpose-built aircraft for Sigint collection.

  ‘Nimrod (R)’, enclosed in Frank Cooper, Ministry of Defence, to Defence Secretary Denis Healey, 30 July 19681

  By the 1960s the British Empire was shrinking fast. Most of Africa was moving rapidly towards independence, and sigint bases in Iraq, Egypt and Ceylon had gone, with much activity being pulled back to Cyprus. Now even Cyprus looked precarious, under pressure from local guerrillas who were campaigning for union with Greece. However, there seemed to be a possible answer to this looming problem: to place one of the large overseas monitoring stations on dedicated sigint ships. This would render it immune to the ploys of post-colonial politicians who wished to twist the tail of the British lion over base rights. It would also permit flexibility, since a floating base could be moved, providing emergency monitoring during the Cold War crises that increasingly erupted in the Third World. These new ‘sigint ships’ would also help Britain to remain one of the world’s leading intelligence powers.

  Sigint ships were not new. The aircraft carrier HMS Albion and the cruiser HMS Superb had been on prolonged sigint cruises in the Baltic as early as 1949; moreover, a number of naval frigates had listening suites installed. Also, as we have seen, several British submarines were modified for intelligence duties and undertook hazardous monitoring operations against Soviet naval bases. There were also listeners on fishery protection vessels and trawlers. However, these were small teams of twos and threes attached to normal ships. Britain had no dedicated sigint ships. By contrast, the Soviets used a flotilla of small spy ships that looked like trawlers. Meanwhile, the American NSA was developing a fleet of Technical Research Ships, consisting of converted supply ships and Second World War Liberty ships (vessels that had been hastily produced for convoy duty); it was these latter developments that had caught Cheltenham’s eye.

  The first vessel in NSA’s Technical Research Ship programme was the USS Oxford, a converted Liberty ship which put to sea in 1961. Its flexibility was immediately apparent as it floated from one trouble spot to another, admittedly at an ambulatory pace of just eleven knots. It proved especially successful during the Cuban Missile Crisis. Accordingly, during the 1960s four more ships in the Oxford class were launched: the Georgetown, the Jamestown, the Belmont and the Liberty. Three smaller ships joined the listening fleet, based on converted military transports: the Valdez, the Muller and finally the Pueblo, which entered service in 1968. However, the American sigint budgets were tight in the 1960s, and while all these vessels were refurbished, they had nevertheless seen better days.2

  In 1964, partly as a result of advice from the Hampshire review, GCHQ decided that it would go one better, creating a purpose-built sigint ship, in contrast to the Americans’ elderly converted transports.3 The cover name chosen for this exciting new project was the ‘Communications Trials Ship’. The Permanent Secretaries Committee on the Intelligence Services (PSIS) approved the construction of the first ship on 19 July 1965. The ambition was to have a fully fitted-out ship by the summer of 1969, allowing time for trials in the autumn before it entered service in early 1970.4 There were hopes for two further ships in the same class.5 In autumn 1965 GCHQ joined the Navy and the Ministry of Transport in sketching out the details.

  The sigint ship was to be built to ‘extreme naval standard’, and capable of worldwide operations, ‘including the Persian Gulf’. It needed to be able to cruise at very slow speeds – about five knots – for long periods, and to have a range of eight thousand miles. Substantial size was important, since ‘spacing of masts’ allowed good direction-finding. The crew was to consist of fourteen officers and fifty-three men. These would be outnumbered by the sigint operators, referred to as the ‘Special Trials Party’. This would normally consist of seven officers, fifteen warrant officers and seventy Radio Operators. In an emergency, twenty further operators might be accommodated. The Radio Operators would work in two large receiving bays, and their facilities would include a trials analysis room, a tape-editing room and a ‘
special facilities’ room. There was also to be a strong room for the most secret material.6 All these ‘special features’ were to be housed in a main superstructure unit dubbed the ‘black box’, which would be prefabricated before delivery of the ship.7

  The GCHQ personnel driving this project were Russell Dudley-Smith from W Division and Michael Herman from S Division, who was effectively the project officer. There was also strong naval representation.8 Staffing was clearly a big issue. It was agreed that the crew that sailed the ship would be civilian, but the biggest challenge was the ‘Special Trials Party’. There were not enough sigint operators available, and no one wanted to pay for them. The Navy asserted that it could only find ‘a small proportion’ of the sigint staff without cutting back on its shore-based listening sites.9 Moreover these personnel were now civilians, and there were obvious problems with telling seventy landlubbers that they were suddenly going to sea.

  The high point of this remarkable project came in March 1966. GCHQ’s plans had grown to envisage a large experimental nuclear-powered sigint vessel. This was intended to be an existing aircraft carrier refitted with a nuclear power plant. It would afford massive space for intercept staff, and would have almost unlimited range for operations. The Atomic Energy Authority was brought in on the planning, and a design study contract was placed with Harland & Wolff of Belfast for ‘finalising the proposal in more detail’ and estimating costs. Because of the sensitive nature of the project, the cover story sought to disguise it as a ‘floating radio transmitting station on behalf of the Diplomatic Wireless Service’.10 This allowed some pre-publicity, and on 5 June 1966 the Sunday Express reported that the ship was designed to allow the Voice of Britain to be heard ‘loud and clear in the world’s most troubled areas’. Journalists were closer to the mark when they argued that the ship would fill a ‘void created by the surrender of British overseas bases’.11