18
Unmasking GCHQ: The ABC Trial
We’ve been to MI5, MI6, Scotland Yard, Parliament and many more. Now we’re going where much of the dirty work goes on – CHELTENHAM!
ABC trial campaign newsletter, 27 May 19781
During the late 1960s and early 1970s, signals intelligence was changing fast. The big players were discovering a whole new world of super-secret interception which provided a different sort of signals intelligence. This new source was telephone calls. As we have already seen, tapping telephones was hardly new, and had boomed in the 1950s in response to problems with reading high-grade codes. However, telephone calls were now increasingly being routed away from old-fashioned cables, which were hard to intercept, especially within secure states like the Soviet Union, and being carried by the more modern means of radio links, using microwave towers and satellites. This was a vastly more efficient system of communication, especially for long-distance or international calls. One of the side-effects was that conversations now spilled freely into the ether, making the possibilities for interception almost limitless. The best thing about this new source was that the material was often not encrypted so it provided a veritable fountain of intelligence virtually for free. A sigint revolution was just around the corner.
The downside of these new developments was that they produced inconceivable volumes of material. Computers were no longer needed just for breaking codes, but also for combing through the intelligence, storing it and distributing it to customers. The volume of sigint that was being collected was too large for any human to read. Moreover, the vast complexes of domes and satellite dishes that now accompanied the supposedly super-secret intelligence activities of NSA and GCHQ meant that they were more and more visible. Sooner rather than later, an enterprising investigative journalist was bound to point to these surreal installations and shout the dreaded words ‘Signals intelligence.’ It is amazing that in the mid-1970s GCHQ was still managing to pass itself off as a glorified communications relay station, hiding its real activities from public view. Anonymity would not last much longer.
What we now recognise as the first glimmerings of a global telecommunications revolution seemed to be in the interests of the world’s major sigint agencies. A fascinating example of this was an operation carried out jointly by the British and Americans in 1969. NSA was gathering a great deal of intelligence from telephone calls between Fidel Castro’s Cuba and the many Cuban exiles living in Florida. Using sigint ships, it was also possible to intercept some calls from Havana to other parts of Cuba. This was of some importance, since an elaborate game of cat and mouse was being played between the CIA and the Cuban intelligence service among the exile communities of Florida. Most of the calls were carried on a radio network called a tropos-pheric scatter system, owned by the Radio Corporation of Cuba, which also carried teletype traffic. The Radio Corporation of Cuba was a subsidiary of the American telecommunications giant ITT which had been gleefully nationalised by Castro when he came to power in 1959, along with all other American businesses on the island. The radio installation in Cuba was in need of a substantial upgrade, requiring new condensers and transistors, leading the Americans to fear that the link would cease to function, and the telephone calls would be sent by a different route, perhaps by cable, which was much harder to intercept.
Because America maintained a strict economic blockade of Cuba, it needed to subvert its own embargo in order for the necessary material to reach the island – besides which, if it came direct from America, the uncharacteristic generosity would alert the Cubans to the interception. The Americans turned to Britain, which had no embargo on Cuba, explaining that ‘The intelligence community regards the maintenance of the link as being of considerable importance.’2 The British, for their part, welcomed the opportunity of placing the Americans in their debt by participating in a covert operation.3 The materials were duly shipped to Britain, and then re-exported. NSA ensured that the suppliers were indemnified against legal action for breaking the embargo, while ITT used its subsidiary in Britain, Standard Telegraph & Cable Ltd, as a cover to make the delivery.
One of GCHQ’s largest ventures into the world of vacuuming up telephone calls was launched in Cornwall in 1967. At Goonhilly Downs on the Lizard peninsula there was a satellite receiving station for one of the world’s first commercial communications satellites, Intelsat. Intelsat was a booming commercial venture that carried a growing proportion of the world’s private communications, and was partly responsible for the explosion of international telephone calls. The system grew from 240 channels when it opened in 1964 to thirty thousand in 1983.4 Displaying a certain amount of barefaced cheek, GCHQ built a duplicate receiving station about sixty miles down the road, near the village of Morwenstow, on the site of a former RAF wartime airfield. Here it could scoop up the same telephone traffic by simply collecting the ‘spillage’ as commercial satellites beamed messages down to earth. This station, with its distinctive domes and satellite dishes littered along the Cornish clifftops, was initially called CSO Morwenstow, and later changed its name to GCHQ Bude. Morwenstow was a classic Anglo–American intelligence venture. NSA paid for most of the infrastructure and the technology, while GCHQ contributed the land and paid for the staff and running costs. The massive flow of intelligence it received was shared and processed jointly.5
In 1969 GCHQ was working hard to develop revolutionary new systems for analysing and distributing the huge volume of intelligence intercepts, with computers being used to search for keywords that indicated subjects of interest. On 3 March Joe Hooper, the Director of GCHQ, explained to Dick White, the Cabinet Office Intelligence Coordinator, that the sigint collected at Bude would shortly be fed into a computer database that would be used by Britain and the USA to select product. The main challenge was restricting access to information at the right level to people with the right clearances. This, Hooper explained, involved ‘a complicated system of “gating” in the computer programmes’. This was the first British venture with secure multi-level access computer systems for delivering intelligence.6
GCHQ was at the cutting edge of what would eventually be a transformative technology for all kinds of researchers. In the past, anyone who wanted to look at large volumes of newspapers would have to trawl laboriously through them physically. The time and effort involved meant that they could focus on only a few chosen titles. Today, accessing the world’s press is effortless and can be done at the click of a mouse, because it is available electronically. The downside, again, is ‘too much information’. Nobody can read all the world’s press, so modern researchers use word-search facilities like Google Alert. In exactly the same way, NSA and GCHQ could not listen in to the entire world’s satellite telephone calls, telexes and faxes. So they fed all the material into computers and built a top-secret equivalent of Google Alert, constructing computers that combed the traffic for keywords and predesignated phrases. This system was called ‘Dictionary’.7
Each new form of interception presents fresh legal challenges. GCHQ’s collection operation at Bude took three forms. At its most expansive it involved baseband trawling, which meant moving through large volumes of traffic from each country that was being monitored to find out whether material of interest was passing through particular channels. Next, there was a system that allowed the interception of all the traffic from a particular telephone dialling code, in other words perhaps one section of a city. Thirdly, there was the monitoring of specific telephone numbers. Initially, the GCHQ staff were uncertain about the legality of some of the broadband trawling, which included some British channels. For reasons that are not entirely clear, managers eventually persuaded them that it was not a legal issue. Allegedly, specific British numbers were targeted, and these were known as ‘P-Numbers’. GCHQ was already beginning to encounter what we now recognise as one of the key characteristics of globalisation – the mixing up of what is inside and outside the nation state. Previously there had been a clear dividing line b
etween domestic and international communications, which made for useful legal distinctions in the realm of interception. During the 1970s these distinctions were beginning to break down.8
Even telephone messages that were not carried by satellite were vulnerable, because they were increasingly being beamed between microwave radio towers. The beams travelled in a straight line, eventually spilling out into space because of the curvature of the earth. This led to yet another collection revolution, since the Americans discovered that much of this telephone traffic was vulnerable to interception by a new generation of sigint satellites. By the 1970s these were scooping enormous amounts of communications from space. One of the first major dividends was an improved ability to listen in to Moscow. NSA began to listen in to telephone calls in the Soviet capital, and also to the radio messages of taxi drivers and the Zil limousines used by the Politburo. The drivers gossiped constantly about their passengers, revealing fascinating insights into life amongst Russia’s elite. Alas, this was revealed publicly in the Washington Post in September 1971 by the renegade American journalist Jack Anderson.9
The ability to listen in to telephone calls carried by radio waves meant a radical shift in GCHQ’s business all around the world. In August 1974, British officials noted with delight that Syrian communications security was unsophisticated, and that ‘In some quarters the open telephone is treated as if it were secure.’10 GCHQ also had to assist friends and allies to protect them against the new sigint techniques: in Oman, for example, speech security equipment was installed.11 The Joint Intelligence Committee also began to rethink Britain’s intelligence targets in Europe because of this telephone sigint bonanza, which it coyly referred to as ‘advances in intelligence-gathering by technical means’.12 By 1973, new subjects had joined Russia as ‘Priority One’. These included the ‘Stability of the UK’, which required intelligence to look inwards. Specific new targets included the IRA and extremist organisations within Britain. The other new priority, which had been growing fast since the 1960s, was Western Europe. Intelligence on West Germany was now far more valuable than intelligence on East Germany. Britain needed to know what Bonn thought about Britain’s diplomatic recognition of East Germany, Britain’s entry into the Common Market and major arms deals such as the new NATO Tornado Multi-Role Combat Aircraft that would eventually equip the air forces of Britain, West Germany and Italy. Britain’s senior policy-makers also wanted better support for the sort of guerrilla wars that had flared in Cyprus, Aden and Oman during the 1960s and 1970s.13
NSA was already experimenting with sophisticated satellites that would revolutionise sigint activity by focusing on signals that were above the high-frequency range.14 On 19 July 1970 America launched its first operational Ryolite satellites, which stayed broadly in the same place above the earth, requiring ground stations in specific locations to receive the voluminous sigint that they beamed down.15 These stations were constructed at Pine Gap in Australia and at Menwith Hill near Harrogate in Yorkshire. Menwith Hill had been a field collection station for the US Army Security Agency since the 1950s, but was later taken over and run directly by NSA. By the 1970s it had been vastly expanded, and had become the largest American overseas intelligence base in the world. Its dozens of space-age domes, each of which hid a satellite dish, looked somewhat incongruous amongst the grazing cattle of the Yorkshire countryside. Now, its main purpose was to download torrents of sigint collected by the new generations of American sigint satellites. About 15 per cent of this ‘overhead’ material was diverted to Cheltenham, where a new super-secret section of GCHQ’s Soviet Section (J Division) – code-named J-Operations or ‘J-Ops’ – was set up to handle this new influx of sigint.16
Much of the popular anxiety about the vast American ground station at Menwith Hill, nominally disguised as an RAF station, has focused on its connection to the nearby Hunters Stones Post Office Tower, which forms part of the British microwave telephone network. For years, campaigners protesting against the American base insisted that this allowed NSA to eavesdrop on British domestic communications. This was fervently denied by Albert Braeuninger, the base commander in the early 1980s, who insisted that NSA was merely a customer of British Telecom. This statement was correct, but it actually hid a different secret. Much of the product from the Ryolite satellites was being routed over this microwave network to GCHQ’s secretive J-Ops section, where it was processed locally on behalf of NSA. This indicated an important change in status for GCHQ in the late 1970s and 1980s, since it increasingly became part of the processing system for what NSA collected.17
NSA’s new satellites were actually a major problem for the Americans by the late 1970s. The ‘take’ from these sources was enormous and still growing, yet funding for NSA was being reduced as a result of post-Vietnam defence budget cuts. NSA’s own historian, Robert Johnson, notes: ‘Scarce resources meant reliance on outside help. As the budget got slimmer, NSA turned increasingly to the help that foreigners could provide. This trend accelerated in the 1970s to a greater degree than at any time in post-World War II cryptological history.’ The collaboration between GCHQ and NSA was ‘almost total’, and at ‘each bend in the road, NSA made the conscious decision to remain engaged’. Indeed, these developments, which also involved other friendly countries, were so important that NSA appointed a ‘Third Party Manager’ to look after the increasingly complex relationship with allies.18
The sigint satellite revolution was an unsettling experience for GCHQ. Although the British were ‘in’ on developments such as Bude and Menwith Hill, they were not ‘of’ them, since they had no satellites of their own. The golden age of high-frequency ground-based interception by manual operators sitting in nissen huts was drawing to a close. Major overseas sigint sites, such as Cyprus, were still vital, especially for missile surveillance. However, broadly speaking, the satellite revolution, together with the possibilities of remote collection, was gradually downgrading the value of GCHQ’s exotic real estate. Meanwhile, the advent of satellites created a new super-club of sigint powers, of which there were only two members, America and Russia. It was a club that GCHQ ached to join.
In one area, GCHQ was joining the satellite club. This was the field of communications satellites for forwarding sigint from collection stations to GCHQ, and for communicating with NSA. In the early 1950s, communications between GCHQ and NSA were poor, due to insufficient cables and inadequate clear radio frequencies.19 In 1956, the transatlantic channels between the two agencies were suffering outages of over four hours a day.20 The main worry was enemy disruption in any future conflict. A secret study had concluded that the Soviets were likely to try a range of sabotage tactics in war, among the easiest of which was trawling up cables and cutting them. Transatlantic cables were at ‘trawl depth’ for long stretches, and the Soviets were expected to deploy specially modified nuclear submarines for the task. Other options were available, including the use of KGB agents to place ‘clandestine, low-powered jammers’ close to the relay stations.21 By the 1970s Britain would be overcoming some of these problems with its own military communications satellite, called ‘Skynet’. Skynet also provided secure encrypted speech facilities to the military wherever a Skynet terminal existed. The first Skynet satellite was expected in 1969, followed by a progressive build-up of further satellites and ground terminals. In obscure locations such as Bahrain, GCHQ was the largest Skynet user. On Cyprus it required no fewer than eight channels, and was responsible for more than 80 per cent of the traffic coming back to Britain. Much of this was data on Soviet missiles from Project Sandra and Project Cobra Shoe.22
Skynet was a major scientific achievement, and represented Britain’s first significant step into space. The launch of the satellite was regarded as a cause for national celebration. In November 1969 the RAF was invited onto the BBC children’s programme Blue Peter to display a mock-up of the satellite which was admired by the presenters Valerie Singleton, John Noakes and Peter Purves. The programme also described Britain’s new s
atellite control centre at RAF Oakhanger. However, no public mention was made on Blue Peter, or indeed anywhere, of Skynet’s biggest single customer, namely GCHQ.23
The capacity provided by Skynet was enormous, yet the planners noted that it was ‘adequate for all users of data except GCHQ’. The unimaginably huge amount of intelligence it was sharing with its main partners meant that GCHQ still required its own dedicated cables to NSA, the Canadian CBNRC and the Australian DSD.24 By February 1972, the Cabinet Office was beginning to look at the next generation of Skynet satellites and its alternatives, at a cost of about £50 million. Again, the biggest customer was GCHQ.25 All of its allies, especially the CIA, were now ‘satellite-conscious’.26 GCHQ pressed for closer integration with the United States, arguing for an off-the-shelf American Type-777 satellite, which was hardened to withstand some of the effects of a nuclear explosion. However, government Ministers were desperate to maintain a British national space programme, so they chose Skynet III.27
Preserving UK national reserves of knowledge and expertise was something that GCHQ touched on in many ways. For example, it had long helped to steer policy on the teaching of languages like Chinese in British universities.28 More importantly, it had a role in the development of British computing. Code-breaking had driven important breakthroughs in computing both during and after the Second World War, led by luminary figures such as Alan Turing. The most famous example is ‘Colossus’, which was used to attack ‘Tunny’, the encyphered teleprinter used by the German High Command. Ten examples of Colossus II were in operation by the end of the war. Other early computers called ‘Robinson’ and ‘Aquarius’ were no less innovative. Both Robinson and Colossus were designed and built at the Post Office Research Station at Dollis Hill by the celebrated Tommy Flowers, now recognised as one of the most enterprising scientists Britain produced during the war. At the end of the war a number of engineers moved from Dollis Hill and the British Tabulating Machine Company at Letchworth to join GCHQ at nearby Eastcote.29
GCHQ Page 37