6 Bases During the Cold War the Bipolar Base Race Before 1940, the U.S

6 Bases during the Cold War The bipolar base race Before 1940, the U.S. basing structure had been restricted to a small number of colonial possessions and satraps, in Puerto Rico, Cuba, the Philippines, Hawaii, Wake Island and the Panama Canal Zone. The Lend-Lease Act added a string of bases in British and Canadian territories along the U.S. Atlantic littoral, in Labrador, Newfoundland, Bermuda, the Bahamas, Antigua, Trinidad and British Guyana. Then at the start of World War II, the U.S. established forward access in Greenland, Iceland, the Azores, Mexico (Acapulco) and Ecuador (the Galapa- gos Islands). By the end of the conflict the U.S. had access to a near-global network of facilities in Europe, Asia and the Middle East. The Soviet Union, by contrast, had basing access only within its contiguous Eurasian empire, added to by the war’s end by access all over Eastern Europe and in Mongolia. Britain, France and the Netherlands – also to a lesser degree World War II neutrals Spain and Portugal – regained control over colonial empires and related basing access after the war, only to see them gradually wither with the decolonization process between 1945 and around 1970. Japan and Germany lost everything they had. By the late 1940s, the U.S. and the USSR were fully engaged in what would become a global struggle for influence and access, in which access to bases was to play a pivotal role, given the expectation of an inevitable “World War III.” That competition accelerated after 1955, outside the bounds of the contiguous Soviet empire and, importantly, had several features, but which were altered across several identifiable, if not neatly bounded, phases of this competition. They were: • The importance of Western, mostly British, colonial possessions for U.S. basing access in the early Cold War period, gradually fading in importance over time up to around 1970. • The increasing importance of alliances, ideological ties and security assistance in determining the patterns of U.S. and Soviet basing access. • The spatial configuration of “heartland” vs. “rimland” as depictive of the U.S.–Soviet “base race,” with the U.S. establishing a ring of bases around the Sino-Soviet Eurasian heartland, and the Soviets gradually leapfrogging the rimland via ideological client states and arms transfers. The configuration İ 2007 Robert E. Harkavy

Page 2 was gradually altered towards a more inter-penetrated basis, with rivals’ bases in proximity in various regions. • The increasing importance in the “base race” of technical or C3I or “ISR” facilities, as supplementary to “traditional” naval and air bases: communications, space-related, ballistic missile defense, maritime detection, nuclear detection, signals and photoreconnaissance intelligence etc. • A reduction in the number of bases required by superpowers as a result of technological development, i.e., longer-range aircraft, tanker refueling, the increased amount of firepower packed into planes and ships, nuclear ship propulsion etc. • Towards the end of the Cold War and previewing the new world after 1991, a newer geopolitical focus on the Persian Gulf and the “arc of crisis” running from the East African Horn to Central and South Asia. • A two-layered situation in which the bipolar rivals required basing access for nuclear deterrence and for conventional power projection, but with these functions often being combined in given basing facilities. In the period immediately following the end of World War II, the U.S. was faced with decisions about how far down to draw the enormous basing network it had developed during World War II. At least one major article in Foreign Affairs during this period pointed out the history of American weakness in this area before World War II, and advocated that in line with its new powers and responsibilities that the U.S. should retain a global basing structure after the war.1 Of course, not all the countries that had hosted U.S. bases during the war were willing to continue to do so; Brazil, for instance, insisted on U.S. withdrawal in consonance with the altered conditions. During the immediate postwar period, however, before the main thrust of decolonization acquired full force, the U.S. was able to rely on the bases provided primarily by the British, but also the French, Dutch and Portuguese empires.2 These empires unraveled only gradually over a 25-year or so period. This provided the U.S. access to a plethora of air and naval bases: Gibraltar, Malta, Cyprus, the Suez Canal Zone, Aden, Libya (Wheelus Air Force Base), Gan in the Maldives Islands, Mombasa, Simonstown in South Africa, Freetown, Singapore, Hong Kong, etc., added to by access to other facilities in Canada and Australia. France’s remnant empire allowed for some U.S. access as well, particularly important in the 1950s, regarding bomber bases and communications facilities in Morocco. During the period after 1945, of course, the U.S. occupied Germany and Japan, and had free access to bases all over occupied Europe, in France, the Netherlands, Belgium, Italy, Greece and Turkey. In the Central Pacific, the U.S. took over the former Japanese League of Nations man- dates in the Caroline, Mariana and Marshall Islands, with few restrictions on their use for bases. In the early 1950s there was the full onset of the Cold War, the Korean War, and NSC-68 as the formalization of the containment policy. Importantly, there was also the construction by the U.S. of a network of formal alliances, both Bases during the Cold War 95 İ 2007 Robert E. Harkavy

Page 3 multilateral and bilateral, that formed a physical barrier around what had come to be called the Sino-Soviet bloc after 1949, a barrier extending around Eurasia from Ireland and Norway around to Japan. There was NATO, CENTO (the Baghdad Pact) and SEATO, the first two hinged on Turkey, the latter two on Pakistan, abetted by ANZUS and the bilateral defense pacts with Japan, South Korea (after 1950) and Taiwan (after 1950).3 Arms transfers and other forms of security and economic assistance underpinned all of these alliances, and with them went a very permissive and comprehensive basing structure for the U.S., in effect in exchange for security against the perceived menace of Soviet and Chinese expansionism, both in terms of possible military attack and internal sub- version. The physical structure of this basing network was neatly reflective of a rimland defense posture around Eurasia à la the geopolitical formulations of Mahan and Spykman. Functions of the U.S. Cold War basing system As the Cold War evolved in the 1950s and 1960s, the functions of the U.S. global basing structure came to be divided along the lines of nuclear deterrence and conventional deterrence and power projection. The latter category further subsumed the use of bases for direct military interventions, arms resupply during conflict, coercive diplomacy (“gunboat diplomacy”) and “presence” (showing the flag).4 The initial use of forward bases came with the stationing of B-29 bombers in the U.K. in the late 1940s. Following that was the stationing, in the 1950s, of B- 47 “Reflex Force” bombers in the U.K., Spain and Morocco.5 Shortly after that, the B-52 bombers were introduced as the backbone of SAC, and they were all based along the northern rim of the continental U.S., in Maine, New Hampshire, upstate New York and Michigan. But, they required tanker refueling, bases for which were established in Greenland (Thule), Canada (Gander), Iceland (Keflavik) and in the U.K. Additionally, numerous bases in Europe and elsewhere were designated as recovery bases for the bombers should they exit the USSR after a bombing raid. Later in the Gulf War, forward based tactical aircraft (FBS) – F-4s, F-16s, and F-15s – were configured for nuclear weapons, based mostly in Germany, but able to mount missions from other bases in Europe, in Iceland, the Netherlands, Italy and Turkey.6 Additionally, F-111E bombers based in the U.K. were deployed as nuclear attack aircraft that could reach into the Soviet Union. Nuclear armed attack aircraft were also based in South Korea and the Philip- pines, capable of reaching the USSR and China. Tanker aircraft based on Guam, Okinawa, Japan and the Philippines could also serve in relation to nuclear-armed aircraft. In the late 1950s, in response to the “missile scare,” i.e., the looming threat of Soviet ICBM deployments ahead of similar U.S. deployments, the U.S. based medium-range ballistic missiles, Thor and Jupiter, in the U.K., Italy and Turkey. In Asia, Matador and Mace land-based missiles were deployed in Okinawa and 96 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 4 Taiwan as a deterrent vis-à-vis China and the Soviet Far East.7 And then, in the 1980s, the U.S. forward deployed Pershing II and ground-launched cruise missiles (GLCMs) in several Western European states – Germany, Belgium, the Netherlands, Italy, U.K. – to counter the menace of Soviet SS-20s being targeted on Western Europe, which deployments were drawn down after the signing of the INF Treaty with the USSR. Nuclear-armed U.S. naval forces also made extensive use of overseas facilities. American “boomers,” submarines mounting long-range SLBMs, were long based at Holy Loch, Scotland and Rota, Spain, in addition to Guam. These forward deployments allowed the U.S. to negotiate a SALT II Treaty with the USSR that gave the latter a larger number of SSBNs, somewhat balanced out by the U.S. having those overseas facilities that allowed for keeping relatively more SSBNs “on station” at a given time. Finally, nuclear armed SSNs, attack submarines, whose main intended function was to hunt Soviet SSBNs, were based overseas at Faslane, Scotland, La Maddalena on Italy’s Sardinia and Sasebo in Japan in proximity to known major transit routes for Soviet submarines. And, lastly, U.S. carriers (CVNs) with nuclear armed aircraft and other surface ships carrying nuclear weapons made frequent stops for replenishing at numerous overseas ports, though that was sometimes a political problem in nuclear-phobic Japan. U.S. nuclear-armed CVNs operated out of Yokosuka in Japan, Subic Bay in the Philippines and Naples, Italy. Electronic intelligence aircraft prowled along the Soviet borders and coasts to “tickle” Soviet radars and to plan possible routes for U.S. bombers. U-2 and SR- 71 surveillance aircraft were used, particularly in the case of the former up to 1960, to monitor Soviet construction of ICBM missile silos and nuclear test sites. The U-2s earlier were based at Mildenhall in the U.K., Bodo in Norway, Wiesbaden in West Germany, Incirlik in Turkey, Peshawar in Pakistan and Atsugi in Japan; the SR-71s mostly in the U.K., but deployable to other bases.8 Some other air and naval basing functions also related to nuclear deterrence. The U.S. had an elaborate network of P-3 Orion naval air bases in relation both to surface and sub-surface detection of Soviet ships, with a focus on ASW.9 Around the Indian Ocean littoral, P-3s were flown out of Masirah in Oman, for instance, Mogadishu in Somalia (after 1978), Singapore and Australia’s west coast. Sigonella in Sicily was used as a major P-3 base for monitoring the movements of Soviet submarines in the Mediterranean. Other P-3 bases have been Embelzebil/ Nairobi (Kenya), Tahkli (Thailand), Djibouti, earlier in the Seychelles and Bandar Abbas in Iran during the Shah’s reign, Misawa and Iwakuni in Japan, Cubi Point in the Philippines and Dakar in Senegal. Then, TACAMO aircraft based in, among other places, Bermuda, were utilized to trawl communications wires through the ocean to communicate with U.S. nuclear submarines.10 Regarding conventional power projection capability, the U.S. also utilized a massive network of overseas bases, again for purposes of conventional deterrence (tripwires), arms resupply, coercive diplomacy and presence. Sometimes but not always, these deployments involved co-location at bases utilized in relation to nuclear deterrence. Bases during the Cold War 97 İ 2007 Robert E. Harkavy

Page 5 First and foremost was the long-term, stable deployment of ground forces in Europe and Asia, primarily in Germany, South Korea and Japan, but also in Italy and Belgium. That involved barracks, training grounds, maintenance depots, hospitals, etc., in relation to large-unit deployments of corps and divisions; like- wise, the deployment of protective surface-to-air missiles to defend these installations. The largest was with the Seventh Army in Germany, amounting at its peak to some 330,000 troops, mostly in Bavaria, Hesse, Baden Wurtemberg and Rhineland Pfaltz, reduced to below 100,000 after the end of the Cold War. In Italy, a combat brigade was long deployed near Vicenza to bolster the Italian army and in relation to a possible Soviet thrust through Austria into Italy. In the Far East, the U.S., after the end of the Korean War, long deployed around 40,000 troops in South Korea, mostly along the DMZ and around the capital city, Seoul. There was a threat by the Carter Administration to reduce or eliminate these deployments, which did not eventuate, in response to the growing disparity in economic might between the two Koreas. In Japan, more than 40,000 troops were long deployed, mostly Army and Marine units on the island of Okinawa, co-located with extensive deployment of combat aircraft, tankers and transports. Later, after the end of the Cold War, this deployment became a hot issue within Japan, and consideration was given to moving the troops to the Japanese mainland. U.S. Navy surface fleet bases11 The U.S. Navy, throughout the Cold War, utilized a number of major and minor bases for its surface fleets, most importantly, for aircraft carrier battlegroups. Several British ports were regularly used. Before France’s withdrawal from NATO’s military structure in the 1960s, Villefranche was a frequent port of call for U.S. ships, as were Barcelona, Livorno, Rota, Piraeus, Souda Bay on Crete and Izmir in Turkey, the latter a regional naval headquarters. Naples and nearby Gaeta provided a homeport and headquarters for the U.S. Sixth Fleet in the Mediterranean Sea. In the Persian Gulf, even well before the Iran–Iraq and Desert Storm con- flicts, the U.S. homeported a small flotilla and had a headquarters for its MIDEASTFOR in Bahrain. Naval deployments in and around the Persian Gulf were upgraded in the late 1980s. Mombasa in Kenya and Djibouti then became ports used frequently by the U.S. Navy. In the Far East, Yokosuka in the Tokyo Bay was long a hub of the U.S. Seventh Fleet, with its extensive drydocking capacity that could handle large nuclear-powered carriers. Sasebo on the Sea of Japan also hosted USN surface vessels as well as attack submarines. And up to the end of the Cold War, Subic Bay in the Philippines was another U.S. Navy main base, which was used to support operations in Vietnam and elsewhere in Southeast Asia. Then, during the 1980s, the Philippines began to eliminate U.S. use of its long-held naval and air bases. The U.S. considered replacement bases variously at Guam, Taiwan, Thailand and Australia, finally settling on extensive utilization of Singapore’s 98 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 6 Bases during the Cold War 99 Table 2 Main and secondary surface-ship and submarine operating bases of the U.S. Navy, as of late 1980s Host nation and base Description For surface ships Japan Yokosuka Major navy base, HQ for U.S. Naval Forces Japan and homeport for aircraft carrier Midway and about ten other 7th Fleet ships, available for contingencies in West Pacific; also used by nuclear attack submarines; extensive dock facilities, naval munitions maintenance and storage, naval hospital, ship repairs including largest USN drydock west of CONUS; supply depot. Sasebo Base used jointly with Japan; naval ordnance facility, docking storage facilities for 7th Fleet; homeport for a nuclear attack submarine, drydock capacity for aircraft carriers, large-scale naval fuel storage, munitions storage for USMC. White Beach (Okinawa) Berthing and storage for 7th Fleet ships, occasional use by SSNs. Taiwan Kaohsiung Occasional port use by U.S. ships. Guam (U.S. overseas possession) Apra Major naval base; ship repairs, logistics wharf, explosives and fuel storage; formerly Polaris homeport for eight SSBNs; patrol boats and mine flotilla based here. Philippines Subic Bay Major USN base, HQ for U.S. Naval Forces, Philippines, major ship repair facility with four floating drydocks which can accommodate all but largest aircraft carriers; piers and other support facilities – support 7th Fleet operations throughout West Pacific and Indian Ocean; 60 percent of all 7th Fleet repairs performed here. Australia Cockburn Sound Australian base, can accommodate four submarines and four destroyers; possible expansion to accommodate carriers; U.S. considered homeporting a destroyer here; mostly potential U.S. base, offered earlier as such by hosts; port calls at Hobart, Melbourne, Sydney, Brisbane, Darwin. Thailand Sattahip Military port constructed by U.S.; major port of entry for military supplies to U.S. bases in Thailand; peaked at end of Vietnam War. Hong Kong Periodic port calls. Singapore Sembawang Some overhauls, reprovisioning for USN ships, potential for expanded use. İ 2007 Robert E. Harkavy

Page 7 100 Bases during the Cold War Table 2 continued Host nation and base Description Sri Lanka Colombo Alleged use by U.S. for R&R. Trincomalee Port calls. Djibouti Port calls by U.S. Indian Ocean task force; refueling and reprovisioning, no shore leave; U.S. leases fuel storage for own use. Reunion Port calls. Somalia Berbera Some use by U.S. Indian Ocean task force; possible storage of matériel for rapid deployment force. Mogadiscio U.S. improved facilities, port visits; possible storage of equipment and supplies for Central Command. Oman Muscat (Mina Qaboos) Restricted USN use by Indian Ocean task force; contingent use for Central Command in Persian Gulf crisis. Mina Raysutt Restricted U.S. use. Masirah Port calls. Bahrain Al Jufair U.S. took over British facilities in 1949; now homeport for “Mideast Force” of four destroyers, communications, storage, barracks, berth, hangars, co-use of adjacent airfield; resupply of Indian Ocean task force; low-key use because of political problems; quiet access for greatly expanded U.S. presence in 1987. Kenya Mombasa U.S. port visits; possible pre-positioning of matériel for use in Southwest Asia. Diego Garcia U.S. naval support facilities; berths Central Command’s matériel storage ships; lagoon dredged to create sufficient anchorage for a carrier battle group. Mauritius Rumored USN port visits; R&R and reprovisioning. Azores (Portugal) Ponta Delgada Fuel storage; breakwaters; frequent visits by NATO warships. Spain Rota Major naval base; also airfield and communications station; major repair capabilities; can berth aircraft carriers; former Polaris SSBN base; fuel depot; weather station; naval hospital. İ 2007 Robert E. Harkavy

Page 8 Bases during the Cold War 101 Table 2 continued Host nation and base Description Italy Naples Major support complex for U.S. 6th feet; HQ for attack submarines; homeport for destroyer tender, communications centre. Gaeta Main base; homeport for flagship of U.S. 6th fleet, refueling facilities. Greece Souda Bay NATO naval base; anchorage large enough for entire 6th fleet; extensive underground fuel and munitions storage. Athens//Piraeus U.S. use of commercial port facilities increasingly in jeopardy during Papandreou regime. Turkey Istanbul USN port visits. Izmir USN port visits. Antalya USN port visits. Portugal Lisbon USN port visits. Israel Haifa Periodic USN port visits. Tunisia Tunis USN port visits. Egypt Alexandria USN port visits, periodically. Sudan Port Sudan USN port visits. Cuba Guantanamo Bay USN port visits, training and exercises; naval air station, drydock, sheltered anchorage; naval hospital; in reality mostly a political bargaining chip. Panama Rodman Naval Station Fleet support, logistics, small craft training facility. Balboa Naval ship repair facility. For submarines United Kingdom Holy Loch SSBN forward base, homeport for ten Poseidon’s; submarine tender permanently berthed; large floating drydock. Japan Yokosuka, Sasebo (See previous mention.) Italy La Maddalena Homeport for submarine tender; base for patrols by SSNs in Mediterranean. Source: Compiled from SIPRI data and the many references cited for this chapter. İ 2007 Robert E. Harkavy

Page 9 Sembawang facility, which provided the U.S. Navy with extensive drydocking capability as well as a good location in relation to possible flashpoints both in East Asia and the Persian Gulf area. Elsewhere in Asia, during the bulk of the Cold War, the U.S. Navy made use of Australian ports on both coasts, visited ports in Indonesia and (earlier on) Taiwan, and made extensive use of the Thai port at Sattahip during the Vietnam War, during which the South Vietnamese port of DaNang was frequently used. The accompanying table details the USN’s basing net as of the mid-to-late 1980s. U.S. Air Force bases12 In relation to possible theaters of combat, throughout the bulk of the Cold War, the U.S. Air Force (USAF) made extensive use of a network of permanent and standby bases, most of them in Western Europe and the Far East in proximity to major ground force deployments. In Germany, that long involved the major air bases at Bitburg, Ramstein, Spangdahlem, Zweibrucken, Sembach et al., at which were deployed the USAFs frontline fighters, earlier on the F-4s, later on the F-15s and F-16s. Additionally in Europe there were the major air bases at Keflavik in Iceland, Sosterberg in the Netherlands, Aviano in Italy, Torrejon, Zaragoza and Moron in Spain, and Incir- lik in Turkey. Additionally, the USAF had deployed during the latter part of the Cold War some 290 combat aircraft in the U.K. In addition to the 150 F111E/Fs deployed as a nuclear strike force, that involved EF-111 Raven electronic warfare aircraft, A-10s, RF-46 reconnaissance aircraft at several sites: Laken- heath, Alconbury, Upper Heyford, Woodbridge and Bentwaters. Finally, tactical aircraft were routinely rotated forward to bases in Italy and Turkey, while a host of other bases in NATO were designated as co-located host bases available for crises, coercive diplomacy or outright war – there were several of these in eastern Turkey in proximity to Middle Eastern danger zones. In Asia, the USAF had forward fighter bases in Japan, the Philippines and South Korea. F-15s, F-16s and RF-4Cs were based at main bases at Yokota and Misawa, Marine Corps Harriers at Iwakuni. There were F-4 E/G deployments at Clark Air Force Base in the Philippines. In South Korea, F-4Es, RF-43s, F-16s and A-10s were stationed at the major air bases at Osan, Kunsan and Taegu. In relation to conventional as well as nuclear scenarios, the USAF permanently stationed KC-135 A/Q and KC-10A tankers in a number of places to deal with interventions or arms resupply operations: in the U.K. (Mildenhall, Fairford), Spain (Zaragoza), the Philippines (Clark AFB), Okinawa (Kadena), Diego Garcia, Iceland (Keflavik) and Guam. Thule in Greenland and Goose Bay and Harmon in Canada were used for similar purposes. Still earlier, up to 1963, SAC’s KC-97s were based in Canada at Namao, Churchill, Cold Lake and Frobischer – some of these later were designated for dispersal and refueling operations. The USAF has also made extensive use of overseas bases for transport aircraft, often times ad hoc in relation either to routine staging of personnel or matériel or to crises. 102 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 10 Most of the primary U.S. aircraft for long-range lift operations – C-141s, C- 5A/B, more latterly C-17s – have been based in the U.S., though they way be dispersed for contingencies. But the USAF also long had numerous squadrons of C-130s designated for tactical purposes (i.e., with shorter ranges and able to utilize shorter runways). Some were based overseas at Rhein-Main (West Germany), Howard AFB (Panama), Mildenhall (U.K.), Clark AFB (the Philip- pines), Yokota (Japan) and Kadena (Okinawa). Sometimes these aircraft may have used overseas facilities for tactical purposes – the Iran hostage raid was conducted by C-130s based in Egypt. In 1973, U.S. transport aircraft engaged in the arms resupply effort on behalf of Israel made critical use of the Lajes air base in the Portuguese-owned Azores Islands. In 1990–1991, USAF transport aircraft had extensive access to airfields along routes to the Middle East both via the Atlantic and Europe, and via the Pacific and Southeast Asia, where that access became domestic political issues in Thailand and India. In the post-9/11 conflict in Afghanistan, U.S. transport aircraft were staged via Eastern Europe and utilized bases in the former Soviet Central Asian states of Uzbekistan, Tadzhikistan and Kyrghizstan. We shall return to the link here with aircraft overflight rights. During the latter part of the Cold War, the U.S. utilized some overseas bases for its Airborne Warning and Control Systems (AWACS), a battle management aircraft often used for coercive diplomacy as a modern form of “gunboat diplomacy.” Earlier in the Cold War, indeed, it had deployed overseas other electronic warfare aircraft (ECM and ECCM), for instance, RC-135 CONVENT/ ELINT aircraft for monitoring radars and telemetry had been based at RAF Mildenhall, Hellenikon in Greece and Kadena in Okinawa. The EC-135, a modi- fied KC-135 Stratotanker, was used as a radio and telemetry intercept aircraft, and four of these “Silk Purse” planes were based at Mildenhall. In the latter part of the Cold War, USAF EF-111 Ravens were based in Spain and Turkey (Diyarbakir). The AWACS themselves were home-based well back from potential lines of confrontation, so as to mitigate the chances of preemptive strike. Some were based at Keflavik in Iceland in relation to the crucial G-I-UK Gap, others at Geilenkirchen in Germany, and several in Japan. Also used as forward AWACS operating bases were Trapani in Italy, Konya in Turkey and Oerland in Norway. A NATO command center at Maisieres in Belgium controlled NATO AWACS and British Nimrod early warning systems. Still other AWACS were sold to Saudi Arabia but operated there with U.S. crews. Aircraft overflights13 One of the less visible forms of foreign military presence, also one which involves movable and transitory presences, is that of aircraft overflight privileges. It is a form of external access. This occasionally crucial matter of aircraft overflight privileges involves a complex range of practices and traditions, some of which were, in an overall Bases during the Cold War 103 İ 2007 Robert E. Harkavy

Page 11 sense, altered by time in an era of increasingly “total” warfare, diplomacy and ideological rivalries. In parallel with – and closely bound up with – what has been wrought by nations’ increasing insistence upon extension of sovereign control further outward from coastlines (now more or less institutionalized by 200-mile Exclusive Economic Zones – EEZs), the trend here during the Cold War was towards tightened restrictions on overflights. In the past – and in some cases continuing to the present – some nations have allowed others more or less full, unhindered and continuous overflight rights (perhaps involving only pro forma short-term notices, i.e., filing of flight plans). In other cases, however, where political relations are weaker or not based on alliances, ad hoc, formal applications for permission to overfly must be made well ahead of time, which may or may not be granted depending upon the purpose and situation, be it routine or crisis. It is to be stressed that the day-to-day diplomacy of overflight rights is a very closed and obscure matter, albeit of often crucial importance. We have little data – the subject periodically emerges to prominence during crises such as the 1986 U.S. raid on Libya. Of course, it is precisely when urgent military operations are involved that the subject acquires the most importance. Nowadays, of course, well past the introduction of radar and its widespread global distribution, few overflights can be made on a covert basis, as was common before World War II, when detection depended primarily on visual observation from the ground. Not only “host nation” radar, but now also the superpowers’ satellite reconnaissance makes such “covert” activities almost impossible, particularly if a small nation has access to information from one of the superpowers, be it on a regular or ad hoc basis. This in turn may have had important ramifications for intra-Third World rivalries, specifically, regarding the balance of diplomatic leverage involved. Nations inclined, for instance, to provide overflight rights in connection with a U.S. airlift to Israel knew that Soviet satellite reconnaissance would provide information about that to Arab governments. That was a powerful deterrent. Some overflights were made without permission (as with the respective use by the United States and the USSR of U-2 and MiG-25 reconnaissance flights), overtly or with a tacit or resigned wink by the overflown nation. Often a nation whose airspace is violated will not openly complain for fear of international or domestic embarrassment over its impotence, or untoward diplomatic repercus- sions with a strong power. Hence, the USSR is thought to have overflown Egypt and Sudan, among others, without permission in supplying arms to Ethiopia during the 1977–1978 Horn War, earlier, its MiG-25 and Tu-95 reconnaissance aircraft apparently flew with impunity over Iran’s airspace. The United States is thought to have threatened overflights in some places for future arms resupply of Israel, if it should be utterly necessary. More recently during the Gulf War, and the Afghan and Iraq wars, this became a big issue in numerous places. During the Gulf War, the U.S. and its allies were allowed overhead access almost everywhere, including ex-Warsaw Pact states in Eastern Europe. In the Afghan war, the U.S. had good overhead 104 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 12 access all over Europe, in the Caucasus and Central Asia and in and around the Persian Gulf excepting, of course, Iraq and Iran. Pakistan, politically cross- pressured, allowed U.S. overflights by bombers coming from Diego Garcia and from aircraft carriers stationed in the Arabian Sea. During the Iraq war, however, the U.S. did have some problems with Switzerland, Syria and Iran. Nuclear-related bases14 The utilization of overseas facilities in connection with the superpowers’ nuclear competition began immediately after the close of World War II. During the late 1940s, prior to the Soviet Union’s development of a deployed nuclear military capability, the U.S. forward-based some nuclear-armed B-29 aircraft in the U.K. in an effort to deter feared Soviet advances in Europe. By 1950, B-29s were based at Brize Norton, Upper Heyford, Mildenhall, Lakenheath, Fairford, Chelveston and Sculthorpe. There were also reserve B-29 bases at other British bases. There were related depots at Burtonwood and Alconbury, and also related LORAN navigational facilities at Angle, Pembrokeshire and in the Hebrides. That provided a clear first-strike deterrent capability for the U.S. well into the 1950s. During the 1950s, the B-47 bombers became the backbone of SAC, and while their effective ranges were greatly extended by the aerial-refueling techniques then emerging, the U.S. then determined on forward deployment to enhance its chances for penetration and to lessen its vulnerability to a Soviet first-strike. This so-called “reflex force” rotated between U.S. home bases and those in the U.K. (Fairford, Upper Heyford), Morocco (Sidi Slimane, Benguerir, Ben Slimane, Nouasseur), Spain (Torrejon, Zaragoza, Moron de la Frontera), Green- land (Thule) and Goose Bay, Labrador. (F-84 fighters used as bomber escorts were also based at Nouasseur until U.S. access to Morocco was lost in 1963.) Related U.S. tankers (then mostly KC-97s) were based primarily at Thule, Greenland and Goose Bay, Labrador, and also at several other Canadian bases: Namao, Churchill, Harmon, Cold Lake and Frobischer. Though the subse- quently deployed B-52s which began entering inventories in 1955 did not require forward main basing, they too utilized trans-Arctic tanker facilities (including one at Sondestrom in Greenland) as well as contingency recovery bases in Spain and elsewhere. The Soviet Union did not utilize forward strategic-bomber facilities during this period. Indeed, early Soviet bombers, such as the 4500-km range Tu-4, could only reach the U.S. Pacific northwest from Siberia and, even then, by con- ceding several hours’ warning time because of U.S. radar coverage in Alaska. During this period, and for a long time thereafter, the U.S. also relied on foreign access for strategic defense, primarily in Canada, Greenland and Iceland – that involved the DEW Line, Mid-Canada and Pinetree Line strings of electronic listening posts, all under the U.S. Air Defense Command, which worked closely with SAC. In addition, some U.S. interceptor aircraft were deployed at Canadian bases such as Goose Bay, and at bases in Greenland and Iceland, for Bases during the Cold War 105 İ 2007 Robert E. Harkavy

Page 13 perimeter early defense, well forward of the large-scale interceptor deployments around major U.S. urban areas. By the late-1950s, Soviet missile developments had rendered somewhat obsolete the three-layered radar early-warning system across the Canadian Arctic, which had been constructed to provide several hours’ warning of bomber attacks. To cope with the missile threat, the U.S. built, beginning around 1958, the Ballistic Missile Early Warning System (BMEWS), the three hinges of which were in Fairbanks (Alaska), Thule (Greenland) and Fylingdales Moor (Yorkshire, U.K.). In the mid- to late-1950s, the U.S. underwent its famed “missile-gap scare,” following the Soviet Union’s initial testing of IRBMs and ICBMs, and the launching of the first “Sputnik” satellite. Coming before the deployments of Atlas, Titan and Minuteman ICBMs and Polaris SLBMs, this created a perceived “window of vulnerability” which, in turn, impelled the short-term solu- tion of U.S. installation of IRBMs in Europe adjacent to the Soviet Union. Specifically, this involved emplacements in 1958 of 60 2400-km range Thor missiles in the U.K. at 20 bases, with headquarters at Great Driffield, North Luffenham, Hemswell and Feltwell, 30 Jupiter missiles in Italy (at Gioia del Colle) and 15 Jupiters in Turkey, installed in 1961 at Cigli Air Base (these were removed as part of the deal in which the USSR removed IRBMs and also IL-28 aircraft from Cuba, after the Cuban Missile Crisis of October 1962 – though the orders for their removal had apparently been given earlier). Late in 1960 the U.S. deployed its first Polaris submarines and then its long- range, counter-value Atlas and Titan ICBMs, thus quickly defusing the missile- gap scare – though the forward-based IRBMs were to remain in place for an additional two to three years. Complementary to BMEWS, the U.S. developed early warning satellites under the MIDAS satellite program. This involved combined use of infrared sensors and telephoto lenses for immediate detection of missile-launching tracks and transmission of this information to U.S. decision-makers. Launched by Atlas/Agena D missiles, advanced MIDAS satellites deployed in 1969 could be “parked” in synchronous orbits, allowing for continuous coverage of the western USSR and the China-Siberia region as well as areas where Soviet submarines lurked in firing positions. This involved the critical data down-link in Australia at Nurrungar, a related control facility in Guam and an underwater cable terminal near Vancouver in Canada. One other key element of the strategic deterrence system came to depend upon overseas access: long-distance and protracted deployment of the Polaris nuclear-submarine force. The Polaris submarines were initially deployed early in the Kennedy Administration. The proportion of that fleet which the U.S. was able to deploy at any given time was enhanced by replenishment and repair facilities at Holy Loch, Scotland; Rota, Spain and at Guam. Indeed, the asymme- tries which these facilities created vis-à-vis subsequent Soviet SSBN deployment allowed the U.S. to negotiate that part of the SALT I Treaty which gave the USSR a 62 to 44 advantage in strategic submarines, but which was claimed 106 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 14 to be counterbalanced by the efficiencies accruing to the U.S. from its overseas replenishment facilities. In the early 1960s, after the brief U.S. missile-gap scare, the Soviet Union underwent a scare period of its own, as several U.S. strategic programs were phased in. To compensate, Moscow gambled with the introduction in 1962 of some 40 MRBMs into Cuba at several installations, precipitating the Cuban Missile Crisis. (One recent report claims that these missiles were not accompan- ied by nuclear warheads.) The history of that crisis bears no repeating here, but it is worth noting that only by the early 1960s did the Cuban revolution avail the USSR of its first valuable overseas assets applicable to the strategic nuclear equation. Henceforth, Cuba would become a very valuable Soviet base, its proximity to the U.S. providing irreplaceable assets related to intelligence, communications, naval replenishment and so on, along with contingent bomber recovery bases in the event of a major war. During the early postwar period, the U.S. made use of numerous nuclear- related intelligence and communications facilities around the Eurasian periphery – directed against the USSR, China and North Korea – mostly in the SIGINT (ELINT and COMINT) categories. Earlier, both U-2 and other aircraft such as the RB-47 were flown from bases in Europe and Asia to “tickle” Soviet early- warning radars and, in the case of the U-2s, to test radars well inside the USSR which might be of different types than the peripheral early-warning systems. By so doing, U.S. planners might ascertain weaknesses and ranges, and scan patterns in the Soviet radar network which could be valuable for planning the penetration routes for a nuclear-bomber attack. These exercises in low-level brinkmanship – apparently involving some mock raids mounted by U.S. units in Turkey and elsewhere – resulted in some serious incidents in which U.S. ferret aircraft were shot down and their crews killed or captured. Some flights originating at Brize Norton in the U.K. apparently traversed the entire Soviet Arctic coastline, emerging at the Barents Sea. The area between the Caspian Sea and the Sea of Azov was also apparently a focal point of U.S. surveillance missions utilizing Turkish and Iranian airspace, some staged originally from West Germany and Cyprus. U.S. use of ground-based SIGINT stations, obviously crucial to various aspects of nuclear deterrence, dates well back into the postwar period. One source reported that this had earlier involved some 40 stations in at least 14 countries, ranging from small, mobile field units to sprawling complexes such as the Air Force Security Headquarters in West Germany. These were said to have involved some 30,000 personnel, with a minimum of 4000 radio-interception consoles operated in such varied locales as northern Japan, the Khyber Pass in Pakistan and an island in the Yellow Sea off the coast of Korea. Further, these COMINT land stations had to be supplemented by numerous airborne and seaborne radio-interception facilities, particularly after Soviet and aligned nations’ military forces switched to VHF radios during the 1950s, after which adequate coverage demanded getting closer to transmitters and overcom- ing terrain features such as mountains. At any time, several dozen airborne Bases during the Cold War 107 İ 2007 Robert E. Harkavy

Page 15 listening posts were said to have been in intermittent operation, flying out of such bases as Kimpo Airfield in Korea, Clark Air Force Base in the Philippines, and many others. Added to these were some 12 to 15 spy ships, such as the ill- fated Pueblo and Liberty, which also presumably required access to foreign ports for replenishment. Strategic nuclear forces: missile launchers and platforms15 Generally speaking, the late Cold War period saw the U.S. and the USSR left with only minimal dependence on foreign facilities – Soviet dependence was relatively less for their strategic nuclear launchers, involving the familiar triad of ICBMs, SLBMs and long-range bombers. Of course, this is not to ignore the fact, particularly as it pertained to U.S. forces, that some launchers designated as theater weapons could serve strategic purposes in that their warheads could have been delivered into the Soviet Union. All of the U.S. ICBMs (450 Minuteman IIs, 550 Minuteman IIIs and a few remnant Titan IIs which were phased out as the new MX-Peacekeeper was phased in) were housed in silos within the continental U.S. – they represented over 2100 accurate warheads. Similarly, the Soviet ICBM forces (448 SS-11s, 60 SS-13s, 150 SS-17s, 308 SS-18s, 360 SS-19s, 72 SS-25s), of some 1398 ICBMs with some 6354 warheads, were sited entirely within the USSR. The Soviet SSBN forces (983 SLBMs in 77 submarines, of which 944 SLBMs and 62 submarines were under the SALT Treaty; and 39 SLBMs on 15 submarines were outside it) were based entirely at Soviet homeland bases, in the Kola Peninsula area at Polyarny and Severomorsk, and at Petropavlovsk and Vladivostok in the Far East. No foreign bases were used for refueling, maintenance or crew changes. By contrast to the U.S. only some 15–20 Soviet submarines were normally away from their bases; perhaps 10–12 on station at any given time. The U.S. SSBN force, which earlier made extensive use of facilities at Holy Loch (Scotland), Rota (Spain) and at U.S.-owned Guam, utilized only the first- named of these, with its other three main bases in the continental U.S. at Kings Bay (Georgia), Bangor (Washington) and Charleston (South Carolina). The development of the longer-range Poseidon and Trident missiles (with ranges of 4000 miles) allowed for utilization of firing stations nearer U.S. bases and further from the USSR, hence reducing requirements for firing stations in the western Pacific (Guam) and in the Mediterranean (Rota). Little was publicly known about where these submarines patrolled, but they are thought to have transited to firing stations in the Arctic, North Atlantic and North Pacific oceans and in the Mediterranean Sea, with about 30 percent of the force on station on day-to-day alert, and a roughly equal proportion in transit or on training missions. The use of Holy Loch as a forward base (for submarines homeported at Groton, Connecticut) allowed more to be on station than otherwise would have been feasible, and it is also to be noted that 400 Poseidon warheads deployed on submarines operating out of Holy Loch were designated for 108 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 16 NATO targeting, presumably either for theater targeting or for battlefield use along the Forward Line of Troops (FLOT) in central Germany. At the strategic level, the U.S. deployed some 260 long-range bombers (some 20 B-1Bs and some 240 B-52G/Hs) and some 55 medium-range FB-111A bombers. Most of the B-52s carried up to 20 ALCMs; others carried a similar number of short-range attack missiles (SRAMs); and some, armed with Harpoon missiles, were operational in a conventional mode. This bomber force, which carried over 5000 nuclear warheads, was main- based at 19 air bases in 13 states and at Andersen AFB in Guam, but numerous dispersal (pre-attack or crisis) and recovery (post-attack) bases, or forward operating bases were involved. Some of these were outside the U.S., for instance in Canada at Cold Lake (Alberta), Goose Bay (Labrador), Namao (Alberta) and Whitehorse (Yukon); in Greenland at Sondestrom; in the U.K. at Brize Norton, Marham and Fairford; in Spain at Moron and Zaragoza. But numerous others, on the basis of ad hoc contingency planning could have been used in connection with numerous tanker bases – or their dispersal bases – at some of the above- mentioned Canadian bases at Namao and Goose Bay; at Mildenhall and Fairford in the U.K.; Zaragoza in Spain; Clark Air Force Base in the Philippines; Kadena in Okinawa; Diego Garcia; Keflavik (Iceland); Guam; Thule (Greenland) and so on. The U.S. forward-based some 156 of the nuclear-armed F-111E/Fs in the U.K., which were actually tactical fighter bombers designated for missions in Central Europe. These were stationed at Upper Heyford and Lakenheath, which hosted some 72 and 84 aircraft respectively, involving in each case the storage of some 300 nuclear bombs. By comparison, the Soviet Union made little use of external facilities for its long-range strategic bomber force, which consisted of some 160 aircraft: 140 Tu-95 Bear A/B/C/G/Hs armed either with ALCMs or air-to-surface missiles, and 20 Mya-4 Bisons (these were superseded by the Blackjack strategic bomber). The same was true of the some 230 Backfire, Badger and Blinder Bombers – some of the former were assigned to Soviet long-range aviation. These strategic bombers utilized five northern staging and dispersal bases within the USSR from which attacks could be mounted on the U.S. Regarding forward emplacement of U.S. nuclear weapons, primarily in Europe, there were several distinct types of circumstance in which they were based, involving the issues of dual control or the two-key system, and the politics of nuclear basing on other nations’ soil. First, there was the actual forward basing of U.S. aircraft, missiles or artillery – fully manned by U.S. personnel involving organic U.S. military units but based on foreign soil and in circumstances in which ad hoc permission for actual combat would have been required. Next, there were the numerous circumstances in which other nations – West Germany, Belgium, the Netherlands, Italy, Turkey, Greece etc. – deployed nuclear weapons on aircraft, missiles, howitzers and so on, but where the stric- tures of the nuclear non-proliferation regime dictated a U.S. custodial presence and the use of a two-key system, further implying a veto on actual use by either Bases during the Cold War 109 İ 2007 Robert E. Harkavy

Page 17 the U.S. or its nuclear partner. Some of these situations involved Allied nuclear systems based within the host nations’ borders as, for example, in Turkey or Germany, as with Belgium or Dutch nuclear-armed artillery. Some nations, such as Norway and Denmark, did not allow nuclear weapons or foreign military bases on their soil during peacetime. Others, such as Iceland, did not allow nuclear weapons on their soil, but had contingency provisions for forward- basing of nuclear weapons for “wartime ASW operations.” The primary land-based U.S. theater weapons were the Pershing II and ground-launched cruise missiles later eliminated in connection with the INF Treaty. These were, of course, all foreign-based, involving the 108 Pershing (plus 12 spares) at several sites in West Germany (Heilbronn, Waldeide, Neckars-Ulm, Schwabisch-Gmund), and the already deployed and planned GLCMs there (at Woescheim), and in the U.K. (Greenham Common, Molesworth), Belgium (Florennes), the Netherlands (Woendsrecht) and Sicily (Comiso). Additionally, launcher repair facilities were identified at EMC Hansen, Frankfurt (West Germany) and at SABCA, Gossens, Belgium, and a missile storage site at Weilerbach in West Germany. Originally, this called for 464 GLCMs overall, based on 116 launchers, with perhaps nearly double that number of warheads, though by the time of the 1987 INF Treaty only some 309 missiles and 109 launchers had been deployed. The Pershing missiles with ranges capable of reaching well into the USSR, indeed, as far as Moscow, may have been perceived as having “strategic” implications, albeit based outside the U.S. Additionally, there were the remnant German Pershings under U.S. custo- dian control (with two warheads per launcher), with shorter reaches well short of being able to target the USSR proper (these too were eliminated under the INF accord). Otherwise, the U.S. based a considerable number of nuclear-capable aircraft forward in Western Europe, that is, those manned directly by USAF personnel, which involved the forward deployment of some 1700 nuclear bombs. These were, obviously, usable for tactical purposes along or directly behind the FLOT, but – particularly with the aid of aerial refueling – also capable of deep interdiction missions throughout Eastern Europe and well into the Soviet Union proper. Some 72 F-16 fighter-attack aircraft and some 140 nuclear bombs were stored at both Hahn and Ramstein air bases; at Spangdahlem, a similar number of F-16s with some 140 weapons were stored. At Aviano in Italy, some 200 weapons were said to have been stored in connection with nuclear-capable F-16s, previously rotated forward from Spain but later based in Italy. In Turkey, the base at Incirlik supported some 36 nuclear capable F-4s or F-16s which could be loaded with the weapons on quick-reaction alert after being rotated forward, while another base at Cigli, Izmir acted as a dispersal base for nuclear-capable aircraft. Of course, land-based aircraft were not the only forward-based U.S. nuclear- armed aircraft. There were also U.S. aircraft carriers in the Mediterranean Sea and the Atlantic Ocean. Typically, there were two carriers on station in the Mediterranean – utilizing ports such as Naples, Souda Bay, Rota – which could launch A-6E, A-7 and F/A-18 aircraft carrying nuclear weapons, with 110 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 18 respective ranges of 3200, 2800 and 1000km. With the capability to strike the Soviet homeland if within range, each carrier deployed over 100 nuclear bombs. As well as nuclear-armed aircraft, there were the numerous U.S. and other NATO nations’ ASW aircraft, involving the U.S. P-3C Orions and the British Nimrods. In a conflict, these would have been vital to NATO efforts at securing North Atlantic SLOCs against interdiction by Soviet attack submarines. U.S. and Dutch P-3Cs and British Nimrods were based at or staged through British bases at St. Mawgan and Machrihanish (the latter in the Strathclyde area of Scotland) – some 63 U.S. nuclear depth bombs were stored at each base. Addi- tional U.S.-manned and operated P-3Cs with forward-stored nuclear depth charges were based at Keflavik, Iceland (48 bombs), Sigonella on Sicily (63 bombs) and Rota, Spain (32 bombs). Other U.S. P-3Cs were rotated through Andaya and Bodo in Norway, Souda Bay on Crete and Montigo, Portugal, while some 32 nuclear depth bombs were stored at Lajes in the Azores for wartime operations. Soviet forward-based nuclear-capable aircraft16 According to the IISS, the USSR had several types of land-based strike aircraft capable of being configured with nuclear weapons, each of which could carry two nuclear bombs. These were the Su-7 Fitter A (80), the MiG-21 Fishbed L (135), the MiG-27 Flogger D/J (810), the Su-17 Fitter D/H (900) and the Su-24 Fencer (700) – a total of 2625 such aircraft. Of the 700 Fencers, 450 fell under the control of Strategic Aviation. Over 1000 Soviet tactical fighter aircraft were forward-based at a large number of facilities in the GDR, Czechoslovakia, Poland and Hungary. Nuclear-capable Su-24 Fencers were stationed at Szprotawa AB and Zagan AB in Poland, at Debrecen in Hungary and at Brand-Briesen AB in the GDR, in each case involving associated nuclear storage sites (the latter base was said to have been converted from Su-7s in 1982). In the GDR, nuclear-capable MiG-27 Flogger D/J regiments and nuclear-bomb storage sites were at Finsterwalde AB and Mirow-Rechlin Larz AB; Su-17 Fitter D regiments were located at Grossen- hain AB, Neuruppin AB and Templin-Gross Dolln AB. Several of the non-Soviet WTO nations also forward deployed nuclear- capable aircraft, perhaps involving something like a two-key system, under Soviet custodianship. Czechoslovakia (50) and Poland (40) had Su-7 Fitters; Poland (40) Su-20 Fitter Cs and Bulgaria (45), Czechoslovakia (40) and the GDR (24) MiG-23 Flogger F/As. The IISS noted that the total actually available as nuclear-strike aircraft may have been lower than the figures shown. It is possible that the Soviet air force could itself have used these aircraft in a nuclear mode if war erupted. Outside the WTO area, the USSR also deployed some 16 Tu-16 Badger bombers in Vietnam, which were capable of nuclear missions, perhaps against China or the U.S. basing structure in the Philippines. Bases during the Cold War 111 İ 2007 Robert E. Harkavy

Page 19 The USSR also utilized several external bases for nuclear-capable ASW aircraft. Bear-F aircraft, for instance, utilized Cuban bases at San Antonio de los Banos and Havana’s Jose Marti airport. Nuclear-capable IL-38 May ASW aircraft regularly operated out of Aden and Al Anad in South Yemen, Asmara in Ethiopia, Okba ben Nafi in Libya and Tiyas in Syria. Overseas “technical” facilities: intelligence, space and communications17 In a relative sense, and increasingly, the U.S.-Soviet global competition for basing access was centered on a variety of what, for want of a better term might be characterized as “technical” facilities, that is, those outside the traditional categories of air and naval bases and land-army garrisons and encampments. Most of these facilities may be subsumed under the broad headings of communications, intelligence and space-related activities. They include such disparate functions as satellite tracking, command and control; signals interception of rivals’ communications, radar signals, missile telemetry, and so on; underwater detection of submarines, accurate positioning of missile-firing submarines; space- based ocean surveillance; nuclear-explosion detection; and a bewildering variety of functionally specific communications systems running along the entire spectrum from extra-low to ultra-high frequencies. The increasing importance of all these systems paralleled the extension of contemporary military activity to an increasingly integrated, three-dimensional game involving outer space, land and sea surfaces and the global undersea realms – submarines communicate via satellite with land-based headquarters; satellites and land-based SIGINT stations locate surface fleets by intercepting their radar emissions; satellite early warning is transmitted, variously, by ground terminals, underwater cables, via other satellites and so on. This advent of the importance of technical facilities gradually picked up speed before and after World War I. Before World War I, Britain developed an elaborate and unrivalled global network of underseas telegraph cables – by far the most important early precursor to modern “tech” facilities. At that time all of the key British overseas possessions were linked together by that network, providing advantages in early warning and command and control of naval forces, and in the ability to control and influence news broadcasting, an earlier form of “public diplomacy” used as an instrument of competition. Britain was not, of course, alone in constructing an elaborate underseas cable network, nor in its efforts at securing the required access points. According to one source, the United States in 1898 annexed Guam and Midway for the specific purpose of providing cable stations en route to the Philippines, decades before those islands would become important U.S. air bases.18 France and Germany also made efforts towards building global systems, but before World War I the latter came to rely more on wireless systems, despite the vulnerability of their communications to interception if decoded. By 1914 the Germans had wireless stations in Togoland, southwest Africa, Tanganyika, Kiung-chow, Yap, 112 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 20 Rabaul, Nauru and Samoa, to abet what they knew would be a very vulnerable cable network should war break out.19 At the outset of World War I, Britain was able rapidly to sever all of the German underseas cables, while losing only a couple of its own cable stations temporarily, at Cocos and Fanning Islands.20 The redundancy built into Britain’s global system, abetted in 1914 by that of France, provided a strong strategic advantage throughout World War I, still one more result of “invisible” global naval superiority. Later, in 1919, the German cables were divided as war spoils among the allies as part of the Versailles settlement. Beyond World War I and up to World War II, increased importance was attached to technical facilities. Long-range radio communications developed apace and with them came the development by major powers of networks of transmitters, receivers, relay stations and so on. With that came the early development of radio interception facilities – before World War II, the U.S. worried about German acquisition of interception facilities in the politically unstable Caribbean area, for instance, in Haiti or Colombia.21 The U.S. breaking of the Japanese code before World War II (which provided what should have been a decisive, timely early warning about Pearl Harbor) serves as an additional precursor of subsequent intelligence activities, now far more institutionalized in organizations such as the U.S. National Security Agency (NSA). Japan, mean- while, in violation of the Washington Naval Agreement, built covert communications facilities on some of its Central Pacific, League of Nations Mandate islands, providing an early problem of arms control verification.22 On the eve of World War II, the U.S. was beginning to install early warning radars in some of the bases acquired from the U.K. as part of the Lend-Lease Agreement.23 Gradu- ally, one could see movement towards the central importance of C3I today – towards quick if not real-time communications and early warning, and the diminishing of human intelligence (HUMINT) as the core of intelligence collection. U.S. communications24 The U.S. used a variety of communications systems and modes, stretching across the frequency spectrum from extremely-low frequency (ELF) to super- high frequency (SHF); these variously utilized satellites, ground terminals, shipboard and submarine terminals and so on.25 These various frequencies involved trade-offs among a number of variables related to rate of data transmission, vulnerability to jamming, size of transmitter, distance capability, etc., and tended to be broadly specific for certain functions, that is, land-based tactical communications, those with submerged submarines, etc. The utilization of various communications modes changed constantly in response to new technological developments; correspondingly, the requirements for FMP access also changed. As ably outlined by Arkin and Fieldhouse, “In the field of electronic communications, each medium and frequency has advantages and disadvantages.” Bases during the Cold War 113 İ 2007 Robert E. Harkavy

Page 21 Varying by wavelength and hence frequency, a number of different paths for the four lowest frequency bands travel what are called “groundwaves” and they are useful for communications with ships far from land. Extremely-low frequency (ELF) waves (below 300 hertz) can penetrate water to hundreds of feet (perfect for submarine communications), while extremely-high frequencies (EHF) (above 30 gigahertz) have difficulty penetrating even a heavy rainstorm.26 Other than the medium (i.e., water or air), still other conditions determine the most suitable frequency to use. Size of transmitter is important – “for frequencies below the HF band, antennas are too large for ships or aircraft” – the ELF antennas or transmitters considered for emplacement in the northern U.S. were many miles long.27 Then, the higher the frequency the higher the data rate. The amount of power required also varies with frequency. For these reasons ELF is not suitable for large-volume commercial communications. There is another variable – reliability – in connection with possible interference, jamming or fasing.28 High frequency (HF) is widely used by the military – it is long-range, cheap, low-power, small and portable, but requires constant adjustments in specific frequencies to deal with atmospheric conditions,29 that is, shifting in the ionos- pheric layers and natural events such as solar flares. This can vary by day, season, location and so on, so that frequencies must be chosen to best suit the prevailing conditions. For these reasons and others – reliability, “crowding” of the frequency spectrum and the advent of computers and satellites (which operate at higher frequencies and data rates) – recent decades have seen a shift away from HF for military purposes. But, the vulnerability of satellites has led to renewed interest, particularly in connection with new technology, that is, “sounders,” solid-state transmitters and microprocessors which can allow for shifting frequencies in response to environmental changes. That interest is underscored by the fact that HF uses the ionosphere for transmission, a medium difficult to permanently interrupt. Finally, in the context of military anxieties about nuclear “black-out” caused by nuclear blasts during war, mobile or proliferated HF systems are considered one of the more robust types of communications.30 Among the HF systems used during the Cold War were SAC’s Giant Talk/Scope Signal III for strategic bombers, the air/ground/air Global Command and Control System network, the Mystic Star Presidential/VIP network and the Defense Communications System (DCS) “entry sites.” Very-low frequency (VLF) and low frequency (LF) are also considered relatively reliable in a nuclear environment and can penetrate sea water as well.31 Hence, for the U.S., a key system was GWEN (Ground Wave Emergency System), a grid of unmanned relay stations with LF transmitters and receivers hardened to withstand electromagnetic pulse (EMP).32 When fully proliferated, it used a system of “automatic diverse routing” so as to maximize imperviousness to interference even by a full-scale nuclear attack. Yet another robust system was “meteor burst communications,” using billions of ionized meteor trails to reflect very-high frequency (VHF) signals.33 This system would apparently benefit from the increased ionization caused by nuclear war. 114 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 22 As an example of what was involved, we may look at some of the earlier global networks utilized by the U.S. for communicating with underwater submarines. Throughout most of the 1970s, the U.S. had eight Omega VLF facilities located overseas. Some of these were phased out beginning in the late 1970s and some were retained, despite their obsolescence, as backup systems. They were located at: Reunion (Mafate), operated by the French Navy; Trinidad and Tobago; Liberia (Paynesword); Australia (Woodside); Argentina (Golfo Nuevo, Trelev): Japan (Tsushima Island); and Norway (Bratland).34 Then there was the far larger global network of LORAN-D/C radio- navigation systems, which were also utilized in connection with aircraft navigation. And as a sub-set, this further involved the Clarinet Pilgrim system in the Pacific, a shore-to-submarine network (four sites in Japan and one on Yap Island) that worked by superimposing data on the waves transmitted by LORAN-C. Some of these were operated by the U.S. Coast Guard, to a degree reflective of the mixed civilian and military navigation aid functions of the LORAN network (still earlier there were some systems designated as LORAN- A). And, as in the case of Omega, some were jointly operated with host-nation personnel. Among the numerous LORAN-C/D transmitters and monitoring stations overseas (there were many others in the U.S., including Alaska and Hawaii) were those shown in the table. In the latter part of the Cold War, as a replacement for Omega and LORAN- C, the U.S. installed its new satellite-based NAVSTAR global positioning system for submarines, which among other things, apparently involved the capability to provide SLBMs with corrective guidance after they surfaced. NAVSTAR’s control segment consisted of five monitor stations to “track passively all satellites in view and accumulate ranging data from the navigation signals.” That information was transmitted to the NAVSTAR Master Control Station at Vandenberg Air Force Base, California. The other stations used for tracking, telemetry, control and passive monitoring were at Ascension, Diego Garcia, Kwajalein and Guam, notably all islands controlled by the U.S. or the U.K. Whereas most communications with submarines are conducted along the LF spectrum, combatants will usually use HF (the corresponding intelligence interception facilities are HF direction finders, that is HF/DF). The U.S. had a considerable number of naval HF transmitters and receivers scattered about the globe, most of them near major naval facilities or near bodies of water heavily traversed by U.S. fleet units. They included those shown in Table 4. Regarding SAC “fail-safe” systems, involving “positive control,” meaning the bombers went ahead in a crisis only if given “executive instructions,” there was a global system of HF communications called “Giant Talk/Scope Signal III,” 14 stations giving flexible approach routes towards the USSR by B-52s or other aircraft with standoff ALCMs. These are shown in Table 5. The Air Force also had its AFSATCOM UHF network devoted particularly to strategic nuclear-related purposes. That utilized several foreign facilities, at Bases during the Cold War 115 İ 2007 Robert E. Harkavy

Page 23 116 Bases during the Cold War Table 3 Location of known Loran-C/D transmitters and monitoring stations overseas Country Location Bermuda Witney’s Bay Canada Cape Race (Newfoundland), Fox Harbor (Labrador), Montagu (Prince Edward Island), Port Handy (British Columbia), St. Anthony (New Brunswick), Sandspit (British Columbia), Williams Lake (British Columbia) Denmark Ejde (Faeroe Islands) Greenland Angissoq Iceland Keflavik, Sandur Italy Crotone, Lampedusa, Sellia Marina Japan Gesaski (Okinawa), Iwo Jima, Marcus Island, Tokachibuto (Hokkaido), Yokota AFB Johnson Atoll (U.S. owned) Norway Jan Mayen Island South Korea Changsan Spain Estartit Turkey Kargabarun U.K. Sullum Voe (Shetland Islands) FR Germany Sylt Micronesia Yap Island Guam (U.S. owned) Anderson AFB Source: SIPRI data, Harkavy, Bases Abroad, p. 161. Table 4 U.S. overseas HF receivers and transmitters Country Location Bermuda South Hampton Diego Garcia Greece Nea Makri and Kato Souli Guam Barrigada and Finnegayan Iceland Grindavik and Sandgerdhi Italy Naples (a master station) and Licola Japan Iruma, Kamiseya, and Totsuka Panama Summit Philippines Capas Tarlac and St. Miguel Portugal (Azores) Cinco Pincos (Terceira) and Vila Nova Puerto Rico Sebana Seca, Isabella, and Aguada Spain Guardemar del Segura and Rota U.K. Edzell and Thurso Source: SIPRI Data. İ 2007 Robert E. Harkavy

Page 24 Landstuhl in Germany, Clark AFB in the Philippines, on Guam and at Bagnoli in Italy and Diego Garcia. The Navajo FLTSATCOM system consisted of seven satellites parked in geosynchronous orbit all around the equator and provided worldwide coverage, except in the polar regions. These provided mostly for the U.S. Navy, communications by digitalized voice, teleprinter and other techniques, and operated at UHF. These satellites were also important hosts for AFSATCOM transponders. Indeed, according to one report, each host for AFSATCOM satellites had 23 channels, ten of which were allotted to the Navy for command of its air, ground and sea force, 12 to AFSATCOM for nuclear- related communications, and one reserved for the National Command Authorities.35 One particularly important function for FLTSATCOM – along with the DSC’s satellites – was the relaying of data from SOSUS and SURTASS (surveillance towed array) hydrophone systems to the Central Shore Station or Acoustic Research Center at Moffett Field, California, “where it is integrated with data from other sources and processed by the ILLIAC 4 computer complex to provide a real-time submarine monitoring capability.”36 There is also real- time transmission of data and displays from ocean-surveillance satellites provided to U.S. surface and submarine fleets. FLTSATCOM utilized control or receiver sites. There were several in the U.S. at Norfolk, Wahiawa (Hawaii), Stockton, California and another at Finnegayan in Guam. Overseas, there were additional stations at Bagnoli, Italy and at Diego Garcia and an AN/MSC-61 system located at Exmouth, Northwest Cape, Australia.37 According to the U.S. Defense Communications Agency, the European AU- TOSEVOCOM system consisted of about 225 wideband subscriber terminals homed on four AN/FTC-31 switches and 16 SECORDS providing secure voice service. Another 85 subscribers were provided worldwide secure voice access Bases during the Cold War 117 Table 5 Giant Talk/Scope Signal III stations Country Location Ascension Azores Lajes, Cinco Pincos Greenland Thule Guam Anderson AFB, Barrigada, Nimitz Hill Japan Owada, Tokorozawa, Yokota Okinawa Kadena Panama Howard and Albrook AFBs Philippines Clark AFB, Cubi Point, Camp O’Donnell Spain Torrejon Turkey Incirlik U.K. RAF Croughton, Mildenhall, Barford St. John Source: SIPRI data, Harkavy, Bases Abroad, p. 164. İ 2007 Robert E. Harkavy

Page 25 through 10 AUTOVON switches.38 Locations of the main switching center for the AUTOVON network are listed in Table 6. One of the major U.S. uses of overseas theater communications was that involved in the highly proliferated microwave/troposcatter systems used to link U.S. and other allied forces within the European and Pacific theaters. This in turn involved a number of sub-systems, perhaps the best known of which was the NATO ACE HIGH system within Europe. According to Jane’s: ACE HIGH is an 80-voice-channel trophospheric scatter/microwave link system which dates back to 1956 when SHAPE developed a plan for an exclusive communications system which would comprise the minimum essential circuits of early warning and alert and implementation of the tripwire retaliation strike plan. The network extends from northern Norway and through Central Europe to Eastern Turkey.39 All of the U.S. NATO allies hosted numerous troposcatter relay links – there were some 40 in West Germany, six in Belgium, eight in Greece, 16 in Italy, 15 in Turkey, and so on. (Earlier there were some 30 such links in France.)40 These types of link also ran from the continental U.S. via Greenland, Iceland, the Faeroes and the U.K. to Europe; indeed, they were originally designed as one link in the Ballistic Missile Early Warning System (BMEWs). In Iceland, it is reported that each such North Atlantic Relay System (NARS) installation consisted of four large “billboard” troposcatter antennas.41 Parts of the troposcatter network were modernized as the Digital European Backbone System (DEBS). The transmission, relay and reception of strictly military and diplomatic mes- sages does not exhaust the uses to which overseas facilities were put within the broad domain of communications. Basing diplomacy also entered the news in connection with broadcast communications. This took any of several forms, for instance, the major powers’ use of foreign territories for clandestine radio 118 Bases during the Cold War Table 6 Location of the main switching centers for the AUTOVON network Country Location Japan Fuchu, AS, Camp Drake (moved to Yokota) Okinawa Grass Mountain or Ft Buckner Philippines Clark AFB Panama Corozal Spain Humosa U.K. Martlesham Heath, Hilingdon, RAF Croughton FDR Germany Schoenfield; Fieldberg, Donnersberg, Pirmasens, Langerkopf Italy Coltano, Mt. Vergine Guam Finnegayan Source: Defense Communications Agency, “Defense Communications System/European Communi- cation Systems: Interoperability Baseline,” Washington, DC, 1 February 1981; and Jane’s Military Communications (Macdonald: London, 1981). İ 2007 Robert E. Harkavy

Page 26 transmitters, particularly adjacent to rivals’ territories or those where civil wars were in progress. (In the 1987 U.S. Congressional Iran/“Contragate” hearings, information emerged about the CIA-run clandestine transmitters in Central America and in the Caribbean directed against Nicaragua and Cuba.) On a more overt basis, this involved, at least as pertained to the U.S. side, the global transmission network of the government’s Voice of America (VOA). There were Soviet, French and British counterparts. Access for VOA transmitters, even despite the absence of obvious military implications, was not always a simple matter. Soviet and other nations’ sensitiv- ities to radio-broadcast intrusion were such that a nation hosting a VOA facility risked a degree of displeasure.42 For its short-wave broadcasts, the VOA had six main 500-kilowatt transmitters; additionally, a variety of some 100 antennas and relay stations in Asia, Africa and Central America. The VOA had overseas radio stations in Antigua, Thailand, Botswana, Greece (two), West Germany, the Philippines (two), Costa Rica, Sri Lanka, Morocco, Belize and the U.K. (and later in Israel). In 1984–1985, as the Central American crisis intensified, it was reported that Costa Rica and Belize had agreed to host VOA broadcast relay stations – in addition, VOA had obtained agreements to construct relay stations in Sri Lanka, Israel, Morocco and Thailand. Intelligence43 During the Cold War the U.S. relied on a variety of technical methods for intelligence collection (i.e., other than HUMINT) which involved the use of overseas facilities. These involved the domains of imaging or photographic reconnaissance, signals intelligence, ocean surveillance, space surveillance and nuclear detonation and monitoring.44 Variously, cutting across these categories, this involved fixed land-based facilities, air bases and naval facilities. In some cases, the utilization of foreign facilities for specific purposes was well-known, as for instance in the cases of large strategic radars or air bases used to stage Bases during the Cold War 119 Table 7 Interconnections between ACE HIGH and DCS DCS ACE HIGH Mormond Hill (U.K.) Mormond Hill (U.K.) Cold Blow Lane (U.K.) Maidstone (U.K.) SHAPE (Belgium) Costeau (Belgium) Bonn (FRG) Kindsbach (FRG) Aviano (Italy) Aviano (Italy) Naples (Bagnoli, Italy) Bagnoli (Italy) Izmir (Turkey Izmir (Turkey) Incirlik (Turkey) Adana (Turkey) Sources: Defense Communications Agency, “Defense Communications System/European Commu- nication Systems; Interoperability Baseline,” Washington, DC, 1 February 1981; and Jane’s Military Communications (Macdonald: London, 1981). İ 2007 Robert E. Harkavy

Page 27 photoreconnaissance flights. In others, however, data were more limited – this was particularly true regarding the relay of data from satellites to major headquarters in the U.S. homeland. The U.S., during the Cold War and beyond, made important use of reconnaissance satellites which conducted area surveillance, close-look and real-time surveillance. Crucially, this involved the Keyhole series, KH-9 (Big Bird) devoted to area-surveillance and KH-8 (Close Look), later superseded (in 1976) by the KH-11 and later by the KH-12 (Ikon). Mostly, these satellites appear to have functioned without the help of overseas downlinks. The U-2 and SR-71 strategic reconnaissance aircraft were usable for a variety of nuclear and non-nuclear related missions. The latter can fly at a speed of Mach 4 (about 4160km per hour), at a height of over 25,000 meters, can track SAM missiles, has radar detectors, a variety of ECMs, and a synthetic-aperture radar for high altitude night imaging. Some 15–19 SR-71s were utilized, co-located at some of the same bases as the U-2s, at Mildenhall in the U.K., Kadena on Okinawa and Akrotiri on Cyprus (the U-2s also utilized Incirlik, Peshawar, Clark AFB, Atsugi and Wiesbaden). The still newer TR-1A reconnaissance aircraft, of which some 14 were deployed, also utilized some of these bases. Other satellites and also land-based facilities were used for SIGINT, an acronym that subsumes several categories of intelligence collection, i.e., COMINT (communications intelligence), ELINT (electronic intelligence), TELINT (telemetry intelligence) and FISINT (foreign instrumentation signals intelligence). The major Cold War SIGINT satellite systems–Chalet, Rhyolite (targeted against telemetry, radar, communications, extending across the VHF, UHF and microwave frequencies), Ferret, Magnum/Aquacade (low orbiting ferret satellites used to map Soviet and Chinese radars) apparently made little use of overseas downlinks. But, the U.S. long made use of a plethora of ground-based SIGINT stations, as identified in Table 8. Although there were diverse types and mixes of these facilities, a few widely deployed types were notable. One involved a combination of AN/FLR-9 HF and VHF interception and direction-finding system (DF) with CDAA (Circularly Disposed Antenna Array) known as an “elephant cage.” Another involved telemetry interception capability with combined VHF-UHF-SHF receivers, used to monitor missile launches.45 Then there were FPS-17 detection radars and FPS-79 tracking radars also used in connection with missile launches. There were also a considerable number of AN/FLR-15 antennas.46 The identifiable land-based SIGINT facilities included those shown in the accompanying table. In some of these cases – Canada, the U.K., Turkey and perhaps Japan – SIGINT stations were jointly operated with host personnel, and the data intake shared to one degree or another, no doubt negotiated on a case- by-case basis and subject to periodic renewal; hence, a function of the state of political relationships and associated reciprocal leverage. 120 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 28

Ocean surface surveillance47 The U.S. had a number of systems – satellites, aircraft, ground stations and ships – for observing the world’s ocean surfaces, that is, for tracking Soviet warships, auxiliary intelligence ships, merchant and fishing vessels etc. Operationally, the goal was to know the location of all Soviet ships at any time. In normal conditions, one major purpose was to track the itineraries of Soviet ships carrying arms to clients – this was a key item of intelligence. In crises or, hypothetically, at the outset of a major war, the hair-trigger, preemptive nature of modern naval warfare – nuclear or non-nuclear – would have put a premium on real-time location and targeting of rival fleets, in all weathers. Contrariwise, both sides would have worked hard to devise methods for eluding detection, again, particularly during wartime conditions. Bases during the Cold War 121 Table 8 Land-based SIGINT facilities Country Location/comments Australia Northwest Cape The Azores Villa Nova Canada Massett, Argentia, Whitehorse, Leetrim China Korla, Qtai Cuba Guantanamo Cyprus Five Stations Denmark Bornholm Diego Garcia Honduras Palmerola Kwajalein Midway FDR Germany Augsburg, Hof, a network called La Faire Vite to monitor WTO communications, and others U.K. Cheltenham, Wincombe, Morwenstow, Kirknewton Greece Iraklion and Nea Makri Iceland Keflavik, Stockknes Italy San Vito, Vicenza, Treviso Japan Misawa, Camp Zama, Hakata, Sakata, Wakkanei, Kamiseya South Korea Yonchon, Camp Humphreys, Pyongtaek, Sinsan-ni, Kangwha Morocco Kenitrar Norway Varda, Vadso, Viksofjellet Oman Al Khasab, Umm Al-Ranam Island Panama Corozol, Fort Clayton, Galeta Island Pakistan Bada Bien Philippines San Miguel, Clark AB, John May Camp Okinawa Tori, Hanza, Sobe, Omna Point Spain Rota, El Casar del Talamanca Taiwan Shou Lin Kou, Tapeh, Nan Szu Pu Turkey Sinop, Dyarbakir, Samsun, Karamursel, Antalya, Agri, Kars, Edirne, Ankara Source: SIPRI data, and J.T. Richelson and D. Ball, The Ties That Bind (Allen and Unwin: Boston, 1985), appendix 1. İ 2007 Robert E. Harkavy

Page 29 Overseas facilities played a major role here. The U.S. utilized its White Cloud satellite system, part of its larger Classic Wizard system, for ocean surveillance, involving a variety of ELINT functions as well as use of interferome- try techniques to locate Soviet or other vessels. This system comprised four satellites. The U.S. Naval Security Group operated ground stations which were part of this system at Diego Garcia and Edzell, Scotland, as well as at Guam, Adak and Winter Harbor, Maine. The P3C Orion, known mostly for its ASW role, was also utilized for ocean surveillance. It had access to bases throughout the world: Clark AFB (the Philip- pines), Misawa (Japan), Kadena (Okinawa), Keflavik (Iceland), Rota (Spain), Sigonella (Italy), Ascension and Diego Garcia islands, Cocos Islands (Aus- tralia), Masirah (Oman), Mogadiscu (Somalia) and several others. For land-based ocean surveillance, considerable use was made of HF/DF systems, which were also mounted on ships which, again, utilized various overseas port-facilities. Among the land-based HF/DF locales were those at Diego Garcia, Rota (Spain), Edzell (Scotland), Keflavik (Iceland), Brawdy (Wales), Japan and Guam. Those in Scotland, Wales and Iceland were located near the crucial GIUK–Gap chokepoint, which would presumably have been a major point of contention at the outset of a major war in relation to North Atlantic sea lines of communication and the Soviet submarine bastions near the Kola Penin- sula. Richelson and Ball actually reported on some 40–50 HF/DF sites for ocean surveillance said to have been operated by the combined assets of the U.S., U.K., Canada, Australia and New Zealand. Space surveillance48 During the Cold War, the proliferation of satellites and other man-made “space objects” made their tracking and identification more vital. The U.S. had an extensive program intended to detect and track its own satellites, but also Soviet and other nations’ space vehicles. In the security realm this had a number of dimensions. Of course, both sides wished to mask some of their ground activities from surveillance and therefore sought the capacity to operate during gaps in surveillance. By detecting and tracking Soviet satellites, the U.S. Satellite Reconnaissance Advance Notice (SATRAN) System allowed the U.S. to avoid Soviet coverage of U.S. military activities. As expectations mounted about a future which might have seen large-scale militarization of space, both sides increasingly perceived an interest in real-time surveillance of each others’ satellites, in the context of possible later hair-trigger preemptive situations as applied to mutual interdiction of satellites. Of course, both sides desired maximally effective intelligence on the others’ various military activities conducted from space: communications, ocean surveillance, SIGINT, nuclear detection, etc. In summary, as stated by Richelson: space surveillance helps provide the United States with intelligence on the characteristics and capabilities of Soviet space systems and their contribu- 122 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 30 tions to overall Soviet military capabilities ... Such data aid the United States in developing counter-measures to Soviet systems, provide a data- base for U.S. ASAT targeting and allow the United States to assess the threat represented by Soviet ASAT systems. For many years the heart of the dedicated sensor system was a group of Baker- Nunn optical cameras, huge cameras which, according to one source, could “photograph, at night, a lighted object the size of a basketball over 20,000 miles in space.” In addition to the two in California and New Mexico, these cameras were located outside the U.S. in New Zealand (Mt. John), South Korea (Pul- mosan), Canada (St. Margarets, New Brunswick) and Italy (San Vito). Earlier there were others on Johnston Island, in Alaska, and in Argentina, Brazil, Chile, Ethiopia, Greece, Iran, South Africa, Upper Volta and Curacao in the Lesser Artilles, among others. Because of its limitations – slowness in data acquisition, processing and response time, absence of all-weather capability and inflexible tracking capability – this system was replaced by the Ground-based Electro- Optical Deep Space Surveillance (GEODSS) system. GEODSS, also with five locations, overcame several of the Baker-Nunn system’s shortcomings by allowing real-time data, better search capability and more rapid coverage of larger areas of space – but was still limited by adverse atmospheric conditions. It was actually a system of three linked telescopes at each site, providing variable coverage by altitude. The five locations were in Hawaii (Maui) and New Mexico (White Sands), within the U.S. and externally in South Korea (Taegu), Diego Garcia and Portugal. Systems used primarily for early warning – BMEWS, FSS-7, PAVE PAWS, Enhanced Perimeter Acquisition Radar Attack Characterization System (EPARCS) and FPS-85 radars – were usable as collateral space-tracking sensors. Of these, BMEWS – based at Thule (Greenland), Fylingdales (U.K.) and Clear (Alaska) – involved extensive use of foreign access. Additionally, COBRA DANE (Shemya Island, Aleutians – 120 arc, 46,000-km range against space targets) and also the AN/FPS-79 (Pincirlik/Diyarbakir, Turkey) radar were usable in a space-surveillance role, as supplementary to the primary missions of monitoring missile-test re-entry trajectories. Numerous other foreign facilities were used as part of the U.S. Satellite Tracking and Data Acquisition Network (STADAN) network of installations used to track and monitor U.S. space activities, including the down-range course of launches. Among these were facilities in: Australia (Orooral Valley, Toowoomba), the U.K. (Winkfield), Ascension, Bermuda, the Canaries (Tener- ife), Spain (Madrid), Brazil (Fernando de Noronha) and Antigua, in connection with space surveillance and under the heading of “miscellaneous radars.” Earlier STADAN tracking facilities were operated, among other places, in Chile, Ecuador, the Malagasy Republic, Grand Turk Islands, South Africa and Zaire. Bases during the Cold War 123 İ 2007 Robert E. Harkavy

Page 31

Satellite control stations49 One of the most secret or classified areas of overseas bases was that of satellite control stations. Ford, in his work on command and control, in analyzing the vulnerable and non-redundant nature of the U.S. early-warning system involving the DSP East satellite, its down-link facility in Australia and the communications link from there to the satellite control facility in Sunnyvale, California, provided some indication of what was involved. Thus, according to him: There are several dozen U.S. defense satellites now in orbit – providing communications, photoreconnaissance, electronic intelligence, navigational, meteorological, and other data-and they require contact with the Sunnyvale ground control station and its seven substations around the globe in order to remain functional. A great deal of fine-tuning, for example, is needed to steer the satellites in precise orbits and to keep their sensors, and antennas aimed properly. . . . A catastrophic loss of this control center would result in a major disruption of communications, tracking, and control of its space systems . . . And, further: Other officials are less optimistic. “We lose the SCF and the satellites basi- cally go haywire,” a Pentagon expert who has studied this subject told me. “The communications satellites drift off to Pluto.” Certain intelligence- gathering satellites in low-earth orbit would be in especially bad shape, he said, since the Sunnyvale facility has to “feed them” with instructions every time they complete an orbit. “You should see them scrambling when one of their satellites comes within range.” Desmond Ball estimated that the typical U.S. defense satellite might be able to remain in operation for three to four days without the Sunnyvale SCF; the most critical satellites, such as DSP East, which require a great deal of caretaking attention from the ground, could go out of service within hours. The seven sub-stations linked to Sunnyvale comprised three within the U.S. – at Manchester AF Station in New Hampshire, Kaena Point in Hawaii and Van- denberg Air Force Base (AFB) in California. Others outside the U.S. were at Thule, Greenland – collocated with various other technical facilities as well as a bomber and tanker base – at Guam at Andersen AFB, at Oakhanger in the U.K. and at Mahe in the Seychelles. Mahe had long hosted a U.S. satellite control facility (SCF) collocated with a DSCS ground terminal. This facility was apparently important in relation to reconnaissance satellites and for monitoring injec- tion into orbit of satellites launched from Cape Canaveral. Earlier, up to 1975, the U.S. also had what apparently was an SCF at Majunga in the Malagasy Republic, one also used to monitor satellites launched from Cape Canaveral. U.S. access to this facility was then lost at a time when Tananarive shifted 124 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 32 towards an arms-supply relationship with the USSR. Still earlier the U.S. apparently had a similar SCF on Zanzibar Island, within Tanzania. Strategic early warning50 One of the most obviously critical areas of intelligence involving foreign bases is strategic early warning, that is, warning of impending or unfolding nuclear attack. Here, as in so many areas of military endeavor, the U.S. was asymmetri- cally dependent on foreign access, and crucially so. Indeed, it was long the case, even before the advent of intercontinental missilery, that the U.S. relied on radars in Canada and Greenland for warning of approaching bombers at a time when such warning could provide several hours of response time. The asymme- tries were partly because of the larger relative size of the USSR and partly because of the location of critical foreign terrain – Canada, Greenland, Iceland – between the U.S. and the trans-Arctic routes that would be traversed by missiles and bombers across the arctic regions between the superpowers. Several key U.S. early-warning systems utilized foreign access; ground stations used to relay data from early-warning satellites; the BMEWS radar system directed against Soviet ICBMs, the several layers of radar pickets used to detect bombers en route to the U.S. from the Arctic region; and a variety of other sensors which might be used in collateral or supplementary roles. Regarding early-warning satellites, the U.S. relied primarily on what have become known as DSP East and DSP West (earlier these were called MIDAS satellites). The former had the primary task of watching for ICBM launches, the latter for SLBM launches in the Atlantic. The Code 647 Defense Support Program satellite – DSP East – sat some 23,000 miles above the Indian Ocean in geosynchronous orbit, monitoring the eastern hemisphere. It contained an infrared telescope equipped with thousands of tiny lead-sulphide detectors designed to pick up the hot exhaust flame produced by large rocket engines during the boost phase of their flights. It had full coverage of Soviet missile fields from an orbit more or less above the equator and, in the case of mass launchings, could tell what kinds of missiles had been launched and from where. Hence, DSP East could also provide valuable information about what kind of attack had been launched, that is, the likely targeting mix. That would in turn have guided the targeting of a U.S. counterforce response. The principal and necessary link between DSP East and NORAD was the down-link facility at Nurrungar, Australia (there were also control and tracking functions at Pine Gap and on Guam) described as a highly vulnerable set of antennas, transmitters and computational facilities. This “readout station” in turn would relay data to NORAD, variously by underseas cable, by HF radio links or via the Defense Satellite Communications System, and in turn via a switching station in Hawaii. These various alternative communications links between Nur- rungar and Buckley AFB in Colorado involved foreign access to New Zealand, Fiji, Norfolk Island and Canada’s British Columbia at Port Alberni. In recent Bases during the Cold War 125 İ 2007 Robert E. Harkavy

Page 33 years, however, another down-link for DSP East (called the Simplified Process- ing Station) had been made operational at Kapaun, FDR Germany. A second major U.S. strategic warning system used to monitor ICBM launches was BMEWS, which comprised three major radars (located at Clear, Alaska, Thule, Greenland and Fylingdales Moor, U.K.). These radars became operational in 1950 and had 4800-km ranges. The BMEWS facility included four AN FPS-50 detection radars and an AN/FPS-49 tracking radar – these also acted as contributing sensors for the SPADATS system. Altogether, there were 12 radars with ranges of 4800 km. Finally, the U.S. had long relied – primarily for warning of the approach of bombers – on the series of radar picket lines across the Arctic known as the Distant Early Warning (DEW) and Continental Air Defense Integration North (CADIN) Pinetree lines. These had been located across Alaska, Canada and Greenland, with a few additional outposts in Iceland and the Faeroe Islands (the 126 Bases during the Cold War Table 9 Location of DEW Line and CADIN Pinetree Line radar sites in Canada and Greenland, 1985 DEW Lines Pinetree Line Canada Broughton Island Alaska Byron Bay Armstrong Cambridge Bay Baldy Hughes Cape Dyer Barrington Cape Hooper Beausejour Cape Perry Beaver Lodge Cape Young Chibougamau Clinon Point Dana Dewar Lakes Falconbridge Gladman Point Gander Air Base Hall Beach Goose Bay AB Jenny Lind Island Gypsumville Komakuk Beach Holberg Lady Franklin Point Kamloops Longstaff Bluff Lac St. Denis Macker Inlet Moisie Nicholson Peninsula Montapica Pelly Bay Moosonee Shepherd Bay Ramore Shingle Point Senneterre Tuktoyaktuk Sioux Lookout Yorkton Greenland Easterly Kulusuk Island (Dye 4) Quiquatoqoq (Holsteinberg-Dye-1) Westerly Source: W.M. Arkin and R.W. Fieldhouse, Nuclear Battlefields: Global Links in the Arms Race (Cambridge: Ballinger, 1985), appendix A. İ 2007 Robert E. Harkavy

Page 34 locations of the sites in Canada and Greenland are listed in Table 9 above). The DEW Line had 31 radars, some 21 of which were in Canada and four in Green- land (several also were used as relays for troposcatter communications relays); in the 1950s submarines on the surface filled gaps in the DEW Line – they could submerge after reporting incoming aircraft. The CADIN Pinetree Line of air surveillance radars in Canada had constituted a second line of warning behind the DEW Line and comprised some 22 stations operated by Canadian personnel for NORAD. The upgrading of the almost 30-year old DEW Line was impelled by cruise- missile developments and by new Soviet Backfire bombers capable of penetrating the old barrier; and it preceded the advent of the Strategic Defense Initiative (SDI). The new system involved a network of 52 new long- and short- range radar stations overlapping Alaska, Canada and Greenland, and was called the North Warning System. It utilized many of the hub facilities of the DEW Line, but with upgraded modern radars and independent power systems. Later, around 2005, the U.S. became involved in the upgrading of the facility at Fylingdales for theater missile defense and, possibly, for the stationing of actual theater defense missiles. Anti-submarine warfare51 During the Cold War, as well as afterwards, the U.S. had perhaps the world’s most extensive and effective global ASW capability, which required access to land facilities around the world: staging bases for aircraft, and processing stations for acoustic and electronic data. Regarding acoustic data, it relied heavily on its SOSUS networks, often alternatively referred to as Caesar. These, going way back to 1954, involved networks of hydrophonic arrays which sent oceanographic and acoustic data to shore processing facilities, that is, large computer analysis centers. These data and others were correlated at regional processing centers (including those in Hawaii, Wales, Newfoundland and Iceland) and then forwarded to a main processing center at Moffett Field, California, via FLTSATCOM and DSCS satellites.52 SOSUS, though augmented by other systems noted below, was the backbone of the U.S. ASW detection capability. It has been described by SIPRI as follows: Each SOSUS installation consists of an array of hundreds of hydrophones laid out on the sea floor, or moored at depths most conducive to sound prop- agation, and connected by submarine cables for transmission of telemetry. In such an array a sound wave arriving from a distant submarine will be successively detected by different hydrophones according to their geometric relationship to the direction from which the wave arrives. This direction can be determined by noting the order in which the wave is detected at the different hydrophones. In practice the sensitivity of the array is enhanced many times by adding the signals from several individual hydrophones after intro- ducing appropriate time delays between them. The result is a listening Bases during the Cold War 127 İ 2007 Robert E. Harkavy

Page 35 “beam” that can be “steered” in various directions towards various sectors of the ocean by varying the pattern of time delays. The distance from the array to the sound source can be calculated by measuring the divergence of the sound rays within the array or by triangulating from adjacent arrays. The first SOSUS systems were completed on the continental shelf off the east coast of the U.S. in 1954.53 Others were later installed off both U.S. shores and at Brawdy, Wales – the Pacific Coast system came to be known by the code- name of Colossus.54 A jointly operated U.S.–Canadian array came to be centered at Argentia, Newfoundland, others at Hawaii, the Bahamas and the Azores.55 By 1974 it was stated that there were 22 SOSUS installations located along the east and west coasts of the U.S. and near various chokepoints around the world – another 14 were identified by Richelson and others.56 Foreign-based SOSUS installations were located at Ascension, in the Azores (Santa Maria), the Bahamas (Andros Island), Barbados, Bermuda, Canada (Argentia), Denmark, Diego Garcia, Gibraltar, Guam (Ratidian Point), Iceland (Keflavik), Italy, Japan (sonar chains across the Tsugaru and Tsushima Straits), Norway, Panama (Galeta), the Philippines, the Ryukyus, Turkey, the U.K. (Scatsa, Shetland and Brawdy, Wales). Others have at times been operated on Grand Turk Island, Antigua, Bahamas (Eleuthera) and Barbados; and maybe on the Canary Islands at Punta de Tero. And the U.S. may possibly also have operated still other barrier sonars, for instance, in the central Mediterranean from Lampedusa and/or Pantelleria Islands, and on Midway Island in the central Pacific Ocean. Burrows described as follows the basic geometry of the U.S. SOSUS network which monitors Soviet egress from the Eurasion bastion: There are actually two SOSUS arrays moored across the approaches to Pol- yarnyy; one between Norway and Bear Island, and the other linking northern Scotland, Iceland, and Greenland. Submarines whose home port is Petropavlovsk are monitored by hydrophones strung from the southeastern tip of Hokkaido, along a line parallel to the Kuriles, and then up toward the northeast, off the Aleutian coast. Still others stretch from southern Japan to the Philippines, covering the approaches to China and Indochina. And there are also SOSUS installations on the Atlantic side of Gibraltar, others about halfway between Italy and Corsica and still others at the mouth of the Bosporus, off Diego Garcia in the Indian Ocean, and not so far from Hawaii. The Navy keeps the precise locations of its SOSUS equipment a closely guarded secret, since interfering with it would be a logical Soviet subject.57 Nuclear detection58 One important, but seldom commented upon aspect of the overall U.S. intelligence effort, was that connected with the detection of nuclear explosions. This involved several separate lines of activity. First, there was the matter of verification of existing arms control treaties, 128 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 36 that is, the Limited Test Ban Treaty of 1963, which bans atmospheric testing, and the Threshold Test Ban Treaty of 1974, never ratified by the U.S. Senate (but tacitly adhered to in the manner of SALT II), which barred underground testing of nuclear devices of over 150 kilotons. Second, there was the monitoring of the horizontal nuclear proliferation activities of hitherto non-nuclear states, as well as of the non-signatory but nuclear states, China and France. Third there was the contingency of protracted nuclear war during which the U.S. would have wanted to determine the locations and frequencies of nuclear deto- nations on both sides and to assess resultant damage, among other reasons, to aid subsequent targeting decisions. Several interrelated systems were used to pursue the above ends, involving satellites, aircraft and ground stations. Use of satellites, in connection with bases, raised the question of external facilities for data down-links and command and control; that of aircraft involved, obviously, bases as well. According to Richelson, the U.S. space-based nuclear-detection system involved, variously, the various components of the VELA satellite program begun in the early 1960s, the previously mentioned DSP satellites primarily intended for early warning of missile launches, and the NAVSTAR global positioning system.59 Numerous aircraft types were used to detect airborne atomic debris left in the wake of explosions (if only the venting of imperfectly conducted underground blasts). One source said that these included the U-2, P-3C, C-135, B-52 and also an HC-130 configured as a sea-water sampler to monitor underwater nuclear tests (monitoring of plutonion-239 separation via kryption-85 analysis was presumably also similarly conducted). For instance, SAC’s U-2s apparently operated out of Australian facilities at Sale and Laverton, gathering radionuclides as part of a High Altitude Sampling Program. These aircraft were operated by the Air Force Tech- nical Application Center (AFTAC) and could presumably have availed themselves of virtually all the airfields normally open to U.S. use throughout the world. Some of these operations no doubt involved ad hoc staging through facilities after an “event,” and the diplomacy of access involved was obscure. But, as there was a general convergence of overall interest by most nations with regard to monitoring others’ nuclear tests, access in these cases was likely to have been permissive. The Soviet Cold War naval basing structure60 Before World War II, the USSR had had only very limited access to external facilities, perhaps solely with the forward deployment of some combat aircraft in Czechoslovakia in the 1930s. In parallel, Moscow was then only a small factor in the international arms trade, so it was not able to bank either on security assistance or ideological ties and alliances to establish forward bases. That situation changed dramatically after World War II, with an additional quantum jump in the late 1950s and on up to 1970s. Directly after the war, of course, Moscow established control over Eastern Europe (earlier also in Mongolia), and these dominated areas became the sites for a massive network of Soviet air, naval and ground facilities, with a heavy Bases during the Cold War 129 İ 2007 Robert E. Harkavy

Page 37 concentration in East Germany, Poland, Hungary and Czechoslovakia as an obvious correlate to Soviet political control over these countries and as a defensive glacis or possible springboard for offensive operations vis-à-vis NATO Western Europe. Again, the basis for basing access was simply brute conquest and imperial control. For some ten years after the end of the Cold War, the USSR had no basing access outside of its huge contiguous Eurasian empire. Indeed, it also had virtually no arms transfer or security assistance relationships outside this area during a period in which the U.S. established its elaborate structure of alliances and arms transfer relationships all around the Eurasian rimland. That all changed around 1955 with the beginning of the Soviet weapons transfers to Egypt and Syria. But then too in the late 1950s and early 1960s, numerous nations in the wake of decolonization and the creation of nominally Marxist regimes, became Soviet arms client states. Along with that, sometimes with a time lag, came the provision of basing access at a time, also, when Moscow was beginning to build a “blue-water navy.” And, some of the new arms recipients and basing hosts involved a “leapfrogging” of the containment rim around the huge Sino-Soviet Eurasian domain. Among the basing hosts developed in addition to Egypt and Syria were Algeria, Guinea, Ghana, Congo- Brazzaville, Angola, Mozambique, Somalia, both Yemens, Iraq, Cuba and Vietnam (Indonesia and India were both major Soviet arms clients by 1965, but provided no significant access to bases). Gradually, the Soviet navy developed an external presence as the Cold War progressed, involving the build-up of a significant blue-water navy, the acquisition of bases and available ports of call, and the compilation of “ship-days” in the major oceans and seas matching that of the U.S. Navy. As recently as the late 1950s, the Soviet navy had had little of “blue-water,” long-range power projection capability nor basing access. Indeed, the early postwar exceptions were use of a Chinese base at Port Arthur, one at Porkalla in Finland, and a three-year interregnum (1958–1961) when Soviet submarines were granted access to Albanian bases in the Mediterranean. After that came the large-scale naval build-up through the late 1980s under the aegis of Admiral Gorshkov, a build-up that produced a navy that, measured by major surface combatants and submarines, was larger than that of the U.S. The Soviet Navy fielded some 269 principal surface combatants (including four Kiev carriers), two ASW helicopter carriers, some 36 cruisers of which two (Kirov-class) were nuclear, 61 destroyers (of which 33 were specialized for ASW) and 167 escorts including 100 corvettes.61 In addition, there were some 762 minor surface combatants: vast numbers of missile patrol boats, fast attack craft, 372 principal auxiliary ships (replenishment, tankers, missile support, supply, cargo, submarine tenders, repair, hospital, submarine rescue, salvage/ rescue, training ships, etc.). There were some 62 intelligence collection vessels (AGI), 456 naval research vessels, 74 tankers, 298 support ships, 1900 merchant ships, and numerous civilian oceanographic, fishery, space-associated and hydrographic research vessels.62 130 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 38 The Soviet submarine fleet was equally formidable, comprising some 360 vessels. It fielded 63 SSBNs and 14 non nuclear-powered ballistic-missile submarines (SSNs), with a total of 983 submarine-launched ballistic missiles (SLBMs) – 39 SLBMs and 15 submarines were outside the SALT limits – some 214 attack submarines of which 70 were nuclear-powered, and 63 cruise-missile submarines, 48 of which were nuclear (SSGNs).63 That Cold War Soviet navy was, obviously huge. But its more limited (relative to the U.S.) external basing network required an outsized force of auxiliary ships to compensate for the shortage of land bases. The same was true for AGIs, fishing and oceanographic vessels, and SIGINT, communications relay, space- tracking, ships etc. In the Mediterranean, the important Soviet presence in Syria included naval access to the port of Tartus, which was the primary maintenance facility for Soviet submarines operating in the area. A Soviet submarine tender, a yard oiler and a water tender were stationed there. Over the years, there had also been varied degrees of naval access to Algeria, Libya and Yugoslavia. Soviet submarines were reported serviced at Annaba in Algeria, and its ships were refueled and maintained at a couple of Yugoslav ports on the Adriatic, at Tivat and Sibenik.64 In Guinea, despite some curtailment of long-maintained access for Soviet Bear reconnaissance aircraft, the USSR routinely used Conakry harbor as a facility for its West African patrol.65 But Luanda in Angola became the most important port for Moscow’s West African naval units – since 1982 that had involved an 8500-ton floating drydock capable of handling most major Soviet naval combatants.66 In Ethiopia, the installation at Dahlak Archipelago was the maintenance facility and supply depot for Soviet naval combatants operating in the Indian Ocean and Red Sea, normally ranging from 20–25 units, including surface ships, attack and cruise-missile submarines and auxiliaries. This facility included an 8500-ton floating drydock, floating piers, helipads, fuel and water storage, a submarine tender and other repair ships. Guided-missile cruisers and nuclear-powered submarines regularly called at Dahlak for repair and supplies.67 In Cuba, in addition to enjoying access for port visits, maintenance, and so on, Moscow permanently based a submarine tender at Cienfuegos, used primarily, if not solely, for servicing attack submarines – access for SSBNs might have been construed as a violation of the agreements emerging out of the Cuban Missile Crisis.68 At another level, Soviet access to Peruvian ports provided logistics support and maintenance for some 200 fishing vessels that operated off the coast of South America. This involved, among other things, extensive rotation of merchant seamen and fishermen.69 Table 10 details the Soviet overseas naval basing structure which, it is important to note, had experienced some major changes in the late Cold War period as reflective of the vicissitudes of external political alignments. Several external main operating bases were crucial to Soviet naval deployments: Cam Ranh Bay (Vietnam) in the Far East/Pacific area; Aden and Socotra (South Yemen) and the Bases during the Cold War 131 İ 2007 Robert E. Harkavy

Page 39 132 Bases during the Cold War Table 10 Main and secondary surface-ship operating bases of the Soviet Navy Host nation base Description Vietnam Cam Ranh Bay Main external Soviet naval base in Far East – guided- missile cruisers, frigates and minesweepers based here; also, attack submarines; on average, deployment was four submarines, two to four combat vessels, ten auxiliaries. Cambodia Kampong Som (Sihanoukville) Reported access for Soviet warships, i.e., replenishment, refueling, etc. North Korea Najin Some port access, earlier reports of submarine base. India Vishakhapatnam Indian naval base built with Soviet assistance, some Soviet port calls, refueling, etc. Cochin Port calls, refueling, etc. reported. Iraq Umm Quasr Soviet assistance in improving facilities here, earlier reported accessible to Soviet warships; access, limited during Iraq–Iran War. Az Zubayr Earlier reported used by Soviet submarines and SIGINT vessels. Al Fao Iraqi port, reported availed to Soviet Union after 1974 agreement. South Yemen Aden Soviet main base for Indian Ocean operations; fuel tanks, replenishment, reports of submarine pens alongside berthing for major surface ships. Socotra Island Anchorage used by Soviet ships, possible shore facilities. Ethiopia Dahlak Archipelago Large anchorage for Soviet Indian Ocean naval squadron. Assab Important Soviet naval facility; floating dry dock formerly moored at Berbera. Massawa Port access, routine. Perim Island Former British facility, reportedly improved by Soviet Union. Mauritius Reported port calls (note concurrently reported U.S. access). Mozambique Nacala Periodic port calls. Maputo Periodic port calls. Tanzania Zanzibar Available, port calls. Angola Luanda Was main Soviet naval base on West African coast, having replaced Conakry; guided-missile destroyer and several accompanying craft stationed here. İ 2007 Robert E. Harkavy

Page 40 Bases during the Cold War 133 Table 10 continued Host nation base Description Madagascar Diego Suarez Available, port calls. Tanative Available, port calls. Benin Cotonou Periodic port calls. Guinea Conakry Formerly hosted small West African flotilla; use later curtailed, if not eliminated. Congo Ponte Noir Reported occasional port calls. Guinea Bissau Geba Estuary Port calls. Algeria Mers El Kebir Port calls. Annaba Soviet repair ships deployed. Submarine repair capabilities reported. Libya Tripoli Regular access, Soviet Mediterranean squadron. Benghazi Regular access, Soviet Mediterranean squadron. Bardia Soviet Union reported constructing naval base here. Syria Latakia Main base for Soviet Mediterranean squadron; fuel, replenishment, etc. Tartus Regular access, maintenance facility for attack submarines, oiler, tender. Ras Shamra Soviet submarine base alleged under construction. Yugoslavia Tivat Repair of Soviet ships and submarines. Rijeka Port calls. Pula Port calls. Sibenik Port calls. Split Port calls. Greece Siros Island Ship repairs, commercial, at Neorian shipyard. Cuba Cienfuegos Replenishment base for Soviet attack submarines, mooring of submarine tender occasionally rumoured. Mariel Port calls. Nipe Bay Port calls, Gulf-class submarines, intelligence collectors. Havana Access for Soviet surface ships. Santiago de Cuba Access for Soviet surface ships. Peru Calleo Occasional ship visits since Soviet–Peruvian arms deal. Romania Mangalia Reported Soviet submarine base on Black Sea. Sulina Forward supply base for Soviet Danube flotilla. Source: Compiled from SIPRI data. İ 2007 Robert E. Harkavy

Page 41 Dahlak Archipelago (Ethiopia) in the western Indian Ocean/Horn of Africa area; Luanda (Angola) in the South Atlantic; Latakia and Tartus (Syria) in the Mediter- ranean; and Havana, Cienfuegos and Mariel (Cuba) in the western North Atlantic. Beyond that, the Soviet Navy had acquired degrees of access – secondary bases, minor facilities, port visits etc. – in numerous other locales (often the subject of debate over facts and interpretations). These included Cambodia, India, Iraq, Mau- ritius, the Seychelles, Madagascar, Mozambique, Angola, Congo, Sao Tome and Principe, Cape Verde, Guinea Bissau, Benin, Guinea, Algeria, Libya, Yugoslavia, Spain (Canary Islands), Nicaragua and Peru. That may not have been a truly global basing structure, but it was something well beyond what would have accorded with a strictly defensive, coastal defense navy, or with the assumptions and expectations of a generation earlier. Soviet air bases70 Numerically speaking, the Cold War Soviet Air Force was huge, comprising – in varying degrees of readiness – some 4000 combat aircraft. More than half of these were primarily configured as interceptors, that is, they were for strategic and/or tactical defense; the remainder were configured mostly as ground-attack craft. Still others had as their primary functions reconnaissance and electronic countermeasures; 540 and 30 craft, respectively. In its strategic forces, the USSR had some 165 long-range, 567 medium-range, and 450 short-range bombers, 68 long-range reconnaissance craft, some 100 ECM machines and (here deficient relative to the U.S.) a then growing force of some 50 Bison and Badger tankers. In line with the facts of Soviet ground and naval deployments, most permanent external deployments of aircraft were in the immediately contiguous areas of Eastern Europe within the WTO, Mongolia and Afghanistan. Otherwise, however, the Cold War years saw the Soviet Air Force break out of the confines of Eurasia to establish more or less permanent bases in Vietnam, South Yemen, Angola and Cuba. The USSR had some 2000 tactical aircraft deployed in Eastern Europe. The MiG-23 Flogger was by far the most numerous fighter-interceptor; followed by late-model MiG-21 Fishbeds and older Su-15 Flagons. Other less numerous fighter-interceptors included the Foxbat, Firebar, Fiddler and the then new MiG- 31 Foxhound and MiG-29 Fulcrum. Among the ground-attack aircraft, the most common were the Su-17 Fitter and MiG-27 Flogger, though reportedly the best interdiction aircraft in the Soviet inventory was the Su-24 Fencer. Other units were comprised of MiG-23 Floggers, the then new Su-24 Frogfoot and older MiG-21 Fishbed and Su-7 Fitter As. Reconnaissance aircraft deployed in Eastern Europe included MiG-21 Fishbeds, Su-17 Fitters, MiG-25 Foxbats and Yak-28 Brewers. In East Germany, there were large numbers of attack and interceptor fighters, a total of 685 combat aircraft, comprising 315 attack aircraft (Su-17s, Su-24s, Su-25s, MiG-27s), 300 fighter interceptors (MiG-21/25/27s), 50 reconnaissance craft (Su-17s, MiG-25s), plus 20 ECM and 40 light transport aircraft. These 134 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 42 were forward-based at some 17 bases; quantitatively speaking, there were both more aircraft and bases than were fielded by the U.S. and British counterparts in West Germany. These were at: Zossen-Wiesdorf, Stralsund, Peenemunde, Parchim (Hind-24 helicopters and long-range transports for troop exchanges), Finow, Werneuchen, Oranienburg, Wittstock, Neuruppin, Zerbst (MiG-25 reconnaissance aircraft), Juterborg, Kothen, Welzow, Finsterwalde (Su-20 Fitter B fighter regiment), Merseburg, Grossenhein and Alternburg. In Czechoslovakia, the USSR regularly deployed some 105 combat aircraft, located mostly in the Bohemian region along the German border and in the area northwest of Prague. There were some 45 MiG-27 Flogger D/Js, 45 Flogger Bs and 14 Su-17 reconnaissance craft. Among the some 30 military airfields in Czechoslovakia, Soviet combat aircraft were reported stationed at Prague’s Ruzyne airport, Milovice, Cheb/Horni Dvory, Dobrany, Karlovy Vary, Zatec, Mimon, Tchorovice and Panensky Tynec. In Hungary, the Soviet Air Force had six major air bases in addition to joint use of some Hungarian bases and also some dispersal strips. There were about 240 combat aircraft deployed, including Su-17 Fitter and Su-24 Fencer attack craft, MiG-23 interceptors and Su-17 reconnaissance aircraft. Tokol in Budapest, a major Soviet base used jointly with Hungary, deployed fighters, bombers, transports and helicopters and was also the headquarters for the Soviet Air Force in Hungary. Other important bases were at Kaposvar, Papa, Veszprem, Debrecen, Mezokovesd, Pecs and Szombathely. There was joint use of other installations at Kalocsa, Szolnok, Kecskemet, Sarmellek and Szeged. In strategically located (relative to the main potential theater of operations along the West German frontier) Poland, the USSR earlier based combat aircraft at Legnica, Gniezno, Pucza Bolimowska, Gdansk (naval air Backfires), Zagan, Brzeg, Opole, Szczecin, Kolobrzeg, Szczecinek and Koszalin. The bases at Zagan and Szprotawa hosted Su-24 Fencer aircraft. In the Far East, again associated with a major ground force deployment, there were in Mongolia about six squadrons of Soviet combat aircraft, including MiG- 21/23/25/27s. They were at Choybalsan northeast, Ulan Bator southwest, Nalayh, Bayan Suma and two facilities around Sayn Shand which fielded MiG- 23s. These were large deployments comparable to those in Eastern Europe, reflecting the size of the Soviet presence vis-à-vis China during the period of tension between the two communist giants. In addition, some 30 MiG-23s were based on Eterofu Island in the Kuriles (disputed with Japan), along with 8000–10,000 troops. During the war in Afghanistan the Soviet Air Force established a major presence in at least five air bases. These was at Kabul, Kandahar (important for airlift and for naval reconnaissance over the Indian Ocean), Bagram (MiG-23 base and reported Tu-95 Bear bomber deployment), Shindand (a squadron each of MiG-21s and MiG-23s, two squadrons of Su-20 fighters and 60 Mi-6 helicopters) and Jalalabad near Pakistan and the Khyber Pass, where 100-plus Soviet helicopters were reported stationed. These bases were later used by the U.S. after the fall of 2001. Bases during the Cold War 135 İ 2007 Robert E. Harkavy

Page 43 Outside of the old (Western-imposed) Eurasian containment rim (or “out of area” relative to the USSR), the Soviet Air Force established bases, ad hoc facilities and staging rights in virtually all of the world’s major regions. Concerning the “permanent” or continuous deployment of aircraft (or sporadic deployments approaching that status), this involved most importantly both combat and naval reconnaissance aircraft. The major deployments were, as one might expect, co- located with the major naval facilities in some of the Soviet Union’s closest allies in the Third World, such as Vietnam, South Yemen, Cuba, Syria, Libya and Angola. Still, relative to USAF access overseas, the paucity of permanent deployment of combat aircraft stood out. Perhaps of greater significance was the considerable access for Soviet naval reconnaissance aircraft. In Vietnam, alongside the major naval deployments at Cam Ranh Bay, the Soviet Union deployed some 24 reconnaissance or combat aircraft, 8 Tu-95 Bears and 16 Tu-16 Badgers D/K, ten of the latter having strike capabilities.71 The Badgers’ ranges extended the Soviet strike capability over the entire Southeast Asian region, notably including the U.S. bases in the Philippines, but also over Guam and the other U.S. facilities in the islands of the Central Pacific.72 In the Southwest Asia/Indian Ocean area, the Soviet Union, after 1978, was provided access to Yemen’s Aden International Airport and to a military airfield at Al-Anad, for IL-38/May naval reconnaissance aircraft, the Soviet equivalent to the U.S. P-3C.73 (These were transited further south to Ethiopia and Mozam- bique.) In the Mediterranean, there was similar access for IL-38s in Libya after 1981, at Okba ben Nafi, the former U.S. Wheelus Air Force Base. Additionally, there were a large number of Soviet Air Force advisers and maintenance personnel in Libya, whose air force comprised MiG-25s, MiG-23s, MiG-21s, Su-22s and Mi-24 Hind helicopters – as well as Tu-22 Blinder bombers and IL-76 Candid and AN-26 Curl transports.74 Soviet ground force bases75 During the peak of the Cold War, the USSR deployed a massive land army and associated facilities in Eastern Europe. At the peak, that involved some 565,000 troops, organized for combat into 30 divisions (16 tank and 14 motorized rifle), plus attached artillery units. The deployments were as follows: 1 GDR: 380,000 troops; one Group and five Army headquarters; ten tank and nine motorized rifle divisions; one artillery division; one air assault division; five attack helicopter regiments with some 500 Mi-8 Hip and 420 Hind attack helicopters. 2 Czechoslovakia: 80,000 troops; one Group and one Army HQ; two tank and three motorized rifle divisions; one air assault battalion; one artillery brigade, 2 attack helicopter regiments with 100 Mi-8 Hip and Mi-24 Hind helicopters. 3 Poland: 40,000 troops; one Group and one Army headquarters; two tank 136 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 44 divisions, one attack helicopter regiment with 120 Mi-8 and Mi-24 helicopters. 4 Hungary: 65,000 troops; one Group and one Army HQ; two tank and two motorized rifle-divisions; one air assault brigade with 65 Mi-8 and Mi-24 helicopters.76 Outside Europe, the only permanent major (peacetime) Soviet ground-force deployment in an allied country was in Mongolia. There, the Red Army deployed two tank and three motorized rifle divisions, 65,000 troops in all (earlier there were 75,000) vis-à-vis China.77 These forces filled a gap in the Sino-Soviet confrontation line amid a much larger overall Soviet deployment in the Far Eastern theatre of some 53 regular divisions (seven tank, 45 motorized rifle, one airborne), abetted by four artillery divisions and two air assault brigades. The Soviet Union’s other main external ground force was, of course, the large army of some 118,000 troops engaged in combat in Afghanistan, which included 10,000 Ministry of Internal Affairs (MVD) and Committee of State Security (KGB) troops. That force remained, numerically speaking, at a fairly constant level from the initiation of hostilities in 1979, to its conclusion in 1989. There are several other locales where clusters of Soviet-bloc advisers and military technicians were significant beyond the “norm” for standard military missions. These were in Algeria (1000), Cuba (8000), Ethiopia (1500), Libya (2000), North Yemen (500), South Yemen, (2500), Syria (4000) and Vietnam (2500), with small numbers in India, Iraq, Cambodia, Laos, Mali, Mozambique, Nicaragua and Peru. Each was a major recipient of Soviet arms. The much- argued “Cuban brigade” – whether defined as a combat formation or as a collection of support troops – achieved some notoriety in 1979 when publicity over its presence (and arguments about whether it represented a violation of agreements made at the close of the 1962 Cuban Missile crisis) was important to the aborting of the SALT II Treaty by opposition in the U.S. Senate. Soviet forces in Syria were important in the wake of the latter’s debacle in the 1982 war with Israel – they manned the some 48 long-range SA-5s which could have contested Israeli air control even over the Mediterranean in the event of renewed hostilities. Earlier, prior to 1972, there was a large force of Soviet troops – some 20,000 – deployed in Egypt’s Suez Canal area, mostly to man air defense installations. In 1977–1978, some Soviet forces aided Ethiopia in its war against Somalia. Soviet technical facilities abroad78 During the Cold War, the USSR made – relative to the U.S. – much less use of foreign facilities for technical functions – communications, space-related, anti- submarine warfare, nuclear detection etc. This was variously due to much greater utilization of shipboard facilities; the larger (relative to the U.S.) Soviet land mass in relation to the major focus of the superpower competition along the Eurasian rim, which allowed many functions to be performed within the USSR; Bases during the Cold War 137 İ 2007 Robert E. Harkavy

Page 45 the lesser number of aligned client states available to the Soviet Union in the Third World; the practice of utilizing a larger numbers of satellites with shorter lives; and the more open nature of Western societies which reduced the (relative) Soviet requirements for proliferated intelligence facilities. The USSR had nothing, for instance, comparable to the U.S. SOSUS network for tracking submarines, though there was a report in 1978 of a Soviet hydrophone apparatus washed ashore in Iceland. The Soviets relied more on surface ships and perhaps also submarines and aircraft-sown sonabuoys for detection of U.S. submarines. It maintained about 50 auxiliary intelligence ships for ASW work, which maintained a constant presence near important continental U.S. bases such as Charleston, South Carolina; Kings Bay, Georgia; Norfolk, Virginia; Mayport, Florida; and Bangor, Washington; as well as at Holy Loch, Scotland. And, unlike the U.S., the USSR apparently made no use of external communications and/or navigation facilities in connection with submarines on patrol. Several sources reported that communications with submarines stationed at great distances from the USSR were handled by a network of some 26 VLF and LF transmitters within the USSR itself, apparently sufficient to cover the patrol areas of Soviet SSBNs and SSNs; in the former case, most were kept close to home in the “bastions,” or on stations in the Atlantic or Pacific oceans within range of the home communications stations. Ford referred to six long-range radio transmitters (at Petropavlovsk, Vladivostok, Dikson Ostrov, Kaliningrad, Matochin Shar and Arkhangelsk) that gave orders to Soviet submarines. In addition, Arkin and Fieldhouse reported a three-station network of “Omega-type” VLF transmitters at Krasnodar, Komsomolsk and Rostov. They also detailed a considerable number of LORAN-C type “Pulsed Phase Radio Navigation System” stations, organized by chains along the western and eastern littorals of the USSR, used to position submarines. Soviet submarines apparently also received communications from satellites during brief surfacing. And, one source indicated the possibility of Soviet use of command and control submarines for relaying communications to other underseas craft within communications distance. The Soviet global ground network of space-tracking and satellite control facilities was, of course, far less extensive than that of the U.S. Again, this was a function of the far more extensive use of ship-borne facilities as well as of the lesser external needs dictated by the larger land mass of the USSR, particularly in relation to many satellite orbits which allowed direct transmission to the USSR. At the core of the Soviet space-surveillance system was a network of at least 12 sites within the USSR claimed to be “equipped with receivers to measure Doppler shifts in radio signals, tracking radars, and photo theodolites and which transmit data to a central computation center.” Additionally, radars associated with anti-ballistic missiles (ABM) – Pushkino, Hen House, Try Add and Dog House – are said to have had space-tracking capabilities, along with the contro- versial (in the context of ABM treaty verification) radar at Abalakova. 138 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 46 Outside the USSR, there were a number of tracking stations in foreign countries. These were reported in Egypt (Helwan and Aswan) before the Egypto– Soviet break, Mali, Guinea, Cuba and Chad, as well as in Czechoslovakia and Poland. At Santiago de Cuba, for instance, there was an Interkosmos laser radar and also a KIM-3 tracking camera, presumably functionally equivalent to the U.S. Baker-Nunn or GEODSS systems. Perhaps overlapping this grouping, there were reports of an Interkosmos laser tracking program (using a laser rangefinder) involving facilities in Egypt, Bolivia, India and Cuba. It is believed that tracking was carried out at Khartoum in the Sudan and Afgoi in Somalia. But just because the Soviet Union was reluctant to become too dependent on foreign land-based stations, it placed considerable emphasis upon shipborne space-tracking (and also missile-tracking) systems. This involved more than ten ships – a Soviet source noted that even despite the nation’s large land mass, space vehicles were within direct visibility from Soviet territory only for about nine hours out of 24. In the field of communications too, the Soviet Union was far less dependent on foreign land bases than the U.S.; correspondingly, far more reliant on ship-borne systems (an exception, of course was the VLF facilities used to communicate with submarines, perhaps too large to be placed aboard ships). As it is, however, relative to what was known about the U.S., there was scant information on how the USSR utilized the whole of the frequency spectrum for various purposes, how it tried to circumvent countermeasures by redundancy, and so forth. For the most part, Soviet use of overseas communications facilities focused on the downlinks for the Molniya communications satellites, of which there were some 40 aloft at a given time. A U.S. government report provided some information (and a map) on the general locations of Soviet tracking ships and, hence, of the locales where they might have sought port access or at least mooring buoys. Among them, off Sable Island, Nova Scotia; in the western Mediterranean near Gibraltar; the Gulf of Guinea; off Mozambique and Madagascar; off Honduras, east of the Philippines; and north of New Zealand. Large tracking ships apparently moored at Havana and/or Santiago while tracking some flights; Trinidad was also mentioned in this context. Some of the other locations provided rationales for periodic access to Conakry or Maputo, maybe also to Nicaragua, as well as providing further indi- cations of the Soviet need for access somewhere in the South Pacific. (The U.S. government report stated that three large Soviet ships took turns serving in the Caribbean area to extend Soviet deep-space coverage.) The USSR depended considerably less on land-based SIGINT collection stations than the U.S., though, it utilized a variety of means; satellites, surface ships and submarines, aircraft, equipment based in embassies and so on. And again, similar to the situation with respect to space tracking and early warning, one had to be aware of the sometimes only indirect importance of external access; for instance, for fueling ships used in lieu of land facilities. The USSR was reported to have had major SIGINT facilities at Lourdes, Cuba; Cam Ranh Bay, Vietnam; Ethiopia (two); South Yemen; Syria; and Bases during the Cold War 139 İ 2007 Robert E. Harkavy

Page 47 Afghanistan. That at Lourdes was reported to be devoted to interpretation of satellite communications. There may have been additional ground stations in Libya and Iraq, though the fate of the latter might be questioned since Moscow had given only modest support to Iraq in its war with Iran, particularly in the early stages, in part because of the cross-pressures from its relationship with Libya. The Lourdes facility was a large one, operated by some 2000 Soviet personnel; it became a prominent issue in 1979 at the time of the imbroglio in the U.S. over the Soviet “Cuban brigade” amid the SALT II confirmation hearings. The facility of 50 buildings housed an antenna field, satellite receiver and so on, and targeted U.S. civilian and military communications, that is, B-52 communications, Fort Benning and Cape Canaveral, the naval headquarters at Norfolk, etc. According to Richelson, the Lourdes facility, complemented by a similar one in the USSR, “gave complete coverage of the global beams of all U.S. geosynchronous communications satellites.” At Cam Ranh Bay, the Soviet Union had an important facility from which to monitor both land and ocean-based emissions – there were two HF/DF sites used to gain locational data on U.S. fleet units in the Pacific. The U.S. bases in the Philippines were obvious targets. The four sites in Afghanistan aided collection, variously, vis-à-vis China, Pakistan, Iran and the Persian Gulf area. And, of course, these sites merely added to the capabilities of hundreds of SIGINT sites located within the USSR. The Soviet Union had some other land-based intelligence-interception facilities located outside the USSR. In Laos, there was an air-surveillance radar, obviously directed against China. According to the IISS, there was a Soviet monitoring station (elsewhere identified as a radar site) in Sao Tome and Principe. In Cuba, along with the Lourdes facility, there were also air-defense surveillance radars, the Tall King system, apparently operated by Soviet personnel. In East Germany, there were HF-finding antennas used in connection with jamming operations, in Poland, a SIGINT station and HF direction-finder at Sinajscie, and at Bierdzany a receiver site for a Soviet OTH radar transmitter located at Kiev. Throughout Eastern Europe, there were large numbers of air- surveillance radars equivalent to the NATO NADGE system – in Hungary, for instance, some 130 sites manned by Soviet personnel were reported. The Soviet AGIs had near global patrolling areas. There was a concentrated effort to monitor off the southeastern coast of the U.S. (where U.S. surface and submarine units were concentrated), in the English Channel, the Norwegian Sea and off Holy Loch. In the Indian Ocean and Mediterranean Sea, Western naval movements were monitored, particularly near the straits of Gibraltar, Hormuz and Bab El Mandeb, and the Suez Canal. In the Pacific Ocean, there was corresponding emphasis off the coasts of China, Japan, Guam and in the waters around Vietnam. The dependence of these “spy ships” on provisioning from foreign ports was difficult to gauge, but must certainly have been considerable. As noted by one 140 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 48 source, the Soviet Union had chosen “not to build an auxiliary fleet of the size necessary to reduce out-of-area base support to a manageable minimum . . . Ship designs, both for ease of maintenance and for reasons of habitability, still are notoriously poor . . . Unlike U.S. ships, most Soviet ships cannot distill enough fresh water and are dependent upon water tankers.” This situation necessitated frequent operational port visits by auxiliaries to take on food and fresh water, which were then transferred to combatants, presumably also AGIs, at roadsteads or at sea. This presumably further directed attention to the main Soviet basing hosts and clients – Cuba, Angola, Syria, Vietnam, South Yemen etc. – as critical to fueling and otherwise provisioning the Soviet Union’s global AGI effort. Soviet forward-based missiles79 By the mid-1980s, of course, both the U.S./NATO and the USSR deployed large numbers of externally based, nuclear-armed missiles in Europe, constituting a massive and critical forward presence. Centrally, this involved Soviet deployment of short-range theater weapons and tactical weapons in Eastern Europe (SS-12/22s, SS-23s and SS-7 “Frogs”) and a countervailing U.S. deployment of Pershings, cruise missiles and Lance battlefield weapons. The Soviet SS-20 theater missiles which were at the center of the INF negoti- ations were first deployed in 1977. These missiles, carrying three nuclear warheads with ranges of 3400 miles (5000km), were phased in to replace the older SS-4s. The latter were first deployed in 1959, with a single warhead platform and a range of 1120 miles. By 1987, it was typically reported that over 300 SS-20s were deployed against NATO west of the Ural Mountains, with another 100 or so in Soviet Asia, for a total of 441. None of these were based outside the USSR (nor were any of the 112 reported SS-4s which were still deployed in the western USSR). They were based in several fields in the western Soviet Union and near the Caspian Sea. The SS-20s threatened the entirety of NATO-Europe with their 5000-km ranges, as well as many other important targets – the Azores, Green- land, Philippines, Guam, Okinawa, etc. Other nations’ external basing: Britain and France80 Somewhat in the face of historical nostrums associated with Mahan and others, it is apparent that the Cold War competition – if that is what it was – for overseas naval access was largely a two-nation game. In that sense at least, bipolarity unquestionably reigned. There were, nonetheless, a few not altogether insignifi- cant instances of naval basing retained by – or recently acquired by – some other nations: France, Britain, Australia, the Netherlands and maybe others. Of course, almost all navies conduct periodic port visits abroad, variously involving “presence,” solidification of political friendships, broadening of horizons for naval personnel, and so on. Bases during the Cold War 141 İ 2007 Robert E. Harkavy

Page 49 France had the most significant external naval presence besides the superpowers, most notably represented by its Indian Ocean Flotilla (Alindien) of five frigates, three minor combatants, two amphibious and four support ships (also a small naval marine detachment). That force was deployed out of Djibouti (hence was within combat range of the Persian Gulf), also making extensive use of bases at Reunion and Mayotte (Mozambique Channel), both French overseas possessions (earlier, up to 1973, France had extensive access to Diego Suarez in Madagascar). There was also a significant naval presence in the Pacific: five frigates, five minor combatants, seven amphibious and 12 support ships. That force operated out of Noumea, New Caledonia and also patrolled via Tahiti (Papeete), Muroroa and other French dependencies in the southwest Pacific. Two small ships were also normally rotated about between the Antilles (Port Lewis, Guadeloupe) and Guyana. The French Navy also made extensive use of a number of other ports, mostly in closely aligned African nations: most notably Dakar (Senegal), Abidjan (Ivory Coast) and Libreville (Gabon). During the events of 1987, France’s access to Djibouti allowed for a significant naval presence in the Gulf of Oman, outside the Straits of Hormuz. That presence was reported as consisting of three minesweepers, three escorts, one anti-submarine ship, the aircraft carrier Clemenceau and two frigates. Great Britain’s once near astonishing network of overseas naval bases and access had by 1985 dwindled to a very small remnant, aside from still extensive routine port calls by the Royal Navy. Small naval forces were still permanently deployed in Belize, Gibraltar and Hong Kong. A relatively large force, including an ASW carrier, an SSN and several escorts and auxiliaries remained in the Falklands in the wake of that (not wholly resolved) dispute. A small naval detachment (one or two destroyers or frigates and a couple of support ships) moved about the Indian Ocean, making use of Diego Garcia, Singapore, Perth and so on, and had quietly been used to escort ships in the Persian Gulf. That presence, which apparently utilized access to Bahrain, was reported in 1987 to consist of two warships, one fleet tanker, four minesweepers and a supply ship. Otherwise, within NATO, only the Netherlands permanently deployed a tiny naval presence outside Europe – the Dutch retained a small presence at Curacao in the Caribbean. (Spain was reported to have established a small “fishing” presence in Equatorial Guinea after the expansion of the Soviet Navy from Luba.) As in so many other ways, the matter of external basing of aircraft remained during the Cold War primarily a two-nation game, reflective of the tenacious hold of bipolarity which had characterized global basing networks in the post- colonial era. Earlier, of course, the primary colonial powers – Britain, France, Italy, the Netherlands, Portugal, Spain – all had had extensive networks of air bases closely associated with colonial garrisons and with colonial rivalries among the European powers themselves. In the 15–20 years after 1945, these assets gradually dwindled, and in the process the number of bases available to the U.S. and, more generally, to the Atlantic alliance contracted. But there were still some remnants, and in the case of France at least, even the hint of a slight expansion of an overseas air power presence. 142 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 50 Of course, there was some intra-NATO alliance forward basing of aircraft in West Germany; again, associated for obvious reasons with land-force deployments and, indeed, located immediately to the rear, westward of those forces. The British Army of the Rhine was backed up by significant forward Royal Air Force (RAF) deployments, involving 12 aircraft and two helicopter squadrons, six deploying nuclear-capable Tornado strike aircraft (co-developed in a consor- tium with the FRG and Italy), one of Jaguar reconnaissance aircraft and two of Phantom fighters; also among these were two squadrons of Harrier jump jets (nuclear-capable) and one of Pembroke communications aircraft. These aircraft were stationed at several main air bases in northern Germany: Laarbruch, Bruggen, Wildenrath and Gutersloh. Canada’s contribution, further south in Baden Wurtemberg, consisted of forward deployment of three squadrons of 36 CF-18s at a base at Baden Sollingen, supported by liaison aircraft and 2700 personnel. In addition, the West German Luftwaffe had training and some support facilities in Portugal, the U.S., the U.K. and Canada, and had permanently based 18 Alpha jets at Beja in Portugal, mostly in connection with training activities. Elsewhere, overseas, some remnants of what obviously once was a much larger RAF presence remained. There were some aircraft or helicopters permanently stationed at the Falkland Islands, Ascension, Belize, Brunei, Cyprus, Gibraltar and Hong Kong. Of these, the RAF maintained only helicopters and/or utility aircraft in Brunei, Cyprus (Akrotiri) and Hong Kong, though Phantoms and Lightning fighters were sometimes deployed to Cyprus. There were apparently Victor bomber and Hercules C-18 tanker detachments on Ascension, in the late 1980s, no doubt to provide the wherewithal for another logistics operation to the Falklands, if that should have been necessary. There was a helicopter squadron at Hong Kong, as the latter’s reversion to China loomed. In Belize, a lingering point of tension in connection with Guatemalan irredentist aims, the supporting British force included four Harriers and also four Puma and four Gazelle helicopters. Gibraltar still saw occasional deployments of Jaguar fighter aircraft. In the Falklands themselves, the U.K. – to deter another invasion which could have seen an enhanced Argentinian air assault – maintained on station a full squadron of nine Phantoms and Harrier vertical take-off and landing (VTOL) aircraft, Hercules tactical transports and several Sea King and Chinook helicopter detachments. This was almost, quixotically, the largest RAF presence outside of Europe. Overall, the RAF had some 17,000 personnel stationed abroad. The U.K. also deployed some ASW aircraft at overseas bases, supplementing the near-global presence of the large U.S. force of P-3C Orions. The British equivalent was the nuclear-capable Nimrod MR2 aircraft. These were periodically deployed at Wideawake Airfield on Ascension; Kindley Naval Air Station, Bermuda; Akrotiri, Cyprus; Stanley Airfield in the Falklands; Gibraltar; Keflavik, Iceland; Sigonella, Sicily and perhaps also Konya Air Base, Turkey. The French Air Force, again in direct association with army detachments, maintained a fairly significant presence in several African states. These forces had been directly engaged in some local wars where, even in small numbers, Bases during the Cold War 143 İ 2007 Robert E. Harkavy

Page 51 they could be decisive or at least telling because of an absence of counter- weights, at least so long as Soviet or Cuban pilots were not directly engaged. At minimum, they acted as tripwire deterrents – for several Francophone regimes – against local aggression or external involvement. The main points of deployment were in Djibouti, the Central African Repub- lic, Chad, Gabon, the Ivory Coast and Senegal. Jaguar fighter-bombers (co- developed with the U.K.) were deployed in the Central African Republic, Chad and Gabon; Mirage F-1C aircraft were also deployed to Chad. Djibouti had a squadron of ten Mirage IIIs. Alouette and Puma helicopter were stationed in all of these countries, in the Ivory Coast they constituted the only French Air Force presence. There were C-160 Noratlas transport aircraft deployed to the Central African Republic, Chad, Djibouti, Gabon and Senegal. In the latter case, France had based at Dakar and also at Djibouti Breguet Atlantique maritime- surveillance aircraft in areas not far from frequent Soviet naval deployments in West Africa and the Indian Ocean. The Noratlases provided for speedy movement of French or other surrogate forces in case of crisis or conflict. Aircraft based in Chad were earlier military engaged, as were those based in Senegal, which flew missions on behalf of Mauritania and Morocco earlier on during the Western Sahara war. Outside of Africa, France had no permanently stationed combat aircraft, though helicopters and utility aircraft were deployed in Martinique, Guadeloupe, French Polynesia, Reunion and the Mozambique Channel Islands. Otherwise, one can point merely to a few scattered external deployments of aircraft represented by still other members of the Western alliance. The Nether- lands had deployed some P-3Cs through Keflavik, Iceland and through British air bases at St. Mawgan and Machrihanish. Australia had kept two squadrons of Mirages in Malaysia as its contribution to the defense of Southeast Asia. New Zealand up to 1982 had also had a small air presence in Singapore. Concerning technical facilities, the U.K., which deployed SSBNs and SSNs in the eastern Atlantic, had LF transmitters at Bermuda and Gibraltar; the latter was capable of reaching across the Mediterranean to the Indian Ocean (until 1976 the U.K. had naval communications facilities at Mauritius and Singapore, when it also still maintained a naval presence east of Suez). It had another at Port Stanley in the Falklands, which would obviously have been of value in case of resumption of hostilities in that area. It also deployed an RAF/UKADGE early-warning radar at Sornfelli in the Danish-controlled Faeroe Islands. Britain also fielded some additional, scattered C3I assets overseas. It had a major SIGINT site on Cyprus at Pergamos/Dhekelia. Elsewhere, on Cyprus, the U.K. had a troposcatter communications relay, a Skynet satellite-communications terminal and an OTH radar in the Troodos Mountains capable of monitoring missile tests within the USSR. There were reported Government Communica- tions Headquarters (GCHQ) SIGINT stations on Ascension (Two Boats) and St. Helena islands; also at Darwin, Australia (earlier, there were others in Botswana, Aden, Bahrain, Malta, Mauritius, Singapore and on Oman’s Masirah Island). There was a COMINT and HF/DF facility at Gibraltar. Two other SIGINT 144 Bases during the Cold War İ 2007 Robert E. Harkavy

Page 52 stations were located at Hong Kong, a major one at Diepholz in West Germany, along with other signals units at Teufelsberg, Jever, Celle, Dornenberg and Gor- leben. A former U.S.-run nuclear-detection site at Pearce, Australia was operated by the British Atomic Energy Authority. Earlier, there was a communications relay facility on Mauritius. Britain also contributed to the overall Western intelligence effort via some jointly operated facilities. In conjunction with Australia, it operated an ocean- surveillance radar at Hong Kong, once directed against the People’s Republic of China, used to monitor Soviet fleet movements in the SLOCs between Siberia and Vietnam. (Australia and New Zealand jointly operated a similar facility at Singapore.) The U.S. and U.K. jointly operated such a facility at Diego Garcia. These and other such facilities – an Australian installation at Darwin, U.S.- operated bases at Edzell, Scotland and Brawdy, Wales, constituted a global system codenamed Bullseye for direction-finding interception of ships at sea. These and related activities were discussed by Richelson and Ball in the context of the multilateral U.K./U.S. arrangement entered into by the U.S., U.K., Canada, Australia and New Zealand in 1947 in the aftermath of World War II. Summary – Cold War basing patterns The 45-year long Cold War presented in some cases patterns that were reminiscent of previous periods, but others that were entirely new, pertaining both to politics and technologies. Reminiscent of the past was the geopolitical heartland/rimland structure of a bipolar struggle. The U.S. rimland basing structure, increasingly leapfrogged as the period progressed, was similar by degree with the earlier Portuguese, Dutch and British basing networks, calling to mind Thompson’s thesis about system leader lineage patterns. But, the ideological nature of the bipolar conflict (earlier, only the religious divide between Islam and Christianity provides a partial analog) resulted in fairly stable alliance and alignment/clientship patterns over several decades. That translated also into stable, long-term basing relationships between sovereign states that had no discernible historical precedent. The basis for basing in this period was also historically unique, increasingly so as the period progressed. As previous imperial control over much of the world by the European powers collapsed, so too collapsed basing access networks based on that imperial control. In its place, both the U.S. and USSR, largely previously bereft of overseas empires, acquired and maintained access to bases largely via a combination of the provision of security to regional states against the rival superpower and its regional clients and with that, the provision of security assistance in the form of arms transfers, training, economic assistance etc. During the Cold War period, the proliferation of new types of basing access was driven by rapid and profound technological change. Up to the interwar period, the story for bases had mostly to do with surface naval bases and related “forts” or ground force deployments utilized mostly for colonial control. By the Bases during the Cold War 145 İ 2007 Robert E. Harkavy

Page 53 1930s, air bases and those for submarines had become important, so too some initial “technical facilities,” such as those for communications and communications intercepts, early-on radars, terminals for underwater communications cables etc. But later in the Cold War, particularly paced by developments in satellites, there was developed a whole range of new basing requirements: various form of communications, satellite downlinks and control stations, satellite surveillance, sonar submarine tracking networks, early warning for ballistic missile attacks, elaborate networks of radars, nuclear detection facilities and many more. Technologically speaking, military operations and the bases that support them had moved into several dimensions and the relations between them, i.e., naval and land surfaces, the underseas, the airspace and outer space. That is, there were many, many new types of bases in addition to traditional naval bases and army installations. Early on in the Cold War, the U.S. and Soviet basing networks tended to be separate and demarcated, but also rivalrous. As the Soviets leapfrogged the rimland alliance structure set up by the U.S. in the 1950s – Syria, Egypt, Libya, Algeria, China, Guinea, Angola, Mozambique, Ethiopia, South Yemen, India, Cuba, Vietnam – the rival basing points came more to be cheek-by-jowl, inter- penetrated. Scenarios for “protracted conventional phase” warfare came to dwell on possible rival efforts to “pick off” the enemy’s bases and, hence, to tilt the balance of power in global conflict. But, that never came to pass, in part because the dangers of escalation of conflict to the nuclear level precluded a more modern version of conflict “beyond the line,” i.e., outside Europe, that was an earlier tradition. The two superpowers eyed each others’ bases, engaged in a cold war of nerves, but never directly attacked their rivals’ bases