Cover Photo:

326th Communications Reconnaissance Company in Korea in 1951 @] @]

I UNirEV srArESCRYPrOLO(iIC HisroRY I I� I I Ser�V I I �Earl:}'P<>Stwar Per� 1945-1952 Ii I Volume,3 iI I iI � � � The,Korearv War: I iI I I The,SI<.iINT Background iI � � � � � � � David A. Hatch � � with I I Robert Louis Benson I i � I I �I Ii � i I� �I � � I I I CENTER FOR CRYPTOLOGIC HISTORY I I NATIONAL SECURITY AGENCY I I 2000 I I� If§j @] @] This page left intentionally blank Introduction peripheral areas occupied by Japan. One of the thorniest problems was the status of Korea. Since the revelation of the vital role of cryp­ tology in World War II, the contribution of The peninsula had been an independent communications intelligence (COMINT) and nation forcenturies beforethe Japanese took it communications security (COMSEC) in post­ as a colony in 1910. In August 1945, Soviet war conflicts has become a frequent question forces were fighting the Japanese military on for many, particularly scholars and veterans' the China-Korea border, and it appeared that groups. the Red Army might occupy all of Korea.

This short summary of the cryptologic back­ The U.S. solutionwas a temporary division ground to the Korean War is intended to pro­ of the country. Americans would take the vide only a general overview of the conflict Japanese surrender in the southern sector, from a cryptologic perspective and give initial Soviet troops in the north. After a suitable - answers to some of the more important ques­ but undefined - period in which Koreans tions about intelligence support. would be prepared for self-rule, both armies would withdTaw. The Soviets agreed to this This paper has been cobbled together from plan, and Korea was divided on either side of summaries prepared during or immediately the 38th parallel. after the period of hostilities, some original documents, and the memories of some of the However, as the Cold War developed, the participants. Some of the materials on which peninsula became a pawn in a larger, interna­ this history is based may not be declassified by tional ideological struggle. After three years, its publication date (June 2000). I have pre­ the United States turned the problem over to pared the booklet in this unusual manner in the United Nations, which mandated elections order to have a general history in time for the to decide on a unified government in Korea. 50th anniversary of the beginning of the war. UN-sponsored elections led to the formationof the Republic of Korea (ROK) on August 15, This booklet is therefore intended only as 1948, under President Syngman Rhee, with its an introduction to the subject, an interim his­ capital in Seoul. North Korea declined to par­ tory until further declassification allows the ticipate in the UN elections and formed its own fuller history of cryptologic support during the government, the Democratic People's Republic Korean War to be written. I would like to of Korea (DPRK), with Kim II-song as its leader emphasize that much remains to be researched and its capital in Pyongyang. and studied about cryptology in Korea and, in fact, I look forwardto more detailed studies of The next two years were marked by struggle cryptology in the Korean War in the near on many levels - military, political, and ideo­ futuTe. logical. Small unit clashes and armed incur­ sions along the 38th parallel were frequent. The Korean War Both the ROK and the DPRK built military forces, but there was a difference: the USSR When it became clear in mid-August 1945 supplied armor and aircraft to Pyongyang, that Japan intended to sunender, U.S. policy while the U.S. denied them to Seoul. makers began to make arrangements for

Page 1 The USSR, as confirmed by the VENONA est feats of American arms ever, an amphibious decrypts, which NSA released to the public in landing behind North Korean lines at the port 1995-97, had stolen the secrets of the atomic of lnch'on. This operation, combined with a bomb through espionage. Without espionage it breakout from the Pusan Perimeter, smashed is inconceivable that the Soviets would have the DPRK's military forces. had their own atomic bomb 1bythe time of the Korean War. Several U.S. and British spies UN forces, primarily American and South were able to keep the Soviets abreast of U.S. Korean troops, crossed into North Korean ter­ and allied diplomatic, military, and intelligence ritory in pursuit of their retreating enemy, activities well into the Korean War period.1 despite warnings from Communist China to remain below the 38th parallel. In November, The United States deliberately excluded as U.S. and South Korean forces approached South Korea from the defensive perimeter it the China-Korea border, the People's was drawing around the Pacific Ocean area. Liberation Army (PLA) struck them in force, The ROK, said U.S. officials, should depend on sending the UN army in a precipitous retreat the United Nations for support. southward.

Finally, in the early hours of June 25, 1950, In the spring of 1951, UN forces reestab­ the Korean People's Army (KPA) crossed the lished a stable line of resistance with the com­ dividing line in strength and began pushing munist armies at roughly the midpoint of the southward toward Seoul. After some initial Korean peninsula. Both sides entrenched. The resistance, the ROK Army gave way before the Korean War continued for more than two larger, stronger KPA, and retreat became rout. years, but consisted largely of limited offensive operations, characterized by only small gains President Harry S. Truman and his advisers and losses, to capture or defend particular assumed the USSR had directed the attack and points of real estate. The strongpoints were that this was the opening move in a wider war. designated officially by their height in meters, At that point, the U.S. reversed its policy and but known popularly by colorful or poignant intervened militarily to support the ROK. The nicknames bestowed by the Gis who fought U.S. persuaded the United Nations to call for over them - The Hook, Old Baldy, Pork Chop assistance in repelling North Korea's aggres­ Hill, and Heartbreak Ridge. sion, and a number of other UN members sent troops or supporting forces. Once the lines had hardened, truce talks opened. The negotiations were first held in the After a period of retreat, General Walton city of Kaesong, behind communist lines. This Walker, in command of the U.S. Eighth Army, was unsatisfactoryto the UN side, so the meet­ stabilized the lines around a defensible area ings were moved to Panmunjom, an obscure that came to be known as the "Pusan village in "no-man's land." Perimeter." Deployed largely along the mean­ dering Naktong River, Walker moved his forces The war ended in August 1953, after more quickly and astutely to blunt repeated North than three years of combat, with the signing of Korean attacks. a truce agreement and the exchange of prison­ ers. On 15 September 1950, a USMC/Army amphibious force, spearheaded by Marines, During the war and in postwar investiga­ striking according to General Douglas tions, there were many charges that U.S. intel­ MacArthur's plans, conducted one of the great- ligence had failedin the Korean War, not once,

Page 2 but twice. Critics charged that American intel­ SecuTity Group in June 1950. During and after ligence organizations had failed to give warn­ World War II, a portion of Army COMINT ings of the initial North Korean attack in June assets were dedicated to support of the U.S. 1950 and failed again when the Chinese Army Air Corps, and, when the independent entered the war in October 1950. Air Force was created in 1947, these cryptolog­ ic assets were resubordinated to the new Background To U.S. Cryptology organization as the Air Force Security Service (AFSS). In the late 1920s and early 1930s, the U.S. Army and Navy for the first time established Many officials favored centralization of and operated permanent codemaking and cryptologic activities, and in 1949 the codebreaking units. The cryptanalytic units Department of Defense created the Armed expanded at the outbreak of World War II, and Forces Security Agency (AFSA) as a national the enhanced activity paid offin plentiful and organization. AFSA, just by its existence, forced high-quality information on the Germans and the Army and Navy to redefine the organiza­ Japanese - their location, armament, and tion and roles of their cryptologic services. The intentions. Working in close cooperation with cryptologic agencies of all three services began Great Britain, U.S. military and civilian deci­ structuring themselves to provide direct sion-makers got accustomed to detailed inside COMINT support for American fightingforces. information about their enemies (and a few neutrals). It was furthermore intended that AFSA eliminate duplication of effort among the Allied exploitation of the German Enigma Service Agencies and get an economy of scale in machine and other high-level German and research and purchasing. As it turned out, Japanese cryptographic systems is well known. however, AFSA did not have sufficient legal Less known but also invaluable to the war authority to provide central direction to crypto­ effort was U.S. and British exploitation of logic work. front-line systems to provide a wealth of tacti­ cal information on their enemies' activities. Because there had been no advance budget­ ing for AFSA in 1949, its financial needs were These decision-makers expected the same met initially by reductions in the cryptologic inside information after the war, but encoun­ budgets of the armed services. Proposals for tered difficulties creating productive and cost­ increases to AFSA's budget or personnel alloca­ effective organizations. The postwar period tions were not approved. was characterized by contradictory problems - escalating requirements for accurate informa­ In those lean budget times, even general tion, rapid demobilization of skilled personnel, requirements for support met with disfavor. In severe budget cuts, the need for expensive pro­ April 1949, the U.S. Communications cessing machines, and a new adversary. Intelligence Board (USCIB) requested $22 mil­ lion in funds, including 1,410 additional civil­ In the period prior to the Korean War, U.S. ian employees, to expand the COMINT effort. communications intelligence underwent major The secretary of defense returned the proposal structural and doctrinal changes. The Army for additional study. In June, the USCIB, not­ Security Agency (ASA) had shared the national ing that intercept and processing resources COMINT m1ss10n with the Navy's were already overburdened even as the military Communications Supplementary Activity required more support, requested an interim (COMMSUPACT) - which became the Naval supplement of $11.6 million, which would

Page 3 include 705 additional civilian employees; this Intercept facilities in the Pacific region were was endorsed by the Joint Chiefs of Staff. The relatively few. All were directed toward higher proposal, however, was disapproved by the sec­ priority targets, primarily Chinese Communist retary of defense because of budget limitations. activities, but also including the Philippine Huk rebellion. Only by diverting collection This situation changed with the corning of fromexisting ones could they cover other inter­ war. Within a month of the North Korean inva­ cept targets. Customers often gave no specific sion, the Joint Chiefs of Staff approved the guidance to AFSA about target priorities, and, transfer of 244 officersand 464 enlisted men to left to its own devices, AFSA sensibly concen­ AFSA. They also recommended a large increase trated on those of obvious great importance, in civilian positions. In August, the Depart­ primarily the USSR and the PRC. ment of Defense comptroller authorized an increase of 1,253 additional civilian COMINT AFSA directed an expanded effort against positions. fue People's Republic of China in early 1950. This included increased intercept and cryptan­ Given the administration's belief that the alytic study. conflict in Korea was merely a part of what could soon be a wider war, only a portion of the ASA in the post-World War II period had increase would go to direct support of the con­ broken messages used by the Soviet armed flict in Korea. But the increase would allow sig­ forces, police and industry, and was building a nificant expansion of the effort to support the remarkably complete picture of the Soviet war. national security posture. It was a situation that compared favorably to the successes of Cryptologyin the Korean War World War II. Then, during 1948, in rapid suc­ cession, every one of these cipher systems went National Priorities dark. Although the loss of these systems occurred over several months (and none hap­ The Monthly Intelligence Requirements pened at the end of a week), U.S. cryptanalysts issued by the U.S. Communications Intelli­ tended to lump the disasters together under gence Board reflectedthe generally low level of the dire designation "Black Friday." government interest in information on Korea. The country was, after all, outside the U.S. Soviet intelligence had had an agent inside defense perii:neterin the Pacificregion. AFSA who had revealed the extent of U.S. pen­ etration of Soviet cipher systems. This was USCIB maintained two requirements lists. William Weisband, who had been recruited by The first consistedof subjects of "greatest con­ the KGB in 1934. During and after World War cern to U.S. policy or security," such as "Soviet II, Weisband was involved in the U.S. COMINT intentions to launch an armed attack." On the efforts, working (as a native speaker of second list were items of "high importance"; for Russian) in the Russian section in ASA and, the month prior to the war, Japan and Korea later, AFSA. Although in 1950 the FBI uncov­ were item number 15 on the second list, but ered information alleging espionage activities this did not focus on Korea itself. The specific by Weisband in the early 1940s, he was never requirements were "Soviet activities in North charged with espionage - Weisband lost his job Korea," "North Korean-Chinese Communist with AFSA and served a year in prison for con­ Relations," and "North Korean-South Korean tempt of a grand jury.2 relations, including activities of armed units in border areas."

Page 4 ,;.. .

The U.S. cryptologic agencies took steps to went from the USSR to North Korea and recover, but this dreary situation continued up Manchuria, starting in February 1950. to the Korean War, denying American policy­ makers access to vital decrypts in this critical These two actions made sense only in hind­ period. This was perhaps the most significant sight, after the invasion of South Korea intelligence loss in U.S. history. occurred in June 1950.

Some North Korean communications were COMINT, supported by information from intercepted between May 1949 and April 1950 other open and secret sources, showed a num­ because the operators were using Soviet com­ ber of other military-related activities, such as munications procedures. Coverage was VIP visits and communications changes, in the dropped once analysts confirmed the non­ Soviet Far East and in the PRC, but none was Soviet origin of the material. These messages, it suspicious in itself. Even when consolidated by should be noted, were not positively identified AFSA in early 1951, these activities as a whole as originating from the DPRK until after the did not provide clear evidence that a significant war began and there was a basis for compari­ event was imminent, much less a North Korean son with confirmed Korean traffic. invasion of the South.

In April 1950, ASA undertook a limited In 1952, when personnel levels and a more "search and development" study of DPRK traf­ static war allowed some retrospective analysis, fic. Two positions were assigned intercept of AFSA reviewed unprocessed intercept fromthe internal North Korean communications, and June 1950 period. Analysts could not find any approximately 200 messages were on hand at message which would have given advance the time the war began, although none had warning of the North Korean invasion. One of been processed. the earliest, if not the earliest, messages relat­ ing to the war, dated June 27 but not translat­ COMMSUPACT monitored communica­ ed until October, referred to division level tions from some North Korean naval bases in movement by North Korean forces. the period before 1950, but only because these were occupied by the Soviet navy. The Initial Responses

As it happened, prior to 1950 there were The outbreak of the Korean War spurred two COMINT hints of more than usual interest significant increases in funding and personnel in the Korean peninsula by communist bloc for the U.S. national security establishment as nations, but neither was sufficient to provide a whole. The cryptologic agencies were no specific warning of a June invasion. In the exception. spring of 1950, a Soviet network in the Vladivostok area greatly increased its targeting In June 1950, prior to the beginning of the of communications in South Korea. Soviet tar­ war, AFSA had had the equivalent of two per­ geting of South Korea was quite low until early sons working North Korean analysis, two half­ February, then rose dramatically afterthe 21st. time cryptanalysts and one linguist. By This coverage continued at a very high level November 1950, AFSA had thirty-six people on until 15 May, when it ceased altogether. the problem, forty-nine by early 1951, and eighty-seven people by March 1953. Prior to In the second case, as revealed in CO MINT, the Korean War, AFSA employed eighty-three large shipments of bandages and medicines analysts against the PRC; by November 1950,

Page 5 the number was 131, and by February 1951 it considered the best prepared of the existing was 156, plus additional part-time assistance. tactical ASA units in the Far Eastern Command, was dispatched to the war zone. It AFSA's chief of the Office of Operations, did not arrive, however, until early October, Captain Redfield Mason, USN, put the process­ three months after hostilities had commenced. ing units for Korean materials on a twenty­ four-hour basis. He also enhanced other oper­ Initial COMINT product was the result of ational areas that might produce information plain text intercepts and trafficanalysis. In fact, to support the war. at many points in the conflict, traffic analysis, that is, the examination of message externals, All available intercept positions in Japan often constituted the only form of COMINT for were redirected to Korean collection. Some Americans. Because of problems with moun­ Navy intercept operators in Japan worked with tainous terrain, there was no steady or reliable ASA Far East on Korean collection. Even the information from direction finding (D/F), 50th Signal Service Detachment, whose mis­ which had been an important source of intelli­ sion was to monitor U.S. forces to ensure com­ gence in World War II. munications security, was diverted to wartime support. CO MINT production was hampered by sup­ ply shortages, outmoded gear, a lack of lin­ A small advance ASA unit arrived in Korea guists, difficulties in determining good inter­ in mid-September 1950 and was assigned to cept sites, and equipment ill-suited to frequent combat support. The 60th Signal Service movement over rough terrain. Until late in the Company, based in Fort Lewis, Washington, war, most ASA trucks and radio receivers were

Two Gls using the M-209 to secure U.S. communi­ cations

Page 6 of World War II vintage. The U.S. military As soon as war broke out in Korea, AFSS fought much of the Korean War with equip­ headqua1ters in Japan dispatched a team to ment from the LastBig One; it is not surprising recruit trusted South Koreans and establish a that intelligence organizations also would have forward suppo1t unit in Korea. They found that to make do. U.S. Air Force personnel in Seoul had already appropriated the services of a South Korean In the earliest period, the intelligence pro­ COMINT unit (see below), the unit was already duced was not appreciated by Eighth Army productive, and there was nothing forthe team officers. COMINT produced by AFSA or ASA to do. The crew returned to Japan. was subject to restrictions on distribution which prevented full exploitation of the infor­ During the Korean War, the Naval Security mation. Officers preferred COMINT produced Group primarily monitored the activities of locally by Korean units which they took into Soviet forcesin the Far East. Since the Soviets service (see below). The distribution problems intervened in the war with air support, there were slowly rectified. remained the possibility they might deploy ground or naval forces, or take advantage of COMINT units moved in support of the U.S. preoccupation to seek advantage else­ fighting forces. The First Communications where in Asia. The possibility of Soviet inter­ Reconnaissance Company, which had arrived vention seemed great in the first days of the in Korea in October 1950, advanced into North war, when elements of the South Korean Navy Korea with the Eighth Army. It became one of fired upon a Soviet auxiliary vessel from the last American units out of Pyongyangdur­ Vladivostok. ing the Chinese counteroffensive. It eventually established a location for itself in Seoul in early Navy intercept also monitored Soviet reac­ 1951. tions to U.S. ferret flights in the North Korean region. This information was passed to U.S. Air The Air Force Security Service (AFSS) also Force units. responded rapidly to the crisis in Korea. The AFSS had only been in existence less than two The Marines who fought in the Pusan years, and had concentrated primarily on orga­ Perimeter, landed at Inch'on, and advanced nizational and doctrinal development; its field deeply into northeast Korea did not have their activities were minimal. It had two mobile own tactical COMINT suppmt. Although sen­ squadrons; the one in the Far East concentrat­ ior commanders likely had access to COMINT ed almost exclusively on targets in the USSR available at higher headquarters, it appears, and was configuredto provide early warning of pending further research, that COMINT did hostile activity rather than provide tactical sup­ not filter down to the Marines who moved port in time of war. northward.

AFSS instn1ctions to its headquarters in The Marine Corps had deployed tactical Japan on June 25 were to devote hvo intercept COMINT units for combat intercept late in the positions to the air activity in the conflict in Pacific campaigns of World War II, but these Korea and increase reporting. On June 27, were demobilized or "downsized" after the war. AFSS instructed its field office to "go into a full A Marine Radio Company, trained for war alert status," with special attention paid to COMINT support in wartim1=,was in existence Soviet actions, particularly any Soviet move­ at Camp Pendleton in 1950, but was not ment against Japan. deployed to Korea because it lacked equipment and was not considered combat ready.

Page 7 The Korean campaigns led to improve­ At the beginning of the war, not only lin­ ments in the 1950s. A USMC study of its guists were in short supply, so were Korean­ Korean War experience recommended language dictionaries. For that matter, no dic­ enhancement of the Corps' tactical COMINT tionaries had listings of North Korean military capabilities. This was done in the years follow­ terms - few linguists knew them, either. ing the war. Working aids were developed over time by con­ textual analysis and by comparison with The Language Problem Japanese and Chinese cognate words.

At the time the war began, only two Korean One solution to the problem for U.S. forces linguists were available to the Army Security was to attach South Korean COMINT units Agency, Youn P. Kim and Richard Chun, both which had lost their parent organizations dur­ assigned to the Army Language School in ing the disorganization in the early period of Monterey, California. Y.P. Kim was from the war. The U.S. Army sponsored an ROK California, the son of Korean immigrants, Navy unit, known fromits leader's name as the while "Dick" Chun had grown up in Hawaii. Kim Unit. The U.S. Air Force Security Service Both had served in World War II and had been sponsored a similar unit from the ROK Air hired by ASA initially because of their Japanese Force, known also from its leader, the Cho language abilities. Unit.

Neither possessed a security clearance in The Kim and Cho Units worked to support June 1950. Y.P. had served as a cryptanalyst the U.S. military for the duration of the war; and translator for the Army at Arlington Hall they were given rations and military supplies in prior to 1945, but had relinquished his clear­ exchange for intercept and translation work. ance when assigned to occupation duty in The Americans at first drew on these units as Japan in the late 1940s. Dick Chun, as a trans­ language resources, but soon were impressed portation sergeant in the Hawaiian National with their discipline in collection and often Guard - with service in Italy and the South were pleasantly surprised by their cryptologic Pacific - had never had a security clearance capabilities. and, in fact, knew nothing about communica­ tions intelligence. It was not until mid-1951, a year after the outbreak of war, that larger numbers of Korean Dick Chun was sent to ASAP AC, first in linguists arrived from the Army Language Japan, then on to the Korean Peninsula. He School. The problem of linguists, however, was tipped ASAPAC off to Lieutenant Y.P. Kim, never adequately solved. who was the more experienced linguist, so when Y.P. arrived in Tokyo, expecting to join The Pusan Perimeter General Dean's headquarters, he was diverted instead to ASA. (This was personally fortunate: Although UN forces equaled or outnum­ the North Koreans mauled General Dean's bered the North Koreans by July 1950, forces, and the general himself was taken pris­ American and South Korean troops were oner). spread thinly over a wide defensive perimeter. Since the KPA could choose the place and time AFSA began a vigorous training program in of its attacks, initially it could put local superi­ the Korean language. A few linguists, stimulat­ ority of force into any battle. More than once ed by the emergency, taught themselves the the North Koreans came close to breaking language. through American lines.

Page 8 At this crucial stage, COMINT identified Milton Zaslow, exploited Chinese civil commu­ North Korean airfields, including timely infor­ nications, i.e., general traffic which included mation on their construction and the disposi­ personal cables as well as unencrypted official tion of aircraft, located distribution centers for messages. This effortwas to assume unexpect­ artillery ammunition, and reported the status ed importance as the possibility of Chinese of the North Korean supply system. COMINT entry into the war loomed. often gave General Walton Walker adequate warning of KP A movements, allowing the Based on translation and analysis of Eighth Army Commander to move his own Chinese civil communications, in July 1950 troops to meet threats. AFSA reported that elements of the Chinese Fourth Field Army had moved from Central As UN forces pressed the KPA northward, China to Manchuria in April and May. In early COMINT followed the progressive collapse of September, AFSA, again basing its reporting on Army and other networks and the relocation of Chinese civil communications, stated that the many air operations. By late October, air force PRC had continued to deploy major militaiy and coastal defense network activity had been units from southern or central China to reduced to "callups." Police networks handled Manchuria. Throughout September and almost all the rest of DPRK government and October AFSA noted continued movement of military communications. It appeared from these and additional forces toward the Sino­ COMINT that Supreme Headquarters had Korean border areas. moved to Sinuiju, near the Manchurian border. A message datelined Shanghai in mid-July The Chinese Enter the War identified General Lin Piao as the commander of Pl.Aforces which would intervene in Korea. ASA and then AFSA had worked the Messages of late September 1950 told how Chinese Communist "problem" since 1945, but, Chou En-lai, the PRC foreign minister, had prior to the intervention of the People's warned diplomats from neutral nations -that Liberation Army in the Korean War, the effort the Pl.A would intervene in Korea if UN forces was largely confined to traffic analysis. With crossed the 38th parallel, the original dividing lives at stake, the need to provide new kinds of line between the two countries. In a radiotele­ tactical and strategic information, and provide phone call, an East European reported from it faster,led to greatly enhanced effort. Beijing in early November that orders had been issued allowing every Chinese soldier to volun­ The cryptologic services had begun enhanc­ teer to fightin Korea, saying, "we are already at ing coverage of mainland Chinese targets fol­ war here." lowing the establishment of the PRC in October 1949. In March 1950, USCIB authorized an It has long been known that military and increase in collection and processing against intelligence officials, in possession of consider­ PRC communications, but it required nearly able warning fromnon-COMINT sources (usu­ two years to develop effective processing of ally referred to by COMINT-cleared readers as PRC military messages. "collateral"), decided either that the PRC was bluffing or that it did not matter, because the Although it took time to develop capabilities time when Chinese intervention could be effec­ against the People's Liberation Army (Pl.A), tive had passed. The reactions of U.S. leaders in one area of exploitation was available. In the Washington, Tokyo, and Seoul to the COMINT postwar period, a team of Chinese linguists and indications of PRC intervention are not known analysts at AFSA, under the leadership of Mr. at this writing, and deserve closer study.

Page 9 The PLA forcesin North Korea attacked UN would be no break in coverage. The second units on 25 October, then unaccountably broke team spent three days in the winter weather off contact for a month, despite large concen­ retreating southward to the capital city, then trations of Chinese troops in North Korea and further south as the Chinese pushed U.S./UN along the border. Many believe that the initial forces back. attack was a warningto UN forces to pull back. COMINT does not shed light on this question. Traffic analysis enabled U.S. analysts to fol­ low the reestablishment of North Korea's However, COMINT in the month between armed forces, first as they centralized at the firstChinese attacks and their all-out offen­ Pyongyang, then as headquarters moved else­ sives which began in late November showed where. By early January, U.S. analysts believed additional movement of Chinese troops toward the North Korean Air forcehad achieved oper­ Manchuria. Messages in November continued ational levels comparable to those in existence to show Beijing in a state of emergency. before it evacuated Pyongyang in October 1950. Messages from PRC civil communications of early to mid-November disclosed an order UN forces recovered from the Chinese drive for 30,000 maps of Korea to be sent from and by spring 1951 had reestablished a defensi­ Shanghai to the forces in Manchuria. Officials ble line in the waist of the Korean peninsula. in the U.S. Army's Milita1y Intelligence calcu­ The line of battle remained there for the rest of lated that many maps would supply thirty divi­ the war. sions. In late November, the PLA attacked U.S. and allied forceswith thirty divisions. The Stalemate

COMINT reports of early July 1950 noted As the line stabilized in mid-1951, and that the Soviet air forces had established a COMINT support became more institutional­ communication net in China to serve military ized, ASA headquarters was established in the and civilian aircraft at airfields in Korea and western suburbs of Seoul on the campus of Manchuria. After March 1951, intercept Ewha College, the largest women's school in showed Soviet control of fighter activity in the Asia.3 The college was selected because it was northernmost regions of Korea, as communist on the periphery of the capital, had Western­ aircraft challenged UN air operations. style buildings, and had suffered comparative­ ly little damage in the liberation of Seoul. Other While in Pyongyang, General Walton COMINT activities were conducted in the Walker visited an ASA Company (where First neighboring Choson Christian College (now Lieutenant Dick Chun mistook him in the dark Yonsei University). for another lieutenant - Walker made light of the error). As Chun remembered, Walker was Advance warning of impending attacks intrigued to read raw intercept; he requested often was derived through analysis of commu­ that it be sent to him instead of the summaries nications associated with PLA artillery prepa­ he had been receiving. rations. Much of the initial reconstruction of the PI.A's order of battle (OB) came from traf­ The 60th ASA, which was accustomed to fic analysis. In May 1952, intercept of a plain­ operating out of trucks, mobilized its vehicles text message allowed ASA analysts to recon­ and departed Pyongyang ahead of the Chinese. struct an almost complete PLA OB. One team was sent ahead to Seoul to begin operations before the other departed, so there

Page 10 In one instance, COMINT answered an rates according to their job and proximity to important OB question. Senior commanders of combat. This rotation policy included ASA per­ the Eighth Army wanted to confirm reports sonnel and created a constant need to find that the Chinese 40th Army, composed largely replacements and conduct training for linguists of combat veterans, had crossed into Korea as well as other specialties. during the winter of 1951. The presence and location of the PLA unit were confirmed by a Despite the problems, COMINT production North Korean message reporting that farmers continued and was appreciated. In encourag­ had complained about soldiers from the 40th ing measures to enhance linguistic support, the Army stealing their rice! It is believed that this Far Eastern Command told AFSA, "Korean message was couriered immediately to the COMINT remains outstanding intel [sic] Eighth Army commander, General Matthew source here for MacArthur and Ridgway." Ridgway. As the war settled into relatively static front The entrance of the Chinese armies in lines and truce talks began, both KPA troops Korea renewed the language problem for the and the People's Liberation Army improved COMINT units. Neither ASA nor AFSS had their communications security procedures. enough Chinese linguists. AFSS began training This resulted in a significant decrease in the airmen in Chinese through a program at Yale quantity and quality of informationavailable to University. UN commanders, although the flow never ceased entirely. It might seem that the large Chinese popu­ lation of the United States would be a natural COMINT supported the UN Command source of linguists for ASA, but this did not when truce negotiations began in July 1951. work out as hoped. One particular problem was For example, intercepts helped identify North the difference in dialect between PLA radio Korean personalities who were participating in operators and American-born Chinese. Some the initial talks in the city of Kaesong. The sup­ help in intercept and translation was obtained port from these communications included by hiring a limited number of Chinese summaries of meetings and communist propa­ Nationalists fromTaiwan as Department of the ganda statements. Army civilians. Some special training was needed here, also, due to differences in military Some reports concerned the frontline situa­ vocabulary between the Communists and the tion and routine administrative matters. Nationalists. Typical was a message to General Nam Il on defensive activities by UN forces. The new war in Korea in 1951 was actually a Sino-Soviet intervention. Soviet pilots fought On the other side of the world from the in the skies over North Korea, although no combat zone in Korea, COMINT assisted the Soviet infantry were committed to the conflict. war effort by exposing Soviet spies in key posi­ This created the need for Russian linguists with tions. In Washington and London, Donald the ability to intercept tactical communica­ Maclean and Guy Burgess, British diplomats, tions. These individuals also were in sh01t sup­ and some colleagues were able to provide the ply in AFSA and AFSS in 1951. Soviets with detailed information from the highest levels about U.S. atomic bomb stock­ Once the battle line stabilized, U.S. forces piles, U.S. and British policy prior to the instituted a rotation policy under which sol­ Korean War, war plans, and - perhaps most diers earned "points" for service at variable important of all - the restrictions on U.S. com-

Page 11 manders in Asia which prevented them from extensively in World War I.5 UN fo rces in carrying the war to Soviet or Chinese territory. Korea commonly planted sound detecting devices forward of their bunkers to give warn­ Maclean was exposed when cryptanalysts ing of approaching enemy troops; it was found working on the VENONA project recovered that these devices also picked up telephone and translated enough messages about his calls. This "ground-return intercept," using the work to identify him. Harold "Kirn" Philby, a principle of induction, enabled collection of co-conspirator with access to VENONA, some Chinese and Korean telephone traffic. warned him and Burgess; the two then fled to the USSR. The bad news was this intercept had to be conducted much closer to enemy positions Despite the failure to arrest the conspira­ than normal intercept, sometimes as close as tors, the leak of vital policy and intelligence thirty-five yards. This risk was assessed care­ secrets was stanched. 4 fully and accepted.

COMINTInnovations Ground-return intercept (GRI) gave UN forces access to information on Chinese or The decline in more traditional methods of North Korean patrols, casualty reports, supply COMINT production forced the services into problems, and evaluations of UN artillery trying new ideas, or, in one case, reverting to an strikes. older one. One colonel who participated in the GRI In late 1951, in conditions reminiscent of program was heard to remark that the infor­ France in 1917, ASA personnel inadvertently mation was so well appreciated by his soldiers rediscovered an intercept technique used that he had little trouble getting volunteers to

Forward site, 326th CRC, Chunchin, 1951

Page 12 go out at night and implant the equipment to reference source on language problems and OB make intercept possible. questions.

A second innovation in COMINT produc­ Afr Fo rce Suppo1·t tion became one of the foremost producers of tactical intelligence for the U.S. military. This The Air Force Security Service continued was low-level intercept (LU). support to the air during the period of stale­ mate. The AFSS also adopted a number of Low-level teams initially consisted of an innovations to provide new kinds of support for officer, driver, and one to three opera­ the air war. tors/translators working out of a jeep; over time the number of operators increased. A good example of AFSS support occurred Although the mobile operations were produc­ in June 1951. Analysts at an Air Force intercept tive, the jeeps were considered too vulnerable, site were able to accurately predict a North and operations were "dug in" in bunkers near Korean bombing raid on UN-held islands. This the main line of resistance, as it was then intelligence enabled the commanding general called. The product was disseminated directly of the U.S. 5th Air Force to ensure that the raid to combat units, usually at regimental level. was met with ample defense.One YA K and two IL-10 bombers were downed, several others The first attempt at front-line LU in July were damaged, and two MiG fighterswere also 1951 proved only partially successful, but, after damaged. It is believed that the commander of some changes in equipment, the program the 5th Air Force may have been aware of the began in earnest in August. Seven LLI teams impending raid before the commander of the were fielded by November 1951. By the follow­ North Korean attacking unit had received his ing May, ten LU teams were in operation, with orders. planning for more. The success of the program is attested by the fact that by October 1952, fif­ In late April 1951, AFSS personnel inter­ teen LU teams were at work, and by the end of cepted messages that indicated aircraft of the the war, twenty-two LLI teams were active. 4th Fighter Squadron were being boxed by Soviet aircraft. The quick relay of this informa­ It was estimated that the tactical value of tion to the flight enabled it to avoid the trap. LLI product lasted from twenty minutes to This kind of warning continued through the three days at best - but, however perishable, it war. paid off. In early September, units in the U.S. 1st Cavalry Division area successfullyrepelled a Soviet Bloc Air defense doctrine called for heavy attack by the PLA. One important ele­ control of local fighter pilots by their tower. ment in this victory was the advance warning These ground control intercept (GCI) commu­ given by the 1st Cav's LLI team. nications were vulnerable to eavesdroppers.

Because the LU teams dealt in perishable At various periods during the air campaign, and current intelligence, not much long-term COMINT units from the AFSS were intercept­ analysis was done - or possible. It thus became ing North Korean, Chinese, or Soviet instruc­ difficult to keep continuity on opposing units. tions to their pilots. These were disguised as These problems were eased somewhat with the "radar plots" and forwarded in near-real time creation of an LLI "control section" at ASA to U.S. pilots operating over North Korean ter­ headquarters in Seoul in late 1951. This section ritory. When this source was exploited, the U.S. collated reports from the fieldand service as a "kill ratio" over MiGs was quite high; during

Page 13 periods of nonexploitation, the ratio was much to well-trained troops, skillfully employed, and lower. backed by coordinated air, armor, and artillery support, demonstrated what might be accom­ Monitoring of North Korean, Chinese, and plished on defense."6 Soviet air communications was done from lis­ tening posts in South Korea, but there were The standard history could not say how hearability problems for certain areas at differ­ good the intelligence really was. ent times of the day. To solve these problems, in mid-1951 the AFSS established an intercept Warning of the battle came first from an site on Pyongyang-do - commonly known as intercepted Chinese message, prompting ASA "P-Y-do" by Americans - a UN-held island to establish a field intercept site for tactical close to the west coast of North Korea. Since COMINT during the battle. Prior to the battle, this was close to enemy territory,the security this site identifiedthe Chinese units assembled regulations had to be relaxed,and 5th Air Force for the attack, then accurately predicted the had to provide special evacuation service. date and time of the firstwave. Low-level inter­ Eventually the Americans abandoned their cept kept the UN forces informed of the loca­ effort on the island. tion of Chinese units during the battle, and artillery fire was targeted on the basis of Once this activity on P-Y-do proved suc­ COMINT. cessful, in the spring of 1952 a similar opera­ tion was undertaken on Cho-do, a UN-held Hill 395 changed hands several times, but, island offthe east coast of Korea, near Wonson. at the end, the ROK Army 9th Division held it. Lieutenant Delmar Lang organized teams of According to ROKA legend, the victors thought linguists and personnel from the Tactical Air that war's devastation had reshaped the hill to Control Center to providenear-real-time infor­ look like a White Horse, and Hill 395 acquired mation to pilots operating over North Korea. tl1e nickname by which it is best knm,vn. Del Lang, by the way, used this operation as a model for similar activity during the Vietnam In March 1953, intercept revealed Chinese War. planning for offensives at Old Baldy and Pork Chop Hill, two UN-held positions in central Security Service also conducted airborne Korea. COMINT revealed troop movements collection operations. In addition to support of and buildups several days prior to the attack. the war effort, these flights were useful in test­ On "D-Day" itself, a low-level intercept gave the ing intercept equipment and general concepts defenders warning that the attack would com­ of operations. mence in fiveminutes. During the battle, inter­ cept continued to provide informationfor U.S. War's End decision-makers.

Hill 395 was located at a strategic point in Similar warnings and battle COMINT were central Korea, and its loss would have endan­ received concerning the all-out attack on Pork gered other UN positions in that region. In Chop Hill in July 1953. October 1952, the Chinese attacked the hill, which was defended largely by ROK Army COMINT provided warning of the final units, with U.S. artillery and air support and Chinese Communist Forces offensive of the French infantry. A standard history of the war, at Kumsong, directed primarily against Korean War notes that "pre-battle preparation, ROK positions. made possible by effective intelligence, added

Page 14 DET D , 3 30'" I ! f -.....- ._.

A detachment of the 330th Communications Reconnaissance Company

The Transition to NSA Despite excellent CO MINT support to com­ bat units, senior commanders, particularly AFSA's role in the Korean War was several­ those who had enjoyed access to COMINT in fold. It first did sophisticated processing not World War II, were dissatisfied. AFSA, which available in the field, conducted long-term was supposed to be setting and maintaining studies, and provided technical assistance. standards of performance, came in formuch of AFSA also coordinated officialtasking with the the blame. General James Van Fleet, com­ services to eliminate duplication of effort in mander of the U.S. Eighth Army, one of the collection and analysis. Finally, AFSA worked principal ground units in the war, put it in writ­ with the services to standardize cryptologic ter­ mg: minologyand reporting vehicles. It has become apparent, that during the The central cryptologic organization had between-wars interim we have lost, through neg­ been established in May 1949, just a year lect, disinterest and possibly jealousy, much of the before the Korean War began. AFSA's rapid effectiveness in intelligence work that we response to wartime needs showed the profes­ acquired so painfully in World War II. Today, our sionalism of its people and its commitment to intelligence operations in Korea have not yet the war effort. However, it should be remem­ approached the standards that we reached in the bered that at the outbreak of war in Korea, finalyear of the last war. AFSA was still in the process of sorting out its relationship with the Service Cryptologic Much of this dissatisfaction centered on Agencies. It is not surprising the relationship AFSA. At the same time, the senior officials of was somewhat uneasy and had some duplica­ the State Department and the Central tion of effort or other inefficiencies. Intelligence Agency also felt AFSA was less

Page 15 responsive to their needs than it should have a sweeping reexamination of doctrine by the been. service agencies, followed by changes to struc­ ture and procedures. Dissatisfactionover AFSA's performance in the Korean War was not the only reason for the Even though fiveyears had passed since the decision to reorganize American cryptology, highly successful cryptologic activities of the but it clearly constituted one of the major fac­ Second World War, little modernization had tors. occurred in tactical support. In a time of lean budgets, priorities were given to development Based on the perceived problems, President and deployment of sophisticated machine sys­ Truman created a committee, headed by New tems for making ciphers and for breaking York lawyer George BrowneU, to study the them. question of proper CO MINT organization. The Brownell Committee Report, submitted in In the short term, after June 25, 1950, all June 1952, noted that four cryptologic agencies had to scramble to provide what the fightingman needed. At the AFSA is dependent on the services forall of its beginning of the war, this was done through the direct interception of CO MINT ... and on Service tested methods from the Second World War. communications for all of its communications As more was needed later in the conflict, mili­ channels. However, none of the three Service tary and civilian analysts innovated and discov­ units is subject to AFSA control, except for the ered new ways to skin old cryptologic cats. intercept positions under AFSA's 'operational direction' by negotiated agreement,and AFSA has The Air Force Security Service was also a no power to compel elimination of duplication of relatively new organization, but set a high stan­ effort between them or to restrain them from dard for support to the effort in its firstwar. engaging in activities thatcould better be central­ ized in AFSA itself... .7 All the SCAs, plus AFSA, benefitedfrom the presence of a nucleus - in some cases, a large The Brownell Committee suggested that the nucleus - of cryptologists who had seen service creation of AFSA could be seen as a "step back­ in World War II. They knew what was expected ward," and recommended that the power of the of them and worked hard to deliver it. director, AFSA, to centralize COMINT be increased. In the long term, all four cryptologic agen­ cies were the beneficiaries of the significant In October, Harry Truman authorized a budget increases to all sectors of the national reorganization and renaming of AFSA, and in security apparatus which the Korean War November, the secretary of defense authorized engendered. And each of the agencies profited the replacement of AFSA by the National from the experience in this war. Security Agency. The cryptologic agencies relearned the tech­ Conclusions niques and skills developed during World War IJ. They also observed and revised their opera­ The Korean War affected the U.S. crypto­ tional doctrines based on the needs of limited logic community in profound ways. When the war. war began, the U.S. government had just estab­ lished its first central cryptologic organization, Finally, perhaps a little sooner than officials the Armed Forces Security Agency. This forced might have liked, the concept of a centralized

Page 16 cryptologic agency was tested under the direst Notes of conditions. The policymakers were able to 1. These and subsequent paragraphs on espi­ observe strong points as well as weaknesses in onage were written by Robert Louis Benson. For fur­ AFSA and, eventually, create a newer, more ther information on Soviet espionage, see Verne W. effectiveinstitution. Newton, The Cambridge Sp ies (New York: 1991) or Alan Weinstein, The Haunted Wo od (New York: As Admiral Joseph Wenger, one of the 1999). architects of centralized cryptology, put it, VENONA is a coverterm of unknown deriva­ I firmly believe that had it not been forthe tion for the ASA/ AFSA project to decrypt Soviet invaluable experience we gained under the joint espionage communications. The Soviet communica­ coordinatingplan in effectprior to the creationof tions occurred during World War II and were AFSA and in the operation of the latter agency, we exploited in the late 1940s and early 1950s. They would have bad far more trouble in solving the provided initial clues that allowed U.S. law enforce­ early problems incident to NSA's establishment ment authorities to arrest a number of important than was actually the case. At the beginning of Soviet spies before and during the Korean War. NSA's existence, we at least knew, fairly certainly, what had to be done. 2. Robert Louis Benson and Michael Warner, VENONA : Soviet Espionage and the American As in all of America's wars, however, the Response, 1939-1957 (NSA/CIA Publication, 1996). story of cryptologyis not the story of brilliance in collecting or processing messages, nor even 3. Interestingly, during World War II, both the about determining the most effective organiza­ Army and Navy cryptologic organizations estab­ tion. The purpose of wartime cryptology is to lished their headquarters in former girls' schools. support the nation's objectives and to save American lives. As in World War II, cryptology 4. Yuri Modin, My Five Cambridge Friends in Korea accomplished these critical goals, and (New York: Farrar Straus Giroux, 1994); Verne W. its contribution to the Korean War still shines Newton, The Ca mbridge Spies (New York, 1991). and inspires. 5. For the World War I application of this inter­ cept technique, see Ernest H. Hinrichs, Listening In: Intercepting German Communications in Wo rld Wa r I (Shippensburg, Pa.: White Mane Books, 1996).

6. Walter G. Hermes, United States Army in the Korean Wa r: Truce Tent and Fighting Front (Washington: Office of the Chief of Military History, U.S. Army, 1966), 307.

?. Report to the Secretary of State and the Secretary of Defense, commonly referred to as the "Brownell Committee Report," Special Research History 123, National Archives and Records Administration, Record Group 457.

Page 17 ACRONYMS

AFSA Armed Forces Security Agency, foundedin 1949

AFSS U.S. Air Force Security Service, the U.S. Air Force COMINT organization

ASA U.S. Army Security Agency, the U.S. Army COMINT organization

ASAPAC Army Security Agency Pacific

COMINT Communications intelligence (now known as SIGINT, i.e., signals intelligence)

COMMSUPACT Communications Supplementary Activity, the U.S. Navy COMINTorganization

DPRK Democratic People's Republic of Korea

KPA Korean People's Army

LLI Low-level intercept

NSA National Security Agency, successor to AFSA

OB Order of Battle

PLA People's Liberation Army

ROK Republic of Korea

ROKA Republic of Korea Army

SCA Service Cryptologic Agency, i.e., ASA, NSG, AFSS

USCIB U.S. Communications Intelligence Board

Page 18