Operation Downfall: The Planned Assault on Japan

The southern-most Japanese island of Kyushu is a mere 350 miles from Okinawa. It contains numerous bays and inlets and has the large Kagoshima Bay on the southern coast that forms a natural refuge and it encircled by two mountainous peninsulas (Osumi to the east and Satsuma to the west), while on the west coast there was Ariake (now Shibushi) Bay. Like much of Japan, it is a mountainous island, with a range of mountains (between three and five thousand feet) that stretch diagonally across the centre of the island separating it into north and south. Moving between north and south involved either transiting a few narrow trails at the bottom of winding defiles and valleys if one moved across the mountains, or a single highway and railway that ran down each coast. In the south, the island is a mixture of beautiful hills and mountains with many falling into the sea, and inlets each hiding a fishing village. The economy is based on agriculture and fishing, being more rural than much of Japan. The only level land consists of three small rolling plains (including the Miyazaki Plain and Miyakonojo-Ariake basin), upon which the Japanese had built some twenty airfields. It is also the nearest part of Japan to the Asian mainland, as the 100-mile wide Tsushima Strait separates the island from the southern tip of Korea. Centuries before, two Mongul armies had been devastated trying to invade the island, the first being defeated by the defenders in 1274, the second being destroyed by a typhoon the Japanese dubbed the 'Kamikaze' (Divine Wind).

Once planning had started seriously on the projected Allied invasion of Japan, the planners looked at three options: the Chinese coast near Shanghai, Korea and Hokkaido. The first two were rejected as they were unlikely to produce a decisive result. Hokkaido was appealing as Japanese defences were weak and the Allies could achieve surprise, although the weather was uncertain at the time of year under consideration, there would be a lack of land-based air coverage for the invasion, it was quite a distance from the strategically important Kanto Plain (the objective for the second phase of the assault on the Home Islands) and it would force both MacArthur and Nimitz off their lines of advance and logistic bases. As planning developed, the appeal of Kyushu gradually grew. The island's location provided a number of important military benefits and appealed to all three services. It would:

The terrain on Kyushu was important for both attackers and defenders alike. The central range of mountains would, for the Japanese, present a major barrier to movement for the attacker, whereas for the Americans, it would be a useful means by which they could bottle up the Japanese garrison in the north of the island. The railways traversed the coast and passed over a large number of bridges and through a number of tunnels, while the coastal highways passed through a number of defiles. These routes would be susceptible to interdiction by artillery, airpower or naval gunfire. If the Americans could cut the coastal routes and quickly seal off the few mountain passes, they could effectively cut off north and south Kyushu from each other. It would deny the movement of Japanese forces between north and south and they would be forced to fight with the forces already stationed there and guess where the assault would fall. In fact, guessing where the assault would fall, was comparatively easy and any competent Japanese staff officer with a cursory map examination could have done so. There were four possible landing beaches in southern Kyushu all backed by the same rugged ridge lines and hills that had made the cave and bunker defences so effective on Luzon, Iwo Jima and Okinawa. A Sixth Army intelligence officer analysed the problems facing the ground forces due to Kyushu's terrain. "The cliff-like terrace fronts, the commanding heights surrounding all lowland areas, and the rugged mountains full of tortuous narrow defiles, are ideal for the construction of extensive underground installations." The broken and fragmented terrain tended to play to the Japanese tactics of suicidal bunker, cave and emplacement defence, and neutralise the American advantages in firepower, airpower and mobility. Movement was barely easier on the lowlands, coastal and interior plains as loose volcanic soil hampered cross-country movement and intensive agriculture in the form of rice paddies created obstacles except during the autumn when they have been drained of water.

Kyushu was larger than Okinawa and Olympic was Iceberg on a far grander scale. Iceberg, the operation to capture Okinawa and the Ryukyus Islands, was conceived to seize them as quickly as possible and turn into a huge staging area, naval and air base for Olympic. Olympic, as it evolved, was planned to capture Kyushu for a similar purpose – to allow the Americans to establish air and naval bases, as well as staging areas for the ground forces. However, Kyushu would be more important in many ways than Okinawa, because although thousands of US and Royal Engineers descended on Okinawa to construct airfields and port facilities, many of those facilities already existed on Kyushu. The huge bay at Kagoshima was a natural anchorage and naval base, as was Ariake Bay to the east of the Osumi Peninsula, where Japanese Naval Aviation had operated out of a large number of airfields around Kanoya. Additional bases could be found near Miyazaki and Miyakonojo. Chiran, on the Satsuma Peninsula was a centre for Army Aviation. Thus the principle driving the planning for Olympic was to seize the southern half of Kyushu for its airfields and naval facilities, that would provide bases for attacking the northern half of the island, as well as Shikoku, Honshu and penetrating the Straits of Tsushima. Kyushu would act as a forward base for Coronet, just as Okinawa was to serve as a forward base for Olympic. From Kyushu, the Air Force could reach every inch of Japan and the Navy could complete its blockade. Initially, US planners looked at having nine divisions in three corps, who would land on three beaches, drive inland to seize the airfields and encircle Kagoshima Bay. A fourth corps of two divisions would form a floating reserve. The ground forces for Olympic would come from formations already in the Pacific. The US Army's I Corps would land at Miyazaki and XI Corps would land at Ariake Bay next to the Osumi Peninsula, while the US Marines of V Amphibious Corps would land on the western side of the Satsuma Peninsula near Kushikino. IX Corps would constitute the reserve, and if not committed elsewhere, would land on the southern tip of the Satsuma Peninsula on X-Day+4. After they landed and secured their initial lodgements, the ground forces would then set about three immediate tasks. The first would be to capture the inland plains on both peninsulas where the majority of Japanese airfields were located; secondly, they would capture the port of Kagoshima as well as secure the shores and entrance to the bay; and thirdly, they would drive northward to the southern edge of the central mountain range to seal off southern Kyushu from the north. US forces would secure a diagonal line across the island from Sendai in the west to Tsuno in the east.

Planning Begins in Earnest

On 3 April 1945, two messages went out to the two principle Pacific commanders from the Joint Chiefs of Staff. Both dealt with the growing awareness that the invasion of Japan was now looking as a distinct possibility – it was now becoming very difficult to continue to put off those hard decisions that had been looming for some time. While King and Leahy were still not completely convinced of the necessity for an invasion, they recognised that contingency planning should begin properly. So as to meet the possibility that the US would be faced with having to conduct an invasion, both MacArthur and Nimitz had to begin the staff work as soon as possible. The Joint Planning Staff had already drafted an initial plan for Operation Olympic and the Joint War Plans Committee had sketched an outline of the assault on southern Kyushu that MacArthur's and Nimitz's staffs would refine and adapt. The first message commanded MacArthur and Nimitz to begin planning, with Nimitz being assisted by Admiral Forrest P Sherman and Admiral Charles 'Soc' McMorris and concentrating on the naval and amphibious phases of the operation, while MacArthur and his staff were to plan for the campaign on land. Both were to 'cooperate' with each other. The second message confirmed the realignment of all the forces in the Pacific according to service, all naval forces would be commanded by Nimitz, while all army forces (except those air forces on strategic bombing missions) were to be commanded by MacArthur. Nimitz already held the posts of Commander-in-Chief, Pacific Fleet (CINCPAC) and Commander-in-Chief, Pacific Ocean Areas (CINCPOA) and so to reflect this realignment, MacArthur, as well as having the post of Commander-in-Chief, Southwest Pacific Area (CINCSPA), he was made Commander-in-Chief, US Army Forces Pacific (CINCAFPAC). While the Joint Chiefs wanted the realignment to take place as quickly as possible, ongoing operations would mean that the transition would have to be handled with practical considerations in mind.

Having failed to achieve a consensus on a Supreme Commander for the Pacific, the Joint Chiefs took it upon themselves to direct strategy, issue operational directives, assign missions and command responsibilities (JCS 1259 / 4). As it happened, both Nimitz and MacArthur maintained small staffs, both of which had hard-won experience of organising and conducting amphibious operations, gradually developing similar techniques and organisations. As their campaigns had progressed, both commanders had preferred their staffs to plan missions, allocate resources, coordinate timings and set the parameters under which the lower echelons of the command structure would operate. Both left detailed planning of the particular mission to the actual operational units but both staffs tended to reflect the personalities of their commanders – Nimitz's staff were affable and more informal, while MacArthur's was more formal and distant. His Chief of Staff, Lt Gen Richard K Sutherland had a particular reputation as being arrogant and impersonal. The failure to designate a Supreme Commander and the directive for Nimitz to draw up the naval and amphibious phases of the operation, while MacArthur drew up the actual land campaign present unique challenges to both commanders. Marshall suggested that a joint planning staff be set up, with representatives from both commands. Another suggestion was that the HQs for both men be collocated next to each other to aid cooperation and coordination, but the egos and personalities of the two men made this almost impossible. MacArthur's solution to the problem eventually won through and the correlation of the two plans would occur through staff conferences, direct teletypes and liaison groups. An initial staff conference was held at Nimitz's HQ on Guam. MacArthur declined to attend in person but sent a delegation headed by Sutherland. The conference had two purposes, the first being to coordinate the realignment of forces as directed by the Joint Chiefs. Both commands had forces that belonged to the other service and so details of the transfers between each command needed to be worked out, the second being to establish a mechanism through which the two staffs could coordinate their respective planning. The three day conference did not have a happy start with Nimitz wishing to talk about the cooperative planning process, and Sutherland, at his most difficult, wanting to discuss the transfer of all Army units under Nimitz. With the campaigns in Luzon and Okinawa still underway, the Joint Chiefs had left the transfer of forces to the respective commanders. With the Okinawa campaign only a few weeks old, Nimitz was reluctant to talk about a specific transfer timetable. MacArthur's staff was reluctant to talk about cooperative planning. After three days, the conference ended with only some general and self-evident conclusions being reached. Within hours of the result being communicated to the Pentagon, Marshall was preparing to send a blistering message to MacArthur that was scathing about Sutherland and his ability to upset everybody he talked to. After reconsideration he decided to advise MacArthur to meet with Nimitz in person as previous staff meetings had raised opposition and friction. MacArthur responded immediately that as he saw it, the negotiations had been amicable but had already issued an invitation to Nimitz to meet personally at his HQ on Manila.

The lower command staffs could not proceed with any planning until the commanders had agreed on the overall planning principles and established a planning machinery. It was important to have the overall responsibilities, command relationships, logistical and territorial parameters set out as soon as possible so that the staffs could begin detailed planning. The two commanders therefore met on 15 May 1945 and in a two-day conference agreed that Nimitz would continue to develop base facilities in the Ryukyus to cater for the redeployment of major army air force units that would be needed to support Olympic. Command of the amphibious forces was invested in Admiral Richmond K 'Terrible' Turner, commander of the Amphibious Forces Pacific Fleet (PHIBSPAC) who would pass command of them over to the ground commander as they established themselves ashore. All ground units, even Marine Corps formations, would then be controlled by MacArthur. Despite MacArthur and Nimitz spent little time on the important subjects of logistics and how to conduct future planning, their meeting boosted the morale of their staff and energised the planning process so that it gave rise to the first working staff conference between the two groups. A team arrived from MacArthur's staff on Guam on 30 May and started work with their counterparts on Nimitz's staff. This was in fact, the first meeting between the two staffs to look at the detailed mechanics of the operation and in five days, the members tackled a wide range of issues. Of utmost importance was their agreement on establishing planning machinery and on the problematic matter of logistics. The most basic problem was the lack of shipping, compounded by vast distances, long turnaround times, competing demands from other theatres and the construction requirements of the infrastructure and basing facilities. Added to this was the redeployment of units from Europe, moving rear area facilities forward as islands were secured and the preparation for Olympic. The control of shipping and port facilities could potentially be even more of a problem as shipping requirements were carefully balanced and a mistake could lead to congested port facilities and chaos. A set of elaborate procedures were worked out between AFPAC and POA for utilising ships as they were needed and to eliminate competing demands on port facilities. The three Marine divisions that were to be assigned to Olympic would be transported to the objective by Nimitz's command and once ashore, items common to both the Marines and the Army would be supplied by AFPAC, while specialised items would be supplied by the POA, who would also fulfil the Marine's construction requirements.

However, the agreement on future planning merely extended the current practice of inter-command conferences. The detailed planning for Olympic would be undertaken by General Walter Kreuger's Sixth Army staff and Admiral Turner's Amphibious Forces Pacific staff. Turner's PHIBSPAC would carry the Sixth Army's assault forces and land them on the beach and would command them until such time as General Kreuger's HQ was established on the island. With that in mind, Admiral Turner sailed to the Philippines and anchored his ship, USS Eldorado, in Manila Bay so as to be close to General Kreuger's Sixth Army HQ. That same day (14 June), a number of Marine Corps officers (specialists in areas such as intelligence, medical, engineering, transportation and signals) arrived at the HQ to establish a liaison for V Amphibious Corps. There, with the headquarters for the Fifth Fleet (which would operate the landing force and the fire support armada) and Far East Air Forces, the four began to work out the detailed plan for implementing Olympic in a series of conferences. The four HQs worked independently but resolved their issues and coordinated their plans in numerous staff conferences which were almost continuous and worked until the last week of the war. Four distinct plans emerged, one each for Sixth Army, Far East Air Forces, Fifth Fleet and Amphibious Forces Pacific. More specialised subjects would be dealt with in annexes to the main text (such as deception, logistics and communications) and were put together by intercommand conferences. Since MacArthur and and Nimitz had decided not to form a permanent planning group, the problem became how to correlate their overall plans. Marshall had already raised the issue of MacArthur relocating his HQ alongside Nimitz for the invasion of Japan – the reply had been negative. On 6 June, Marshall relayed Admiral King’s suggestion that he relocate near to Nimitz’s HQ on Guam during the preparations for Olympic.

The forces that were to be used in the operation were organised almost identically in the individual staff studies. Nimitz designated two fleets to undertake the naval and amphibious phases of Operation Olympic - Admiral Halsey's Third Fleet and Admiral Spruance's Fifth Fleet. Halsey's Third Fleet had a preponderance of attack carriers and so would provide almost 2000 planes with which to damage Japanese transport and communications capabilities before the attack and cover the landings themselves. Spruance's Fifth Fleet would carry the troops to the designated landing areas and provide gunfire support to help get them ashore and stay there. Admiral Turner, Commander Amphibious Forces Pacific Fleet under Spruance, would command the amphibious landing. Under him were Third, Fifth and Seventh Amphibious Forces (as well as a Reserve Amphibious Force), who were to land the assault forces at Ariake Bay, Kushikino and Miyazaki respectively. As for MacArthur, his main commands were General Kreuger's Sixth Army and General Kenney's Far East Air Forces. The Sixth Army would be the premier ground formation that would conduct the land campaign on Kyushu while the FEAF would support the campaign from bases on Okinawa with the expectation of moving forces onto Kyushu by X-Day plus 2. The Sixth Army was split into four Corps, each of which was to be carried by one of Turner's Amphibious Forces. I Corps under Maj Gen Innis P Swift would be landed at Miyazaki by Seventh Amphibious Force; XI Corps under Lt Gen Charles P Hall would be put ashore by Third Amphibious Force at Ariake Bay; while Maj Gen Harry Schmidt's V Amphibious Corps would be transported by Fifth Amphibious Force and land near Kushikino on the Satsuma Peninsula. IX Corps was to act as the army reserve - if not needed to reinforce one of the main landings, it would itself land near the tip of the Satsuma Peninsula on X-Day plus 4. As these plans were completed, the component commands began to draft their own plans and undertake preparations. These were published in the last days of the war, just as the corps staff were drafting Olympic orientations for the divisional units.

The relations that seemed to be reasonably amicable between MacArthur and Nimitz after the Manila Conference were still plagued by the odd disagreement and squabble between their subordinates and mild suspicions between the two senior commanders. The old Army verses Navy rivalry that had plagued American military planning from before Pearl Harbor was still present. For example, one of the main points of contention was the island of Okinawa. The island was in Nimitz's domain but was coveted by MacArthur as a base for the redeployment of the FEAF from Luzon and the additional air units coming from Europe. Towards the end of May, he outlined the agreements that had been reached in a report to Marshall. His only comment in an otherwise bland factual report was that the views contained therein were not entirely that of CINCAFPAC especially with regards the methods of control and coordination at Okinawa but did represent a solution to which CINCPAC would agree. The campaigns for both Luzon and Okinawa lasted far longer than originally planned, and complicated the transfer of forces as requested in the 3rd April directive from the Joint Chiefs of Staff. This issue caused more problems and friction than the planning for Olympic. MacArthur was more eager to gather all the army forces under his command and take control of the facilities on Okinawa than Nimitz was prepared to relinquish them. Sutherland had demanded the release of all army forces under Nimitz's command at the conference on Guam, even the Tenth Army who was just beginning the long and bloody campaign on Okinawa. Nimitz merely promised that he would release formations as they came out of combat, as immediate and rash changes would invite disaster. BY July, the issue had come to a head, with MacArthur radioing Nimitz a message of dissatisfaction with the progress that had been made on transferring army units over to him. He complained that it was difficult for him to plan the forthcoming invasion of Japan with little knowledge as to the current and future status of army units as he had been forced to negotiate individually with each unit. A few days after MacArthur's sour note, the Joint Chiefs of Staff ordered Nimitz to transfer the Tenth Army and its facilities in the Ryukyus Islands over to MacArthur.

Command arrangements between the army and the navy continued to plague planning for the invasion of Japan. Indeed, the subject of 'unity of command', something that had been settled in Europe much earlier in the war, had upset relations since the start of the Pacific War. The Navy took the term to mean generals commanding fleets (i.e. supreme command in the Pacific for MacArthur), while the army saw a single commander in the Pacific as vital for the effective control of such a huge and complex theatre. The command arrangements for Olympic had been 'sorted out' between Marshall and King, in a compromise that saw Nimitz being responsible for the naval and amphibious aspects of the campaign while MacArthur would plan and command the land campaign. This therefore ensured the development of two plans, although the two commanders were supposed to consult with the other on their respective parts of the others' plan and that an invasion of Japan would have two commanders, an army general and a navy admiral.

The Origins of Pastel

As the plans for Downfall (and therefore Olympic and Coronet) were progressing, the plans for the deception measures to accompany them were progressing a month or so behind them. Until mid-1944, the Allies assumed that southern China and Taiwan would be potential targets for the advancing Allied forces as the war was likely to last well into 1946 and possibly into 1948. However, in June 1944, staff of the Joint War Plans Committee (under the Joint Planning Staff) came up with a new idea, one that advocated a rapid advance directly at Japan itself. Intermediate objectives included the Bonin Islands, the Ryukyus and the Chinese coast near Shanghai, which were to be secured between April and June 1945, followed by a landing on Kyushu on 1 October 1945. All this however, was to follow on from the operations against the Chinese coast and Taiwan. These plans were outlined in a document entitled 'Operations Against Japan Subsequent to Formosa' and in fact marked a transition in US thinking towards an invasion of the Japanese Home Islands while still retaining operations against the Chinese coast. The next step in the evolutionary process followed a number of American triumphs in the Pacific, which led planners to eliminate the intermediate step of attacking targets on the coast of China and aim the American advance straight at Japan. In March 1945, the Joint Chiefs of Staff set down a tentative timetable for the invasion of Japan, codenamed Downfall, with an invasion of the island of Kyushu on 1 December 1945 although the US Navy (Nimitz and King) were still considering their encirclement strategy with landings at Shanghai, Shantung and Korea. Marshal however, opposed this and after discussions with MacArthur (who recommended a direct advance on Japan in April 1945) pushed the JCS to accept these proposals, which were accepted in principle at the end of the month. After the preliminary conference between AFPAC and CINCPAC in mid-May, a directive was issued for Olympic soon afterwards.

With the issue of the directive for Olympic, the Navy plans for continuing operations around the periphery of Japan came to an end, but in many ways lived on in the deception plans for Olympic and Coronet. CINCPAC produced an early draft of Olympic operations in May 1945 in which the main premise of Pastel was already there - a fictitious attack near Shanghai followed by one on Shikoku. Including Shanghai as a target area allowed CINCPAC to utilise the plans for the already drafted attack on the Chusan-Shanghai area (Operation Longtom), which had been drafted in April. CINCPAC's early plan also alluded to Operation Bluebird, a fictional attack on Taiwan and southern China that covered the Okinawa campaign, which had also been based on a recently cancelled genuine operation. Pastel therefore originated from a long line JWPC staff who had concentrated on planning an invasion of the Chinese coast, as well as CINCPAC staff whose commander had been an advocate of Chinese operations as recently as April 1945. The idea was also influenced by the success of Operation Bluebird. Although the US Navy's version eventually dominated, other agencies were drawing up plans for Pastel, such as the JWPC and AFPAC. At the conference on Guam at the end of May at which AFPAC and CINCPAC staff were addressing Olympic, it was decided that AFPAC staff would prepare plans based upon CINCPAC's concept, after which CINCPAC staff would then visit AFPAC headquarters in mid-June to coordinate ideas and take the plan back to CINCPAC headquarters for concurrence. The meeting between the two groups of planners produced an initial draft of Pastel on 13 June 1945. This draft of Pastel resembled the mature Pastel Two except that Pastel Two specified changing the objective from Shanghai to Shikoku on 7 September, not 1 October and had few detailed plans for any deception at the operational level.

Meanwhile, The JWPC, which was part of the Joint Planning Staff, an agency subordinate to the Joint Chiefs of Staff, was working on a deception plan known as Broadaxe. The JWPC often cooperated with Joint Security Control, an agency that coordinated inter-theatre deception activities and answered directly to the JCS, as did the JPS. The JWPC and Joint Security Control prepared a report entitled 'General Directive for Deception Measures Against Japan' that was adopted by the JCS on 16 June 1945. This plan later became known as Broadaxe and sent to the Theatre Commands in order to help guide their planning, however, AFPAC and CINCPAC had already created Pastel without knowing this had been in the pipeline. Broadaxe tried to create the impression that the invasion of the Home Islands was waiting for the USA to seize areas on the periphery in order to tighten the air and naval blockade around Japan. It called for a fictional assault on Taiwan in the late summer of 1945, an invasion of Hokkaido, landings in Southeast Asia and an attack on Sumatra in the autumn, followed by an advance into the Yellow Sea in the Winter of 1945 - 6 to secure additional air and sea bases. With the exception of the assault on Hokkaido, the various fictional operations envisaged in Broadaxe seemed to concentrate on operations around the Chinese coast - US forces would attack Taiwan, then move northwards, with Shanghai and southern Korea following. These fictitious landings would cover the real landings at Okinawa, Kyushu and Honshu. The one major difference was that Broadaxe did not feature an attack on Shikoku as the planners were caught in the dilemma that after Okinawa, attacks on the Asian mainland may not be believed and attacks on the Japanese Home Islands may draw forces towards the intended target areas. They would eventually solve this by assigning fictitious attacks to both the Asian mainland and Japanese Home Islands that were separated from the intended target areas by bodies of water.

After Broadaxe had come out, the parties involved (AFPAC, CINCPAC, JWPC and JPS) quickly realised that the plans had crossed, so a meeting was arranged for 27 June 1945, so that the Theatre and Washington based staffs could reconcile their versions. Planners from CINCPAC travelled to Washington to meet with the Joint Security Control. Together, they revised the Pastel plan to bring into line with the language used in Broadaxe. The proposed landings in Southeast Asia and Sumatra were not mentioned and in fact, very little was revised in the Pastel plan. This suggests that while the Pacific headquarters paid lip service to the Washington plans, the Pastel deception plan was formed first and foremost by the Pacific headquarters, with CINCPAC providing the overall concept and AFPAC filling in the details. Although there were no real changes to the plan, the conference meant that the Joint Chiefs of Staff to adopt the revised Pastel and issue it as JCS Directive 1410 on 9 July 1945. Added to this, according to JSC Serial 6117 (Enclosure B), AFPAC and CINCPAC were to prepare an implementation annex to show who had responsibility for the plan's implementation and its timings, which resulted in Pastel Two and was issued on 30 July 1945. Pastel Two included significant detail on the operational deception measures to be used (from AFPAC) including those of fictional airborne assaults over Kyushu, a concept that had not been in any previous plan. Therefore Pastel Two, with extensive details on how, when, where and who was to implement it, and with little controversy. Discussions continued however, with Douglas Fairbanks Jr, then a naval officer, complained to Admiral King's staff that Shantung was a better target than Shanghai as it would require a more urgent response from the Japanese and that Hokkaido should be the home island target rather than Shikoku as it would keep Japanese forces more dispersed and Major general Clayton Bissell, the senior Army member of the Joint Security Control, pointed out to AFPAC and CINCPAC staff that the 1 December date for the fictitious Shikoku assault was actually quite close to the actual date for the launch of Operation Olympic. These ideas did not sway the planners however, and Pastel two proved to be the final evolution of the deception plans covering Operation Olympic.

The Origins of the Coronet Deception Plan

The evolution of the deception plans for Coronet were less well-advanced than that for Olympic, probably because the assault was not scheduled until the beginning of March 1946 and had not gone through the many rounds of consultation and negotiation that Pastel had done. Indeed, AFPAC would not finish the detailed plans for Coronet until 15th August 1945, the day the Japanese surrendered. On that day, the deception plans for Coronet were still at the staff study stage, in a document entitled "Cover and Deception Plan for 'Coronet', JWPC 190/16" that had been issued on 17 July 1945 by the Joint War Plans Committee and Joint Security Control. The plans outlined in Coronet Deception was detailed and closely followed the pattern developed for Pastel, which may mean that the Pacific Theatre planning staffs had provided material for use by the JWPC. As such, the Theatre staffs would have found little to argue with in Coronet Deception as it was very similar to Pastel, which the parties had all agreed to in late June 1945. While it is probable that while working on pastel, the main issues for the invasion of Honshu were worked out to the satisfaction to most of the participants. There was however, one objection from an Army Air Force planning group (AFAEP) that had not featured prominently in the discussions over Pastel. AFAEP asked that the plan not be forwarded to the Theatre commanders for consideration (which was the next step) but be redrafted according to its own recommendations, which were logical. The group opposed the selection of Korea as a target for a deception operation as after the landing on Kyushu, the Japanese may not believe that the Americans would land on Korea and even if they did, they would not alter their strategy of moving as many troops as possible back to the Home Islands, if they reinforced Korea, it would be by moving troops down from Manchuria. In addition, the diversion of American attention away from the Japanese Home islands may help Japanese morale. Instead, the AFAEP advised an alternative - a fictitious drive made up of two prongs, leapfrogging up and along the northern and southern coasts of Honshu away from the Kyushu landing area, with each leap remaining within the range of the available fighter cover. The northern prong would aim for the Fukui-Kanazawa area, while the southern advance would aim for the area between Osaka and Nagoya. This might seem more plausible than a Korean objective, given the real lodgement already on Kyushu. Also, it followed MacArthur's well-known leapfrog strategy, which would be easier to believe. The Hokkaido and Shikoku landings would also be more compatible with this strategy, rather than Korea, and moving along the coast would be explained by MacArthur's desire to rely as little as possible on US Navy help, due to his rivalry with Nimitz.

While the Air Force ideas were eminently plausible, on 7 August 1945, Capt H R Thurber, Admiral King's Chief of Staff responded in a memorandum. He criticised the plan as it may well draw forces up the island of Honshu rather than keeping them dispersed and away from the actual landing area. An emphasis on the rivalry may hurt Allied morale and that keeping Korea as an objective may prevent Japanese forces from being transferred from there to the Home Islands. The Japanese might try and hold Korea, even in an emergency, to prevent the Allies from gaining a foothold and would focus Japanese attention towards the Southwestern portion of the Home Islands instead of Honshu and the Tokyo Plain. Besides, actual intelligence had indicated that the Japanese had believed the fictitious assaults on Taiwan and southern China to cover the Okinawa landings, the Chusan-Shanghai assault to cover Olympic was starting to be believed and the Japanese were already concerned over an attack on Korea. Thurber therefore urged that the JPS approve the Coronet Deception plan as it stood, emphasising the around-the-China Sea concept that had already been central to deception proposals since Okinawa and part of naval planning up until April 1945. In addition, any sudden change to the basic concepts used in the deception plan might have caused confusion, not only to the Japanese but to the American staff who were to implement it. Before the JPS could make a decision on any modifications to Coronet Deception, the war effectively ended on 15 August 1945, but given that the JCS, JPS, JWPC, JSC, AFPAC and CINCPAC had all agreed on the outline to Pastel, given that Coronet Deception was very similar in structure, as AFAEP were the only ones to have objected to it, it is very likely that Coronet Deception would have been adopted and passed to the Theatre commanders. Overall, the evolution of the deception plans progressed smoothly and the discussions and few minor disagreements led to a refinement of the plans rather than a wholesale rewrite.

The Deception Plan for Olympic

As American forces under both MacArthur and Nimitz continued on their respective axes of advance towards the Japanese Home Islands, US planners started to think about ways of reducing the potential casualties that would accrue from invading the Japanese Home Islands. One such method was deception and the deception plan for the first operation, targeted at Kyushu and scheduled for the beginning of November 1945 was developed under the name of Pastel. The plan was first sketched by Nimitz's staff in May 1945 and by the end of July, had become a mature operational order. By then, the deception plan for the second operation, Coronet, had been outlined and was in the process of development. The final deception arrangements were developed by both MacArthur's and Nimitz's staffs and the Joint War Plans Committee (JWPC), from operations that had taken place, or those that had been cancelled for one reason or another. Those that related to the Kyushu landing, were finalised by US Army Forces, Pacific (AFPAC) in a document entitled "Staff Study: 'Pastel Two', Cover and Deception, OLYMPIC Operations", completed on 30 July 1945 in Manila. The final plan proposed two fictitious assaults at the strategic level. The first was to be directed against China's Chusan-Shanghai area, with a fictitious landing date of 1 October 1945. The second was to be directed against Shikoku (the small home island just to the south of Honshu and to the northeast of Kyushu) with the false landing set for 1 December 1945. At what is today called the operational level, Pastel Two proposed several large-scale airborne assaults in the interior of Kyushu either side of the real assault date. The assault on the Chusan-Shanghai area was planned to help explain away the planned force build-up throughout the Pacific and was the farthest reasonable target from the staging areas planned for the real assault. It was obvious that as the campaign got closer to the genuine landing date, the staging of US forces would fail to match the intended target area and so it was planned that the operation be cancelled around the 7 September and the new target be set as Shikoku. There could be a number of reasons 'leaked' for this, such as the deterioration in the Japanese position in China, satisfaction at the results of the bombing campaign and the success of the accelerated reinforcement and redeployment activity prior to the invasion.

While Pastel assumed that several kinds of operational deception measures were undertaken, the most positive ploy was the development of one or more large-scale, fictitious airborne landings that would support the main Olympic landings. The preparations for such an operation would have commenced in mid-August and lasted until just after X-Day (the day of the invasion). This hoax was specifically meant to exploit Japanese fears over US airborne operations. On 20 August, six gliders and their pilots were to be sent to Okinawa to train in the most conspicuous ways possible. After this, FEAF were to construct six dummy gliders per week until they reached a target of 100, all of which were to be displayed at or near existing military airfields. In addition, a fictional airborne corps headquarters and a divisional headquarters were to be set up on the island by the beginning of September. The 11th Airborne Division, a real unit that was stationed in the Philippines, was to be designated the second component formation of the fictional corps. As an added touch, 1,000 arm patches for the fictional division were to be made (under the auspices of the Joint Security Control, an agency under the Joint Chiefs of Staff) and shipped to Okinawa. On top of that, there were to be parachute drops of equipment across Kyushu on the night before and night after the invasion. On the night before X-Day, one parachute drop was to occur behind each of the three landing beaches: at the Kagoshima airfield behind the Kushikino beach, at the Shibushi airfield behind the Ariake Bay beach and at the Nittagahara airfield behind the Miyazaki beach. Three to four planes were to be used in each drop, each plane dropping material and supplies as if to supply an airborne unit with the objective being to pin down Japanese forces in the interior on X Day. Pastel Two outlined plans for the use of radio traffic as a means of deception. Such deception was to be practised at both the strategic and operational levels, with AFPAC, CINCPAC, USFCT (US Forces, China Theatre) and USASTAF (US Army Strategic Air Forces) all taking part. They would be assisted by the Signal Security Agency in Washington DC who would undertake analysis of all supra-theatre traffic to keep everyone informed as to the overall pattern of traffic.

Strategically, radio traffic to the major Pacific Headquarters would increase to indicate the Chusan-Shanghai assault on 1st October 1945. After 7 September, this traffic would decrease slightly but still be high enough to disguise the variations caused by the preparations for Operation Olympic. This would encourage the perception that the radio traffic was just an anomaly or that there were no large-scale preparations for an assault after the cancellation of the Chusan-Shanghai operation. Communications between Washington and the major Theatre Commands would be increased by the insertion of 'control messages' that would include so-called 'book messages', the messages often relayed after a major high-level decision had been taken, between 4th and 7th September to coincide with the story that would leak the decision to cancel the Chusan-Shanghai operation. Operationally, the Sixth Army would provide radio deception measures during and after the build-up and staging period for Olympic at levels established before the build-up period. Naval point-to-point channels were to be carefully managed in order to disguise when the naval and assault forces had set sail and when they were en route. In addition, there was to be media deception, as Pastel Two set out an agenda to be put into operation by Joint Security Control that outlined a series of press releases and media reports that were to be disseminated to sources that the Japanese were known to use. These events were to be carefully managed so as to suggest and then confirm the story in an indirect manner. The plan was laid out in a timetable of two-week blocks, where all the fictional elements previously outlined, including the Chusan-Shanghai assault, the Shikoku assault and the airborne drops, were to be mentioned between the 30th July and 1st November. For example, as regards the airborne assaults, the plan was to reveal the creation of 'training areas' for 'rehabilitating airborne units' in the northern USA. Further articles in September were to describe how the airborne units were getting on with their training and reconstruction and that gliders had passed through ports on the West Coast and in October, were to mention airborne troops moving to ports on the west coast and a particular company’s glider production.

The strategic deception was to be maintained in early August with press releases that shipping to China was being done directly and that a large number of selected Chinese were being sent to the Pacific to act as interpreters, government officials and municipal administrators. This was to follow on from press releases in August about Shikoku being the least defended of the Japanese Home Islands, having a number of useful airfields and (in September) being under the watchful eye of the High Command. Then, in October, a press spokesman would announce that General MacArthur had been unable to improve on his original 1 December 1945 date for his next offensive. The planners did not want to divulge anything specific about the offensive, merely reinforcing the deception by referring to things in an indirect, roundabout way or undertaking certain preliminary operations, such as beach penetration parties and leaflet drops. If the information was too specific or there was too much of it, the Japanese may have become sceptical. Overall therefore, Pastel Two planned to carry out deception at both the operational and strategic level with several different elements - concrete (including preliminary operations such as beach landings, air reconnaissance and dummy gliders); radio deception; and media deception (where deliberately misleading information was given to media sources or other sources that the Japanese monitored in an indirect way). The common theme of this carefully managed effort were the Chusan-Shanghai assault, the Shikoku assault and the airborne landings on Kyushu.

The responsibility for Pastel Two's implementation was outlined by AFPAC Headquarters who allocated those responsibilities within the command. The G3 (the Assistant Chief of Staff for Operations) was to help direct all the AFPAC commands in terms of where the plan was in terms of its execution as well as coordinate its implementation with the Joint Security Control and the other major theatre headquarters in the Pacific. The G2 (Assistant Chief of Staff for Intelligence) in AFPAC was responsible for the command's deception measures using its own resources (with the single exception being having to arrange the Chusan area beach penetration mission with CINCPAC). They had their own list and timetable as regards media deception measures in addition to that entrusted to the Joint Security Control. As well as items such as straight press releases, there were activities planned such as spreading rumours among 6th Army troops that there was a landing planned in the Chosun area and canvassing for local Chinese who may be familiar with the Shanghai area. The Chief Signal Officer for AFPAC was to direct false radio traffic and coordinate this with other Pacific Commands. They were to be helped by the Psychological Warfare Branch who were to assist in the leaflet drops (with FEAF) and radio broadcasts to the Shanghai area. FEAF also had other duties under Pastel Two such as leaflet drops in other areas, reconnaissance, bombing, organising gliders on Okinawa and drops of airborne supplies on Kyushu, both before and after the main landing. The other forces involved (6th Army, 10th Army, Pacific Fleet) were to assist in communications deception and providing cover for tactical deployments.

The Deception Plan for Coronet

Meanwhile, the plans for Coronet continued to progress a couple of months behind that of Olympic. As they stood, they had no cover name and had just reached the point where the Joint War Plans Committee had released a draft version entitled 'Staff Study: Cover and Deception for Coronet (Coronet Deception), prepared 17 - 23 July 1945' and forwarded it to theatre planning staffs. The plans for media and communications deception had not been developed in the same detail as Pastel Two, neither were responsibilities for the various elements outlined. The story and its fictitious elements were clearly laid out however, and split into both operational and strategic elements. The Coronet deception plan outlined a number of fictitious assaults, on Shikoku, southern Korea and Hokkaido as well as a number of feint attacks on Y-Day (the day of the Coronet landings) itself as tactical deceptions. The story surrounding these assaults was that, before invading Honshu itself, the USA needed to tighten the naval and aerial blockade around Japan. This translated into seizing Shikoku for bomber bases and as an advanced staging area, seizing a lodgement on the southern coast of Korea in order to interdict the movement of Japanese forces from mainland Asia to the Home Islands and to complete the naval blockade of Honshu and then invading Hokkaido in order to stop food supplies to Honshu, secure additional air and naval bases and control the Tsugaru Strait.

The fictitious assault on Shikoku was to take place on 1 April 1946, a month after the actual Coronet landings were planned to take place, thus holding forces there even after the landing had taken place. All the fictional assault areas were separated from the actual target area by water thus making harder to reinforce the troops on Honshu once the battle had begun. Ground forces were to be instrumental in giving the impression that an attack was going to occur on Shikoku. A ranking officer was to be assigned the command of the amphibious assault with the five infantry divisions training in the Philippines and single division on Okinawa were to mislead that they were to operate on Shikoku. Fictional numbers were to be given to the units in the Philippines so as to maintain a fictional presence even after the units had left for Coronet, reinforced by a flurry of activity by service troops. Staff in the Philippines were to prepare intelligence studies on Shikoku and officers of the divisions involved were to study maps of both Shikoku and Honshu. Other suggestions included having a truck spill a load of maps in a crowded Manila street combined with an obvious effort to hide the maps, troops were to see movies featuring cities on the Inland Sea, civil government officials and airfield construction units were to be briefed on working on Shikoku and the division in Okinawa was to be given amphibious training using Shikoku beach names that could be witnessed by Japanese PoWs. The FEAF (Far East Air Force) were to drop leaflets telling civilians to avoid certain areas, such as airfields and beaches, after 1 March 1946. Parachute dropped supplies and pre-damaged material would be dropped near Kochi on 1 January and then again in March. Time delay charges and rubber boats would be dropped along the coast to simulate underwater demolition work. This would be combined with operations by CINCPAC who were to attack targets all along the Japanese coast. The plan also envisaged other deception measures such as extensive radio traffic, media reports and news stories, rumours and bogus build-ups.

The fictitious assault on Korea was timetabled to 'occur' on 1 May 1946, a month after the Shikoku operation and two months after the actual Honshu (Coronet) operation to allow for additional staging time between the 'assaults'. The Korean operation was supposedly designed to stop Japanese forces being moved from the Asian mainland to Honshu. It was also meant to fit in with the deception regarding a possible American attack on the Asian mainland that had been in the works since the assault on the island of Okinawa. As in the Shikoku deception, an actual commander would be named and 11 of the divisions staging in the Philippines (and another on Okinawa) would be told they were bound for Korea. Fictional numbers would again be assigned and continue to be used in radio traffic after the troops had left for Honshu, with the troops having seen movies about Korea and the staff officers studied models of both Honshu and Korea. As in the Shikoku operation, FEAF were to drop leaflets, parachutes, supplies and pre-damaged materials on Korea on 1 November 1945 (the date of the actual landing on Kyushu for Olympic) and to simulate the insertion of agents on a periodic basis after 1 February 1946 with photo reconnaissance starting after 1 December 1945 and naval deception measures including simulated submarine-launched commando raids and false radio traffic starting about the same time. Direct operations against Korean ports were to enhance this deception and serve as decoy operations for the convoys travelling to Vladivostok. The deception measures designed for Korea were to include special measures, such as interpreters and government officials leaving the USA on 15 April 1946, US currency being overprinted for use in Korea and the Red Cross were to be alerted to be ready to transfer personnel to the peninsula after the operation had begun. To increase the chances of the Japanese learning of this, the Red Cross offices in Chungking and Kunming were to be informed of this as well. A variety of notes, maps and briefing documents were to be prepared, along with a full campaign plan prepared by CINCPAC, were to be 'lost' and planted in Japanese-held territory by the US China theatre commander (COMGENCHINA). Lastly, irregulars were to be landed in Korea in order to conduct hit-and-run operations against suitable targets.

The fictional Hokkaido assault was timetabled last on 1 June 1946 and meant to pin Japanese forces there and in the Kuril Islands in place, even after Coronet had begun three months earlier. Although the deceptions involving US ground forces was less developed than in the other scenarios, a number of divisions bound for Coronet who were to embark from Seattle, were to be assigned fictional numbers and assignments in the Aleutians. There was also the provision to simulate six divisions that were to be based at a number of actual facilities in the Aleutians in late 1945. During the winter of 1945 - 6, radio traffic would maintain the illusion of four divisions still being stationed there, with another three divisions passing through there during the spring of 1946 and a Canadian division bound for Hokkaido via the Aleutians. This would be reinforced by photo reconnaissance flights of Hokkaido and northern Honshu starting in February 1946. CINCPAC and COMNORPAC (Commander, Northern Pacific) were to coordinate operations such as landing patrols by submarine in the Kurils, landing agents and capturing prisoners. There were to be leaflet drops in the Kurils and on Hokkaido, warning fisherman to avoid certain waters as well as radio broadcasts to Hokkaido for the benefit of the civilian population, telling them to avoid certain coastal areas and messages in a (compromised) code telling of the construction of rocket launchers to assist in the pre-invasion bombardment. Joint Security Control were to coordinate additional radio traffic and media disinformation programmes, although their task was made more difficult by having to coordinate a number of other Theatre commands, besides the usual CINCPAC and AFPAC ones that included COMNORPAC and COMGENALASKA (Commanding General, Alaska).

Operational deception for Coronet was meant to disguise the fact that the main assault was to be conducted at Sagami Bay, by hinting that the operation was indeed a supporting or diversionary one and by suggesting progressive landings running from south to north. The first genuine landing would have taken place on the Katakai-Choshi beaches on 1 March (Y-Day), with the main landings following on the Sagami beaches just south of Tokyo on Y+10. The sequence of real and fictitious landings were designed to draw attention away from the genuine intentions of the Coronet plan. The preliminary operations were to begin at Sagami on Y-7 but after that, no more was to be done until seventeen days later after five other coastal operations had taken place, thus reinforcing the notion that the earlier operations had just been a feint. These preliminary operations would include heavy bombardments by air and naval power, clearing beach obstacles, minesweeping, demolition of underwater obstacles and for the areas associated with the fictitious assaults, dropping time-delay explosives, rubber boats and other equipment associated with preparatory activity, as well as simulating beach reconnaissance, minesweeping, use of smoke, the movement of naval and amphibious forces, fireworks and pyrotechnics, radio and radar deception and jamming. The second operation was to be the preparation of the Katakai-Choshi area beginning on Y-4, followed by the actual landing at Katakai on Y-Day. Simultaneously, preliminaries were to begin at Kashima, to the northeast of Tokyo and then in the Sendai area, 200 miles northeast of the capital on Y+4. On Y+8, simulated landings were to be held at Kashima, utilising two divisions still in transports. After all this, the main landings were to begin on Y+10 with three divisions, expanding to eleven divisions by Y+30, with another demonstration by empty transports off Sendai on Y+10. This pattern of activity was so that the attention of the Japanese would be drawn from south to north away from the genuine landings, with each preliminary activity followed by an assault eight to ten days later. This impression was to be reinforced by tactical radio deception measures that were to reinforce the premise that the bombarding and assault forces were moving north rather than south.

The exception to this, naturally, was the main landings at Sagami Bay. The time interval between the initial preparation and the actual assault was seventeen days instead of between four and eight, so as to give the impression that they had been a feint. In addition, Sagami took place late in the sequence at the most southerly point even though all the other preparations had moved northwards. The aim of the carefully worked out sequence was to make it difficult for the Japanese to work out from the preliminary operations, the existence, location and timing of the main landing at Sagami. In looking at all these activities, it may seem a bit superfluous to have a demonstration off Sendai on Y+10, especially as the main landings had already started but it was intended to pin potential reinforcements as far from the Tokyo beaches as possible, and as such was worthwhile. Tactical deception for Coronet was an elaborate dance involving timing and location, while measures at the operational and strategic levels would contribute in their own way.

World War Two: U.S. Military Plans for the Invasion of Japan, Thomas Fensch (Editor) This is a very useful collection of official American documents relating to Operation Downfall, the plan for the invasion of Japan. The invasion never needed to be carried out, but the plans had reached a very advanced stage by the time the two Atomic bombs ended the war. [see more]

How to cite this article: Antill, P. (10 October 2006), Operation Downfall: The Planned Assault on Japan, 27 April 1296, http://www.historyofwar.org/articles/wars_downfall2.html

Operation Downfall: The Planned Assault on Japan

The Initial Planning for Olympic
Planning Begins in Earnest
The Origins of Pastel
The Origins of the Coronet Deception Plan
The Deception Plan for Olympic
The Deception Plan for Coronet

The Initial Planning for Olympic

Planning Begins in Earnest

The Origins of Pastel

The Origins of the Coronet Deception Plan

The Deception Plan for Olympic

The Deception Plan for Coronet

Part One | Part Two | Part Three | Part Four