Deterring a Taiwan Invasion

Lessons from Imperial Japan

Last year, ChinaTalk covered Operation Causeway — the WWII-era US military plan to invade Japan-occupied Taiwan and use the island as a launching base for an assault against Japan. As US Army LTC Kevin McKittrick explained, Causeway was ultimately scrapped because, “the enemy always has a vote.”

Today, we’re exploring how Imperial Japan fended off Operation Causeway — in the form of its own large-scale Taiwan-Okinawa defensive plan, dubbed Operation Sho-2Go 捷2号作戦 (“Operation Victory No. 2”). Executed along side Operation Sho-1Go (Philippines), Operation Sho-3Go (Japan mainland), and Operation Sho-4Go (Hokkaido), Sho-2Go involved rapidly fortifying Taiwan with five divisions — approximately 165,600 troops — and employing a strategy of “vertical resilience” by constructing extensive underground tunnel networks and fortifications to leverage Taiwan’s mountainous terrain. Sho-2Go’s key innovation was to thoroughly disperse and conceal air and naval forces in protected positions, preserving them from initial US attacks, and then concentrate these forces in asymmetric kamikaze strikes against US landing fleets at the critical moment of invasion. Sho-2Go’s successful deterrence of a Taiwan invasion led American troops to target Okinawa instead.

This article is the product of fantastic archival research — using Japanese-language source materials — conducted by Japan Air Self-Defense Force Colonel Hirokazu Honda, who just finished a year-long stint at the Air War College (AWC) at Maxwell Air Force Base, Alabama. Colonel Honda earned his commission upon graduating from the National Defense Academy of Japan 防衛大学校 in 2004; he also has a master’s degree in aerospace engineering. He has served as A6, A5, and A1 staff for the Air Staff Office; Commander of the 36th Air Control and Warning Squadron; and defense strategy planner at the Joint Staff Office (J5). Colonel Honda was advised by AWC professor and US Air Force Colonel Jared D. Paslay.¹

Thanks to the US-Japan Foundation for sponsoring this article.

The strategic importance of Taiwan to Imperial Japan

One of the earliest Japanese leaders to focus on Taiwan’s importance to Japan was then–Minister of Education Kowashi Inoue 井上 毅. In a letter to Prime Minister Hirobumi Itō 伊藤 博文 in 1894 — just after the outbreak of the First Sino-Japanese War — Inoue observed,

By controlling the right of navigation in the Yellow Sea, the Sea of Korea, and the Sea of Japan through the possession of Taiwan, we can open and close the gates to the Orient, and by extending our communication with Okinawa and the Yaeyama Archipelago, we can control the entry and exit of others.

Inoue also posited that, should Taiwan fall into the control of another nation, “the security of the Okinawan islands would be disturbed.”

A year later, in 1895, Japan acquired the island of Taiwan as a colony, per the Treaty of Shimonoseki 下関条約, and in 1941, Japan invaded Southeast Asia in search of resources. This expansion made Taiwan indispensable to Imperial Japan as a logistics and trade-route node — and militarily, it made Taiwan indispensable as a naval base, logistics base, air base, information base, and communications relay station for deployment to the south. The colony’s strategic importance was well-summarized in the 10th Area Army Operation Record: “Taiwan, as the southernmost territory of the Empire, has long had significance in imperial policy toward the south.”

Even so, as late as 1940, Imperial Japan maintained only the minimum forces necessary to guarantee Taiwan’s security. The forces on the island consisted of two garrison regiments, one anti-aircraft gun regiment (with fortifications in Keelung 基隆, Kaohsiung 高雄, and Penghu 澎湖), and one flight squadron. Pre-WWII Taiwan was considered relatively secure from external threats — China was preoccupied in its own civil war, and neither Chiang Kai-shek’s KMT forces nor Mao Zedong’s CCP forces had the will or ability to reclaim Japan-occupied Taiwan — and as such, the island primarily served as a logistics base for Imperial Japan.

The prelude

In November 1940, in light of the intensifying Sino-Japanese War, the Imperial Japanese Army decided to expand its deployed forces in Taiwan — but only gradually, with a completion date of the expansion was set for 1943. At the Pacific War’s onset in December 1941, only a few air-defense units, a security force for Kaohsiung Harbor, a supplemental force of one division (without command staff), and an air-education unit were stationed on the island.

Three years later, and Japan’s advance had slowed significantly, while US forces began counterattacks that threatened both the Japanese mainland and its Southeast Asian territories. As of March 1944, the Imperial Japanese Army estimated that, if enemy forces were concentrated in the South Pacific in the future, a landing force of seven to ten divisions would invade Taiwan and Okinawa from the Philippines. By then, Colonel Jihei Ichikawa 市川 治平, the Taiwan Army’s Chief of Staff for Operations, was requesting seven divisions for Taiwan, while Lieutenant General Shūichi Miyazaki 宮崎周一, Director of Operations at the Imperial Japanese Army Headquarters, wanted even more: his memoirs mentioned a desire to deploy nine to ten divisions to Taiwan.

On March 15, 1944, then-Colonel Kazushi Sugita 杉田一次 led an Imperial Japanese Army field inspection team to Taiwan to research how to strengthen the defense of Taiwan and Okinawa. Meanwhile, the Commander-in-Chief of the Sasebo Naval Base submitted a report to the Commander-in-Chief of the Navy, observing that the island of Taiwan was “defenseless against air attacks and enemy landing operations,” and as such, proposed increasing air bases, installing anti-aircraft radar, and strengthening base facilities.

A week later, on March 22, 1944, the Imperial General Headquarters ordered the reinforcement of Taiwan and Okinawa by way of issuing “Outline of Preparations for Operation No. 10,” which aimed to “strengthen operational readiness and prepare for enemy surprise attacks in order to defend the mainland and secure transportation to and from the southern areas.” The initial emphasis on guarding against small-scale intermittent attacks — as opposed to defending against full-scale landing operations — focused on strengthening defense facilities, like air bases, so they could respond to enemy surprise attacks as well as secure maritime traffic and critical lines of communication. For this reason, the outline stated that preparation for air operations support would be the highest priority, and all “other matters shall be subordinate to it.”

After losing many aircraft and ships to US forces, the Japanese finally decided to abandon Saipan in June 1944. At that time, only one division and one mixed brigade were deployed to the island of Taiwan — an “overwhelmingly insufficient” force posture.

A strategy emerges

Following Japan’s defeats in Saipan as well as the Battle of the Marianas, the Taiwan-Okinawa area and the Philippines moved from the rear echelon to the front lines of national defense. Imperial General Headquarters made two initial orders in this regard:

• First, they wanted to triple the number of deployed forces on the island of Taiwan by diverting three divisions, one brigade, and one mixed brigade from mainland China.

• Second, they wanted to treat the defense of Taiwan and Okinawa as defending a single area. As Lieutenant General Miyazaki wrote in his diary,

  Taiwan is a strategic point that will serve as a base for the US invasion of Japan, and Okinawa will serve as a base for air raids on the Japanese mainland. For this reason, the area including Taiwan and Okinawa must be regarded and defended as one area.

  To achieve that end, the Taiwan Army assumed control of both the 32nd Army 第32軍 in Okinawa and the 8th Air Division; the Taiwan Army, on July 11, became a “district army”; and then on September 22, its named changed from “Taiwan Army” to “10th Area Army” 第10方面軍.

Establishment of an air base in 1944 in Taiwan and Yaeyama Islands Area (source)

Beyond force reassignments, the Japanese leadership also needed a new strategy to defend themselves against the enemy’s overwhelming air and sea power.

At a meeting in Tokyo on June 25, 1944, analyzing their staggering losses in Saipan, the leadership made the deliberate decision to shift from a proactive engagement model to an entrapment strategy, in which, after attracting the enemy to within an ideal range, Japan would launch a concentrated counterattack (including asymmetric, self-destructive kamikaze attacks) just before an invasion-force landing. The leadership acknowledged the overwhelming gap in conventional troop strength between the invading US and defending Japanese airpower, and, rather than pursue advance attacks on enemy aircraft carriers, shifted to a “Fleet in Being” philosophy, wherein Japan would avoid being pulled into a decisive battle and instead conserve airpower in dispersed and concealed locations until a critical window of operational opportunity.

In line with Tokyo’s strategic shift, the 10th Area Army began approaching how it established air bases with a new mindset. The most distinctive feature of the 10th Area Army’s air-base setup in Taiwan was its thorough dispersal of aircraft, fuel, and other facilities. The mastermind of this formulation was Lieutenant Colonel Kan’ichi Ishikawa 石川 寛一, the chief of staff for operations of the 8th Air Division, who stated,

We will thoroughly disperse and conceal planes, fuel, and ammunition against the enemy’s advance air attack, and preserve and secure our forces until the decisive battle. When the enemy lands, we will launch an attack with the division’s frontline forces and thoroughly destroy the enemy’s landing craft.

On July 24, 1944, the Imperial General Headquarters formally issued orders to rapidly strengthen Taiwan-Okinawa defenses in preparation for a decisive battle against US forces, should they attempt to invade either island.

Operation Sho-2Go was born.

Redeployment jockeying

Fulfilling the directive to triple the number of forces on Japan-occupied Taiwan proved easier said than done. In particular, military leaders were jockeying over how to divide limited resources between engagements in the Philippines (Sho-1Go), Taiwan-Okinawa (Sho-2Go), the Japan mainland (Sho-3Go), and Manchuria (Sino-Japanese War). In which area could Imperial Japanese forces most likely win a decisive battle against US forces?

Following the Formosa Air Battle — a major aerial battle from October 12 to 16, 1944, in which Japan lost over 300 aircraft — the Imperial Japanese Army believed it could defeat US forces in a decisive battle by committing additional forces to the Philippines, ie. Operation Sho-1Go. (The basis for this belief, as it turned out, was surviving Japanese pilots’ exaggerated reports that US carrier forces had been decimated in that battle.) On October 18, 1944, following the landing of US forces on Leyte, Operation Sho-1Go was launched. Taiwan’s defending forces at that point consisted of three divisions and one brigade — less than forces of the 32nd Army (four divisions, five mixed brigades) defending Okinawa, an island approximately thirty times smaller than Taiwan. Colonel Ichikawa hoped to extract at least one division from Okinawa and assign it to Taiwan.

Tensions over force redeployments continued into November 1944. On November 4, Colonel Takujiro Hattori 服部 卓四郎 — Director of the Operations Division of the Imperial Japanese Army — traveled to Taipei to hold a meeting on the future distribution of forces between Taiwan, Okinawa, and the Philippines. Participants in that meeting included Lieutenant General Haruki Isayama 諌山春樹, Major General Kiyomizu Kitagawa 北川潔水, Colonel Hisashi Kisaki 木佐木 久, Colonel Hiromichi Yahara 八原博通, and Colonel Jihei Ichikawa. The competing arguments of the Imperial Japanese Army, which recognized the need to divert forces to the Philippines, and those of the 10th Area Army, which insisted on strengthening the defense of Taiwan and Okinawa, ended in a stalemate, with no resolution at the meeting.

Eventually, on November 10, 1944, the Imperial General Headquarters decided to dispatch one division from Manchuria to the Philippines — aiming to defeat the United States via Sho-1Go — and to incorporate one division from Okinawa into Taiwan, to balance the defensive forces between Taiwan and Okinawa.

From the standpoint of defending colonial Taiwan, then, the redeployment jockeying was ultimately successful. By December 1944, Taiwan was garrisoned by four divisions, and in February 1945, another division was transferred from Manchuria, resulting in a garrison of five divisions. One study estimates that these five divisions represented approximately 165,600 troops — an eightfold increase, from just 20,000 troops, in under one year. Though the five divisions ultimately stationed on the island of Taiwan were less than the seven divisions originally claimed to be necessary by the 10th Area Army, they were enough to deter US Admiral Nimitz from launching Operation Causeway: US war planners estimated that Imperial Japanese forces on Taiwan would increase by about 70,000 between July 1944 and March 1945 (in fact, Japanese forces increased by well over 100,000) — which meant that, using a 3:1 attacker-defender ratio, the US offensive force estimate had to be three times larger, at least 210,000 troops.

On the other hand, Japan’s decision to reduce troop strength from Okinawa by diverting one division there to Taiwan may have sealed the fate of the Okinawa defenders. General Buckner came to realize that an invasion of Okinawa — with only three defensive divisions — was more feasible than an invasion of the island of Taiwan, with its five divisions. Okinawa was invaded by the United States at the end of March 1945.

How Imperial Japan would have defended colonial Taiwan

If Admiral Nimitz had given the green light for Operation Causeway, Imperial Japan may well have been ready to defend its colony. Manipulating force ratios was only one part of Sho-2Go.

From August 12 to 16, 1944, the general staff of the Imperial General Headquarters, the commander of the 32nd Army (Okinawa), the commander of the 8th Air Division, and other commanders in Taipei convened to conduct a tabletop exercise to prepare for Operation Sho-2Go. Although many details of the discussions during this exercise are unclear, a skeleton of the Taiwan Army’s operational plan still exists. According to the skeleton, the main battle areas for Taiwan’s defense were projected to be the Kaohsiung-Pingtung 高雄-屏東 area in southern Taiwan and, depending on the situation, Taoyuan 桃園 in the north. Fortifying these defensive positions was the top priority. In addition, since weather conditions such as typhoons led the Japanese to predict the enemy might land as early as mid-September 1944, all initial operational preparations were ordered to be completed by the end of August, ie. in less than a month.

The Imperial Japanese Navy ordered the laying of underwater mines (500 ordinary mines and 4,000 small mines) in the waters surrounding Taiwan — part of an effort to prevent enemy landing operations. Should US landing forces make it through the mines, the 10th Area Army’s “Plan for the Fortification of Taiwan Island” (depicted below) detailed fortification methods along with an analysis of invasion patterns, assessed region by region.

The blue numbers in the map above correspond to the following list:

1. Yilan 宜蘭

• Area characteristics:

  • Could be used as an air base for attack on Okinawa

  • Possibility of surprise landing

• Fortification focus:

  • Airfield-defense facilities

  • Coastal-defense facilities

  • Liaison road defense to Taipei

  • Deepening of rice paddies [ie. fill them with more water, so they are harder to walk across]

2. Hualien Port and Taitung 花蓮港, 台東

• Area characteristics:

  • Isolated from other areas

• Fortification focus:

  • Facilities for independent endurance operations

  • Coordination capability between districts

3. Hengchun 恒春

• Area characteristics:

  • Isolated from other areas

• Fortification focus:

  • Preservation of Pingtung Valley

  • River flooding capability using Pingtung Valley

4. Pingtung 屏東

• Area characteristics:

  • Interception area after enemy landing

  • High east, low west topography

  • Many rice paddies in west

• Fortification focus:

  • Mountain positions in east to lure enemy

  • Tactical advantage from elevated positions

  • Deepened rice paddies as obstacles

5. Fengbitou 鳳鼻頭

• Area characteristics:

  • Interception area but not suitable for defense

• Fortification focus:

  • Extra-solid facilities to withstand bombardment

  • Thickened underground tunnel fill

  • Positions for bombarding Kaohsiung Harbor

6. Kaohsiung 高雄

• Area characteristics:

  • Critical interception point for landing forces

  • Requires coordination with Fengbitou

• Fortification focus:

  • Solid facilities connected with Fengbitou

  • All batteries placed underground

  • Reinforced rear defense

7. Takangshan/Siaogangshan 大小崗山

• Area characteristics:

  • Target for enemy to isolate Pingtung Plains

• Fortification focus:

  • Positions in mountainous areas

  • Capability for bombardment from multiple directions

8. Tainan 台南

• Area characteristics:

  • Strategic point for launching troops north to south

• Fortification focus:

  • Solid bunker near Tainan Airfield

  • Deep vertical position

  • Field obstacles through deepening

9. Taichung 台中

• Area characteristics:

  • Vulnerable to paratrooper attacks

• Fortification focus:

  • Underground facilities 30 meters below ground

  • Accommodation for one battalion

10. Houlong 後龍

• Area characteristics:

  • Possible enemy landing to block north-south traffic

• Fortification focus:

  • Facilities at each highland site

  • Maximum endurance capability

11. Hsinchu 新竹

• Area characteristics:

  • Suitable for obstructing landings in Taoyuan and other areas

• Fortification focus:

  • Solid facilities on multiple plateaus

  • Support for mutual defense

12. Taoyuan 桃園

• Area characteristics:

  • Possible landing site to capture Taoyuan Airfield

• Fortification focus:

  • Multiple reinforced positions for mobile defense

  • Numerous battery positions around airfield

13. Tamsui 淡水

• Area characteristics:

  • Possible landing site for attack on Taipei

  • Could be invaded simultaneously with Taoyuan

• Fortification focus:

  • Positions for endurance warfare

As illustrated above, Imperial Japan did not prepare for a conflict of symmetric attrition, but rather planned to fight via a tenacious strategy of “vertical fortification” that took advantage of the island’s topography. They constructed extensive underground tunnel networks, including 6.7 km near Hualien Harbor, 10 km in the Taitung area, 15 km in the Fengbitou area, 23 km in the Pingtung area, and 39 km in the Kaohsiung-Pingtung area; these tunnels were used for discreet, unexposed troop and supply movements. Imperial Japanese forces also built underground fortified positions — some as deep as 30 meters below ground (Taitung) — developed battery positions completely underground (Kaohsiung), and created positions in mountainous areas to gain height advantage (Pingtung) while reinforcing vulnerable positions in plateau areas (Hsinchu).

Cross-sectional view of fortifications, from Taiwan Area Army Operational Record

Cijin Star Tunnel 旗津星空隧道, near Kaohsiung

Making these extensive, island-wide fortifications so rapidly was a whole-of-nation effort. According to the Fortification Plan, military personnel from regular army units and flight divisions were involved in the construction work, due to the lack of dedicated construction units. Further, Taiwan Area Army records show that many Taiwanese civilians were mobilized for this construction work, alongside Imperial Japanese forces. Many Taiwanese civilians also actively cooperated in the operation of air-defense organizations, such as air-defense surveillance and communications. Assistance from the Taiwanese populace manipulated the force ratios even more to Japan’s favor.

Takeaways for defending Taiwan today

To be sure, Taiwan’s strategic environment has changed significantly from 1945 to 2025. In 1945, the United States intended to use occupied Taiwan only as a staging base for a full assault on Japan; today, China seeks the full annexation of Taiwan. In 1945, Taiwan was a Japanese colony; today, Taiwan is free to focus on its own self-defense, free from regional complexities. In 1945, the availability of mass-produced, asymmetric wartime technology was limited (hence kamikaze attacks); today, asymmetric tech is more abundant.

But the similarities which have survived the past eighty years are also striking: there is a significant power imbalance between the attacker and defender; the defenders will likely have very little advance notice of an invasion; and the defenders must fight despite the attacker’s air and sea superiority. As such, Taiwan and allied militaries have much to learn from Japan’s successful defense of Taiwan during WWII.

First, rapid force expansion deters invasion. Japan’s eightfold increase from 20,000 to 165,600 troops in under a year raised the cost of a US invasion and was a significant part of the calculus which led to the abandonment of Operation Causeway. Thus ground-force readiness is essential: Taiwan should enhance both active and reserve force readiness to raise the perceived cost of invading. One way to optimize Taiwan’s reserve system would be to implement a three-stage reserve system with a “Quick Reaction Reserve,” similar to Japan’s model today. In Japan, there are two types of reserves: a Quick Reaction Reserve, obligated to respond immediately, and the General Reserve. The Quick Reaction Reserve receives more training days per year than the General Reserve, are obliged to respond to call-ups more quickly, and are required to work in departments similar to their active-duty counterparts (and get additional pay incentives). Taiwan could adapt the Japanese model to their domestic needs by establishing a Quick Reaction Reserve quota — say, 150,000 troops.

Second, Taiwan should adopt a modern-day “vertical resilience” strategy. Taiwan should use its mountainous terrain for concealment and protection against overwhelming firepower. Underground infrastructure is vital: Taiwan should modernize and expand existing tunnel networks to protect weapons, communications, and personnel. And vertical resiliency can help establish redundancy in communication networks — from cables deep underground to space-based systems.

Third, asymmetric capabilities are critical to fighting lopsided battles. Taiwan cannot afford to invest in high-end, exquisite capabilities that will take ten years to acquire. Instead, Taiwan is now in a situation where, like WWII-era Japan, it must do everything in its power to strengthen its defenses in a compressed time. Therefore, Taiwan should focus on acquiring asymmetric military capabilities over conventional ones. Drones are an excellent example. They can be mass-produced, stored in ordinary (and concealed) cargo containers, and launched in swarms. And not all drones require expensive guidance systems: many can follow remotely controlled lead drones, while other drones can even be operated by civilians with relatively minimal training.

Fourth, just as important as having deterrence is messaging deterrence. Imperial Japan successfully telegraphed to its would-be US invaders that Taiwan’s force posture was changing rapidly, and that any battle over Taiwan would be horrific. Taiwan today can likewise clearly communicate its ability and willingness to defend. For instance, Taiwan could consider holding public competitions showcasing its military capabilities. What would China think if a world-class drone-piloting student competition were held in Taiwan, where award-winning students told media, “We would like to use our skills to support military operations in the event of a Taiwan contingency?”

And finally, if nothing else, Taiwan and allied nations should feel confident that deterrence is possible. Eighty years ago, Operation Sho-2Go worked: the United States was deterred from invading Taiwan. Deterrence can be achieved again today.

Primary sources

“台湾方面軍作戦記録  昭２１年８月,” Taiwan Area Army Operation Record August 1946, Center for Asian History, National Archives of Japan. [link]

“第八飛行師団沖縄航空作戦記録,” 沖縄作戦記録（改訂版） 昭和２４年１１月,  8th Flight Division Air Operation Record in Okinawa, Okinawa Operation Record (Revised Edition) November 1949, Center for Asian History, National Archives of Japan. [link]

“第１０方面軍態勢要図  昭２０年８月１５日,” 10th Army Situation Map, August 15, 1945, Center for Asian History, National Archives of Japan. [link]

“米機動部隊本土来襲に関する戦訓, 戦訓特報第４１号  昭和２０年２月１６、１７日,” Lessons learned for the US task force’s attack on the mainland, War Lessons Special Report, No. 41, February 16th and 17th, 1945, Center for Asian History, National Archives of Japan. [link]

“昭和19年3月以降台湾軍の実施せる築城に関する重要施策,” Important measures regarding fortifications implemented by the Taiwan Area Army since March 1944, Center for Asian History, National Archives of Japan. [link]

1. The conclusions and opinions expressed in this article are those of the author, and do not necessarily reflect the official policy or position of the US Air Force, the Air University, or the Japan Air Self-Defense Force.