Whirlwind - Barrett Tillman [49]
The most obvious factor affecting the command’s performance was totally beyond its control: the weather. The alliance of Pacific currents and frigid winter winds from the Asian landmass produced almost perennial clouds over Japan. Bombardiers could not hit what they could not see, even with radar, which was largely used for navigation. Better equipment was forthcoming, but meanwhile there was another inescapable problem—the damnable jet stream. Little known before the war (it had been discovered by Japanese scientists in the 1920s), the stream’s hellacious winds strewed bombs far and wide when dropped from the doctrinal 27,000 to 30,000 feet. Though a few targets had been hard hit, those missions were the result of unusual circumstances: manageable winds aloft and rare days of good visibility, which occurred perhaps five days a month. Operations analysts computed that 400 B-29s were necessary to destroy an industrial target via high-altitude bombing but LeMay had not yet been able to launch 200. Even if he had more, many were always down for maintenance or repair.
What to do?
LeMay put his brain trust to work. The staff sought all manner of information, especially target intelligence, enemy order of battle, and weather data. The latter was hard to come by, and at one point the AAF tried to decipher coded Soviet weather reports, as the Russians seldom shared information with their allies. In any case, cryptanalysis was only marginally useful because, in LeMay’s mind, the Kremlin changed its codes “with diabolical frequency.”
In China, more than two-thirds of the bombs loaded in B-29s had been high explosives. That figure had remained fairly steady in the first three months of Marianas operations, but in February LeMay and some staffers began questioning the conventional wisdom. He knew that incendiaries had proven unexpectedly effective in the December strike that razed Hankow, and that Tokyo’s construction was much the same—mainly wooden structures.
Firebombing Japan had been discussed as far back as Billy Mitchell’s heyday in the 1920s and confirmed by prewar attaché reports (see Chapter One). If in fact 90 percent of Tokyo consisted of wood buildings, the entire city was vulnerable to incendiaries.
After consulting his subordinates, LeMay approved a change of tactics. Small-scale fire raids had been flown against Nagoya (fifty-seven planes on January 3) and Kobe (sixty-nine bombers on February 4). Results were encouraging enough to warrant a bigger test: Mission 38 was scheduled for February 25: the largest incendiary raid yet launched from the Marianas.
Somehow, the Japanese learned of the forthcoming mission in astonishing detail. The female radio personalities collectively known as Tokyo Rose had long provided unwitting entertainment to American servicemen, but in February the airmen got a shock. Rose astonished the 313th Wing by welcoming one of the 6th Group fliers by name: Captain Edgar McElroy, assuring him a warm welcome from Japan’s finest fighter pilots.
McElroy was unique in the 20th Air Force in that his first Tokyo mission from Tinian would be his second to the enemy capital. He had flown the Doolittle Raid in 1942 and sustained a serious back injury but returned to flight status. After a practice mission to Truk, McElroy realized that he might be unable to stand fifteen-hour homeland missions but he refused to be put off by Tokyo Rose. The Texan holstered his pearl-handled .45 automatic and prepared to fly Mission 38.
The command’s three wings established a new record on February 25: 231 Superforts in a daylight high-altitude attack. At the assembly point off the enemy coast, fliers gawked at the sight of so many B-29s inbound, representing an aerial cavalcade of kinetic destruction aimed at the Tokyo urban area. Here was Giulio Douhet’s vision embodied in chilling reality: an unstoppable air fleet aimed directly at the core of the enemy homeland.
In all, 202 bombers unloaded 454 tons of fire bombs, including thirty planes that attacked alternate targets. The result surpassed