A New View

For years, forensic historians relied on survivor and eyewitness testimonies and, if available, material testing or intelligence reports to help gather evidence to piece together an event. In the early 1990's, seduced by computer technology, photo imagery analysis gained a tremendous uplift enabling imagery scientists to use special algorithms to enhance and sharpen images; changing the way we view the world today. History recorded that five midget submarines deployed as advance forces to attack capital ships in Pearl Harbor had failed to inflict any damage. By the use of digital photo imagery combined with forensic engineering analysis technology, the results of an analysis on a single photograph taken during the 1941 attack will change that view.

The Setting In the early pre-dawn hours of December 7, 1941, five Japanese Type-A midget submarines quietly launched from the various mother subs less than 10 miles outside of Pearl Harbor. They were the Advance Force planned to enter the harbor and attack in concert with the air raid, to sink capital ships and to block the narrow harbor entrance. The narrow entrance channel of Pearl Harbor was blasted out of coral to a depth of 65 ft. at the center. Anti-torpedo nets were strung up one backing the other at the entrance throat from shore to shore. Each of the anti-torpedo nets hung down 35 ft. as a primary defense against torpedo shots up into the harbor. The nets were also intended to prevent submarines from entering the harbor. The approach to Pearl Harbor was heavily guarded. The torpedo nets were open during the day but closed from nightfall to sunrise.

The Secret Weapon The type A midget submarines that attacked Pearl Harbor were secret weapons kept not only from the Japanese public, but also from nearly the entire Japanese navy as well.

In 1934 under the direction of Kishimoto, two midget submarines were built at Kure. They were an experimental craft without conning towers. It was designed as an auxiliary weapon to be carried by fast surface vessels. The two crafts were coded as "Metal Fitting, Type A" were built in secrecy and tested as "Target A" — achieving an underwater speed of 25 knots.

After testing, conning towers were added to the design. In 1936 two more were built. The following year, the midgets were tested and launched from seaplane tenders. Although they had to be launched close to the target, the test pilots realized that these midgets could be used as a stealthy attack craft. Mass production of midget submarines was underway in strict secrecy. Training — also conducted in the strictest of secrecy — commenced at Ourazaki for selected elite crews chosen to operate these attack midgets designated as HA3-44 and to be carried by mother submarines or surface vessels. At the end of October 1941, Admiral Yamamoto at the urging of Captain Hankyu Sasaki finally consented to the use of Type-A midget submarines in the Pearl Harbor attack operation.

The Type-A Imperial Japanese midget submarine, as built in 1938, was a two-man complement sub — a junior officer who conned the boat and a petty officer who manipulated the control valves and ballast for diving and trim.

Structurally, the midgets were connected by four longitudinally welded, cold-rolled 10-in. steel strakes reinforced by welded transverse angle-iron frames. Its length measured 78.5 ft. overall with a 6 ft. diameter hull and displaced 46 tons while submerged. Two bolted hull joints permitted the midget to be separated into three sections.

The hydrodynamic characteristics of the Type-A midget submarine provided little reduction in speed even when submerged. The drag resistance of a full conning tower, strategically located, was kept to a minimum only achievable in Japanese midget submarines due to the smaller physical size and height of the pilot and crew. The midget's length to diameter ratio was 13.3 to 1, closely resembling its weapon, the Type-97, 45 knot, torpedo whose ratio is 12.3 to 1. Their ability to navigate through small entrances was attributed to its slim design, distinguishing itself from many of its counterparts, including the European and Allied designs.

The midget was propelled by a single 600-hp electric motor via single geared shaft with contra-rotating propellers. Power was supplied from acid-cell batteries but without on-board recharging generators. Re-charging could only be accomplished by a mother sub or by a tender. She had a top speed of 23 knots surfaced and 19 knots submerged for 55 minutes of battery power. However, at a submerged speed of 2 knots, the midget had an effective range of 100 miles.

The midget submarine carried two 18-in. torpedo tubes mounted one over another. For the Pearl Harbor attack operation, it was armed with two Type-97 torpedoes, each with an approximately 800-lb. explosive warhead, twice the amount in the Type-91 aerial torpedo. The Japanese torpedoes fueled by oxygen, leaving hardly any wake, are capable of hitting targets faster from longer range. They were considered more advanced than the allied forces.

The Attack At 0755 the first wave of the Japanese carrier planes commenced their attack at their targets. The first wave of torpedo bombers "Kate" B2N5 attacked in groups, flying in an opposite direction attack pattern. The light cruiser USS HELENA was torpedoed while the other group attacked Battleship row. The USS Helena was hit at 0758, recorded in her ship's log. This time was a critical datum for establishing a torpedo attack timeline in the forensic analysis. The rear gunner of a "Kate" took a series of photographs as she flew over the target. One of which was released and shown to the world as propaganda by Japan. This perhaps was one of the most important combat photos of its time.

Presence of a Submarine In 1994, Historian and National Park service of the USS Arizona (BB38) memorial, Daniel Martinez and noted historian and journalist, Burl Burlingham, suggested that the combat photos contained an image of a Japanese midget submarine near Battleship row. An Arlington photo-imagery company digitized the photo and by use of computer algorithm technology, enhanced and sharpened the image of the suspected submarine. Measurements of the submarine were obtained for comparison by triangulation and special imagery techniques employed. Measurements were computed by referencing a scaled stereo imagery of the harbor taken by the Army Air Corp at the end of November. A detailed version of the photogrammetry analysis was published by The U.S. Naval Institute in the December 1999 issue of Naval History magazine.

Image Source and Identification The combat photograph, figure 4, for the forensic analysis was provided by the Naval Historical Center. It was taken during the first wave of the attack by the Japanese Navy carrier based Nakajima B5N2 "Kate" torpedo bomber. This aerial photograph was published in the Japanese press shortly after the Pearl Harbor attack. This photographic image is considered one of the most dramatic combat photographs taken during the 20th Century. Shortly after the attack on Pearl Harbor, an American citizen living in Spain obtained the original positive photo print. It was reported that he received the photograph from the Japanese Embassy in Madrid. Since the original negative has never been found, all subsequent reproductions have come from the lone photograph kept at the Navy Historical Center in the Navy Yard. Imperial Japanese naval history recorded that five Type-A class midget submarines were planned in concert with the main force to attack Pearl Harbor. An anomaly was captured on the surface in the photograph taken by an Imperial Japanese Navy flight crew. With this premise it formed the basis for the photo imagery analysis.

Photo Imagery Analysis(by Autometric, 1994) The photographic imagery analysis considered five phases to be conducted on the photograph.

*The first phase involved photo scientists who digitally reproduced the image, then digitally enhanced it by applying computer algorithms and techniques. The photographic imagery obtained from the U.S. Navy Historical Center was digitized into a computer software image file. Next an edge-sharpening algorithm was applied to the overall image scene. The subject area and the anomaly were then digitally magnified by a factor of 200, 400 and 800 percent. Each magnification was analyzed to determine which magnification would best maintain image integrity.

*The second phase was to establish the chronology of the photograph. Smoke clouds and debris can be seen rising from Hickam Airfield obscuring most of it. This indicates the photograph was taken shortly after one initial attack run on Hickam, which occurred at 0755 hours on December 7, 1941. Smoke can be seen rising from the USS HELENA and the USS OGLALA. This indicates that at least one, if not both ships had been hit. The log of the Helena was very specific about the time she was struck by a torpedo at 0758 hours. Dan Martinez, the Park Historian at the USS ARIZONA Memorial, provided the time of the torpedo strike on the HELENA. History documented the first wave of torpedo bombers attacking Battleship Row within minutes of the attack on Hickam Airfield. Photo evidence shows that the first torpedo strike against Battleship Row (USS CALIFORNIA, USS OKLAHOMA, USS WEST VIRGINIA) has been completed. The CALIFORNIA had been struck by one torpedo and was leaking oil, the OKLAHOMA had been torpedoed and was listening to Port and the WEST VIRGINIA had been struck with at least one torpedo as indicated by several torpedo wakes. The first attack on Battleship Row took place at 0758 hours.

*The third phase was to investigate the photograph for observable objects, which would lend themselves to the presence of a midget submarine. The digital image was examined and digitally interpreted for visual observables to support the premise that a Japanese Type-A class midget submarine was present in the area during the attack. Figure 5 shows that a rectangular, black object is present in the area. The object appears to be sitting on top of a dark linear structure. The composite shape is surrounded by white water area, which looks as though a wave is about to cover the rectangular object. Two distinct water spray rooster tails appear behind the first two. Each rooster tail appears to coincide with the passing of each shock wave. The height characteristic of these rooster tails grows toward the anomaly with the smallest tail occurring first, or being the oldest.

*The fourth phase involved the precise set of photographic measurements of the anomaly in the image. Two photo objectives must be achieved by a photogrammetric approach. First, the location of the anomaly in the geographic reference; second, the position of the anomaly relative to the battleships in the field view. The second objective was to determine the height and size of the anomaly. To achieve these objectives, the position and altitude of the camera used by the gunner/observer in the "Kate" must be defined by using at least photogrammetric triangulation. The location of the anomaly was established by measuring the pixel location of the anomaly. Several objects with known location were also measured to establish the validity of the triangulation resection. The geographic coordinates of the anomaly were calculated as 21º 21' 04.2" North latitude by 157 º 57' 12.5" West longitude. The minimum height and length dimension of the anomaly were thus determined by the same technique and calculated to be 1.25 m and 18.2 m, respectively.

*The fifth and final phase was to compare the size ratio of the anomaly with the specification of the Japanese Type-A class midget submarine. The vertical height of the feature, as computed to be 1.3 m, compares favorably with the height of a Type-A class midget submarine sail, which measures to be 1.7 m. The length dimensions fall within the design length overall of the Type-A midget submarine.

Forensic Engineering Imagery Analysis "FEIA": To establish a methodology for forensic analysis, a positive photo print of the combat photo was obtained from the Naval Historical Center (figure 4). Imagery enhancement and sharpening techniques were applied to key areas of interest established from chronology of images deserved and activity scenario postulated. The approach of the forensic engineering imagery analysis differed from the Autometric imagery analysis in that FEIA emphasized on "how the images of interest got to look the way it was, frozen in time." In simple terms — what caused it to look like that?

History documented that the first attack wave occurred at 0758, by Rudimentary Shadow analysis, it was verified that the combat photo was taken between 0801-0803, thus establishing the latest time at about 0803. In the absence of both a huge waterspout associated with torpedo detonations against the side of the ship and torpedo bombers flying over the battleships, it was concluded that the attack group had completed its attack run.

Six torpedo tracks can be seen in the photo. Four appear partially erased by the traveling surface concussion waves. Two tracks appear to be thinner and show less alteration by the wave concussion. Four of the torpedo tracks can be seen in alignment with respective water splashes associated with aerial torpedo drops. The two thinner tracks that were less affected by the wave concussions converged at a point coinciding with the linear object but without water splashes. One of the thinner tracks terminated with a small water plume characterized by the beginning of an underwater explosive detonation against the side of the WEST VIRGINIA hull.

The observed aerial torpedo trades were measured from their entry splash points to the targets. Torpedo running times were computed and chronologically summarized. The surface concussion rings propagated radially from the WEST VIRGINIA and OKLAHOMA were the result of the torpedo detonation, its gas bubble in expansion and contraction, a phenomenon of underwater explosion. The magnitude and propagation velocity of the concussion waves can be measured. Thus total elapsed time intervals were computed to be 174 seconds (approximately three minutes) after the HELENA was hit. This substantiated that the combat photo was taken at approximately 0801. (Rudimentary shadow analysis indicated the time was 0801-0803).

As the concussion waves approached the advancing midget, they crashed passed its conning tower onto the contra-rotating propellers.

Generating three distinct rooster tail sprays, various progressively increase in size from the earliest spray, the farthest and smallest. The distance between each of the rooster tails was computed and found to coincide with the pattern of the propagating waves. Thus enabling to compute the distance between the observed vertical edges of the rooster tail to the midget conning tower and found to be 38 ft., match half of the midget's overall length of 80 ft.

The thinner tracks were a typical characteristic of the Japanese Type-97 submarine torpedo. Test reports of U.S. conducted torpedo run trials after the war indicated only minimal wake was visible to the overhead flying observation aircraft during the test runs. The thinner tracks characteristically are also indicative that they were more recent with older track defused in time. Thus their convergence in the absence of water splashes and the thinner tracks characteristic lend evidence that a submarine was present and fired two torpedoes at the targets.

During an underwater explosion event, the chemical reaction of the solid explosive yields gaseous products and exerts high explosive pressure on the surrounding water. The disturbance (compression shock wave) near the charge propagates radially at three to five times the speed of sound. The surrounding water is compressed and attains an extremely high radial velocity. The compression shock wave reflecting from the free surface results in a tensile reflected wave, causing a region of water to cavitate, thus forming a bulk cavitations zone. The water refraction near the surface wave has a vertical velocity. The midget submarine is within the cavitation zone and was buffeted by a vertical motion, exposing the conning tower and its contra-rotating propellers.

The magnitude of the vertical velocity by explosive gas pulsation will lift an object vertically in the cavitation region. Extrapolations of the bulk cavitation zone calculated from selected charge size and attack geometry indicated that the midget submarine was within the cavitation region causing the submarine to be launched upward by the explosion.

As the submarine broached near the surface and headed into the oncoming waves caused by the concussion. Water crashed into the propeller disk, forming three distinct rooster-tail sprays synchronous with each passing wave. The third smallest rooster tail had dissipated somewhat by the time the photo was taken. This underwater explosion analysis has substantiated the action the midget took and strengthened photo-imagery evidence of the existence of a midget submarine through a positive correlation between the motion history of the midget submarine and the observed surface disturbance generated. The observed torpedo track indicated alignment with the water plume on the sport side of the WEST VIRGINIA. Due to the water's optical reflection and parallax effect, the torpedo was determined at a depth of about 15 ft. beneath the water surface. This alignment triangulates the impact point to be at the ship's lower armor edge but where the vertical plane intersects with the observed plume.

Conclusion At least two, possibly three midget submarines had entered Pearl Harbor. In the early hours of Dec. 7, due to naval activities, the nets were open from 0200 and reported closed at 0846 after the air attack had begun, providing free passage to the midgets. Out of the five midgets launched, Destroyer WARD sunk one at the harbor entrance; one was beached at Bellows Field; one was sunk west of Ford Island inside the harbor; the fourth midget launched both torpedoes at the cruiser ST. LOUIS near the harbor entrance and was reported sunk. Photo imagery analysis suggested sufficient evidence and substantiated by forensic analysis that the midget shown on the combat photo was the fifth midget; penetrated Pearl Harbor, released two torpedoes, one struck and detonated against the battleship WEST VIRGINIA. The other torpedo track was seen running towards the battleship OKLAHOMA. These controversial findings and also the combined analytical approach on what many currently regard as historic fact.

The fate of the fifth midget remains a mystery. She could have escaped during the confusion in the heat of the air raid or "crabbed" under the torpedo nets at nightfall and made it to open sea. In the evening hours of December 8, the mother sub I-16 received a message that the midget was experiencing "navigational problems". The midget could have slipped out to sea in a failed attempt to rendezvous with the mother sub. Given the ninja ideology of the time, it could have self destructed in deep water in order to keep the secret weapon from falling into enemy hands.

About the Author Peter K. Hsu is a Technical Director in the Ship System Group of Anteon, and a member of the Marine Forensic Panel (SD-7) of the Society of Naval Architects and Marine Engineers (SNAME).