Appendix 1

Technical Analysis

Martin Shough


16th January 1947:   The general conclusion of FO Sewart's report on the events of Operation Charlie was that the targets were probably weather balloons from Downham Market driven back inshore by abnormal NE winds. The unidentified target X-362 on Jan 16 covered the roughly 100 miles from acquisition to signal loss over Norfolk in less than 30 minutes, implying a displayed mean speed in excess of 200 mph, even without taking into account episodes of 'evasive action'. A balloon seems a desperate theory in this case. Nevertheless evidence of a 'real object' is quite strong.

The GCI radar in use at this time was the metric (200 MHz) AMES Type 7. The AI radar fitted in the Mosquito was the 9 cm. helical-scan Mk.10. Thus ground and airborne radars had totally dissimilar timebases, wavelengths and pulse-repetition frequencies (250 & 1500 pps respectively) which makes explanations such as internal noise and mutual interference very unattractive. Anomalous propagation is also sensitive to wavelength, and typically the stable stratified conditions required for AP do not develop in very disturbed winter weather. The probability of AP echoes is thus remote to begin with, and since echoes on these two different fixed and mobile radars separated in altitude by 22,000' would be generated by very different ray paths through different regions of the atmosphere there is no common ducting mechanism to explain correlating target behaviour. Moreover a consistent 30-minute ground track is not at all typical of sporadic AP.

It is true that consistent mobile echoes can be caused by partial reflection from moving waves on an inversion layer, or by scattering from domains of turbulence. These would have been entirely mysterious in 1947 as the theory behind these effects was not fully developed until the mid-'fifties. However the efficiency of scattering layers is quite poor at the metric wavelengths used by the Type 7, and partially forward-scattered echoes would imply severe-hurricane force easterly winds of significantly more than 100 mph at 11,000' - conditions which may well have obtained but which are, again, inconsistent with the stable stratification required for a scattering layer to form. Also this mechanism offers no plausible explanation of concurrent AI radar contacts and cannot explain target motions which would be described by the phrase "efficient controlled evasive action."

Neither birds, insects, balloons nor other windborne objects can explain such a target. Multiple-trip effects can only reduce the displayed speeds of targets detected beyond the unambiguous range of the set and are thus unhelpful. Sidelobe returns likewise offer no useful explanation. A "ghost" reflection from the Mosquito would appear on the same azimuth as the aircraft and at greater range proportional to the added trip-time to a secondary reflector and back, and it would not be possible for the aircraft to "pursue" its own ghost inbound towards the radar site. Moreover, the target was detected in the area before the Mosquito arrived; and this in turn suggests that GCI controllers would know which target was which, so that there appears little likelihood of a trailing inbound ghost being confused with the aircraft echo. A consistent ghost echo from a secondary surface reflector is scarcely conceivable over a 100-mile track beginning over the North Sea and ending inland, with aircraft altitudes up to 22,000'. A ghost echo involving another moving aircraft is possible, and bombers of Exercise Bullseye were currently in the radar pattern at or below about 20,000'. A ghost at twice the range and twice the altitude could be consistent with a target detected farther out at sea and at 38,000', as reported. But such ghosts are fleeting, very sensitive to small changes in the reflection geometry, and a simulated 'chase' from the Dutch coast to Norfolk lasting half an hour would be truly astonishing. And once again, the AI radar contacts made by the interceptor are unexplained by any such mechanism.

By far the most likely explanation of such a target would be a high-performance military aircraft - presumably a domestic flight-plan mix up considering the intruder's apparently reckless 'escape' inbound into UK airspace at a mere 5000'. However, weight has to be given to the fact that seven months later the Air Ministry still logged the incident as unexplained.

17th January 1947:   The events of the next day are more amenable to interpretation in terms of balloons. Nevertheless U-306 was tracked by Humberston CHL for 90 miles on a north-south heading at a mean groundspeed of 180 mph and a constant altitude of 10,000'. The constant altitude during 30 minutes of tracking and the extreme average windspeed implied together argue against a balloon.

No support for the balloon theory can be gained by linking U-306 to the later GCI target that was intercepted several hours later, because that target was at 18,000', whereas if U-306 was a balloon then it probably had a slow leak (sticking at 10,000') and is more likely to have lost buoyancy than to have begun climbing. Moreover the same extreme winds required to explain a 180-mph balloon would make it unlikely that such a balloon would still be in the Norfolk area several hours later, yet the 2000' altitude at which the GCI target and the pursuing Mosquito finally dropped out of the bottom edge of the radar cover 'going out to sea' indicates that the range of the target from Neatishead at this time was probably only about 20 miles.

The details of the interception by the Mosquito have some similarity to cases that are known to have been balloons, typified by the 1948 Fargo N.Dakota case. Repeated passes on a target in the same general area are a typical feature, and Gp. Capt. Kent (who had combat experience against enemy aircraft) confirms that the rate of closure seemed very rapid, suggesting that the target could have been slow-moving or nearly stationary. However it is also true that at no time did Kent acquire a visual silhouette of a balloon - which would have been a fairly large object – or see a tracking light, and he expresses mild surprise that they didn't fly straight into it. Moreover, the ground radar report describes the target Kent was vectored to as "an unidentified high-flying aircraft". Disregarding any possible continuity with earlier trackings, it remains true that this target must have been observed for a good many minutes before Kent's aircraft was scrambled and finally climbed into AI radar range at 18,000'. The question is, if it was a balloon why didn't it appear to be "an unidentified high-flying balloon"?

This implies high displayed speed, which appears to conflict with the "near-stationary balloon" scenario for the interception. But it is not necessary for extreme winds to coincidentally cease just as the Mosquito arrives, since classic interception technique is to come in behind and under the tail of your target which means that Kent could be vectored into the same airstream as a balloon and their relative airspeeds would be preserved. In other words the balloon could appear on ground radar as a speeding target with the Mosquito in hot pursuit, whilst those on the plane would observe the rate of closure of the AI radar target relative to the indicated airspeed of their Mosquito and conclude, correctly, that the speed of the balloon relative to the airstream was near zero.

Coincidences are still required: Two quite distinct balloon "UFOs" on the same day, and the fortuitous bursting of the second balloon just about the time that Kent arrived so that he then followed it down to the deck. But it is just possible that both objects were balloons, given some very extreme middle-tropospheric winds. For illustration, the average N hemisphere winter windspeeds at 18,000' are only about 20 knots, on the order of 1/10th of the speeds implied. But it is true that the weather of January 1947 was unusually severe. It also true that an RAF evaluation concluded that balloons had been dragged back inshore by unusual NW gales.

Perhaps the main difficulty is that both the Operations Record Book and the Sector Log agree that AI radar contacts were repeatedly broken by "evasive action" or "jerking" of the target, described as "violent", at ranges between 1500 yards and two miles. Air Publication 1093D Vol.1 notes that the wider coverage of the relatively new Mk 10 radar was more effective against evasive aircraft targets than the Mk. 8 which crews had used during the war. Even using the narrowest scan limits the altitude and azimuth coverages at such ranges would be hundreds of yards across. The implied angular velocities are therefore excessive and the similarity to balloon encounters becomes strained.

      Copyright © 2002 Martin Shough

Read more of Martin's detailed analysis in his expanded and updated article: A New Study of the British “Ghost Aeroplanes” of 1947