The UHF frequency in question was around 270 MHz. Given dark sky conditions one would have to assume a considerable depth of weakly-ionised air in the line of sight in order to achieve a significant radio opacity at the same time as zero effective optical emission from recombination. It seems conceivable that a certain critical electron density might create a "one way" radio mirror, if the aircraft UHF transmissions were of lower power than the ground UHF transmissions, explaining the loss of all transmissions from the aircraft to RAPCON during the incident whilst the aircraft could receive RAPCON transmissions perfectly. There is presently no information on radio power outputs however.
The main problem with this model would then be how to account for the origin of a large volume of ionised air in the absence of any atmospheric electrical activity. It would be reasonable to speculate that this could be related to the presence of the unidentified radar target and associated visual reports of a luminous object or objects; but as already discussed there is presently no available physical model of such a phenomenon.
Also the ionisation theory does not explain in a very natural way why the UHF transmission cut out suddenly in the "middle of a word", implying a very rapid rise in the electron density along the radio line of sight some 90 degrees of azimuth and some 6000 ft away from the presumptive source. It would be interesting to investigate the decibel attenuation required for signal strength to go from receiver saturation to receiver noise level. Some estimate of the rate of rise of electron density in the line of sight could then be derived, leading possibly to limits on ionising particle energies and an overall mean power output.
Such speculation aside, this model is fairly consistent with the report. Short waves are less subject to scatter from ionisation than are long. This dependency has been measured as proportional to the cube of the wavelength for electron densities typical of meteor trail ionisation, ~10^12 cm^-3, and proportional to the power 7.7 of wavelength for auroral ionisation with electron densities around 10^5 or 10^6cm^-3. It is reasonable to assume that densities in the present case would be lower still and the wavelength dependency of a higher power than 7.7. This fits the sequence of events in the RAPCON transcript: Ground reception of the B-52 was initially weak on 271.3 MHz but immediately loud and clear when they switched to 326.2 MHz. Conservatively assuming a power 7.7 wavelength dependency, ionisation would be more transparent by a factor of about 4.0 or better at the higher frequency. Meanwhile the B-52's IFF signal and raw radar paint (probably nearer 3 GHz) were both apparently detected without difficulty on the ground; and the B-52's radar also continued to paint a constant ground echo at 3cm (10 GHz), because attenuation at these wavelengths would be in the order of 10^10 less than at one metre and therefore negligible.
If the asymmetry between the two radio responses remains too much of a problem to be explained by a power differential and a "one-way radio mirror", one other possibility that might be investigated is icing-up of the UHF transmitter power relay. Unlike some other types of relay failure this could be spontaneously self-correcting. Radiosonde temperature readings given in the file (for Glasgow, Mont., Table 4) show 1.0 degree C above freezing at the highest level recorded, 11,500 ft above terrain. Subfreezing temperatures are very likely at higher levels of the B-52 approach. However one would have to demonstrate the unusual vulnerability of this crucial power relay to icing (unlikely given the sealed components used at this date); explain why no other systems on the plane appear to have been affected; and explain the fact that the relay had been operating perfectly normally for some time at FL200 and higher, and would be expected to be warm, which seems inconsistent with sudden freezing in the middle of a transmission just when the B-52 started descent. In summary, there seems to be no wholly satisfactory explanation of the UHF failure.
However Claude Poher has proposed a very interesting hypothesis to explain the asymmetry of this situation. He proposes that a region of weak air ionisation is associated with the presence of the unidentified object, just below a threshold electron density at which it would begin to noticeably affect UHF transmission and reception. However each time the copilot depresses his transmit switch to energise the radio transmitter the electric field strength around the antenna is increased and free electrons are shunted along the field lines. Obviously the energy used in this redistribution of charge is energy stolen from the transmitter power, so the output radio signal is effectively 'shorted'. It's possible that this could neatly explain why only air-to-ground UHF transmissions were affected.