BA038 and EMC - a thought experiment
It was a long flight. Most of the passengers had been in China on business. As the a/c came into land, they were very aware that half the UK working day had already passed and it would be some time before they were back in the office. Some had forgotten to turn off their mobile phones, others had turned them on in expectation of picking up important messages as soon as they were on the ground, still others had been playing games or writing emails without flight mode enabled and had just slipped their mobiles into their pockets when the seatbelt light came on.
Among the active devices, five or six were 3G/HSPA, the rest GSM/2.5G. On late finals, the a/c was finally low and slow enough for the mobiles to make contact with the base stations that clustered around LHR; the Faraday cage of the hull no longer providing enough shielding to keep them isolated. But there was enough attenuation to make the handsets push their power up to compensate. Simultaneously, all the active devices in the cabin started link negotiation across three bands and many tens of MHz.
Whether the aviation cable shielding was inadequate by design or by degradation over time, it wasn't up to the job of keeping the high energy pulses out of the electronics. Bad data appeared, which the engine control computers refused to accept - or perhaps it was any one of a number of other interesting fault conditions that happen when the analogue reality underlying digital systems stops behaving itself. Unable to distinguish the valid commands from the front of the a/c, the controllers stuck to their last setting and waited for the situation to resolve itself. Which, shortly afterwards, it did...
------------------------------------------
Fictional and simplistic, to be sure, and there's no indication that anything like this actually occurred with BA038. I think it's highly improbable. But however unlikely it is, the chances of something like it can only increase. There have been plenty of anectodal and more-or-less technical discussions about these issues already, but precious few solid conclusions.
So if it did happen, how could we tell? As far as I know, the passengers weren't asked about their mobile usage after the event - and it's doubtful that much useful would be learned had they been. "Yes, sir, I had been disobeying the safety rules about using my mobile shortly before the plane fell out of the sky, putting hundreds of people in mortal danger and writing off umpteen million pounds-worth of tin, to say nothing of the cost of subsequent disruption. Why do you ask?"
There will be interesting data in the records of the mobile operators, who will know what terminals registered with which cells along the flightpath and when. The signature of an airborne terminal will be distinctive and easy to extract from ground-based users, just from hand-off records alone.
That sort of information will be useful in general, for characterising the extent of illicit mobile operation, and I _think_ it doesn't need a warrant to be collected. If the mobile operators anonymise it, I'm certain there are no data protection issues, and the decision can then be made if an incident has occurred that the full details should be provided to investigators. It would certainly be relatively cheap to acquire.
Also, monitoring band usage on the ground near to the approaches would collect examples of bursts of activity which could be correlated with events elsewhere.
Finally, on-board parametric spectrum monitoring may provide more evidence, but that would be expensive and intrusive.
R