It may be of interest to other jet pilots, so I reproduce here a relevant passage from Checkboard's and Yaw String's posts regarding 'warm air pools' - a new concept to me. It does not offer any advance on all the 'computer' glitches, I fear, but may explain the TAT changes quoted on various threads. It is from Appendix 1 to the report quoted.
The phenomenon is almost certainly the result of folding of stratospheric air back into the troposphere with concomitant warming as it descends. The most likely mechanism for this to occur is strong convection (and for the pilot, strong returns on his weather radar). It is reasonable to assume that, where there is strong thunderstorm activity, there is a fair likelihood that an area of anomalous warming will be present somewhere around the storm cell. The vertical extent of the anomaly is difficult to ascertain and is probably dependent on the strength of the outflow from the storm and its interaction with the environmental flow, but in this case, with the tropopause at 41000 feet, observations from the aircraft indicated that the warming stopped abruptly at 28000 feet. The horizontal distribution of the anomaly is most likely to be oncentrated in a zone of confluence between the environmental flow and the thunderstorm outflow, as that would be the area where maximum descent was occurring, associated with the largest perturbation of the tropopause. In this case the anomaly took place on the western flank of the main convective activity. It is unlikely that any anomalous warming would have occurred on the eastern side of the storm as there would have been no confluence and no tropopausal perturbation. The temperature anomaly was observed over a horizontal distance of approximately 60 nm. By comparison, the incident in 1979 took place within a distance of near 40 nm. It is appropriate to relate some of the pilots’ comments here. His normal practice (and it is understood that this is standard practice) when diverting around radar echoes was to ensure a buffer of 10 nm between the radar echo and the aircraft. As an extra safety margin on this occasion he diverted 30 nm from the strongest echo and in doing so encountered the phenomenon. Perhaps this is a reason why it is not encountered regularly by aircraft - most pilots (in maintaining the normal 10 nm buffer) may travel between the anomalous zone and the thunderstorm. It is estimated that in the ncident associated with tropical cyclone Kerry in Queensland, the aircraft encountered the temperature anomaly about 40 nm on the western flank of the main convection area. The BAe 146 pilot also commented that he had been flying in an area of innocuous radar returns - visible as a “green fuzz” on his radar screen. He noticed that the warming commenced as they flew along the boundary between the weak returns and the clear air. It is suggested that that boundary was an indication of the delineation between air masses of differing characteristics. Another interesting comment was that prior to the incident there had been an incredible display of St Elmo’s Fire, the best that he had ever seen. Whether this was a precursor to this type of event or merely coincidence is conjectural at this stage.
Disregarding, of course, the type involved etc etc, it is significant that the apparent rise in SAT caused a large degredation in the altitude/weight capability of the engines.