Over the last three years, Inmarsat, the ATSB, and the DSTG have been commendably proactive in explaining the mathematical process by which they deduced MH370’s most likely endpoint from the Inmarsat data. The most recent installment, “The Use of Burst Frequency Offsets in the Search for MH370″ by Ian D. Holland, continues that tradition by shedding light specifically on the BFO analysis. All told, it reinforces the impression that the team had good reasons for thinking that the plane would be found in the 120,000 sq km search box. This, of course, only deepens the riddle of why it wasn’t.
Holland’s paper gets into some pretty dense math but a couple of points stand out. On page 6 we read:
Since the mean of the BFOs from the 18:39- 18:41Z call attempt are in broad agreement with the linear trend observed in the BFOs from 19:41Z to 00:11Z (for which the BTOs themselves were consistent with straight and level flight ), this supports the finding in  that there were most- likely no major turns after the unanswered call attempt (see (, Fig. 10.5)).
This is a familiar idea: that the BFO value at 18:40 indicates that the plane, if flying level, was already heading south at that time, and that the FMT had already occurred. This was long accepted as being almost certainly true. But as Victor Iannello recently pointed out on his newly started blog, if the plane turned before 18:40, then its wreckage would have been found in the SIO search area. Therefore, he proposes an alternative that originally seemed less likely: that the plane was in a steady, 2900 feet per minute descent.
Why would whoever was flying MH370 want to descend in such a fashion? Iannello proposes that they might have been setting up for a landing at Car Nicobar airstrip in the Andaman Islands. Perhaps the plane descended for a while — at this rate, getting down from 35,000 o 20,000 would take about five minutes — and then the pilot changed his mind and flew south into the SIO instead.
However, as I’ve pointed out before, if the plane flew straight after 18:40 then the geometry of the BTO rings itself suggests what speed the plane was flying at. The reason is that the 19:40 and 20:40 rings are quite close together, and so there is a small angular distance between, say, a 400 knot path and a 500 knot path. This small angular distance means that the intersection of these paths with the 22:40 and 00:11 rings are spaced at very similar distances. The upshot is that the BTO rings themselves imply a speed of about 480 knots. This is not dissimilar to the speed seen on the radar track, about 500 knots.
To fly this fast without burning all its fuel before 0:11, 9M-MRO must necessarily have been flying at or close to normal cruise altitude.
What Iannello is effectively proposing, then, is that MH370 flew fast and high, then descended, then climbed and flew fast and high again. It is not easy to see what such a dive-and-climb might have accomplished. That is not to say it didn’t happen. But it does run counter to the behavior that the plane otherwise exhibited, which seems to have been geared toward getting where it was going rather quickly.
Another thing I found interesting about the paper was the amount of attention given to the question of the anomalous BFO value at 18:25:27. Apparently a body within the official search effort called the “MH370 Flight Path Reconstruction Group – SATCOM Subgroup” produced a whole paper on this topic; it has not been released to the public but Holland refers to it no less than 7 times, as sk999 has pointed out. We’ve known for some time that the ATSB was unable to find a reason for this value. This struck me as suspicious and I wondered if it might be evidence that the SDU has been tampered with. In this paper we learn for the first time that a study of 20 previous 9M-MRO logins found that one similar anomalous value. (Previous reports have stated that login requests in mid-flight are extremely rare, so we can assume these occurred during normal power-up sequences on the ground.) Unfortunately, Holland doesn’t say anything about what the circumstances were. The implication however is that this kind of anomaly can arise innocently.
Finally, Holland also touches upon the issue of the rate of acceleration implied by the final two BFO data points: 0.68 g. Again, we’ve discussed this before on this blog, but since Holland revisits it I think it bears repeating. This is an extremely high rate of acceleration — two-thirds of what the plane would experience if it was free-falling in a vacuum. With no engines to hasten its descent, the plane must have been pointed nearly vertically down. With its velocity increasing at such a rate, the plane must have impacted the surface quite soon after, therefore it couldn’t be very far from the 7th arc.
My overall impression upon reading this report was: Wow, this is extremely solid work. The DSTG’s analysis of the Inmarsat provides a very compelling case for where the plane went in the southern Indian Ocean. I wouldn’t want to bet against this, I thought.
Then I remembered that their predicted area has already been searched and nothing was found.