Hello,

I was waiting for several days that someone more qualified than me would post something about the SAR but, as nothing is comming, I'll post a few maps to illustrate my findings. Beforehand, I would like to point out that the Press releases so far are fairly inacurate or contradictory about what actually happened to AF 447. If you want to cross-check my data below, please, go to the BEA:
Press Releases
Information on Investigation
and FAB
FORÇA AÉREA BRASILEIRA - Asas de um povo soberano

So, we'll start with the few (real) facts we already had on hand and everybody should be able to reconstruct the maps as I did. For doing so, we need to place a few points on Google Earth (coordinates below are in decimal degrees):


1. Waypoints & Flight Plan
During the first hours of June 1st, AF 447 flight on its way between the waypoints INTOL and TASIL:
- INTOL (-1.362, -32.832)
- TASIL (4.005, -29.990)

AF 447 automaticaly reported its position to Air France HQ (ACARS) every ten minutes. This is shown in this BEA Reconstituted flight path which use the same AF ACARS source: http://www.bea.aero/anglaise/actuali...light.path.jpg
http://takata1940.free.fr/images1/Crashsite0.jpg

2. Last Position, Turbulences & ACARS
The last auto-report was sent at 0210Z and revealed that AF 447 was in cruise (Mach 0.82) at FL350 (BEA).
- Last Position Reported was (2.98, -30.59)

I found this position, page 2, in the 6 June SHOM pdf (Services Hydrographiques et Océanographiques de la Marine,- SHOM, 6 juin 2009) from the French BEA site which is is missing in the English part:
http://www.bea.aero/fr/enquetes/vol....hom.050609.pdf

I cross-checked this position with the BEA presentations and it is always the same spot matching those coordinates. For unknow reason, the Brazilian "Ultimo reporte" is off 20 km (longitude) in their powerpoints.

Flight AF 447 was on track (slighty west 3NM) at 0210Z heading 30 to TASIL. The throttle settings (AUTO) and speed are showing that it was not in "turbulence" mode (A/THR OFF, Mach 0.80) at 0210Z. It might have encountered turbulences @ 0200Z but it cannot be acertained at the moment: basicaly this point is extrapolated from AF 1st press release:
Press Releases
From the above press release, it is not clear if the 0200Z situation was:
a) an AF assumption because of the meteo reports;
b) an AF assumption extracted from the automated ACARS which included also the aircraft performances beside the positional report @ 0200Z.

So far, I tend to believe that no direct report was made by the crew about the so-called "fortes turbulences". Never, so far, the BEA did imply such a fact. Beside, this AF press release is counterfactual about the "electrical failure" pointed by the ACARS as this problem wasn't reported.

Between 0210Z and 0214Z, 24 maintenance ACARS were automaticaly sent by AF447. They are time stamped but unsorted in the listing. Those maintenance ACARS do not transmit the position of the aircraft => The real position of F-GZCP @ 0214Z is unknow. By flightplan extrapolation, She should have been about 31 NM further on its way to TASIL (3.432, -30.328).

http://takata1940.free.fr/images1/Crashsite1.jpg

5. Search & Rescue, recovery:
The BEA pdf "Sea search operations" page 10: http://www.bea.aero/anglaise/actuali...search.ops.pdf
is showing a clear partern of the body distribution (red dots) along a South-North axis. By Picking up the northern red dot each day, it would give roughly the daily speed of the sea drift.
- 6 June (3.566, -30.458) -> dot 06
- 7 June (3.810, -30.485) -> dot 07
- 8 June (4.045, -30.460) -> dot 08
- 9 June (4.415, -30.525) -> dot 09
- 10 June (4.730, -30.505) -> dot 10

There is no need [for this presentation purpose] to consider the other parts of the aircraft as the bodies are very unlikely to be affected by the wind like the other materials from the airframe. From d06 to d10, the distance is about 128 km covered in four days, then an average drift of 0.37 m/sec, with an heading of 357.

A quick Verification using the SHOM pdf linked above, there is a map, page 8, of the surface drift forecast (right map) in this area (red circle) for the period 6-12 June. The right gauge indicate that a 0.40+ m/s drift is pretty close and the heading is also North. Then, this rounded measure is correct enough for an estimation.


7. Probable Crash Zone
By doing an extrapolation of the surface sea drift allocated to the period following the crash we'll find that the recovery of d06 is dated ~0900Z on 6 June. Hence, 126.5 hrs after the crash. During this period, assuming that the drift speed (0.40 m/s) and heading (357) was the same, the bodies would have covered 182 km, which point to about 100 NM south of d06.
- Estimated Crash Area (1.915, -30.390)

http://takata1940.free.fr/images1/Crashsite2.jpg

8. Hypothesis and course post 0214Z
Considering that d06 was recovered ~127 hrs after the supposed crash hour at only 11 nautic miles North-West of the planned AF 447 position @ 0214Z, it seems very unlikely that F-GZCP crashed anywhere near this place if one is taking into account the surface speed of the drift and its heading to the North. The pattern of the recovered bodies seems very regular and is pointing directly to the crash zone and its surroundings.

The only possibility for F-GZCP to reach such a remote place, considering her flightplan, was to make a 180 turn East. After having suffered whaterver crippling instrumentation/airframe damages due to the weather or other cause. Her crew certainly attempted to divert for an emergency landing solution in Fernando de Noronha, at any time past 0210Z (last know position).

Depending at what time the Captain decided to turn back (even after 0214Z as, from the ACARS, no catastrophic failures occured until then), the flight might have lasted much longer than we thought, involving a cripled aircraft unable to keep its cruise speed and certainly also its altitude.

Therefore, the lack of subsequent ACARS past 0214Z should be considered as the consequence of the failure/destruction of the SATCOM system (i.e. ice/hail) rather than the consequence of an immediate fatal dive/break up. Moreover, F-GZCP had a COM3 inop when she took off from Rio. This was revealed by Eurocockpit.com yesterday.

Of course, this is a rough estimation of either the crash zone and the real weather situation. A much more detailed analysis of the actual surface sea drift between 1-6 june would be very usefull to confirm this estimation but, even if the drift was actually half my figure (it may also be twice), AFF 447 would have turned back and not crashed close to 2014Z. I tend to believe also that such a move concured a lot to the delays before the SAR could find any wreckage close to the supposed crash zone around TASIL as they reached it only 5 days later.

S~
Olivier

. . .

The only possibility for F-GZCP to reach such a remote place, considering her flightplan, was to make a 180 turn East. [/QUOTE]

I agree.

I made a similar assumption some days ago, either in a PM or on the board. The only thing I would add is:

a second plot analysis for wind affected debris might give an additional coroborating position.

takata,

Very good investigative work! I think you also need to take into account that the human body does not float in seawater until about 72 hours after death. Then there should be some allowance for the time from establishing positive buoyancy to float from the wreckage at the bottom of the ocean up to the surface. Again some allowance should be made for the average sea current during the ascent time (unlikely to be the same as the surface current). So finally it could be possible that the surface drift time before discovery of the first body could have been less than 24 hours...this might put the wreckage very close to the flight plan track?

takata - congratulations on some excellent and thorough work. I did ask 'the experts' a day or so ago which choice - continuing ahead or turning back through the weather - would be likely vis a vis ETOPS/fuel etc. and have had nothing back. Instinctively I think your 'turn-back' looks like a very likely scenario. Even your projected radius of turn is plausible.

I'd still like to focus on what was LEFT on the instrument panel for them following the known ACARS messages; whether there is some common electrical supply issue that could have caused these failures and whether there is a common 'software junction' in it all. What did the ?Air Caribe? have left in their incident? I cannot find any detailed report.

Very impressive Takata, but your plot assumes that bodies float on the surface. My understanding is that the buoyancy of bodies is highly variable, depending on the quantity of air in the lungs, and the ambient temperature. Often we sink, then, as our bacteria in the digestive tract and elsewhere continue their work, and generate gases, we float. So bodies aren't necessarily a faithful reflection of the surface current either.

--Edit: Flexible Response beat me to it!--

ttcse

I think that ttcse is correct. Beyond a certain depth, a body is unlikely to ever return to the surface, due to the very high pressure exerted by the depth of water.

Bodies don't always come back up
 

Here in Ontario, Lk. Superior is known to keep its bodies because cold water temperature prevents gas formation.

Bodies do tend to come up in Lk. Ontario when drowned near shore, but those lost well offshore may not. One sad Spring night several youths took a small overloaded boat out of a marina for a joyride. Not a single body ever came up -- even though the Ontario and New York shorelines were thoroughly searched and several items from the lost boat were washed to the NY shore.

As to the present case, people strapped in in fuselage sections that went to the bottom would be in water too cold for gases to form.

In the cases of bodies that have appeared on the surface, I suspect that they initially sank to an intermediate depth where the temperature was warm enough for gases to form.

@takata
 

@takata

Nice work putting together the graphics. In your analysis you discounted the effects of wind condition on the drift of the bodies. I think this is a miscalculation. While wind conditions won't affect flat objects in the sea, like it would a floating galley for example, the wind does influence the wave action. Wave action, in addition to current, would determine direction and distance of floating objects. Thus, your conclusion of probable crash site can't really by determined from the floating bodies position at time of recovery. Theoretically, a computer program might be able to calculate float direction and speed from many samples of wind direction/force and current condition over 5 days. Often the wind direction is opposite of current, creating different wave forces then when wind and current are in the same direction. It's too complicated a model to figure out without the help of supercomputers.

ELAC

Amen to that.

That's the fullfillment of being on scene. Most of the questions raised in two days of posts here (and stilll unanswered) could be answered in an hour on scene

Spot on, but it probably helps us to reject ignorant speculation from others that doesn't fit the little facts we can discern. Ask yourself how many times have we been asked by our friends if this or that they heard or read is true. Do we just brush them off even as friends or do we throw a fact or two at them to politely end their speculation.

ttcse

So far it only suggests that the pieces cited impacted the ocean in relatively free fall.

Most of the discussions in the last days have been about the tertiary events and understandably not about the missing links.

lhp, initially I agreed with you. Then I realised that takata was actually comparing like with like. He has a series of plots all similarly affected by current and modified by wind driven surface motion.

By taking one common type of object then he was able to deduce the combined drift rate. He parts of the wreckage been included then the plot would not have been as pure. It would be possible to do a similar exercise with debris alone and compare the two plots.

RECAP!! Do we ACTUALLY have any 'ACARS message' about waste water other than a possible vaccuum flush fault?

drift
 

My previoous analysis was that the drift increased from day 6 to day 16 thus the trend was much less in the first few days leaving impact somewhere between last report point and 50 odd km north north east of it, (if there was no turn to reciprocal followed by ditching)

Takata, in your analysis of the sequence and date of body recoveries, and the drift of the bodies over time (from ocean currents and/or prevailing surface winds) how do you account for the significant amount of wreckage found on June 7 that is well to the west and south of the bodies recovered that day?

http://i136.photobucket.com/albums/q.../SNAG-0289.jpg

The bodies (and wreckage) recovered on subsequent days suggest a northward drift, which then became NW or W last week.

Of course, if one had the index to which parts of the plane (or passenger items) the white numbered dots represent, that would probably help. To date, neither BEA nor Brazil have released even the approximate coordinates for where significant parts of the wreckage were retrieved. What was retrieved at dots numbered 9 and 11 may be key. There are no items numbered 1-8, or 10, that I have seen on the BEA charts. There is a 12. It may be that the early numbers were assigned to flotsam mistakenly identified as being from AF447 in the earliest days of the search. Either that, or only the larger items are shown on the maps that were publicly released.

PJ2 oxygen
 

According to my info fixed oxygen cylinders on the A330 are to be found in one other place than the one below the flightdeck reffered to: for cabin oxigen supply there are 5 + up to another 5 cylinders situated in the r/h sidewall lining in the fwrd cargo hold (source FCOM 1.35.20)

PJ2,

Just to let you know gaseous oxygen is an option for the passenger system on the A340/340 aircraft. If it is fitted the oxygen bottles are fitted aft of the forward cargo door behind the cargo bay lining panels. I have worked on aircraft with generators and the gaseous systems fitted. I am not aware of what option Air France have on their longhaul airbus fleet.

Jimpy

jimpy
 

AF uses gas cylinders for passenger oxigen. See also the accident report on AF's A340 F-GLZQ: http://www.tsb.gc.ca/eng/rapports-re...2/a05h0002.pdf

"The forward cargo door was found in place and closed. Next to the cargo
door, coincident with the location of the passenger oxygen cylinders, was a large hole in the fuselage wall that bore signs of explosive force."

BOAC;
No, we do not.

The early, Class 3 maintenance message "LAV" would have nothing to do with any water/overflow issues. The "waste-water" theory is emerging in the absence of evidence. It is almost certainly a non-starter.
If that's the post, I do have a few comments:

Your post doesn't ask about a turn-back, it only suggests the possibility of a descent as may be indicated by other, associated messages (which, as you say, may or may not exist).

First, a bit of background on the ACARS ADVISORY message:

When a system requires the crews' attention but isn't serious enough to warrant formal master caution/master warning indications plus other associated ECAM actions, the affected system's page is brought up on the lower ECAM, displaying a pulsing green 'advisory' message of the relevant parameter, in this case, the "CAB PRESS" text display on the pressurization page.

An advisory message to the crew does not necessarily require immediate crew action but advises that system parameters have been exceeded, not by a lot but crew attention/advice is required.

The ACARS cabin "ADVISORY" message refers to a pulsing, advisory 'CAB PRESS' message displayed on the pressurization system page on the lower ECAM display.

From the QRH, the "CAB PRESS" pulsing green ADVISORY message is displayed when:

- the cabin vertical speed is greater than, or less than, 1800fpm, or
- the cabin altitude is equal to/greater than 8800ft, or
- the cabin differential pressure (delta 'p') is equal to/greater than 1.5psi in flight phase 7.

Flight phase 7 refers to the approach phase below 800ft AGL to touchdown.

In each case of an advisory message there are recommended crew actions listed in the QRH. These action items are, as described, not of an emergency nature.

That doesn't mean that there was no emergency - it just means we don't have further ACARS messages.

This in and of itself wouldn't indicate either an aircraft descent or a turn-back. There would not likely be any ACARS messages associated with a descent per se, (and certainly none automatically associated with a turn-back), although a "top of descent" message may be automatically sent as a matter of individual airline ACARS design requirements - we don't know.

I understand the theory behind takata's fine sleuthing work and hope it bears fruit but as we can see, there are other considerations emerging which may place the aircraft differently. Very frustrating for all, as we know.

It is impossible to say whether the crew turned back and it is equally impossible to say whether such a decision would have been made or by who, (depending upon who was on the flight deck - we may surmise it would "definitely" be the captain, but we do not know this).

A turn-back at that point in the flight is a highly unusual response - it is essentially an emergency response made out of dire need. I can see a 90deg turn to the east to avoid Cu but that is not the maneuver contemplated by takata's work.

Clearly, there is a pattern, though "the data" is granular in nature, to where the wreckage and most bodies were found. I think takata's work is on the right track and (as I believe s/he him/herself would say), may require fine-tuning. I wouldn't be prepared to say there was a turn-back but such a notion teased as it is out of what little is known, is now a factor which must be proven/discounted.

Pontius Navigator;
The key is what takata and many other are attempting to do: pinpoint the actual position of the aircraft at 02:10. All the rest flows from that one piece of information.

A normal descent for a normal approach/landing will take, at the most, 160nm, with an average of 130nm. That would be the greatest "radius of action" in terms of the location of the larger sections.

The winds in the ITCZ are almost always light and variable, 20, 30kts or so, at cruise altitudes; the "polar jet" and "subtropical jet" are further north and south.

Somewhere, someone has these winds for the time of the accident.

Falling (not in a high speed dive) parts would be exposed to such winds for between roughly three and ten minutes depending upon weight and shape Even assuming the winds remain at, say, thirty knots all the way to the surface, (not likely), that translates to very little lateral distance covered in relation to the position of the aircraft at 02:10 - less than 20nm.

Ocean currents, as has been posited, will account for the rest of the displacement from original impact point. I don't think it takes a supercomputer to work out a reasonable radius of action - it takes the assumptions as demonstrated by takata's methods, which may then be fine-tuned by the availability of more accurate data, (as is stated).

Dutch Bru, jimpy1979uk;

Thanks, I wasn't aware there is an oxygen-tank option for the 340/330 - it isn't in the AMM - only the O2 generation system and the F/A portable O2 installations are.
To be completely accurate, we could reasonably assume from the quoted report, (thank you) that AF 340's have the cylinders installed. We need to confirm that the same installation has been made for the 330's.

In any case, the statement that the oxygen cylinder in the photograph IS from the cockpit, is in question.

ACLS65;
The key in this is the statement, "all this assumes"; this is precisely the situation where a little knowledge (and no experience on the airplane), is a dangerous thing. I think it is wise to suspend judgement on this risk and wait.

As I said a few pages ago, backups/flooding can and do occur on rare occasion but the F/As are very swift at ensuring any such overflow of water is mopped up. The concern is that the forward lavs or galleys may leak into the EE Bay under the cockpit. The concern is legitimate in the sense that it is good, cautious thinking on board an aircraft but is not legitimate in the sense that there is a problem with flooding the EE Bay.

The "LAV" ACARS message is not an indication of an abnormality; it is a Class 3 maintenance message which appears quite often, usually in climb.

The comments about "16,000ft altitude" and vacuum generators being driven by a system that "thought it was below 16,000ft", while possible, is highly improbable and would not result in flooding anyway. A galley tap which can't be shut off is more likely a source of local (I emphasize "local" to forestall further launching in new directions!), flooding.

Very nice work, takata.

I was unaware of the SHOM PowerPoint and had not seen the last known position coordinates in print before - thanks. I have spent a bit of time re-plotting the BEA reconstructed flight path, and it is reassuring that the locations derived from reverse engineering the BEA flight path points in Google Earth are a good match for the GPS coordinates obtained from the ACARS 0210Z position report. My earlier analysis is in the Tim Vasquez forum:

Jet Crash forum • View topic - Crash location

The BEA flight path points are labeled to the second (e.g. 02:00:00, probably the GPS time stamp), so that by reproducing the coordinates of the points in Google Earth by moving them around until they are a good match for details in the sea floor bathymetry in BEA plot it is possible to calculate speed and distance covered. Here is the result:

0120Z - 2° 48.000'S, 33° 35.000'W

77.0 nm, 462 kt

0130Z - 1° 39.500'S, 32° 59.000'W

78.9 nm, 473 kt

0140Z - 0° 29.421'S, 32° 22.002'W

78.0 nm, 468 kt

0150Z - 0° 39.943'N, 31° 45.392'W

77.4 nm, 464 kt

0200Z - 1° 48.654'N, 31° 9.109'W

77.2 nm, 463 kt

0210Z - last known position - 2° 58.700'N, 30° 35.800'W

This last known position derived independently by reverse engineering the BEA plot is 2.978°N, 30.597°W in decimal degrees, a very close match to the location published by SHOM of 2.98°N, 30.59° W, which gives me some confidence in the others. Incidentally, this position is also a good match for the last reported position plotted on one of the actual FAB SAR maps (the red airplane symbol) used by the search teams, as I posted earlier:

http://www.pprune.org/rumours-news/3...ml#post4982877

Variation in the locations, and therefore the speeds and distances, could well be due to small errors in reproducing the BEA flight path positions in Google Earth. It will be interesting to see how close these are to the actual GPS coordinates that hopefully will be in the BEA accident report. The calculated speed between 0120Z and 0140Z is 468 kt, a match for the ground speed of 467 kt calculated by Barry Carlson from the flight plan speed of Mach 0.82:

Final Route of AF447 - an Analysis

One key result of my analysis is that there evidently was no change in speed prior to 0210Z.

Much of takata's analysis concerns the speed and direction of the drifting debris. There are actually some data on drift in this area that cover the period 1 June to 6 June from oceanographic drifting buoys, although they are not all that close to the last known position. Buoy 31919 NE of TASIL drifted 46 nm to the WSW during this period, and Buoy 31256 SE of TASIL drifted 64 nm to the NE. These are more or less in opposite directions, so I suspect the actual drift of the debris during this period was complex.

Attachments do not seem to be enabled, or I would post a map of the flight path and drifter buoy paths. You can view it in the Sun Jun 21, 2009 6:53 pm post at:

Jet Crash forum • View topic - Crash location

-rer47

qantas report - normal to alternate law - PRIM 1 reset
 

From the plots in the qantas preliminary report

http://www.atsb.gov.au/publications/...070_prelim.pdf


a) IR1 fault is detected but the A/C continues in normal law even after master PRIM becomes PRIM2;

b) When master PRIM is switched back to PRIM1 then the a/c gets into Alternate Law and the crew regains control.