The final report was released on March 30, 2016. On the 29th November 2013 at 09:26 UTC, an Embraer ERJ 190-100 IGW aircraft with registration number C9-EMC departed Maputo International Airport on a scheduled flight to Luanda, Angola. On board the aircraft were a total of 33 people comprised of the following: Six (6) crew members (two pilots, one engineer and three flight attendants) plus 27 passengers. The flight operations were normal and the aircraft was in radio communication with Gaborone Area Control Centre (ACC) on frequency 126.1 MHz and was cruising at FL380 (38000 ft). At position EXEDU, a mandatory reporting point in Gaborone FIR (Flight Information Region) which is at 72 Nautical Miles (nm) south of the point called AGRAM in the Zambezi strip, the Namibian radar data revealed that the aircraft commenced a sudden descent from the normal cruising level of FL 380. Radar contact and voice contact were lost with ATS (Air Traffic Services). Search and Rescue was instituted but could not locate the accident site the same day of the crash due to bad weather that developed in the region and nightfall. The Search and Rescue team only managed to locate the accident site the following day (30 November 2013) at around 08:00 UTC, in Bwabwata National Park. Due to the high rate of impact with terrain and post- impact fire, the aircraft was totally destroyed and there were no survivors.
Photo : © carlos oliveira reis
CIVIL AIRCRAFT ACCIDENT REPORT
Name of Owner : ALC E190 581 LLC
Name of Operator : Mozambique Airline (LAM)
Manufacturer : Embraer
Model : ERJ 190-100IGW
Nationality : Mozambican
Registration Marks : C9-EMC
Place : Bwabwata National Park, Namibia
Date : 29 November 2013
Time : 11:16
All times given in this report are in Coordinate Universal Time (UTC).(Namibian daytime +1 GMT)
History of Flight
On the 29th November 2013 at 09:26, an Embraer ERJ 190-100 IGW aircraft with registration number C9-EMC departed Maputo International Airport on a scheduled flight to Luanda, Angola. On board the aircraft were a total of 33 people comprised of the following: Six (6) crew members (two pilots, one engineer and three flight attendants) plus 27 passengers. The flight operations were normal and the aircraft was in radio communication with Gaborone Area Control Centre on frequency 126.1 MHz and was cruising at FL380 (38000 ft).
At position EXEDU, a mandatory reporting point in the Gaborone FIR (Flight Information Region) which is at 72 nautical miles (nm) south of the point AGRAM which is the boundary position between Gaborone FIR and Luanda FIR (18’ 56S 02’ 228E), the Namibian Radar Data (after it was played back) revealed that the aircraft commenced a sudden descent from the normal cruising level of FL380 at 11:09:07. Radar footage from the Namibian Air Traffic Services (ATS) during the investigation indicated that the target was lost at 11:15:49 at an altitude of 6 600 ft AMSL while on its abrupt descent. The aircraft impacted the ground at 11:16:04 at an altitude of 3 390ft Above Mean Sea Level (AMSL) as indicated by the Flight Data Recorder (FDR).
There were no distress calls made by the pilot to declare an emergency nor was there any signal transmitted from the Emergency Locator Transmitter (ELT) due to the fact that its antenna cable was cut during the time the aircraft impacted the terrain.
Namibian authorities (police) at the Eastern Kavango and Zambezi regions were informed about the missing aircraft at around 12:00.
Search and rescue operation was instituted by Namibia Air Traffic Services but could not locate the accident site the same day due to bad weather and darkness in the area. Eyewitnesses from villagers on the Botswana side near the border heard explosions and observed smoke coming from the Namibian territory then informed the Namibian Authorities.
The accident site was located the following day (30 November 2013) at around 07h00 , in Bwabwata National Park (Namibia). The Namibian Investigators responded immediately after the crash site was located.
The Government of Mozambique and the Mozambique Airline were informed by the Namibian Government on Saturday 30 November 2013 that the wreckage has been located and identified as that of C9-EMC, an Embraer ERJ 190-100IGW and that there were no survivors. Both CVR/FDR recorders were retrieved from the accident site on Saturday 30th November 2013.
DAAI lead a team of accredited representatives and advisors from Mozambique, the state of operator and registry; Brazil, the state of design and manufacture; Botswana, the state that was controlling the accident aircraft; Angola, the states of final destination; and the USA, the state of power plant manufacturer.
Based on the outcome of the CVR/FDR readout, the Directorate of Aircraft Accident Investigation (DAAI) in Namibia issued a Preliminary report on 18 December 2013 which stated as follow;
a) The aircraft was operating at normal conditions and no mechanical faults were detected.
b) Minutes before the crash, the F/O (first officer) left the cockpit for the lavatory and only the Captain remained on the Flight Deck.
c) The aircraft flight altitude was manually selected three (3) times from 38000 feet where the aircraft was cruising, to 592 feet (below ground elevation).
d) The autothrottle was manually re-engaged and throttle level automatically retarded and set to idle.
e) The airspeed was manually changed several times until the end of the recording, which remained close to Vmo (maximum certified operating speed limit).
f) The speed brake handle parameter indicates it was commanded to open the spoiler panels and remained in that position until the end of the recording. This was manually commanded as the parameter monitors the handle position.
g) During all these actions there was audible low and high chimes as well as repeated banging, an indication for a call to enter the cockpit.
According to the statement in LAM’s Manual of Flight Operation Chapter 10.1.4, Page 5 of 36, Edition 3 Revision 8, (Absence from Flight Deck) which is in Portuguese but was translated to English as follows;- “A pilot can only leave the Flight Deck when the aircraft is above 10 000 feet, for physiological reason or when carrying out his operational task. He or she shall, for that matter, first call a cabin attendant who will remain in the Flight Deck until his return. At the moment of the pilot’s exit, the other (pilot) shall lock the door, to only re-open it for the return of the absent pilot”.
“The pilot remaining in the Flight Deck should be in a state of alert and situational awareness, have free and clear access to the commands of the flight”. (Refer to Appendix 1, full final report).
All passengers in Mozambique flight 470 were issued with passenger tickets.
The weather at the altitude where the aircraft was cruising was fine.
Figure 7. Illustrates the nominal angle of impact.From Directorate Of Aircraft Accident Investigation Accident Report
Organizational and Management Information
1. LAM 470 was on a scheduled flight operating on the TM 470 route.
2. The aircraft was owned by ALC E190 581 LLC and was operated by Mozambique Airline (LAM).
3. The last Inspection was certified on 28 November 2013 at 2902.00 airframe hours. At the time of the accident, the aircraft had accumulated a further 3.0 hour since its last Inspection was certified.
4. The Aircraft Maintenance Organization (AMO) which carried out the last Inspection to the aircraft prior to the accident was in possession of a valid AMO Approval Number 02/OMA/2011 issued on 20 September 2013 and having an expiry date of 19 September 2014.
5. LAM had advanced proactive safety procedures that were way above the international minimum standards. Such procedures such as “more than one person in the cockpit” are only being adopted by most airlines after 2014.
Additional Information Human Factors
During the investigation, the team traveled to Maputo in Mozambique looked into Mozambique Airline, the Institute of Civil Aviation of Mozambique (IACM), the Immigration department at Maputo International Airport and also interviewed the family members and friends of the two pilots.
The investigation team found out that all passengers were issued with passenger tickets and no evidence shows that there was any passenger who did not pass through security screening.
The Investigation team also discovered through the interview that the captain went through numerous life experiences ranging from:
a) The separation from the first wife on which the divorce process had not been dissolved to almost (10) years after separation.
b) The death of a son who passed away in a car accident on a suspected suicide on the 21st of November 2012.
c) The captain was reported as not to have attended his son’s funeral.
d) The captain’s youngest daughter underwent heart surgery in one of the hospitals in South Africa not long time ago.
Flight Path Reconstruction
Flight path reconstruction that utilizes flight data to construct 3D animation incorporating flight deck instruments is an essential investigation technique that enables an investigator to quickly understand exactly what transpired overcoming weakness in traditional methods such as graphs charts and tables that require considerable amount of time and know how to absorb the extensive information and show an accurate picture of what happened. Interpretation can dilute the content of the information which is why this method is most useful.
DAAI utilized the Embraer FLYBACK animation software. This enabled the investigation to visualize the last critical moment of the occurrence and with the combinations of other methods have a better understanding of the last phases of the flight. This clearly shows how the controls were manipulated and the reaction of the aircraft to those inputs.
Below is a snapshot from the Animation which depicted the ‘controlled’ descent of the aircraft auto flight system. This snapshot illustrates that the altitude was selected at 592ft which is below the area’s field elevation of around 3600ft AMSL. It also displays the excessive descent of the vertical speed at 10158ft/minute.
The speed was selected at 309 knots and because of the high rate of descent the actual speed was way higher than the selected speed. The cabin altitude remained at a comfortable 6300ft ruling out any rapid decompression as a motive to descend. At this point the aircraft was still in an altitude flight level change mode meaning the aircraft is trying to achieve selected altitude of 592 ft given the selected speed of 309 knots.
Speed brakes are displayed as being deployed, an indication of the desire to achieve the highest rate of descent in autopilot mode to counter corrective autopilot maneuver that would level the aircraft as a speed protection to prevent the aircraft reaching Vmo (over speed).
The throttle level at this point was at flight idle which confirmed by the N1 (engine fan speed) N2 engine rotor speed indication. The angle of attack was normal ruling any possibilities of stall.
Figure 19. LAM 470 at peak of descent with the VS of -10158 ft/m. From Directorate Of Aircraft Accident Investigation Accident Report
Flight crew qualifications
The flight crew was suitably qualified and experienced in their respective roles. At the time of the accident they were within their allowable duty time and they were suitably rested.
Conduct of the Flight
LAM 470 departed Maputo at 09:26 for Luanda estimated their arrival at 13:10. The Flight operations were normal and the aircraft was cleared by Gaborone ATC for FL380 (38000ft) which was maintained from 9:55 till 11:09 just before the position EXEDU.
Most of the conversation in the cockpit for the first one hour and fifty minutes of the flight was dominated by general discussion about the country’s politics and social activities. There was a cordial, if not pleasant, conversation between the two crew members in the cockpit, at no point was there a hint of any un-procedural activities or other deviation.
After 1hour 50 minutes into the flight the First Officer stated he had to go to the toilet and asked the captain if he had controls to which he responded “no problem” and thereafter sounds similar to the door unlock jingle are heard and then immediately after, the electromechanical door latch are closing.
From this time, there was no other intra-cockpit conversation recorded on the CVR indicating that the Captain remained alone in the cockpit which is not in line with the company procedures.
At 11:04 The aircraft was flying at FL380, with ALTITUDE PRESELECTOR set to 38000 ft. The autopilot was engaged (and remained engaged until the end of recording) with flight director vertical mode VALT (FMS altitude hold).
At 1:53:31 hours into the flight the CVR picked click sound of the altitude preselect rolling, these sounds are consistent to the FDR outputs which showed that at;
11:06:36 – the altitude preselect adjustment (from FL380 to 4,288 ft, 11:06:52 – the altitude preselect adjustment (from 4,288 to 1,888 ft, 11:07:08 – the altitude preselect adjustment (from 1,888 ft to 592 ft,
At 11:07:41 the FDR recorded the AUTOTHROTTLE ENGAGE parameter being transitioned from ENGAGED to NOT ENGAGED at this time the MASTER CAUTION remained inactive. This can be attributed to the action being performed manually i.e. through commanded input rather than inadvertently ((In case of an auto-throttle failure, a caution CAS message would be displayed to the crew and would be recorded in the DFDR. As the parameter MASTER CAUTION remained not active by the time of the disengagement, it is possible to infer that this action was manually performed.)
11:08:31 The engaged autopilot vertical mode transitioned from FMS altitude hold (VALT) to FMS flight level change (VFLCH) and thereafter flight level change (FLCH) (There are two conditions to consider: an FMS failure or a manual selection. In an FMS failure, the vertical mode would drop to the AFCS basic mode (FPA), which was not observed. Moreover, the lateral mode remained engaged in the LNAV, what indicates that the FMS was available along the whole descent. Therefore, it is possible to infer that these transitions were manually commanded by pressing the FLCH pushbutton and subsequently the VNAV pushbutton on the guidance panel).
11:08:42 – The autothrottle was manually reengaged and the throttle levers were automatically retarded. This is the expected behavior as the FLCH mode was engaged and the desired altitude (altitude pre-selector) was below the current aircraft altitude.
11:09:01 – After the throttle levers were reduced, the autothrottle was disengaged. The MASTER CAUTION parameter remained off, this can be attributed as the disengagement was manually performed.
11:09:26 – The TLA (throttle lever angle) parameters indicate an advance and subsequent retard back to IDLE at 11:09:35. This action was manually accomplished as the autothrottle was disengaged.
11:09:52 – The parameter SELECTED AIRSPEED AUTO transitioned from ACTIVE to INACTIVE at the same time that the SELECTED AIRSPEED MANUAL transitioned from INACTIVE to ACTIVE (The AFCS (auto flight control system) automatically limits the manually selected speed to VMO. This is evident from the Speed Brake Handle parameter indicator which gets its pick up from the proximity sensor on the Handle as it transition from one position to next.).
11:10:54 – The SPEED BRAKE HANDLE parameter indicates that it was commanded to open the spoiler panels and remained in this position until the end of the recording. This was manually commanded as the parameter monitors the handle position.
After the speed brakes were commanded open, the vertical speed rises reaching a maximum value of 10560 fpm at 11:11:34. Also, the indicated airspeed rises, leading to the automatic transition of the flight director vertical mode from FLCH to OVSP (over-speed) at 11:11:34; 11:13:57; 11:15:01; 11:15:08 and 11:15:15.
At 11:12:52, the BLEED 1 PRESS and BLEED 2 PRESS drop simultaneously to near zero. The fact that there is no MASTER CAUTION parameter activation indicates that both bleeds were intentionally deactivated.
Between 11:13:27 and 11:13:33, while the aircraft was crossing 17,000 ft, the parameter MASTER CAUTION was activated. Technical analysis revealed that this MASTER CAUTION is associated with the detection of icing conditions. In this situation, the anti-ice system will be automatically activated. However, since the pilot turned off both bleeds, the anti-ice system activation was not feasible. In response, the A-I WING FAIL caution message was generated.
At 11:16:01 (25 seconds before the end of recording), the GPWS CAUTION parameter indicates that the first GPWS alert was triggered as the aircraft crossed 2,010ft AGL (5,150ft ASL) followed by a second activation at 11:16:06.
At 11:16:14, the parameter GPWS WARNING was activated and at 11:16:24 the GPWS TERRAIN PULL UP parameter was activated.
The pilot switched the Speed outer knob from AUTO to MAN in the guidance panel then rotated the inner knob to desired airspeed (SELECTED SPEED).This is evident by the SELECTED AIRSPEED AUTO transitioned from ACTIVE to INACTIVE at the same time the SELECTED AIRSPEED AUTOSPEED MANUAL transitioned from INACTIVE to ACTIVE. The SELECTED AIRSPEED parameters indicated that the desired speed was manually selected several times until the end of the FDR recording. This speed remained close to VMO8 throughout the remaining portion of the descent.
At around 1:56:46 into the flight, sounds were heard of someone trying to open the door. Spoiler panels were then commanded open and remained in this condition for the rest of descent.
This action (deploying speed brakes) raised the vertical speed to maximum value of 10560ft/m seen at 11:11:34 as well as increasing the vertical mode from FLCH to OVSP (over speed).
The MASTER CAUTION was activated at 17000 ft between 11:13:27 and 11:13:33.
Technical analysis revealed that this MASTER CAUTION is associated with the detection of icing conditions. In this situation, the anti-ice system will be automatically activated. However, since the pilot turned off both bleed, the anti-ice system activation was not feasible. In response, the A-I WING FAIL caution message was generated.
EGPWS WARNINGS were heard and parameter warnings activated at 11:16:06 and “TERRAIN
PULL- UP” sounded but no action by the captain was evident. Test results from the simulator indicate even at this point the aircraft would have been recovered
The Captain’s 72 hrs history before the accident
Information provided to the investigation team by Mozambique Airline stated that only information related to the captain that pertain the airline could be obtained and listed as follow:
♦ 26th November flew on the route Maputo – Tete – Maputo (TM 136 – 137)
♦ 27th November flew Maputo – Johannesburg – Maputo ( TM 315 – 306)
♦ 28th November – Leave
♦ 29th November flew TM 470
NB: Although information on the Financial and Insurance position of the Captain was requested from Mozambique through their accredited representative, DAAI could not obtain this information due to huge bureaucratic and legal/judicial hurdles until the finalization of the final report.
Photo: © Globespotter https://c1.staticflickr.com/9/8442/7930280030_d3694905b6_b.jpg
1. The flight crew members were licensed and qualified for the flight in accordance with the Mozambique Institute of Civil Aviation’s existing Regulations.
2. Their medical certificates were valid and the aircraft type endorsed into their licenses.
3. The aircraft’s Certificate of Airworthiness (C of A) and that of Registration (C of R) were valid at the time of the accident.
4. The crew was properly rested and well within their duty times.
5. The en-route weather at 38000 ft which was the selected aircraft cruising level was fine and was not considered to be a contributory factor to the accident. However, the deteriorating weather at the accident site due to rain and thunderstorms that started building up late in the afternoon causing search and rescue to be called off until the next morning.
6. There was no evidence to indicate that some passengers didn’t pass through security screening as all passengers were issued with passenger tickets and had boarding passes.
7. 1hour 50 minutes into the flight, the first Officer (F/O) left the cockpit for the lavatory and only the Captain remained on the Flight Deck. The cockpit door was then electronically locked and there was no further conversation in the cockpit.
8. The Captain manually selected the ALT SEL three (3) times from 38000 feet cruising altitude, to the final 592 feet setting.
9. The Captain manually disengaged the Auto throttle then later manually reengaged it which caused the throttle levers to retard automatically. The speed was then altered manually, several times until the end of the recording, which remained close to Vmo (Maximum Operating limit speed).
10. The speed brake handle parameter that monitors the handle position indicates it was manually commanded to open the spoiler panels and remained in that position until the end of the recording.
11. Actions performed by the Captain indicate explicit knowledge of the EMB-190 Systems and specifically the automatic flight control system that is evident as the entire descent was conducted with the Autopilot engaged and no force applied to the control columns.
12. The Emergency Locator Transmitter (ELT) was found relatively intact but due to the fact that its antenna cable was cut off from the ELT during the aircraft impacted the terrain, there was no reception of ELT transmission by the global emergency centers which operates on satellite relays.
13. Alerting services were not activated as per procedure by Gaborone ACC i.e. when aircraft failed to report at AGRAM , Gaborone ACC did not commence Uncertainty Phase as per procedure but rather waited for Luanda ACC to report when they have established contact.
14. There were no communication difficulties reported. The pilot last communicated with Gaborone Area Control Centre on frequency 126.1 MHz at around 10:19:32 as the aircraft was cruising at FL380 before it made an abrupt descent.
15. Gaborone Air Traffic Controller (ATC), who was alone on duty and operating a combined Area East 126.1 MHz and Area West 127.1 MHz which is the entire upper airspace of Botswana, could have coupled the two frequencies together for him to be able to control these sectors effectively.
16. The Botswana Radar failed the controller by not flushing a red light or setting of an alarm for him to signal that the aircraft has started commencing a sudden descent and deviated from it flight plan route, therefore, could not execute an emergency or any alerting alarm.
SIMULATOR INVESTIGATIONS FINDINGS
17. The behavior of the simulator was deemed identical as the actual flight although the flight dynamics revealed very small differences which can be attributed to the limitations of replicating all environmental conditions as they prevailed at the time such as actual avionics loads, wind component at the actual area, systems logics etc.
18. It was also evident that actions recorded on the FDR were commanded rather than inadvertent as it required cognitive effort to input.
19. Logical sequences of inputs were commanded as a counter response to the corrective actions initiated by the Auto-Pilot.
1. The inputs to the auto flight systems by the person believed to be the Captain, who remained alone on the flight deck when the person believed to be the co-pilot requested to go to the lavatory, caused the aircraft to departure from cruise flight to a sustained controlled descent and subsequent collision with the terrain.
2. Contributing factors
2.1 The non-compliance to company procedures that resulted in a sole crew member occupying the flight compartment.
As a result of the investigation of this accident the DAAI make the following Safety Recommendations:
To the Mozambique CIVIL AVIAIATION AUTHORITY:
Safety recommendation number 001/2015 LAM
DAAI recommends that Mozambique Civil Aviation Authority should come up with a mechanism to ensure that the procedure of two people on the flight deck is adhered to at all times as laid out in LAM’s Manual of Flight Operation Chapter 10.1.4, Page 5 of 36, Edition 3 Revision 8, (Absence from Flight Deck).
1.Safety recommendation number 002/2015 LAM
DAAI recommends that ICAO should establish a working group that should look into the operation and the threat management emanating from both sides of the cockpit door.
2.Safety recommendation number 003/2015 LAM
DAAI recommends that ICAO should establish standards that implement recommendations of the working group, formed under safety recommendations number 002/2015 LAM to suitably avert the locking out of the cockpit of authorized crew members.
3.Safety recommendation number 004/2015 LAM
DAAI recommends that ICAO should establish a working group to review the installation of visual recording inside and outside the cockpit that should provide information on who was in the cabin, who exactly was controlling the plane at the time of the accident and even where their hands were in relation to the plane’s controls.
4.Safety recommendation number 005/2015 LAM
DAAI recommends that ICAO should expedite the implementation of international requirements on global tracking of airline flights providing early warning of, and response to, abnormal flight behavior information to ensure that search and rescue services, recovery and accident investigation activities are conducted timely.
5.Safety recommendation number 006/2015 LAM
DAAI recommends that ICAO working group (Global Tracking 2014-WP/6) speeds up the research and implementation of aircraft tracking and localization other than ELT system.
Excerpted from Namibia Directorate of Aircraft Accident Investigation Accident Report http://www.mwt.gov.na/published-reports
Do you really believe that the presence of a cabin crew member in the cockpit, while the copilot left to the lavatory, would have prevented this accident? Does a person that has decided to execute actions like the ones described in this report could be stopped by a subordinate? I don’t think so
By Laura Victoria Duque Arrubla, a medical doctor with postgraduate studies in Aviation Medicine, Human Factors and Aviation Safety. In the aviation field since 1988, Human Factors instructor since 1994. Follow me on facebook Living Safely with Human Error and twitter@dralaurita. Human Factors information almost every day