On 19 March 2016, a Boeing 737-800 (A6-FDN) being operated by Flydubai on a scheduled international passenger flight from Dubai to Rostov-on-Don as FDB 981 failed to complete a second approach at destination and, following an initial climb, transitioned to a steep descent whilst in night IMC and subsequently crashed onto the intended landing runway. The aircraft was destroyed by the high energy impact and the explosion which followed and all 55 passengers and 7 crew members were killed. Damage was caused to the runway surface and to runway lighting and the airport remained closed for two days to allow wreckage recovery and infrastructure repairs.
A general view of the crash site after the removal of larger pieces of wreckage. [Reproduced from the Official Report]
An Investigation was carried out by the Air Accident Investigation Commission of the Interstate Aviation Committee (MAK). Both the FDR and the two-hour CVR were recovered from the wreckage having sustained "substantial mechanical damage" but the data from both was successfully downloaded. Recorded ATC radar and communications data were also available.
An Interim Report with some initial factual information and including five "Prompt Safety Recommendations" was issued on 20 April 2016. These Recommendations were intended to direct attention to what appeared to be key aspects of the accident at that time and were as follows:
- Inform the flight and maintenance personnel operating Boeing В737-800 aircraft about the accident.
- Have additional training on elements of go-arounds in various conditions, in manual control mode with two engines operative from various heights and at low aircraft weights.
- Study the possibility of introducing into the Full Flight Simulator (FFS) training program scenarios which involve go arounds in various conditions, including under manual control with two engines operative from various commencement heights.
- Carefully study and analyse the implementation of Safety recommendations issued by investigation teams which investigated the 2013 Boeing 737-500 accident near Kazan Airport and the 2006 A320 accident near Sochi Airport.
- Carefully study the potential applicability of the Safety Recommendations to prevent accidents and incidents during go-around developed in the BEA Study on Aeroplane state awareness during go-around (ASAGA) and take any applicable safety measures.
Upon completion of the Investigation, a fully detailed reconstruction of what had happened and the most likely explanations provided by the available evidence was made available and a summary of this now follows.
The Flight Crew
It was noted that the 37 year-old Captain had a total of 5,965 flying hours of which 4,692 hours had been on type and 1,056 hours of those had been in command. The 36 year-old First Officer had a total of 5,767 total flying hours of which 1,100 hours were on type. Both pilots had begun their professional flying careers in Spain and by the time of the accident, their current ATPLs were ones issued by the UAE GCAA. Prior to 2004, the Captain had been flying light aircraft (under 5,700 kg) but had then qualified as a 737 Classic pilot and worked for Cyprus-based Helios Airways as a First Officer between 2004 and 2006 before moving to do the same job at UK-based XL Airways between 2006 and 2008. In 2011 he obtained a 737 Classic First Officer job with Malaysian Airlines System before joining Flydubai as a First Officer in 2012. His 737 NG transition training had then been completed at the CAE flight training facility in Dubai prior to commencing line flying. He had completed command up grade training for the 737-800 in January 2015. The First Officer had also initially flown professionally in light aircraft before working as a First Officer on the ATR 42/72 for a Spanish regional airline between 2006 and 2008 and then, after almost a year unemployed, worked as an ATR 42/72 pilot between 2009 and 2013 during which time he achieved a command on type. He then qualified as a First Officer in the A320 and worked as a First Officer on that type before joining Flydubai in August 2014 and, after type training covering the 737-300/900, flying as a First Officer on their 737-800 fleet.
It was established that the two pilots had not flown together previously nor had either of them flown to Rostov-on-Don before, although the Captain had made 14 previous flights to Russian Federation airports. Eight of those flights had been made since his promotion to Captain. The operator’s Operations Manual (OM) noted the relatively shallow glidepath angle at Rostov (2° 40’) and the necessity to confirm whether QNH or QFE would be used for airport vertical clearances by ATC and required the Captain to perform the approach, landing and takeoff. It was noted from the CVR that the two pilots had conversed operationally exclusively in English but non-operationally, including with a member of the cabin crew, at times in Spanish.
The History of the Flight
Pre-flight information provided to the crew was as would be expected with the exception that a SIGMET for the Rostov-on-Don FIR which forecast severe turbulence south of 48° N and west of 48° E from ground level up to FL150, an area including the airport, was not provided. The nominated alternates were Volgograd and Trabzon (Turkey).
With indications that the First Officer had acted as PF until almost the end of the cruise, made at FL 360, the flight appeared to have proceeded uneventfully. The first relay of the actual weather and runway in use at Rostov was provided by Rostov Control whilst the flight was still at FL 360 and this gave relatively normal conditions except for BKN CB at (the equivalent of) 2,950 feet and a TEMPO of a reduction in visibility from 5km to 1000 metres in a rain shower with runway 22 in use. Still at FL 360, the flight was transferred to another Rostov Control frequency where the controller advised them of the severe turbulence SIGMET but on query from the crew also stated that “no aircraft in the specified area had reported turbulence”. A few minutes later, the flight requested descent and was cleared to FL 190, which was commenced. Due to the time which elapsed due to holding after an initial unsuccessful approach at Rostov (detailed below) it was noted that the time period covered by the CVR did not include the crew briefing for approach and landing.
On requesting further descent as the aircraft passed FL 220, the flight was transferred to Rostov Radar and on checking in there was re-cleared to FL 060 and then, as FL 060 was approached was transferred to Rostov APP where re-clearance to 600 metres on QFE 990 - the equivalent of 2250 feet on QNH (the airfield elevation being 259 feet) - was given continuing on the STAR already being followed. Since the operator’s altitude reference system when below altitudes which use the standard pressure setting is based on the use of QNH and feet, pilots must use conversion tables provided. The Investigation noted that, although after this re-clearance, the crew had mistakenly set the QFE on their altimeter subscales, this had no actual effect on the flight before it was corrected to the QNH equivalent of 1000 other than a slightly higher glideslope interception than intended but concluded that this served to show that “the use of QFE instead of QNH at Russian Federation airports as used in practically all other countries, presents extra risks”.
Clearance for an ILS approach to runway 22 was subsequently given by Radar and the aircraft became fully established with evidence that the Captain was using his HUD monitored by the First Officer using conventional instruments. The go around altitude of 600 metres QFE was set (correctly) to the equivalent QNH of 2300 feet. Initially the AP and A/T remained engaged but as the aircraft entered increasing turbulence, the AP disconnected (it was unclear whether this was due to an intentional action on the Captain’s part or a result of the “intensified turbulence”). FDR data showed that the required flight path was maintained accurately with the A/T still engaged. The Captain was recorded as having the runway in sight at a range of just inside 4 nm but this was followed almost immediately by a predictive windshear warning - ‘GO AROUND, WINDSHEAR AHEAD’. Since this was not a warning that windshear had actually been encountered, the QRH-required response was to perform either the ‘Windshear Escape Manoeuvre’ (in which the configuration remains unchanged and maximum thrust is “aggressively applied”) or the normal go around procedure in which it was possible to use a reduced thrust when initiating the go around as well as making the usual configuration changes.
At 1,100 feet aal with a recorded 160 KCAS, the Captain initiated a ‘Windshear Escape Manoeuvre’ and as the First Officer advised ATC of the go around, with only 1200 feet to climb before the go around stop altitude of 2,300 feet was reached, the ALT ACQUIRE mode was almost immediately activated. With the ‘windshear ahead’ alert still displayed, thrust was not reduced and the Captain began to move the control column forward to reduce pitch and made a corresponding 4 second stabiliser trim input. The effect was an increase in speed which soon reached the VFE (maximum flap extended airspeed) for flap 30°, the unchanged approach setting, which triggered the automatic load relief retraction of the flaps to 25°.
Climb continued above 2,300 feet without clearance and the Captain reduced thrust from 101/102% N1 to 83% N1 in response to the exceedance. With the windshear warning still displayed, further climb was requested from ATC and the flight was re-cleared to (the equivalent of) 2,950 feet QNH and after a further crew request mentioning the windshear warning to (the equivalent of) 5,000 feet QNH. As the aircraft climbed through 3,750 feet with a recorded 176 KCAS with the flap lever still at 30° (although with the flaps at 25° due to the automatic load relief), the Captain re-engaged the AP and A/T and the landing gear and flaps were selected up.
On transfer to Radar, the flight was instructed to maintain FL 050 and informed that it was now number 2 for another approach. Weather avoidance then followed and the Captain eventually decided, since there was considerable fuel remaining consequent on the operator’s fuel tankering policy, that rather than make a further approach straight away, he would wait to see if the aircraft ahead, an Aeroflot SU95, landed. This flight then made a go around from 1970 feet due to windshear. The Captain decided to take up a hold and await an improvement in the conditions, aware that the available fuel was considerably greater than that required for the preferred destination alternates of Volgograd and Mineralnye Vody, both about 45 minutes flight time away and where a check had confirmed that the present and forecast weather conditions were fine. Whilst manoeuvring under radar control to take up a hold clear of adverse flying conditions, ATC advised that the Aeroflot SU95 had made a second approach which had also ended in a go around due to windshear and then a little later that it had made a third approach and go around and followed this by diverting to Krasnodar.
On reaching a suitable holding position the crew requested a climb to FL150 so as to remain clear of what were reported to be moderate icing conditions and once there set the best endurance holding speed of 210 knots. Whilst in the hold, the Captain contacted Company by SATCOM and obtained concurrence with his proposed choice of Mineralnye Vody as the most suitable diversion should this be necessary after one further approach and subject to continuing to hold as long as possible before making this approach if required. The crew then prepared for a further approach and possible go around and for the diversion as agreed if necessary.
Various flight deck discussions on what might be described as underlying concerns were evident from the CVR. These included the fact that the flap speed exceedance during the go around meant that a special inspection would be required before further flight and the prospect of being out of hours to operate the return flight to Dubai and having to nightstop in either Rostov or Mineralnye Vody as well as the possibility that if they diverted, they might be required to position empty back to Rostov. The Investigation formed the view that the Captain was strongly motivated to land at Rostov despite being aware that another predictive (or reactive) windshear warning on approach would prevent this.
After approximately 1 hour 45 minutes holding the Captain decided, having assessed the apparent weather conditions, which included no further pilot reports of turbulence (although the latest Rostov METAR obtained from the TWR controller had a plain language addition of “on final severe turbulence and moderate windshear”), to make a second approach. With the corresponding preparations, including a landing briefing having been completed, ATC was advised accordingly and a descent clearance was given. As descent began, the crew noted that the reduced VREF compared to the first approach would give a consequently bigger gap between it and the approach flap 30 VFE if a Windshear Escape Manoeuvre was flown and the Captain said that if another windshear warning occurred, he would respond with the Escape Manoeuvre (unchanged configuration and maximum thrust) rather than fly a normal go around. It was also agreed that ATC would be advised that if the approach was discontinued, a climb to FL 080 would be required and this was done and acknowledged.
The descent and vectoring onto the runway 22 ILS LLZ was uneventful and the aircraft became established at a range of 11.5 nm level at an altitude of 2250 feet QNH and was transferred from Radar to TWR. A landing clearance was given with the surface wind velocity given as from 230° at a speed (equivalent to) 23-35 knots. Descent on the ILS GS began at a range of 7nm and after 26 seconds, at an altitude of 2165 feet QNH, the Captain disengaged the AP without giving a reason (which was not required under operator SOPs based on Boeing documentation). Thereafter, he flew the aircraft manually by reference to the HUD whilst the First Officer monitored his actions on his EFIS displays. At 1960 feet QNH, the A/T was disengaged too and landing flap 30 was selected, following which the landing checks were run.
As the aircraft passed the 1000 feet QNH, the Captain’s response to the auto-callout was “stabilising now” with the speed slightly above target but reducing and with ILS deviations within the required tolerance. However, as soon as the target approach speed of 153 knots was reached, the combination of an increase in thrust, the Captain’s forward pressure on the control column and gust of wind caused the airspeed to jump 15 knots in one second and in the next two seconds to 23 knots. The First Officer called “check the speed” and the Captain followed this immediately with the call “go around” and set TO/GA (full) thrust as the First Officer acknowledged the call.
The go around was initiated from (the equivalent of) 1100 feet QNH with the speed just within flap 30 limits. It was noted by the Investigation that “it cannot be clearly established if the Captain perceived the abrupt increase of speed readings as one of the signs of the windshear encounter”. However, it was noted that according to the QRH, a 15 knot increase in IAS when below 1000 feet represented “an unacceptable flight path deviation” which might be “one of the signs of a windshear encounter or as the ordinary sign of an unstabilised approach”. The Investigation “assumed that in advancing thrust levers to full thrust the Captain had ‘in his mind’ initiated the Windshear Escape Manoeuvre, particularly given that, if he perceived the speed jump in question as associated with an actual windshear encounter, this very manoeuvre is prescribed in the QRH without an alternative”.
However, the evidence suggested that in the absence of any clarification beyond the Captain’s “go around” call, it was possible that in almost immediately offering, in a questioning voice “flaps 15?”, the First Officer may have believed that it was his overspeed alert alone which had led to the Captain deciding to go around, especially as the Captain (who should have made the call himself if he was making a normal go around) immediately accepted. This point of confusion in the flight was identified as “most probably” having been “a turning point in the chain of the events”.
TWR were immediately advised of the go around and instructed the flight to contact Radar although no contact was subsequently made. The maintenance of maximum thrust as gear was retracted and flaps set to 15° at what was a reduced aircraft weight compared to the first go around led to an even greater increase in the pitch-up moment than had occurred in the earlier go around. This required a pushing force on the control column of up to 23 kg which was then continued for more than 40 seconds whilst the aircraft remained out of trim. FDR data showed that the required 15° pitch up had not been reached initially and that thereafter the achieved pitch had ceased to follow the pitch command bar. The aircraft continued to be flown out of trim indicating that the Captain was “getting behind the aircraft”. Within less than 20 seconds from the beginning of the go around, the First Officer had begun repeatedly prompting the Captain to fly the required 15° nose up.
However, “having focused the attention on maintaining pitch, both pilots had completely lost the awareness of the airspeed” which gradually increased to 200 knots, the limit for flaps 15º at which point the load relief system set them to 10º. It was concluded that they were effectively demonstrating ‘tunnel vision’ with the number of parameters being monitored and analysed at a time rapidly decreasing. It was considered that a “critical situation” had existed within 40 seconds of the go around beginning. The aircraft reached a maximum of around 3,350 feet completely out of trim at which point the Captain began an exceptionally sustained (12-second) nose-down electric stabiliser trim input. The Investigation’s recreation of the HUD image as this trimming input commenced showed that the TO/GA pitch target line would have been higher than the aircraft reference symbol and that “most probably, the Captain was seeking to “relieve” his pushing force on the control column, rather than to transition the aircraft to nose down” although he did concurrently return the control column to the neutral position. The result was the rapid (around 6°/second) decrease of pitch and the aircraft quickly transitioning from climb to descent with significant negative ‘g’ (up to - 1.07 g) followed by the near-zero and negative g as a steep (40°) descent developed. After 500 feet of this descent, the airspeed had increased to 280 knots and was still rising.
The CVR data at this time showed that the First Officer realised that the hazardous situation was emerging and had “desperately, with the increasing anxiety in his voice”, tried to get the Captain to pull back on the control column to recover the situation and prevent an accident. However in the absence of any response, the Investigation concluded that by this time, the Captain had no longer heard the First Officer’s calls and “had fully lost the control of the situation, as well as the capability to control the aircraft”. For a few more seconds, it was calculated that had the control column been placed in a fully back position and kept there, then “the aircraft could have been recovered from descent with a sufficient height margin” but although the First Officer eventually began to pull on the control column, ground impact was only 6 seconds away. An EGPWS PULL UP warning occurred although too late for any response to have been effective at the prevailing 18,000 fpm rate of descent. Impact occurred close to and on the intended landing runway with a nose down pitch of about 50° and a forward speed of approximately 340 knots just one minute after the go around had been initiated. The resultant explosion and fire which occurred completely destroyed the aircraft and killed all on board.
The Cause of the accident was determined to have been “an incorrect aircraft configuration and piloting which was followed by the aircraft commander’s loss of situational awareness in night IMC and resulted in a loss of control of the aircraft and its impact with the ground”.
A total of 10 Probable Contributory Factors were identified as follows:
- the presence of turbulence and gusty wind with the parameters, classified as a moderate-to-strong “windshear” that resulted in the need to perform two go-arounds;
- the lack of psychological readiness (not go-around minded) of the Captain to perform the second go-around as he had the dominant mindset of landing at the destination aerodrome, formed out of the “emotional distress” following the first unsuccessful approach (despite the runway having been in sight and the aircraft stabilised on the glide path, the Captain had been forced to initiate go-around due to the windshear warning activation), concern on the potential exceedance of the duty time to perform the return flight and the recommendation of the airline on the priority of landing at the destination aerodrome;
- the loss of the Captain's leadership of the crew after the initiation of go-around and his “confusion” that led to the impossibility of the on-time transition of the flight mental mode from “approach with landing” into “go-around”;
- the absence of a requirement to announce the type of manoeuvre intended when initiating a go-around in the aircraft manufacturer documentation and the airline OM;
- the crew’s uncoordinated actions during the second go-around - despite the low weight of the aircraft, the crew was performing a standard go-around procedure (with the retraction of landing gear and flaps), but using the maximum available thrust consistent with the Windshear Escape Manoeuvre procedure that had led to the generation of a substantial excessive nose-up moment and significant (up to 23 kg) ‘pushing’ force on the control column to counteract it;
- the failure of the aircraft commander over a long time to fly the pitch required to perform a go around and maintain the required climb profile while piloting an aircraft unbalanced in forces;
- the insufficient knowledge and skills of the aircraft commander in respect of the operation of the manual stabiliser trim which led to the long period (12 seconds) of continuous stabiliser nose-down trim with the subsequent substantial imbalance of the aircraft and its upset encounter with the generation of the negative ‘g’ which the crew had not been prepared for. The potential impact of a somatogravic ‘pitch-up illusion’ on the aircraft commander might have contributed to the long time he kept the stabiliser trim switches pressed;
- the psychological incapacitation of the aircraft commander resulting in his total spatial disorientation which did not allow him to respond to the correct prompts of the First Officer;
- the absence of criteria for assuming the psychological incapacitation of a pilot in the airline OM which prevented the First Officer recognising the situation in time to take more decisive action;
- the possible ‘operational’ tiredness of the crew since by the time of the accident the crew had been in flight for 6 hours including 2 hours under a high workload that involved the need to make non-standard decisions; in this context the accident occurred at the worst possible time in terms of the circadian rhythms - at a time when human performance is severely degraded to its lowest level along with an increase in the risk of error.
It was also noted that “the lack of the objective information on the operation of the HUD” during non-normal circumstances (flight testing of the HUD did not cover the entire range of ‘g’ which might be experienced, including negative ‘g’) and the impossibility of reproducing the HUD display as it would have been seen when reconstructing the progress of the accident flight (as also affected by the pilot’s posture in his seat at the time) did not allow conclusions to be drawn in respect of its possible effect on the outcome of the flight”.
The Investigation team also concluded that “the specific features of the HUD indication and display in the conditions which existed during the final phase of the accident flight (severe turbulence, the aircraft upset encounter with the resulting negative ‘g’ and the significant difference between the actual and the target flight path) that generally do not occur during standard simulator sessions, could have affected the situational awareness of the aircraft commander having (already) been in a highly stressed state”.
A total of 26 Safety Recommendations were issued at the conclusion of the Investigation as follows:
- that the Federal Air Transport Agency (FATA) inform flight personnel, the air training centre staff and ATC officers of the results of this investigation at their subject-oriented debriefing meetings.
- that the Federal Air Transport Agency (FATA) considers the possibility of expediting the Russian Federation’s transition to flight operations on QNH.
- that the Federal Air Transport Agency (FATA) draws the attention of organisations engaged in the testing of aviation personnel for their compliance with the ICAO Language Proficiency Requirements to the need for enhanced monitoring of these rating / examining these activities, as well as the need to further develop measures to exclude the assignment of the ICAO language proficiency levels to persons, whose competence does not meet the corresponding requirements.
- that the Federal Air Transport Agency (FATA) draws the attention of the air navigation service providers to the need for enhanced monitoring of responsibilities of instructor personnel who are engaged in the officers’ simulator training, inter alia, of the adherence to the R/T rules in English.
- that the Federal Air Transport Agency (FATA), in association with Roshydromet and the State ATM Corporation, organise training for ATC and meteorological services officers on the procedure for communicating information on windshear to flight crews.
- that the Federal Air Transport Agency (FATA), in view of the position of the aircraft manufacturer that the Boeing 737-800 operational documentation assumes that pilots’ previous experience of operating multiengine jet aircraft means that they are familiar with the basic systems and basic airmanship relevant to such aircraft, assesses the risks of pilot type rating approval if such an approval is the first jet multi-engine aircraft in the pilot’s career (and) if required amend the current regulations (and) assess the applicability of this safety recommendation in respect of other aircraft types.
- that Flydubai conduct pilot training on the specific aspects of the operation of the manual stabiliser trim.
- that Flydubai considers the practicability of changing the HUD Model 4000 changing to the Model 6000 STC ST02522SE) to equip both pilot positions.
- that Flydubai, in association with the aircraft and HGS manufacturer, considers the practicability of developing additional instructional guidelines on use of the HUD at the different stages of flight.
- that Flydubai considers the practicability of elaborating their pilot training programmes to allow for the incorporation of practical familiarisation (training) in upset conditions, including zero and negative g states.
- that Flydubai considers the practicability of amending their OM to include the criteria for determining psychological incapacitation of a pilot (in flight) and the respective recommended actions in such circumstances.
- that Flydubai considers the practicability of elaborating their SOPs to include a requirement for the PF to callout the next intended manoeuvre (for example Go-Around, Windshear Escape Manoeuvre).
- that Flydubai considers the practicability of elaborating their SOPs to include a requirement to monitor of the duration of stabiliser trim inputs and the current stabiliser position.
- that Flydubai evaluates the possible risks associated with the partial blocking of the PFD when there is significant forward deflection of the control column and takes measures in mitigation (if required).
- that the FAA and Rockwell-Collins consider the practicability of conducting additional flight testing of HGS to include all the anticipated operating conditions and the entire range of g of the aircraft with these systems installed.
- that the FAA and Rockwell-Collins, taking into account the views of experts in ergonomics and aviation psychology, consider the practicability of improving the presentation of information on the HUD in order to mitigate the risk of its erroneous interpretation.
- that the FAA and Rockwell-Collins, in association with the designers and manufacturers of aircraft equipped with the HGS, consider the practicability of development of the additional guidance on the use of HUD at different stages of flight.
- that the FAA and other certification authorities (EASA, IAC Aviation Register, FATA etc.) consider the practicability of amending the applicable regulations with provisions on the mandatory flight assessment of pilot flight parameter indication systems to cover the entire operating range of aircraft which have such systems installed.
- that the FAA and other certification authorities (EASA, IAC Aviation Register, FATA etc.) consider the practicability of amending the applicable regulations that determine the procedure for STC issue for systems providing the indication systems to pilots, with the requirement to the manufacturer of the equipment in question to have the hardware/software package available to reproduce the indication as per the FDR data in real time and in the scope, sufficient for the investigation of the aircraft accidents and incidents.
- that the Boeing Company considers the practicability of amending the 737NG FCTM Low Altitude Level Off - Low Gross Weight section with more detailed information on the criteria that the pilots should follow to determine the point, when the maximum thrust should be reduced, including go around performance in windshear.
- that the Boeing Company considers the practicability of implementing design changes of the stabiliser control system to reduce the risk of pilots setting the stabiliser in-flight to an out-of-trim position.
- that the Boeing Company considers the practicability of elaborating SOPs to require the PF to specify the type of go around manoeuvre to be flown (for example ‘go-around’ or ‘windshear escape manoeuvre’).
- that the Boeing Company, in association with FAA, assess the possible risks arising if sight of the PFD is partially restricted when there is considerable forward deflection of the control column and take measures to mitigate such risks if necessary.
- that the Boeing Company, taking into consideration the information given in the FCOM in respect of the statement that the control column actuated stabilizer trim cutout switches stop operation of the main electric and autopilot trim when control column movement opposes trim direction, in which the term “movement” is ambiguous, considers the practicability of introducing additions and amendments to the FCOM and/or the FCTM to explain the stabiliser control sequence on the Boeing 737 aircraft under different conditions and considers the applicability of this safety recommendation for other aircraft families.
- that the Boeing Company consider the practicability of the introducing additions and amendments to the FCOM and/or the FCTM explaining the general principles of stabiliser use and the trimming out of forces as well as the need to monitor the current stabiliser setting.
- that the ICAO considers the practicability of establishing a working group to study the issue of psychological incapacitation of flight crew members and the elaboration of relevant recommendations to provide guidance to operators and States in the drawing up and approval of the OM.
The Investigation was completed on 25 November 2019 and the Final Report in an English language translation was released the following day.