Indonesia Air Asia 8501: A Meteorological Analysis

Updated: 10 January 2015 (updated upper air analysis and summary).

In the early morning of December 28th, 2014, an Indonesia Air Asia A320, enroute from Surabaya, Indonesia, to Singapore, disappeared from radar screens and radio contact was lost part of the way through its cruise over the Java Sea, just southeast of Pulau Belitung Island. All 162 people on board perished. The crew had performed a deviation off course due to weather, and had requested a climb from FL320 (32,000 ft) to FL380. ATC radar showed the aircraft passing through FL363 (36,300 ft) when radar contact was lost. See here for factual updates on the flight and investigation.

Although at present there is no confirmation that weather had any part to play in this accident, and we must wait for the full details to come out in the official investigation, I have written an analysis of the meteorological conditions in the area at the time. It is purely a statement of observations as I see them and is in no way aimed at interfering with the official investigation.

The time of the accident is around 23:16 GMT on 27th December, 2014 (06:16 local time on the 28th). This is near enough to 00:00 GMT, which is one of the main synoptic hours, when many upper air soundings are taken around the world and the main numerical model analyses take place. This means we have a wide array of meteorological data available to enable us to paint a picture of what conditions the plane may have encountered. It should be pointed out that there were seven other aircraft in the area at the time, none of which reported any problems (as reported in the AvHerald).

Indonesia Air Asia 8501 route from Surabaya  (SUB) to Singapore (SIN). Original map from www.flightaware.com
Indonesia Air Asia 8501 route from Surabaya (SUB) to Singapore (SIN). Original map from http://www.flightaware.com
Navigational map from www.skyvector.com, showing the approximate final radar position of the flight relative to Airway M635.
Navigational map from http://www.skyvector.com, showing the approximate final radar position of the flight relative to Airway M635.

Below are satellite images taken round 00:00 GMT, showing vigorous convection in much of southeast Asia, including the Java Sea. In the visible image the low dawn sun angle picks out some overshooting tops showing up as small white specks. These are updrafts that are powerful enough to punch through the tropopause and are a sign of a very unstable and turbulent atmosphere. The MTSAT enhanced IR images from 22:32 and 01:32 GMT show the development of the convection, with extremely cold (high) cloudtops of well below -70 °C. The SigWx forecast chart underneath for the time, which is what the crew would have been using as a reference, showed isolated embedded cumulonimbus (Cb) with tops of 48,000 ft in the general area, with occasional Cb in the area just to the southwest of the accident site.

Visible image from 0000Z, 28 December, 2014, showing overshooting tops, a sign of vigorous updrafts. Image from JMA.
Visible image from 0000Z, 28 December, 2014, showing overshooting tops, a sign of vigorous updrafts. Image from JMA.
Infrared image at 0000 GMT.
Infrared image at 0000 GMT. From JMA
WV00Z
Water vapour satellite image from 0000Z, showing strong convection. Image from JMA.
MTSAT21Z
MTSAT enhanced IR image from 22:32 GMT, showing strong convection just to the east of the flightpath. Image from CIMSS
MTSAT00Z
MTSAT enhanced IR image from 01:32 GMT, showing the strong convection moving westwards over the flightpath. Image from CIMSS
WAFC SigWx chart from 0000 GMT, 28 December 2014.
WAFC SigWx chart from 0000 GMT, 28 December 2014.

The GFS analysis from the same time shows strong upward motion at 500 hPa (around 5.8 km altitude), especially in the Java Sea, corresponding with CAPE values of 1500-2000 J/kg. Such values would normally suggest some potential updraft velocities of up to 100 km/hr (60 mph) in the strongest cells, which would mean severe turbulence for any aircraft flying through one of these updrafts. In equatorial oceanic thunderstorms, however, updrafts can be enhanced above FL200 due to latent heat of sublimation (see here), and as this flight had the lowest cruising level of all traffic in the area, it is possible that it encountered stronger updrafts than the others. The crew had begun a diversion to the left around some bad weather, therefore their onboard radar was most likely showing strong returns in the strong cells just off to their right (north).

GFS 500 hPa vertical velocity analysis for 0000 GMT, 28 December 2014. Image from www.wetter3.de
GFS 500 hPa vertical velocity analysis for 0000 GMT, 28 December 2014. Image from http://www.wetter3.de

Below are actual upper air observations from the area for both the 300 hPa (~FL320) and 250 hPa (~FL360) levels . Each station plot shows the temperature (red), the dewpoint depression (green), wind vectors (light blue) and the geopotential height of that pressure level (dark blue), in tens of metres, e.g. “971” means 9710 metres (~31,800 ft).

The temperature at its original cruising level FL320 (300 hPa) was generally -29 °C, with a dewpoint depression ranging from 15 degrees at Surabaya (WRSJ) to 7 degrees at Pangkalpinang (WIKK). GFS model data for the location of the incident show the temperature and dewpoint at FL320 were both around -29 °C, with high relative humidity (>90%) over much of the area. This would tie in with the presence of the clouds seen on satellite, which at that temperature could still contain some supercooled water droplets in the strongest updrafts, yielding the possibility of airframe or engine icing.

At FL360 (250 hPa) the observed temperature was -40 to -41 °C, with a dewpoint depression of 7 degrees at Surabaya. Model data show temperature and dewpoint of -39 °C at the incident location, which tie in pretty well with observations, and would indicate cloud at this level also. At this lower temperature, however, supercooled water (and therefore icing) is much less likely, but if there was already accumulated icing from lower levels then the thinner air at FL360 would seriously reduce aircraft performance. Ice crystal ingestion has been known to cause engine shutdown in some situations, so this is another factor to consider.

Upper observations at 300 hPa (~FL320). Original map from University of Wyoming
Upper analysis at 300 hPa (~FL320). Original map from University of Wyoming
Upper observations at 250 hPa (~FL360). Original map from University of Wyoming.
Upper observations at 250 hPa (~FL360). Original map from University of Wyoming.
sounding text
GFS sounding data for the accident location

One interesting point is that the screenshot of the controller’s radar screen in the AvHerald article (linked above) showed a groundspeed of only 353 knots at FL363. The wind forecast chart for that time shows that winds at FL340 were forecast to be easterly at 20 knots, and at FL390 were to be easterly at 15 knots. The GFS sounding data show it as eastsoutheasterly at 12 m/s (23 knots) and the upper air observations show 15-20 knots between Surabaya and Pulua Belitung. In any case, as the aircraft was flying in a northwesterly direction it should have been experiencing a small tailwind of around 10-15 knots, and as it was climbing the total groundspeed would normally have been around 50-75 knots more than shown. If they were experiencing turbulence then they may have slowed down slightly, however 353 knots seems an extremely slow speed for normal operations and may hold a clue in the investigation.

Indeed, taking a screenshot from http://www.flightradar24.com at 23:14 GMT we can see this flight and the other traffic in the area. AWQ8501 is shown in red, flying level at 32,000 ft and with a groundspeed of 469 knots. Just ahead of it is Emirates flight UAE409 (36,000 ft/506 knots), having taken an even wider diversion off Airway M635, and just behind it is Indonesia Air Asia flight 502 (38,000 ft/466 knots), also seen turning on a more westerly heading off the airway. There are also other flights, including another Indonesia Air Asia flight 550 ahead at 34,000 ft, having taken an identical path as the ill-fated flight 8501. All flights are showing groundspeeds consistent with easterly winds of around 15-20 knots. In such a dynamic and quickly-evolving environment, small-scale updrafts can catch a plane unawares, so this could be one possibile source of investigation.

flightradar24
Screenshot from http://www.flightradar24.com, showing flight 8501 and other traffic in the area.
FL340winds
Wind and temperature forecast chart for FL340.
FL390winds
Wind and temperature forecast for FL390.

To summarise:

– The flight was in a northwesterly heading and probably climbing from its original cruising level of FL320 (32,000 ft).

– They had already initiated a deviation to the left of bad weather just before the accident, and meteorological data show some vigorous convection – and therefore moderate to severe turbulence – in the immediate vicinity. At least two other flights also appear to have taken a (wider) deviation. As this flight was at a lower level than the other traffic in the area it is feasible that it encountered stronger updrafts than the others.

– Upper air observations and model data suggest that some airframe icing and ice crystal engine-shutdown were also possible.

– ATC radar data show that the aircraft’s groundspeed at FL363 was unusually slow for the conditions observed.

The above analysis is purely my personal opinion, based on the available data. The official investigation will analyse in more depth all possible contributors to this accident, including meterological factors. Until this final report is released we should not speculate on the possible causes but remember the lives of the 162 people and their families, friends and colleagues.

Fergal Tierney

Advertisements

4 thoughts on “Indonesia Air Asia 8501: A Meteorological Analysis

  1. An excellent and thought provoking analysis, very objective and professional.

    Thank you for taking the time to put it together.

  2. Excellent write-up and meteorological discussion! Any thoughts to the occurrence of a rare positive lightning strike due to the strong convection around the flight path? From what I understand, aircraft usually are alright with taking a negative strikes, while a positive strike can have devastating and possibly catastrophic effects…

    • Nothing can be ruled out, but equatorial oceanic thunderstorms usually have very little lightning. I haven’t found lightning data for then but I would say that a very low chance that this was the cause.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s