The Search for MH370
On March 8 at 01:20 MYT Malaysian Airlines flight MH370 departed Kuala Lumpur International Airport bound for Beijing, China. On board the Boeing 777-200ER was 12 crew members and 227 passengers. The planned route was north-east over Malaysia, the Gulf of Thailand and Vietnam, arriving in Beijing at 06:30 MYT. At 1:21 the aircraft had its last recorded sighting with air traffic control. The plane was last seen on civilian radar above the waters where the South China Sea meets the Gulf of Thailand, which was detected by secondary radar near waypoint IGARI. At 1:25 the aircraft deviated from its planned route. At 6:30 MYT the aircraft missed its arrival at Beijing Capital International Airport.
MH370 is believed to have turned sharply westward, heading towards waypoint VAMPI in the Strait of Malacca. At this point the transponder signal was lost. The aircraft is believed to have traveled over the Andaman Islands.
Figure 1: Map of the ATSB search area
Based on transmissions between the aircraft and an Inmarsat satellite it is believed the aircraft continued flying on a southern track for up to seven hours, finishing in the southern Indian Ocean. The potential splash down region shown in Figure 1, supplied by the Australian Transport Safety Bureau (ATSB) was based on calculations of a range of likely aircraft speeds and associated fuel consumption.
The potential splash down region was within the Search and Rescue management area of the Australian Maritime Safety Authority (AMSA) and so AMSA took responsibility for coordinating the Indian Ocean search for evidence of the crash of MH370. The search involved ships and a number of aircraft including the RAAF P3 Orion search aircraft. Aircraft from China, Japan, Korea, USA, and Malaysia were also involved in the search.
On March 28th Dr. Graeme Hubbert and Sam Blake from Global Environmental Modelling Systems (GEMS) joined Dr David Griffin of CSIRO and representatives from AMSA and APASA to form an expert group to advise AMSA. The group’s major task was to provide daily advice as to the appropriate search area, given estimates of the aircraft splash down location and, if wreckage was found by the searching aircraft, to hindcast an approximate splash point of MH370.
GEMMS developed the AMSA Search and Rescue prediction software, known as Net Water Movement (NWM). NWM was first installed and became operational in 1998 and consists of:
1) GCOM3D (the GEMS 3D Coastal Ocean Model) which predicts ocean currents based on the following inputs:
- Marine meteorology data from the operational Australian Bureau of Meteorology (BoM) forecast models, downloaded to AMSA twice daily
- Large scale ocean currents from the joint BoM and CSIRO OceanMaps model (Bluelink)
- Tidal data in the AMSA region from extensive tidal modelling carried out by GEMS
- Bathymetric data from GEBCO, Geosciences Australia and other data sourced by GEMS
- Currents from GCOM3D
- Winds from the BoM forecast models (for leeway calculations)
- Leeway data for different objects from the US Coast Guard studies
- Estimate of splash down location
- 1000 simulations randomly selecting reasonable errors on all major inputs (current speed/direction, wind speed/direction, splash down location, splash down time etc.).
In the 16 years since NWM became operational it has been used almost daily for search and rescue operations conducted by AMSA. The difference between the MH370 search and normal NWM operations are significant however and required deeper investigations, namely:
- The search was in deep waters in the Indian Ocean and not on the continental shelf where most AMSA searches are conducted;
- The duration required for the prediction of the path of drifting debris increased to well beyond the normal period needed by AMSA (2-3 days)
- As a result of the search location being so far out in the Indian Ocean the tidal forcing would have a minimal contribution to the drift path of debris and the major contribution would come from the geostrophic and wind driven currents;
- Therefore the main contribution to the prediction of drift paths was coming from Bluelink and the BoM atmospheric forecast model (acces_r for that region)
In order to investigate errors in the predictions, GEMS undertook a comparison of predictions of the tracks of the self-locating datum marking buoys (SLDMB) dropped by AMSA during the search with the actual tracks.
This analysis provided a critical validation of the large-scale current predictions from Bluelink which drove all drift modeling undertaken by GEMS, and other agencies, throughout this incident.
The final task undertaken by GEMS was to undertake a stochastic study of the possible destination of debris from aircraft splashdown anywhere in the possible search region identified in Figure 2.