Mathematics vs. Meteorologists

Last month, the US Coast Guard Marine Board of Investigation concluded a two-week hearing into the disappearance of the Tote-operated MV El Faro, which sank off the Bahamas on October 1st after encountering Hurricane Joaquin, leading to the tragic loss of 33 crew members.

As has been reported in the , the findings of this latest hearing have revealed that the vessel was receiving weather information that was 10 hours out of date around the time that Joaquin intensified. This means that the crew did not have an accurate track of the storm and may explain why from the ship tracking data indicated that the El Faro was at near full speed going into the centre of the storm before it lost propulsion.

The news agency Reuters had reported that by the evening the ship was trapped between the track of the storm, which had shifted further southwest during the day, and the Bahamas, directly to the west of the vessel. This revelation that the ship was in stormy waters even before it lost power – waters that any normal sailing plan would avoid routinely under a mariner’s ‘ guideline in order to keep out of a tropical storm or hurricane’s potential danger area – was, until these recent findings, mystifying.


Hurricane Joaquin: Problematic To Predict

Through its developmental period, it proved especially problematic to predict the forecast track and intensity of Hurricane Joaquin, a point that was made clear by the World Meteorological Organisation’s Regional Specialised Meteorological Centre in Miami, the National Hurricane Centre (NHC) after the event. Up until September 29th, Hurricane Joaquin had been moving slowly south-west and was expected to continue westward before changing direction and heading north.

At the time, global numerical models were split between two differing tracks: the low resolution Global Forecast System (GFS) predicted that the hurricane would take a more northerly track. Conversely, the high resolution European Centre for Medium Weather Forecasts (ECMWF) predicted that the hurricane would be continuing south-west. The ECMWF forecast proved to be correct in this event and the NHC quickly shifted from a consensus of the two tracks to one that followed the ECMWF prediction. The timing for these changes, together with delays in getting this updated information to the vessel, remain key concerns for now.

Hurricane Joaquin

Image Source: NOAA


The Risk Of Forecasting By Numbers

Aside from the potential for delays in data being updated and received, the above example highlights a key concern in the use of on-board data displays and their use of Numerical Weather Predictions at sea. Unlike managed services, they do not always come with expert meteorological guidance as standard and rely on the end user to interpret the data themselves. Moreover, there is just as much chance that in another situation, a GFS forecast could be more accurate than ECMWF predictions. But, without expert analysis, mariners may still find themselves at risk of conflicting data and interpretation. And while some weather providers do provide some quality assurance over data being pushed to sea, this is not the case for everyone.

Although it is acknowledged that Numerical Weather Prediction has significantly improved the accuracy of weather forecasts in recent times – forecasting at 5 days has improved by over 20% between 1996-2014 – and has enabled the development of various services for mariners, from the International Union of Marine Insurance (IUMI) indicate that the leading cause of total losses of ships (all vessel types) between 2001-2015 was still weather with an attention-grabbing 45% of casualties between 2006-2010 and rising to 47% between 2011-2015.

With modern meteorological forecasts shifting to the provision of ‘numerical’ data capable of being displayed on computers, there has been a prolific increase in the number of on-bridge weather display systems, on-board automated routing systems and weather overlays on ECDIS. However, all of these systems rely on the accuracy of the information within the supply chain and not directly on expert meteorological expertise. In itself, this may not be the greatest concern but the issue is exacerbated by the onerous duties already expected of a ship’s master and busy bridge teams who are not, and cannot, be expected to be experts in meteorology to the extent needed to maintain safety in the circumstances encountered with El Faro.

Automation of weather forecasting services will undoubtedly continue as in many other areas of shipping but incorporating the advice and guidance from shore based experts should remain part of the process – before and during transit. If onshore teams lack the skills required, outsourcing this function to accredited marine meteorologists who have both access to more timely and reliable information as well as the maritime knowledge and experience will help to minimise the risk weather poses to ships and crew.