How open data can improve ship arrival times
In the present global economy, goods produced in one country are sold in any market across the world. This means that these goods need to be transported all over the world. Many of these transports, especially long-distant transports, are done by ships. These long-distance carriers are large, several hundreds of metres long, and provide a huge amount of storage. When the ship arrives at the port of destination the cargo must be unloaded to smaller ships, trains and trucks for further distribution. Subsequently, new cargo must be brought on board, again from various different sources. The unloading and loading of large ships is a complex logistic operation that needs to be planned days in advance. When a ship arrives later or earlier than expected a lot of personnel and equipment will have to wait. Better information on the arrival time makes logistic operations in harbours more efficient and helps to reduce costs.
Within the MELODIES project we are developing a service to provide improved Expected Time of Arrival (ETA) estimates for transport ships during their trip. To do this we track the ship's progress during the trip and use forecasts of wind, wave, and surface currents along the route to estimate delays.
Almost all ships nowadays are equipped with an Automatic Identification System (AIS). This system transmits the ship's identity, location, speed, and heading every 2-10 seconds. The primary purpose is to allow captains to view the traffic in a wider area around their own ship, thereby increasing safety at sea. But the AIS signals are also picked up by stations along the coasts, and publicly made available on Internet. The image below shows a snapshot of the traffic in the Channel area based on AIS signals.
AIS signals also make it possible to track the position and speed of an individual ship in time. The map below shows a ship's route from the Gibraltar area to the port of Rotterdam. The colour of the location markers show the ship's speed (to be more precise: the speed over ground). This speed is fairly constant over the whole trip but shows strong variations in two different regions. The speed variations in the Channel area are consistent with the tidal currents that the ship encounters there and can be explained this way. Off the coast of Portugal the ship's speed is also strongly reduced. It turns out that here the ship was caught in a storm with high waves.
As well as tracking the ships using open AIS data, we can also assess what wave conditions the ship will encounter further down the route, and estimate what the effect on the speed will be. Wave models are continuously computing wave conditions worldwide and the results from these models as well as from many higher resolution regional models are published as open data (e.g. http://marine.copernicus.eu/, http://www.ecmwf.int/en/forecasts/charts, and http://polar.ncep.noaa.gov/waves/viewer.shtml?-multi_1).
The image below shows a wave map for the time that the ship was halfway to Portugal - this is produced from a model hindcast. The significant wave height near the ship was over 6 metres. Such high waves add to the ship's resistance and slow it down.
As well as hindcasts of wave-height like this, these MetOcean models also produce forecasts for up to five days ahead for a range of different output parameters - for example wind and currents, which also influence a ship's speed.
Improved estimation of ship arrival times is similar to the methods used in car navigation systems. These systems know your present position, from GPS, and calculate the distance to the destination. The speed is measured and extrapolated along the route. This way the trip duration can be estimated and thus the arrival time. These estimates are continuously updated until the destination is reached. Smart car navigation systems can improve the estimated travel time by taking into account information on traffic jams further down the road.
Ship delays are not caused by traffic jams but by bad weather. Additional wind and wave resistance and the effect of currents on the speed can be modelled and calibrated using earlier AIS data. With the help of MetOcean model forecasts future conditions along the route can be assessed and the vessel speed can be estimated. From this the travel time can be computed and an ETA assessmentm made. This process is repeated every few hours providing arrival time updates that will be more and more accurate as the ship comes closer to the port of destination.
This ship arrival time estimation is an example of a service that is made possible by open data that are made available on Internet. AIS data make it possible to track ships and the MetOcean model forecasts allow the prediction of their speeds along the route.