143
1 INTRODUCTION
On the morning of 1 October 2015, the U.S.-flagged
containership El Faro foundered and sank
approximately 40 nautical miles northeast of Acklins
and Crooked Island, Bahamas, while sailing close to
the eye of hurricane Joaquin. The U.S. National
Transportation Safety Board (NTSB) and U.S. Coast
Guard (USCG) investigated the sinking; the NTSB's
final report on the sinking [1] stated: "collection and
near real-time dissemination of meteorological and
oceanographic data is vital to produce the best
Sharing Ships’ Weather Data via AIS
B
. Tetreault
USACE Engine
er Research and Development Center, Vicksburg, Mississippi, United States
G.W. Johnson
Alion Science and Technology, McLean, Virginia, United States
ABSTRACT: In the aftermath of the sinking of the US-flagged containership El Faro in October 2015, one of the
U.S. National Transportation Safety Board’s (NTSB) recommendations was for the National Oceanic and
Atmospheric Administration (NOAA) to explore increasing the collection of weather data from ships in order
to improve the weather forecast products that they distribute. Currently NOAA runs the Voluntary Observing
Ship (VOS) program, where ships voluntarily submit weather observations to NOAA. However, only a small
fraction of the thousands of vessels sailing worldwide participate in this program, and VOS weather
observations are submitted infrequently, typically four times per day via a mainly manual process. One method
being explored to automate and increase the frequency of data submittal is by using the existing Automatic
identification System (AIS) equipment installed aboard vessels.
Most commercial ships (in particular those subject to the International Maritime Organization's (IMO) Safety of
Life and Sea (SOLAS) convention) have a Class A AIS transceiver installed to comply with mandatory carriage
requirements. Many vessels not required to carry AIS equipment voluntarily install AIS transceivers (either
Class A or B). All of these AIS transceivers (except for the Class B “CS”) can be used to transmit an AIS message
8 (broadcast binary message) by sending the transceiver an appropriately formatted NMEA sentence (BBM).
Weather data can be embedded in an AIS application-specific message (ASM) carried by the AIS message 8 and
automatically transmitted by the ship. This transmission can be received by terrestrial AIS stations (when in
range) or by satellite AIS receivers. The AIS weather data can then be converted into the appropriate format and
forwarded to the weather forecasting offices for use in models and weather predictions. This data may also be
of use to other researchers monitoring climate change or other environmental factors. By leveraging this
existing base of AIS transmitters, the volume of weather data being sent to weather forecast offices and others
could be greatly increased.
The U.S. Army Corps of Engineers (USACE) and U.S. Maritime Administration (MARAD) have been exploring
the feasibility of this concept. Following a simple bench test performed by Alion Science in September 2018, an
initial proof-of-concept was tested aboard the MARAD vessel Cape Wrath while moored in Baltimore in
October 2018. After this successful demonstration, a prototype was installed on the Massachusetts Maritime
Academy training ship TS Kennedy during her training cruise Jan-Feb 2019. During this cruise, the AIS
equipment aboard the ship transmitted weather data at 3-minute intervals. Several different ASM formats were
tested, including two developed specifically for this test to improve satellite reception. This report will discuss
the concept, the demonstrations, and the results to date including the efficacy of the various ASM formats.
http://www.transnav.eu
the
International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 14
Number 1
March 2020
DOI:
10.12716/1001.14.01.17
144
possible weather forecasts and advisories to keep
mariners safe" and included the following
recommendation addressed to the National Oceanic
and Atmospheric Administration (NOAA):
Coordinate with the National Weather Service, vessel
operators, automatic identification system (AIS) service
providers, and required onboard technology vendors to
perform a “proof-of-concept” project to establish whether
AIS, or another suitable alternative, can practically deliver,
in a single message, (1) meteorological and oceanographic
data obtained directly from automated instrumentation and
manual observation on board vessels at sea, (2) vessel
position and time of observation, and (3) other important
metadata, by satellite and land-based receivers, to global
meteorological authorities via the Global
Telecommunication System with acceptable time delay.
Currently NOAA runs the U.S. Voluntary
Observing Ship (VOS) program [2, 3] as the U.S.
component of the World Meteorological Organization
(WMO) VOS [4]. Under this program ships
voluntarily submit weather observations to NOAA.
However, this data is submitted infrequently,
typically four times per day and is primarily a manual
process. Many ships participating in the VOS
program use the TurboWin software [5, 6] to enter
weather observations which are then sent to NOAA
(or other meteorological offices) via email. This
software allows for all observation data to be entered;
however, it still requires the mariner to manually
collect and enter it. One method being explored to
automate and increase the frequency of data submittal
is by using weather sensors, connected to an
application that receives sensor data and converts it
onto AIS messages, which are then transmitted using
the ship's installed AIS transceiver.
AIS was originally designed primarily as a
situational awareness tool, aiding in ship-to-ship
collision avoidance, provision of vessel traffic services
from shore authorities, and allowing coastal states to
monitor their waters [7]. However, in addition to
allowing the monitoring of ship locations, AIS may
also be used for the transmission of other information
between AIS stations onboard and ashore. This has
been used extensively in the U.S. for the transmission
of navigation safety information from shore to ship
including weather, water levels, lock queues,
navigation restrictions, etc. [8-10]. It could also be
used to transmit weather information from ship to
shore.
2 AIS WEATHER TRANSMISSIONS CONCEPT
There are three main components needed on the ship
to generate and send an AIS weather message:
1 A source of weather data presented in a
recognized industry standard.
2 An AIS transceiver.
3 A processor to take the sensor’s weather data and
format it for transmission.
To receive and use the messages on shore there are
two main components:
1 An AIS receiver network.
2 A process to convert the received AIS messages
into the desired format for input into the NOAA
database.
This complete system is shown in Figure 1.
Figure 1. System Diagram.
The best source of real-time weather data is from a
digital weather station because it generates the data
automatically at up to a 1 Hz interval. There are
numerous commercial options on the market that
measure barometric pressure, wind (speed and
direction), air temperature, dew point, and humidity
and output the data in National Marine Electronics
Association (NMEA) 0183 sentences [11]. Equipment
used during our tests were the Airmar 220WX
weather station and the Mintaka Duo weather station.
The Airmar weather station contains a GPS and a
variety of sensors to provide position, roll, pitch, true
wind speed and direction, pressure, air temperature,
and relative humidity data as NMEA 0183 messages
via a RS-422 serial connection. The Mintaka weather
station provides pressure data from dual sensors as
TurboWin sentences via a USB connection. The
weather station data can be directly read by a
computer process and converted into an AIS message.
Data from the weather station can be accumulated
(averaged) and sent automatically without mariner
intervention, allowing for more frequent weather
updates.
Since some ships in the VOS program use
TurboWin to enter marine weather observations, this
can also be used as a source of data to be included in
AIS transmissions. All of the data entered in
TurboWin is encoded into ship’s synoptic code (FM
13-X) [12, 13] and written to a text file for later
transmission. This text file can be read by a computer
process and converted into an AIS message. The data
generated by TurboWin is more complete (includes
sea state, wave heights, visibility, etc.); however, it is
manually entered and typically only entered every 6
hours.
To maximize the use of the available data, both
options can be used simultaneously. The manual
TurboWin data can be combined with the automatic
weather station data and coded into the message for
transmission. The resulting message will thus be a full
data set made up of a combination of data from
TurboWin (which may be older) and up-to-date
weather station data. Some of the ASMs allow for
different time stamps for the different data fields so
that this difference in age of the data can be
communicated to the shore user.
145
Most commercial ships (in particular those subject
to the International Maritime Organization's (IMO)
Safety of Life and Sea (SOLAS) convention) have a
Class A AIS transceiver [13] installed to comply with
mandatory carriage requirements. Many vessels not
required to carry AIS equipment voluntarily install
AIS transceivers (either Class A or B). All of these AIS
transceivers (except for the Class B “CS”) can be used
to transmit an AIS message 8 (broadcast binary
message) by sending the transceiver an appropriately
formatted NMEA sentence (BBM) [11].
The “glue” between the weather data and the AIS
is a computer process. This can be an application
running on the same computer as TurboWin, or if
TurboWin is not used, it can be a separate “black box”
processor. The computer process takes the weather
data (both automatic and manually-entered) and
creates the AIS application-specific message (ASM)
[14-16] and embeds it in the NMEA BBM sentence for
transmission. This can be done as often as desired
(e.g., intervals of minutes rather than hours). The
process can also average the real-time data from the
weather station over whatever interval desired (10
minutes is what is defined in the VOS Handbook
[12]). This process can also monitor the data
transmission and push the message to the AIS
transponder again if it fails to transmit the first time
(the Class A will attempt to transmit a broadcast
binary message using Random Access Time Division
Multiple Access (RATDMA) within 4 seconds of
receiving the BBM sentence, but transmission is not
guaranteed). The complete ship-side data flow is
shown in Figure 2.
Figure 2. Ship-side data flow for AIS weather transmission.
On the shore side, these AIS messages will be
received by terrestrial shore receiver networks when
within range, as well as by satellite AIS networks
(worldwide). For our tests, reception was via the
USCG’s Nationwide AIS (NAIS) network for coastal
coverage and 2 commercial satellite AIS providers for
offshore coverage. The key component on the shore
side is the software processor that takes in the AIS
data feed (in NMEA 0183 format) from the receiver
network, isolates the AIS message 8s, parses the
sentences to retrieve the data, and then reformats it
into the desired format (BBXX [12] or BUFR [17]
messages) for sending to NOAA over the GTS (see
Figure 3). For the proof-of-concept, there is no error-
checking of the weather data; however, this could be
added to the software on the shore-side to discard
data that is out of normal parameters.
Figure 3. Shore-side data flow for weather messages via
AIS.
3 PROOF-OF-CONCEPT TESTING
The concept described above has been tested on two
ships to date, the MV Cape Wrath and the TS Kennedy.
Figure 4. MV CAPE WRATH. Equipment was installed on
bridge (see arrow).
3.1 MV CAPE WRATH
The first test was conducted on the MV Cape Wrath
(see Figure 4) in October 2018. The MV Cape Wrath
was moored in Baltimore harbor for the duration of
the test. For this installation an Airmar 220WX was
used as the weather station; it was installed on the
flying bridge with the cable running through a
stuffing tube in the pilot house window (see Figure 5).
The WeatherTransmitter and TurboWin software
were running on a laptop computer. The laptop was
placed on a ledge adjacent to the AIS pilot port; a
cable from the pilot port was run to the laptop. A
power strip plugged into the one available outlet was
used to power the laptop and weather station (see
Figure 6).
146
Figure 5. Airmar weather station mounted to rail (left) and
cable coming through stuffing tube to bridge (right).
Figure 6. Left: Laptop (orange arrow) on ledge to the right
of the pilot port (green arrow). Right: close-up of pilot port.
Figure 7. Screen shot of Laptop. TurboWin is in the back
(wave picture) and AIS Weather Transmitter is the black
window to the right. The configuration file for AIS Ship
Weather is open as well (white window). You can see the
messages being successfully transmitted in the AIS Weather
Transmitter window.
Figure 8. Shore side software (Ship Weather Monitor) is
running in upper right window. The CSV file of logged data
is shown in the background with reports from the CAPE
WRATH (MMSI 303940000) showing.
The system successfully created and transmitted
AIS messages using data automatically generated by
the weather station and manually entered using
TurboWin. Figure 7 shows the software running on
the laptop. The messages were received by the USCG
NAIS network and routed to the Ship Weather
Monitor process running on a computer at Alion
Science in New London. The received messages were
parsed and put into a text file (see Figure 8).
Some lessons learned from this initial prototype:
The system worked as planned.
The installation of the weather station and laptop
was very simple and not time-consuming.
We found that the ship's pilot port was not
correctly wired. This is a common problem
according to ship’s pilots. We thought we had this
solved with a pilot port cable that was auto-
sensing; however, it turns out that it only accounts
for transmit and receive being swapped, not the
individual transmit or receive pairs being
reversed. Since most pilots are not transmitting
with the AIS, most ships have never tested this
part of the connection so any installation needs to
be prepared for the possibility that the wires are
reversed. It might be worth creating an inline wire-
swapping box to enable quick correction of these
types of errors in the future.
3.2 KENNEDY
The second and longer test was on the Massachusetts
Maritime Academy training ship Kennedy (see Figure
9) during their winter cruise from 12 January through
26 February 2019. This was the first real-world test of
the ability to collect, process and transmit weather
observations from the ship via the ship's installed AIS
transceiver of a ship sailing offshore. The goals of the
test were to:
demonstrate transmission of both automated and
manually entered weather data from ship;
demonstrate the ability to receive the data from
terrestrial and satellite receivers; and
demonstrate the ability to decode and save the
data in various formats.
147
Figure 9. Training Ship Kennedy.
A laptop with the Weather Transmitter and
TurboWin software was installed on the bridge and
connected to the AIS via a pilot port installed near the
chart table (see Figure 10). The laptop was connected
to the ship’s Mintaka Duo weather station via USB.
Weather data was collected automatically from the
weather station as well from the data file created by
the TurboWin application. During the cruise the
bridge crew and cadets entered the data hourly. This
weather information was compiled every three
minutes into several different AIS messages and sent
to the ship's AIS Class A transceiver via the pilot port
for transmission every three minutes. The software
creating the messages created a log file of every
attempt to transmit messages, a copy of the messages,
and whether the transmission from the ship's
transceiver was successful. See Figure 11 for a picture
of the software running on the laptop.
Figure 10. Equipment installation on the TS KENNEDY.
Figure 11. Weather Transmitter and Electronic Charting
software running on the laptop on the KENNEDY.
Messages were received via the USCG NAIS
network of terrestrial transceivers when Kennedy was
within range. The USCG provides the U.S. Army
Corps of Engineers (USACE) with a live feed of their
data; this was filtered to only collect messages
transmitted by TS Kennedy and sent to Alion Science,
the USACE contractor for this effort, where they were
logged and parsed to extract the weather observation
data, which was also logged. ExactEarth, a provider
of satellite-based AIS information provided a live feed
of the data they received from TS Kennedy. Another
satellite AIS provider, Spire Global, agreed to provide
a data file at the conclusion of the test of data their
system received from TS Kennedy.
4 DATA RESULTS
4.1 ASMs
Initially three different ASMs were transmitted (see
Table 1) to assess the differences between them. These
three ASMs are all part of existing standards. Mid-
way through the cruise, two additional (shorter)
ASMs were developed to try to improve satellite
reception. A software update was pushed to the ship
and starting on 29 January these additional two ASMs
were transmitted along with the initial three.
Table 1. ASMs Used During KENNEDY test.
_______________________________________________
DAC FI Title Marker Notes
Color
_______________________________________________
1 21 Weather Blue Circ 289 ASM,
observation report circle 2 slots (360
from ship bits)
1 31 Meteorological Cyan Circ 289 ASM,
And Hydrographic diamond 2 slots (360
data bits)
367 33 Environmental Black + ASM developed
Message by USCG RDC
and used by the
USACE, multiple
(typically 3) slots
(470 bits)
367 23 Satellite Ship Red + Test ASM
Weather single slot
weather data
(168 bits)
367 24 Satellite Ship Magenta + Test ASM less
Weather Small than single slot
weather data
(128 bits)
_______________________________________________
4.2 Transmit
Figure 12 shows the locations of all transmissions
from the KENNEDY. The ship regularly transmitted
AIS messages 1 and 3 (position reports) and message
5 (static information) per the ITU Standard [18]. The
message 1s are plotted as yellow •s (but not visible at
this scale). These could have been as often as every 2
seconds. The ship additionally transmitted AIS
message 27s (Long-range AIS broadcast message).
These were transmitted at 3 minute intervals during
148
the entire cruise, and are plotted as black •’s.
1
The
ASMs (see Table 1) were also transmitted at 3-minute
intervals. These are indicated in Figure 12 with the
colors indicated in Table 1. There was a large gap in
the southbound transit where the software was not
working (due to a Windows 10 glitch that was
resolved mid-way through the cruise) and thus there
are no ASMs transmitted. In Figure 13, which is
zoomed in on the ship's track north of Cuba, the
different reports are more visible (except for the red
and magenta + which are under the black +).
Figure 12. KENNEDY transit, southbound (on right) and
back north (on left).
Figure 13. Close-up of part of the KENNEDY transit; north
of Cuba.
1
Typically message 27s would not be transmitted within range of a
base station (the base stations received by the KENNEDY are
marked on Figure 12 with red triangles); however, the AIS Class A
must receive both message 4 and message 23 from the base station
in order to turn off the message 27s and the USCG is only transmit-
ting message 4s.
4.3 Terrestrial Reception
2
When the ship was within range of a base station in
the USCG NAIS network, the ASMs were all received
at fairly high percentages (see Table 2). The lower
percentage on day 53 is due to the ship not being in
reception range of the terrestrial network the entire
day. There also appears to be some correlation
between number of slots and reception percentage;
the shorter messages are received with a higher
percentage than the longer. This is to be expected
since the ASMs are sent in RATDMA mode.
Table 2. Terrestrial Reception Percentages (slots in parens).
_______________________________________________
Day FI 21 FI 23 FI 24 FI 31 FI 33
(2) (1) (<1) (2) (3)
_______________________________________________
051 2019 89.96% 90.79% 92.90% 89.73% 89.33%
052 2019 85.36% 88.10% 87.87% 86.01% 83.89%
053 2019 65.62% 69.25% 70.08% 65.63% 65.00%
054 2019 97.90% 98.53% 98.11% 98.74% 96.44%
055 2019 97.28% 98.54% 96.65% 97.28% 96.23%
056 2019 94.76% 95.80% 96.01% 92.66% 92.45%
057 2019 95.36% 96.83% 98.31% 95.15% 92.81%
_______________________________________________
4.4 Satellite Reception
The percentage of messages received via satellite was
much lower than originally expected. Although the
authors have not seen any reception statistics from the
satellite providers, the impression given is that most
messages are received. This was definitely not our
experience. Figure 14 shows all of the AIS messages
received via satellite from the KENNEDY. The color-
coding is the same as listed in Table 1, except that +’s
are used for all of the ASMs. In the first part of the
transit, it was noticed that the reception of the ASMs
was very poor, so two additional ASMs were quickly
developed: one that was a single slot (168 bits), and
one that was less than a single slot (128 bits). These
were added into the transmit software and were
broadcast starting on day 29. The long southbound
stretch with no transmissions is when the software
was offline due to the Windows 10 issue discussed
previously.
2
While not part of this test, messages transmitted by the TS Kenne-
dy could also be received aboard other vessels. It is assumed that
most vessels do not have the capability to parse and display the re-
ceived messages. However, on at least one occasion TS Kennedy
was contacted by a nearby ship and asked about the weather data
that appeared on the nearby ship's navigation display near the
Kennedy. It is assumed that the manufacturer of the navigation
equipment used aboard the nearby vessel incorporated the ability to
display at least one of the ASMs (likely DAC 1, FI 21 or 31) on their
systems. This may be a consideration for wider implementation of
this capability, to address concerns such as "screen clutter" and "in-
formation overload."
149
Table 3. Sample Satellite AIS Reception Percentages.
___________________________________________________________________________________
Day % Msg 1 % Msg 3 % Msg 27 % F21 % F23 % FI24 % FI31 % FI33
___________________________________________________________________________________
029 2019 0.40% 5.2% 64.4% 0.64% 2.38% 7.14% 1.07% 0.00%
030 2019 0.43% 5.7% 75.0% 0.47% 3.53% 6.13% 1.41% 0.00%
031 2019 0.32% 4.1% 62.8% 0.27% 1.89% 4.05% 1.62% 0.00%
032 2019 0.14% 1.7% 61.5% 0.21% 2.08% 2.51% 1.04% 0.00%
033 2019 0.00% 3.1% 63.1% 1.05% 2.73% 2.94% 0.21% 0.00%
034 2019 0.00% 3.1% 66.7% 0.42% 3.35% 4.61% 0.42% 0.00%
035 2019 0.07% 2.3% 65.8% 0.64% 1.91% 6.37% 0.21% 0.00%
___________________________________________________________________________________
Table 4. Additional Satellite AIS Reception Data.
___________________________________________________________________________________
Day % Msg 1 % Msg 3 % Msg 27 % F21 % F23 % FI24 % FI31 % FI33
___________________________________________________________________________________
051 2019 0.09% 0.6% 68.5% 0.21% 0.41% 0.63% 0.00% 0.00%
052 2019 0.08% 7.0% 67.7% 0.21% 1.88% 9.60% 0.21% 0.00%
053 2019 0.06% 8.0% 71.9% 0.21% 2.92% 15.27% 0.21% 0.00%
054 2019 0.19% 3.8% 75.2% 0.42% 3.35% 2.93% 0.00% 0.42%
055 2019 0.01% 18.9% 77.3% 0.21% 3.98% 24.11% 0.21% 0.00%
056 2019 0.00% 24.6% 63.8% 0.00% 5.02% 21.97% 0.63% 0.00%
057 2019 0.08% 18.8% 70.2% 0.84% 6.93% 22.90% 0.84% 0.63%
058 2019 0.33% 20.7% 68.3% 0.42% 3.59% 22.78% 1.27% 0.21%
___________________________________________________________________________________
Figure 14. AIS reports received via satellite.
Table 3 is a sampling of satellite reception results
for when the KENNEDY was in transit between
Barbados and the British Virgin Islands (after the new
ASMs were added). The reception percentage for all
of the ASMs was very low. There is a very strong
correlation with message length: the longest ASM
(FI33) was not received at all, and the shortest (FI24)
was received the most. The AIS message 1’s were
received less than 1% of the time. Interestingly the
message 3’s were received at a higher percentage
(though still low). After consultation with the satellite
provider, we determined that this was due to their
filtering of the data to 1 message of each type (1,3,5) in
every 10 minutes. Also of note is that the reception of
AIS message 27 was the best by far; this is because
these messages are only 96 bits long and more
importantly, are transmitted on two different VHF
channels (75 and 76) and thus avoid any conflict with
AIS traffic on channels AIS1 and AIS2.
Data from the end of the transit is shown in Table
4. Of interest are the times that the vessel was in port.
On day 53 the Kennedy was in New York and days 55-
58 she was back at homeport (Bourne, MA). On these
days the reception of most messages was significantly
higher. After consultation with the satellite provider
we determined that we had also been receiving some
messages via terrestrial receivers, which skewed these
results (except for the message 27s which are only
received via satellite). We also discovered that due to
overlapping satellite coverage, at times we received
duplicate message 27s; these duplicates have been
removed from the data before the percentages in
Table 3 and Table 4 were calculated.
5 CONCLUSIONS
It is clear from the data reviewed thus far from this
single voyage that reception of weather ASMs when
in range of shore stations is very reliable, on the order
of 90%. This was not a surprising result. Of more
interest are the results for satellite reception. In
particular, that smaller ASMs are received much more
reliably via satellite. The downside of the smaller
messages is less data. The largest message (ASM FI
33) carries the most met/hydro data and also has time
stamps for each data element allowing for the most
information transfer. This message however, does not
work well with the satellite link. It should also be
noted that although the percentage of messages
received via satellite is low, it is still more messages
per day than the number of synoptic reports that a
vessel would typically send (i.e., 4).
6 RECOMMENDATIONS/FUTURE
Some avenues for further research are:
Another test ASM will be developed to reduce the
size to 96 bits (the same as the AIS message 27). To
reach this size, the position will be dropped from
the weather ASM, and we will assess the feasibility
of linking the received weather ASM to the
position from the AIS message 27s.
150
Additional ship testing is planned, using a larger
pool of vessels (perhaps 10). Testing satellite
reception on trans-oceanic voyages may also
increase satellite reception for multi-slot messages,
due to fewer vessels being in the satellite footprint.
Expanded tests aboard commercial vessels would
be beneficial from several aspects:
Evaluating this capability in various environments
(coastal, deep sea, Great Lakes, transoceanic, etc.)
Gathering information about the different weather
collection capabilities and AIS equipment aboard
commercial vessels
Gathering feedback and concerns from vessel and
shipping company personnel
Gathering more data to test reception from other
terrestrial and satellite sources
For the additional ship testing, a small embedded
processor will be used to assess the feasibility of
small, low-cost, hands-off kits for wider
installation.
Beginning work on the back end requirements for
decoding, routing, quality checking, and using the
received observations in modelling, forecasting,
and other systems.
ACKNOWLEDGEMENTS
This work has been a partnership between USACE, U.S.
Maritime Administration (MARAD), and NOAA. The
shipboard data collection was supported by the Chief
Engineer on MV Cape Wrath, and by the Captain and crew of
TS Kennedy, especially Arthur Levine and Steve Gardiner.
Evaluation of the feasibility of satellite reception of the AIS
weather messages would have been impossible without the
in-kind contributions of satellite AIS data by ExactEarth and
Spire Global.
REFERENCES
[1] "Sinking of US Cargo Vessel SS El Faro, Atlantic
Ocean, Northeast of Acklins and Crooked Island,
Bahamas, October 1, 2015," National Transportation
Safety Board (NTSB),, Washington, DC, Accident
Report NTSB/MAR-17/01, December 2017.
[2] E. C. Kent, G. Ball, D. I. Berry, J. Fletcher, A. Hall, S.
North, et al., "The Voluntary Observing Ship (VOS)
Scheme," in OceanObs ‘09, Venice, Italy, 2009.
[3] (8 April). United States Voluntary Observing Ship
Program. Available: https://vos.noaa.gov
[4] (10 April). Observation Components of the Global
Observing System. Available: http://www.wmo.int/
pages/prog/www/OSY/Gos-components.html
[5] "TURBOWIN Plus Version 3.1.2, Installation and
User’s Manual," ed: National Oceanic and
Atmospheric Administration, 2018.
[6] (9 April). TurboWin Software. Available:
http://projects.knmi.nl/turbowin/
[7] "Recommendation on Performance Standards for an
Universal Shipborne Automatic Identification System
(AIS)," International Maritime Organization, London,
IMO Resolution MSC.74(69), Annex 3, 12 May 1998.
[8] G. W. Johnson, M. Winkler, and B. Tetreault,
"Dissemination of enhanced Marine Safety Information
(eMSI) via AIS," in Radio Technical Commission for
Maritime Services Annual Conference, St Petersburg, FL,
2016.
[9] G. W. Johnson, B. J. Tetreault, and I. M. Gonin,
"Development of an AIS Transmit Architecture to
support the dissemination of electronic Marine Safety
Information (eMSI)," presented at the European
Navigation Conference (ENC2014), Rotterdam,
Netherlands, 2014.
[10] G. W. Johnson and M. Wiggins, "Providing e-
Navigation on Inland Rivers with AIS," presented at
the Smart Rivers Conference 2017, Pittsburgh, PA,
2017.
[11] "NMEA 0183: Standard for Interfacing Marine
Electronic Devices, Version 4.1," National Marine
Electronics Association, Severna Park, MD, Standard
NMEA 0183 Version 4.1, June 2012.
[12] "National Weather Service Observing Handbook No.
1: Marine Surface Weather Observations," U.S. Dept of
Commerce, National Oceanic and Atmospheric
Administration, Stennis Space Center, MS, Handbook
No. 1, May 2010.
[13] "Maritime navigation and radiocommunication
equipment and systems Automatic identification
systems (AIS) Part 2: Class A shipborne equipment
of the universal automatic identification system (AIS)
Operational and performance requirements, methods
of test and required test results," International
Electrotechnical Commission (IEC), Geneva,
Switzerland, International Standard 61993-2 Ed. 3, July
2018.
[14] "Guidance on the Use of AIS Application-Specific
Messages," IMO, Maritime Safety Committee, London
SN.1/Circ. 289, 2 June 2010.
[15] "Guidance for the Presentation and Display of AIS
Application-Specific Messages Information," IMO,
Maritime Safety Committee, London, IMO SN.1/Circ.
290, 2 June 2010.
[16] "Harmonized implementation of Application Specific
Messages (ASM)," International Association of Marine
Aids to Navigation and Lighthouse Authorities, Saint
Germain en Laye, FR, IALA Recommendation e-NAV-
144, June 2011.
[17] "Manual on Codes - International Codes Vol 1.2,"
World Meteorological Organization, Geneva,
Switzerland, Annex II to the WMO Technical
Regulations, WMO-No. 306, 2017
[18] "Technical Characteristics for an Automatic
Identification System using Time Division Multiple
Access in the VHF Maritime Mobile Frequency Band,"
International Telecommunications Union,
Radiocommunication Sector, ITU-R Standard M.1371-
5, February 2014.