Intelligent geospatial maritime risk analytics using the Discrete Global Grid System

Figures & data

Figure 1. Comparison between representations of AIS data.

Table 1. Significant causes of maritime accidents.

Category	Cause	Description	Availability of data
Human and Organisational Factors	Inattention and Fatigue	Impacts decision making and position monitoring.	Not directly measurable.
	Bridge Resource Management	Break down of communication onboard results in misunderstandings/poor decision making.	Not directly measurable.
	Communication	Communication between vessels can result in misunderstandings of intentions.	Not directly measurable.
	Position Monitoring	Failure to monitor position can result in vessel off course.	Not directly measurable.
	Training and experience	Inexperience can impact decision making.	Not directly measurable.
	Regulation	Regulatory regime can require enhanced safety regimen onboard.	Not directly measurable.
Vessel and Mechanical Factors	Ship Dimensions and Manoeuvrability Characteristics	Less manoeuvrable vessels are constrained in ability to take evasive action.	AIS static data includes length, offsets and draughts. AIS MMSI/IMO Number can be linked to more detailed vessel databases such as IHS Fairplay.
	Vessel Age	Older vessels may be less well operated or have mechanical issues.	AIS MMSI/IMO Number can be linked to more detailed vessel databases such as IHS Fairplay.
	Vessel Flag State and Safety Regime	Some flag states have greater safety regimes than others.	AIS data includes vessel flag, this can be linked to Paris MoU or Flag of Convenience datasets.
	Vessel Speed	Faster navigating vessels have reduced time to take evasive action.	AIS data includes vessel speed.
External Factors	Traffic Density and Distribution	Higher traffic density increases interactions between vessels.	AIS data analysis to derive density/distribution measures.
	Waterway Geometry	More complex waterways require finer navigation skills than open sea.	Topography/geometry of waterways from charts/digital landmass models.
	Depth	Shallow waterways have more potential for vessels to run aground.	Nautical charts (Paper/ENC), digital bathymetry models.
	Weather (Visibility, wave, ice, darkness etc.)	Wind and waves can overwhelm/capsize a vessel or offset the vessel from its intended course. Visibility impacts hazard perception.	Earth observation datasets, time of transit, ice charts.
	Hydrodynamic Effects (E.g. Tidal, Bank Effect)	Impacts course-keeping ability in constrained waters.	Tidal atlases, hydrodynamic models.
	Support Availability (VTS, Tugs, TSS, Pilotage, Aids to Navigation etc.)	Risk controls can monitor/manage passage safely or respond quickly following a mechanical failure.	Nautical charts, port directions.

Figure 2. Intelligent geospatial ship grounding model framework.

Figure 3. Methods of assigning DGGS cells to geospatial data.

Figure 4. Ship grounding accident distribution.

Figure 5. AIS data processing overview.

Figure 6. Vessel traffic data.

Figure 7. Bathymetric and topographic datasets.

Figure 8. Navigational complexity.

Figure 9. Wind speeds and wave heights.

Figure 10. Tidal flows and ice thickness.

Figure 11. Presence of infrastructure.

Table 2. Summary statistics of each DGGS cell.

#	Name	Type	Min	Mean	Max	Description
1	Grounding Frequency/year	Cont.	0	0.004	6.44	Dependent variable
2	Total Vessel Exposure (hours)	Cont.	0	198	72,634	More vessel activity increases likelihood of incidents.
3	Average Wind Speed (m/s)	Cont.	2.0	6.1	10.0	Greater average wind speed reduces ship handling capability.
4	Average Wave Heights (m)	Cont.	0.0	1.24	3.3	Greater average wave size reduces ship handling capability.
5	Average Tidal Flows (m/s)	Cont.	0.00	0.03	1.27	Greater average tidal flow reduces ship handling capability.
6	Average Ice Thickness (m)	Cont.	0.0	0.1	4.8	Greater ice thickness reduces ship maneuverability.
7	Average Depth (m)	Cont.	0	2,524	7,908	Shallower depths increase risk of grounding.
8	Distance from Shore (km)	Cont.	0	3,955	15,000	Closer navigation to shore increases risk of grounding.
9	Navigational Complexity – Topography (Ratio)	Cont.	0.0	4.3	14.0	More complex waterways increase likelihood of grounding.
10	Navigational Complexity – Course Changes (degrees)	Cont.	0	0.31	35.4	More complex waterways increase likelihood of grounding.
11	Presence of TSS (Binary)	Binary	0	0.02	1	Presence of TSS reduces risk of grounding.
12	Presence of Pilot (Binary)	Binary	0	0.02	1	Presence of Pilot reduces risk of grounding.
13	Presence of VTS (Binary)	Binary	0	0.00	1	Presence of VTS reduces risk of grounding.
14	Presence of Key Ports (Binary)	Binary	0	0.02	1	Presence of Ports increases risk of grounding.

Figure 12. Ship grounding risk model.

Figure 13. Feature importance.

Figure 14. Distortion of cell area across study area.

Figure 15. Comparison of vessel traffic exposure at different DGGS resolutions.

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request. These data were derived from the following resources available in the public domain:

• https://marinecadastre.gov/

• https://marine.copernicus.eu/

• https://www.ngdc.noaa.gov/mgg/global/

• https://gadm.org/data.html

• https://homeport.uscg.mil/Lists/Content/Attachments/211/MISLE DATA.zip

• https://www.tsb.gc.ca/eng/stats/marine/index.html

Aggregated data of the above is available from: https://doi.org/10.7910/DVN/ZZHBFD